MBL U.S. Department of Commerce Volume 101 Number 1 January 2003 Fishery Bulletin U.S. Department of Commerce Donald L Evans Secretary National Oceanic and Atmospheric Administration Vice Admiral Conrad C. Lautenbacher Jr., USN (ret.) Under Secretary for Oceans and Atnnosphere National Marine Fisheries Service William T. Hogarth Assistant Administrator for Fisheries lOOFc The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fish- eries Service, NOAA, 7600 Sand Point Way NE, BIN C 15700, Seattle, WA 981 15-0070. Periodicals postage is paid at Seattle, WA, and at additional mailing offices. POST- MASTER: Send address changes for sub- scriptions to Fishery Bulletin, Superin- tendent of Documents, Attn.: Chief, Mail List Branch, Mail Stop SSOM, Washing- ton, DC 20402-9373. Although the contents of this publica- tion have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $45.00 domestic and $.56.25 foreign. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Scientific Editor Dr. Norman Bartoo Editorial Assistant Sarah Shoffler National Marine Fisheries Service, NOAA 8604 La Jolla Shores Drive La Jolla, California 92037 Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 981 15-0070 Editorial Committee Dr. Harlyn O. Halvorson Dr. Ronald W. Hardy Dr. Richard D. Methot Dr. Theodore W. Pletsch Dr. Joseph E. Powers Dr. Harald Rosenthal Dr. Fredric M. Serchuk Dr. George Watters University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Service Universitat Kiel, Germany National Marine Fisheries Service National Marine Fisheries Service Fishery Bulletin web site: fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new .system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 101 Number 1 January 2003 Fishery Bulletin JAN 2 1 2003 Contents Articles 1 -9 Blaylock, Reginald B., Leo Margolis, and John C. Holmes The use of parasites in discriminating stocks of Pacific halibut (Hippoglossus stenolepis) in the northeast Pacific 10-21 Comyns, Bruce H., Richard F. Shaw, and Joanne Lyczkowski-Shultz Small-scale spatial and temporal variability in growth and mortality of fish larvae in the subtropical northcentral Gulf of Mexico: implications for assessing recruitment success 22-31 DeMartini, Edward E., Gerard T. DiNardo, and Happy A. Williams Temporal changes in population density, fecundity, and egg size of the Hawaiian spiny lobster (Panulirus marginatus) at Necker Bank, Northwestern Hawaiian Islands The conclusions and opinions expressed in Fishery Bullelm are solely those of the authors and do not represent the official position of the National Marine Fisher- ies Service (NOAA) or any other agency or institution- The National Marine Fisheries Service iNMFS) does not approve, recommend, or endorse any propnetary product or pro- pnetary material mentioned in this pub- lication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends, or endorses any proprietary product or pro- prietary matenal mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. 32-43 Friedlander, Alan M., and David A. Ziemann Impact of hatchery releases on the recreational fishery for Pacific threadfin (Polydactylus sexfilis) in Hawaii 44-57 Hart, Deborah R. Yield- and biomass-per-recruit analysis for rotational fisheries, with an application to the Atlantic sea scallop (Placopecten magellanicus) 58-74 Hearn, William S., and Thomas Polacheck Estimating long-term growth-rate changes of southern bluefin tuna (Thunnus maccoyii) from two periods of tag-return data 75-88 Loefer, Joshua K., and George R. Sedberry Life history of the Atlantic sharpnose shark (Rhizoprionodon terraenovae) (Richardson, 1836) off the southeastern United States Fishery Bulletin 101(1) 89-99 Maunder, Mark N., and George M. Walters A general framework for integrating environmental time series into stock assessment models: model description, simulation testing, and example 100-115 Miller, Michael J., David M. Nemerson, and Kenneth W. Able Seasonal distribution, abundance, and growth of young-of-the-year Atlantic croaker (Micropogonias undulatus) in Delaware Bay and adjacent marshes 116-128 Newman, Stephen J., and lain J. Dunk Age validation, growth, mortality, and additional population parameters of the goldband snapper (Pristlpomoldes multidens) off the Kimberley coast of northwestern Australia 129-146 Ralston, Stephen, James R. Bence, Maxwell B. Eldridge, and William H. Lenarz An approach to estimating rockfish biomass based on larval production, with application to Sebastes jordani 147-167 Winship, Arliss J., and Andrew W. Trites Prey consumption of Steller sea lions (Eumetopias jubatus) off Alaska: How much prey do they require? Notes 168-174 Beerkircher, Lawrence, Mahmood Shivji, and Enric Cortes A Monte Carlo demographic analysis of the silky shark (Carcharhlnus falciformis): implications of gear selectivity 175-182 Gunderson, Donald R., Mark Zimmermann, Daniel G. Nichol, and Katherine Pearson Indirect estimates of natural mortality rate for arrowtooth flounder (Atheresthes stomias) and darkblotched rockfish (Sebastes cramerl) 183-188 Marcogliese, David J., Elaine Albert, Pierre Gagnon, and Jean-Marie Sevigny Use of parasites in stock identification of the deepwater redfish (Sebastes mentella) in the Northwest Atlantic 189-193 Polovina, Jeffrey J., Evan Howell, Denise M. Parker, and George H. Balazs Dive-depth distribution of loggerhead (Caretta caretta) and olive ridley (Lepidochelys olivacea) sea turtles in the central North Pacific: Might deep longline sets catch fewer turtles? 194-198 Smith, Susan E., Robert A. Mitchell, and Dan Fuller Age-validation of a leopard shark (Triakis semifasciata) recaptured after 20 years 199 Subscription form Abstract — The use of parasites as in- dicators of the stock structure of Pacific hahbut {Hippoglossus stenolepis) in the northeast Pacific was investigated by using 328 adult (>55 cm fork length) hahbut from 15 composite localities ranging from northern California to the northern Bering Sea and 96 ju- venile (10-55 cm) halibut from five localities ranging from the northern Queen Charlotte Islands to the Bering Sea. Counts of eight selected parasite species (the juvenile acanthocephalans Corynosoma strumostim and C. vil- losum, the metacestode Nyhelinia sur- menicola, the digenean metacercaria Otodistnmum sp., and the larval nema- todes Anisakis simplex, Pseudoterra- noi'a decipiens, Contracaecum sp., and Spirurid gen. sp.) that produce infec- tions of long duration, do not multiply in the host, and that have a relatively high abundance in at least one geo- graphic locality were subjected to dis- criminant function analysis. Juvenile Pacific halibut showed no separation and, even though they were not heav- ily infected with parasites, the analysis suggested that juveniles could be a mixed stock. Three groups of adults were identified: fish from California to the southern Queen Charlotte Islands, those from the northern Queen Char- lotte Islands to the central Bering Sea, and those from the central and north- ern Bering Sea. These groups suggest that the single stock concept be more thoroughly evaluated. The use of parasites in discriminating stocks of Pacific halibut (Hippogiossus stenolepis) in the northeast Pacific Reginald B. Blaylock Department of Biological Sciences University of Alberta Edmonton, AB T6G 2E9, Canada and Department of Fisheries and Oceans Pacific Biological Station Nanaimo, B.C. V9R 5K6, Canada Present address: College of Manne Sciences The University of Southern Mississippi 703 East Beach Blvd PO Box 7000 Ocean Spnngs, Mississippi 39566-7000 E-mail address: reg blaylock@usm.edu Leo Margolis (deceased) Department of Fishenes and Oceans Pacific Biological Station Nanaimo, B.C. V9R 5K6, Canada John C. Holmes Department of Biological Sciences University of Alberta Edmonton, AB T6G 2E9, Canada Manuscript accepted 10 July 2002. Fish. Bull. 101: 1-9(2003). ' The Pacific halibut (Hippogiossus steno- lepis ) is an Arctic-Boreal Pacific pleuro- nectid flatfish ranging throughout the North Pacific from southern California to northern Japan, but is most abun- dant in the Gulf of Alaska, The halibut supports one of the top five commercial fisheries in North America, with aver- age annual landings of approximately 25,000 metric tons from 1991 to 1995 (IPHC, 1996), and is also widely sought in the sport fishery, thus contributing significantly to west coast economies. The International Pacific Halibut Com- mission (IPHC) is responsible for man- agement of the resource. From the 1930s through the 1950s the IPHC re- cognized at least three stocks of hali- but from tagging experiments, egg and larval drift, anatomical differences, and differences in growth rate: 1) those in the Bering Sea; 2) those from the Gulf of Alaska south to Cape Spencer, Alaska; and 3) those south of Cape Spencer (Skud, 1977). These bound- aries roughly followed the zoogeo- graphic zonation in the North Pacific. Skud (1977) re-analyzed the data and concluded that there was extensive intermingling of fish among areas and that there was no evidence to indicate that fish north and south of Cape Spencer, Alaska, constituted different stocks. Available biochemical evidence (Tsuyuki et al., 1969; Grant et al., 1984), although limited in scope and by sampling effort, suggests little genetic variation throughout the northeast Pacific. As a result, the IPHC manages halibut as a single population, but with statistical divisions for management of data. Parasites have been used successfully to distinguish populations or stocks of fishes and, as a result, provide informa- tion useful in fisheries management ( see Fishery Bulletin 101(1) 180 Figure 1 Sampling localities for 328 adult (circles) and 96 juvenile (squares) Pacific halibut, Hippoglossus stenolepis, in the northeast Pacific. OR = Oregon-northern California, WA = Washington, VI = Van- couver Island, SQC = southern Queen Charlotte Islands, NQC = northern Queen Charlotte Islands, SAl = southeast Alaska site 1, SA2 = southeast Alaska site 2, KP = Kenai Peninsula, KI = Kodiak Island, NI = Nagai Island, UP = Unimak Pass, WAL = western Aleutian Islands, SB = southern Bering Sea, PI = Pribilof Island, SMI = St. Matthew Island. Individual hauls with at least 10 fish (for a total of 202 fish) are shown as solid circles. Other collection sites are shown as stippled circles. See Table 1 for sample sizes. reviews by Lester, 1990;Moser, 1991;Williamset al., 1992). With respect to flatfish, Gibson ( 1972) used parasitological data to distinguish three groups of Platichthys flesus and Krzykawski and Wierzbicka (1982) used parasitological data and other information to distinguish between Bar- ents Sea and Labrador stocks of Greenland hahbut, Rein- hardtius hippoglossoides. Khan et al. (1982) and Arthur and Albert (1993) used parasites to distinguish between Atlantic and Gulf of St. Lawrence stocks of/?, hippoglos- soides, and Boje et al. (1997) used parasites to indicate differences among Greenland stocks of Greenland halibut and stocks from the western Atlantic. No similar work on flatfishes has been done in the Pacific and, with the excep- tion of Krzykawski and Wierzbicka ( 1982) and Boje et al. (1997), there has been no attempt to distinguish between stocks of a species across a significant portion of the spe- cies' range. In this article, we use discriminant analysis on counts of some of the parasites from adult Pacific halibut to deter- mine if they form discrete groups or stocks in the north- east Pacific. We do a similar analysis on the juvenile fish and compare the results to the adult analysis to determine when separation is likely to occur Materials and methods A total of 328 adults ( >55 cm fork length) from 15 composite localities, ranging from northern California to the vicinity of St. Matthew's Island in the Bering Sea and 96 juveniles (10-55 cm) from five localities ranging from the north- ern Queen Charlotte Islands to the Bering Sea (Fig. 1), were caught by staffs of the IPHC and the U.S. National Marine Fisheries Service during the summers of 1990-92 (using longlines and trawls). Most localities (for the adult samples) included fish taken from several hauls; however, 202 fish came from 13 individual hauls, each of which con- tained at least 10 fish. Fish were bagged individually and immediately frozen at sea for later examination. Fish and parasites were processed by using standard parasitological techniques (see Blaylock et al., 1998a). We followed Bush et al.'s (1997) definitions for prevalence, abundance, and intensity. Parasites used in the analyses were chosen according to the guidelines of Arthur and Albert (1993). Only those species with infections of long duration, that do not multiply in the host, and that have a relatively high abundance in at least one geographic locality were used. Of the 59 parasite taxa identified from Blaylock et al : Use of parasites In discriminating stocks of Hippoglossus stenolepis Pacific halibut (Blaylock et al., 1998a), eight taxa met these criteria: the juvenile acanthocephalans Corynosoma strumosum (body cavity) and C. villosum (body cavity), the metacestode A^v6e/;>;(a surmenicola (stomach wall), the di- genean metacercaria Otodistomum sp. (stomach wall), and the larval nematodes Anisakis simplex (body cavity, or- gans, musculature), Pseudoterranova decipiens (body cav- ity, organs, musculature), Contracaecum sp.(body cavity), and Spirurid gen. sp. (stomach wall). A ninth taxon, the larval nematode Hysterothylaciun7 adiincum (body cavity and organs) was included for the analysis of juveniles. Because individual fish varied extensively in size (fork length), and the number of a parasite individuals was strong- ly correlated with fish size (Blaylock et al., 1998a), parasite numbers were corrected for differences in host size. Counts of individual parasites were first log-transformed (ln(A:-i-l)). To adjust for the effect offish length, a regression of the trans- formed parasite numbers on fish length for each species in each locality (and haul) was calculated. This relationship was then used to adjust the number of parasite individuals within each fish in each locality (and haul) to that expected for the average-size fish in the overall sample (80.9 cm for adults, 39.2 cm for juveniles). These data were then used in discrimi- nant function analyses. We applied the most widely used (and available) method of discriminant function analysis, in which the data were divided into training and test sets, and a dis- criminant function calculated on the training set was used to classify the test set. Interpretations were based on patterns in the test sets. To insure that any identified patterns were due to differences among localities rather than simply differ- ences among individual hauls, we performed the same analy- sis on both the locality and the individual haul data. Our training set consisted of six fish randomly selected from each haul ("haul" training set) or these fish plus four from the northern Queen Charlotte Islands and six from Unimak Pass ("locality" training set). Discriminant func- tions calculated from data on these "training" fish were used to classify each of the remaining fish from each haul ("haul" test set) or those fish plus all remaining fish ("local- ity" test set). The test set fish were first classified into one of the 13 hauls or 15 localities. Classification matrices were examined for the degree of misclassification. Hauls or local- ities were then grouped and regrouped into four and three groups based on patterns in the 13 or 15 category analyses and the zoogeographic zones from Blaylock et al. (1998b). Analyses were then repeated. Classifications were exam- ined for misclassification, and boundaries adjusted for re- testing. Results presented are those from the best fit "test" classifications. Statistical analyses were performed in SYS- TAT for Windows version 5.05 (Wilkinson et al., 19921. The entire data set from which the data for this analysis came is available for purchase from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Coun- cil of Canada, Ottawa, ON KIA 0S2, Canada. Results Of the taxa that met the Arthur and Albert ( 1993) criteria, A^. surmenicola was most common and abundant in north- ern localities and fairly common and abundant in central localities. Corynosoma strumosum, although variable in prevalence and abundance, was much more common in the northernmost localities. Corynosoma villosum, although prevalent everywhere, was more abundant in northern fish. Otodistomum sp. and Spirurid gen. sp. were more common and abundant in southern localities. Anisakis simplex, although present in virtually every fish from every locality, was more abundant in southern fish. Pseu- doterranova decipiens and Contracaecum sp. appeared to be more common in central areas (Table 1). In the juve- niles, A. simplex and P. decipiens were more common in central localities, whereas C. villosum, C. strumosum, and Hysterothylacium aduncum were more common in north- ern localities (Table 2). The haul analyses indicated that the majority of fish from some hauls (12/14 Vancouver Island |VI] fish, 3/4 Southeast Alaska 1 |SA1| fish, 3/5 from the Pribilof Islands |PI|, and all 4 from St. Matthew's Island ISMI]) could be correctly classified but that fish from surrounding areas also were incorrectly classified to these hauls. Moreover, the percentage offish correctly classified by the haul func- tions was, in all cases, within only a few percentage points of that correctly classified by the equivalent locality func- tion. Thus, patterns do not appear to be associated with independent hauls. Therefore, we present only the results of the locality analyses. Fifteen category discriminant analyses revealed severe misclassification in most areas. Only 39% were correctly classified to locality (Table 3). The functions did assign correctly the majority of test fish from two localities ( 19/26 from Vancouver Island [VI] and 14/22 from the southern Bering Sea [SBl ). However, misclassification of fish from surrounding areas to these localities indicated less than accurate discrimination. The clearest indications from these analyses were that localities from the vicinity of the Queen Charlotte Islands south should be grouped together and that there is a suggestion that the two northern Ber- ing Sea locations (PI and SMI) should be grouped. Regrouping the localities into four categories by using boundaries from zoogeographic analyses (Blaylock et al., 1998b) plus the apparent northern Bering Sea grouping (PI-SMI), considerably improved the predictive ability of the functions. The "best fit" four-category grouping gave ap- proximately 62% correct classification at the locality level (Table 4). The four-category functions were good predictors for the California-Oregon (OR) to southern Queen Char- lotte Islands (SQC) fish; over 80% of these southern fish were correctly classified, and only about 6% of the other fish were misclassified to this group. Over 70% of the Pribilof-St. Matthew Island (PI-SMI) fish were correctly classified, and only 7% of the other fish were incorrectly classified to this group. There was much misclassification in the two central groups, and adjustment of the boundary between these two groups did not produce marked improvement (not shown). Grouping into three categories by combining the two central groups resulted in substantial improvement in discrimination (83% correct) (Table 5). Shifting of the boundary between the northern and central group re- vealed that discrimination broke down when the southern Fishery Bulletin 101(1) Table 1 Summary of parasites used for discriminatioi SA1= southeast Alaska 1, SA2 = southeast Al be = body cavity, o = organs, m = musculature. of stocks of adult Pacific halibut by locality. OR = Oregon-northern California, WA = aska 2, KP = Kenai Peninsula, KI = Kodiak Island. NI = Nagai Island. WAI = western sw = stomach wall. Intensity = mean number of parasites per infected host. Parasite Site Stage OR(n=23) WA(/i = 14) % Intensity % Intensity Anisakis simplex be, 0, m larva 100 258.2 ±520.2 100 122.2 ±101.0 Corynosoma villosum be juvenile 74 7.9 ±5.4 71 6.3 ±7.3 Corynosoma strtimosum be juvenile 52 5.6 ±6.4 50 6.3 ±5.2 Nybelinia surmenicola sw metacestode 17 2.8 ±2.9 14 4.0 +4.2 Otodistomum sp. sw metacercaria 44 14.3 ±13.8 36 32 ±55.7 Pseudoterranoua decipiens be, 0, m larva 44 2.5 ±2.1 21 2.3 ±1.5 Contracaecum sp. be larva Spirurid gen. sp. sw larva 22 1.4+0.5 50 11 ±24.7 Parasite KP(n=21) KI(«=26) NI(?! = 131 % Intensity % Intensity % Intensity Anisakis simplex 100 3.3.6+22.1 100 29.5 ±27.3 100 80.3 ±62.0 Corynosoma villosum 95 11.1 ±12.4 100 11.2 ±18.4 85 12.6 ±15.0 Corynosoma strumosum 19 1 ±0.0 27 1.3 ±0.5 8 2.0 ±0.0 Nybelinia surmenicola 43 4.2 ±3.9 65 29.5 +106.6 39 42.2 ±89.0 Otodistomum sp. 14 2.0 ±1.7 4 1±0 8 1±0.0 Pseudoterranoua decipiens 29 1.5+1.2 50 1.9 ±1.4 61 2.5 ±2.0 Contracaecum sp. 62 3.0 +3.08 50 3.2 ±3.2 Spirurid gen. sp. 9 1.0 ±0 4 1 ±0.0 Table 2 Summary of parasites used for discrimination of stocks of juvenile Pacific halibut. NQC = northern Queen Charlotte Islands, NI = Nagai Island, UP = Unimak Pass, PI = Nunivak Island (central Bering Sea), be = body cavity, o = organs, m = musculature, sw = stomach wall. Intensity = mean number of parasites per infected host. Parasite Site Stage NQC (n=20) KI(f!=13) NI (n= 20) UP(fi=20) PI (n=23) % Intensity % Intensity % Intensity '7f Intensity 9, Intensity Anisakis simplex be, 0, m larva 25 1.2 8 1.0 ±0 65 3.1 ±1.9 70 3.6 ±3.1 30 1.7 ±0.8 Corynosoma villosum be juvenile 5 1 ±0.0 46 1.3 ±2.4 75 3.4 ±2.6 80 5.8 ±7.4 9 1.5 ±0.7 Corynosoma strumosum be juvenile 8 1.0 ±0 5 1 ±0.0 48 1.2 ±0.4 Hysterolhylacium aduncum be, juvenile 15 1 ±0.0 92 4.4 ±3.2 95 8.7 ±10.7 85 5.4 ±8.8 74 2.8 ±1.5 Nybelinia surmenicola sw metacestode 5 1 ±0.0 26 1.3 ±0.8 Pseudoterranova be, 0, m larva 8 1.0 ±0 25 1.4 ±0.9 15 1 ±0.0 4 6 ±0.0 decipiens Contracaecum sp. be larva 15 1.0 ±0 10 5 ±5.7 Spirurid gen. sp. sw larva 5 2.0 ±0.0 Blaylock et al.: Use of parasites in discriminating stocks of Hippog/ossus slenolepis Table 1 Washington, VI = southern Vancouver Island, SQC = Aleutian Islands. SB = southern Bering Sea, PI, Pribi southern Queen Charlotte Islands, NQC = northern Queen Charlotte Islands, of Island (central Bering Sea), SMI = St. Matthew Island (northern Bering Seal. VI(/i=32) 3QC(n=31) NQC(n=8) SAl (/i=20) SA2 (n=29) % Intensity % Intensity % Intensity '7, Intensity % Intensity 100 381.4 ±357.1 100 167.8 ±101.4 100 76.1 ±47.6 100 81.8±141.1 100 44.0 ±57.0 94 8.8 ±7.4 94 13.7 ±24.3 100 5.9 ±7.4 90 10.9 ±16.6 93 16.1 ±32.0 44 3.6 ±5.2 39 1.8 ±1.5 25 2.0 ±0.9 15 1.7 ±0.6 52 1.9 ±1.0 19 16.0 ±31.9 3 1.0 ±0.0 50 13.3 ±23.8 30 11.8 ±25.1 52 4.1 ±2.0 9.4 18.3 ±30.0 39 8.3 ±13.9 38 4.7 ±2.3 35 13.0 ±11.0 21 8.3 ±11.0 16 1.2 ±0.4 13 1.5 ±0.6 25 2.0 ±0.0 25 3.2 ±4.4 69 3.1+3.0 3 1.0+0 50 3.8 ±4.9 40 3.5 ±2.2 45 2.0 ±2.0 3 1.0 ±0 15 2.3 ±1.2 10 2.8 ±2.0 UP(n=20) WAI(«=20) SB(n=29) PI(n = 14l SMI (n=28) % Intensity % Intensity % Intensity 7, Intensity % Intensity 100 53.3 ±43.8 100 41.5 ±55.0 100 40.6 ±33.6 100 21.5 ±22.3 84 10.9 ±10.0 85 29.6 ±57.4 95 13.8 ±14.0 97 16.8 ±18.3 86 34.4 ±37.0 90 34.4 ±55.9 30 2.0 ±1.5 40 2.1 ±2.0 38 8.5 ±20.2 79 9.0 ±9.3 77 23.2 ±25.4 55 2.7 ±2.1 35 16.1 ±33.8 52 6.7 ±13.1 57 34.6 ±69.2 58 25.0 +69.8 5 1±0 45 1.4 ±0.7 45 2.3 ±2.0 38 2.3 ±1.5 29 2.3 ±1.3 23 2.4 ±1.8 15 2.0 ±1.0 30 1.3 ±1.0 38 3.0 ±4.5 14 1.0 ±0 7 4.0 ±3.5 5 1.0 ±0.0 4 1 ±0.0 Bering Sea (SB) was included in the northern group (not shown). Inclusion of the northern Queen Charlotte Islands (NQC) in the southern group had little effect (81% cor- rect classification) (not shown). These analyses indicated a southern (OR-SQC) group, a central (NQC-SB) group, and a northern (PI-SMI) group. Classification into two categories (with SQC as the di- viding line) provided no substantial improvement (87'^ correct) (not shown). Inclusion of NQC in the southern group had little effect (88% correct classification). Discrimination of juveniles was poor with any organi- zation of localities. The "best fit" classification correctly classified only 66% of the fish and there was substantial misclassification among the localities (Tables 6 and 7). Fish from the northern Queen Charlotte Islands (NQC) through Nagai Island (NI) separated reasonably well, but the majority of fish from the northernmost locality were also misclassified to this group. Note that parasite num- bers and prevalences were low in the juveniles (Table 2). Discussion Our results show four things: 1) parasites clearly differen- tiate a group of southern adults; 2) parasites provide some evidence for a separation of the northernmost adults; 3) the differentiation is not always unequivocal; and 4) para- sites do not differentiate groups of juvenile fish. Skud (1977) concluded that southern and northern groups mixed extensively at all ages of their life history and that, although populations of adults may be largely discrete in the summer, any such discreteness was tempo- rary because tagging evidence suggested more extensive winter migrations associated with spawning. Our data, on the other hand, suggest that there is some merit to the IPHC's early recognition of three stocks of adult halibut. Parasite data support the existence of two major groups of halibut and suggest the possibility of a third group in the central and northern Bering Sea. The high proportion of correct classifications based on parasites suggest that these differences are well established. Recognition of three such groups is also supported by several of Skuds (1977) observations. He presented data suggesting that after fish home to spawning areas, southern and northern fish maintain reasonably separate migration circuits between feeding and spawning grounds. Data from Skud (1977) and more recent tagging data (Geernaert, 1996) also suggest that southern fish move less than their northern counterparts. Skud also recog- nized a resident population in the Bering Sea. These con- Fishery Bulletin 101(1) Table 3 Cross validation results of a 15-category discriminant function classification of adult Pacific halibut in the northeast Pacific based on parasite data. Numbers offish assigned to each locality and the corresponding percentage of the sample assigned to that cat- egory are shown. See Table 1 legend for key to abbreviations. Correct classifications are shown in bold ('289r of 240). True category Assigned category OR WA VI SQC NQC SAl SA2 KP KI NI UP WAL SB PI SMI OR 4 17% 2 9% 1 4% 2 9% 2 9% 3 13% WA 2 25% 2 3% 1 13% 1 13% 2 25% VI 1 4% 19 73% 2 8% 3 12% 1 4% SQC 1 4% 1 4% 10 40% 11 44% 2 8% NQC 2 50% 2 50% SAl 1 7% 1 7% 4 29% 1 7% 1 7% 7% 3 21% SA2 1 4% 1 4% 2 9% 1 4% 3 13% 3 13% 4% 1 4% 8 35% 1 4% 1 4% KP 7 50% 1 7%: 1 7% 1 7% 7% 2 14% 1 7% KI 7 35% 2 10% 1 5% 2 10% 5% 2 10% 3 15% 2 10% Nl 1 14% 1 14% 2 29% 1 14% 1 14% UP 2 14% 1 7% 2 14% 2 14% 1 7% 2 14% 3 21% 1 7% WAI 3 21% 1 7%' 1 7% 2 14% 2 14% 5 36% SB 1 7% 1 7% 4 29% 3 21% 1 7% 2 14% 1 7% 2 14% 5 36% 1 7% 2 14% PI 1 13% 3 38% 4 50% SMI 1 5% 3 14% 3 14% 1 5% 14 64% elusions pose two questions. First, do fish from different groups mix extensively? Second, do such groups represent reproductive units or stocks? Our analysis was based on a small set of larval para- sites, all of which are known to be long-lived and do not multiply in the host. Other long-lived parasites such as the myxosporeans have been used in stock discrimination but were not included here because of a lack of abundance data. However, the decreased ability to detect differences because of the small data set was offset by an increased ability to detect the host's past activities. Most of these parasites live for at least several years; therefore, the presence and abundance of these parasites may indicate where the host has been over that time period. At least some of the individuals of each of the parasite species, however, were probably short-term acquisitions (lasting a few years); thus, there may be some bias in the data of the recent past. Our data suggest less extensive movement of Pacific halibut in southern areas. Because parasites are generally more abundant in the south, southern fish may be more easily classified. Nevertheless, if the southern fish mingle extensively with more northern fish, there should be more similarity in the parasite faunas. In particular, central area fish should develop characteristics of southern fish. This did not happen, as is shown by the very low propor- tion of central fish misclassified as southern fish (Table 5). Our information cannot completely rule out the move- ment of southern fish to central areas during the spawn- ing season, and then back to southern areas for the feed- ing season. Their long-lived parasite fauna, having been established in the distinct southern areas, would probably Blaylock et al : Use of parasites in discriminating stocks of Htppoglossus stenolepis Table 4 Cross validation results of a four-category discriminant function classification of adult Pacific halibut in the northeast Pacific based on parasite data. Numbers of fish assigned to a category and the corresponding percentage of the sample in that category are shown. Correct classifica- tions are shown in bold (63% of 240). OR-SQC = Oregon- northern California to southern Queen Charlotte Islands, NQC-KP = northern Queen Charlotte Islands to Kenai Peninsula, KI-SB = Kodiak Island to southern Bering Sea, PI-SMI = Pnbilof Islands to St. Matthew Island. True category OR-SQC NQC-KP KI-SB PI-SMI Assigned category OR-SQC NQC-KP KI-SB PI-SMI 60 79% 3 5% 7 9% 3 4% 28 50% 23 30% 1 3% 7 9% 21 42% 41 53% 6 20% 6 8% 4 7% 7 9%. 23 77% Table 6 Cross validation results of five-category discriminant function classification for juvenile Pacific halibut in the northeast Pacific based on parasite data. Numbers of fish assigned to a category and the corresponding percentage of the sample in that category are shown. Correct classi- fications are shown in bold (44% of 62). NQC = northern Queen Charlotte Islands, NI = Nagai Island, UP = Unimak Pass, and PI = Nunivak Island (central Bering Sea). True category NQC KI NGI UP PI Assigned category NQC KI NI 9 69% 1 14% 5 39% 5 39% 6 38% 4 31% 5 71% 3 23% 4 25% 3 23% 2 15% UP 2 15% 5 39% 1 6%' PI 1 8% 5 31% Table 5 Cross validation results of a three-category discriminant function classification for adult Pacific halibut in the northeast Pacific based on parasite data. Numbers offish assigned to a category and the corresponding percentage of the sample in that category are shown. Correct clas- sifications are shown in bold (83% of 240). OR-SQC = Oregon-northern California to southern Queen Charlotte Islands, NQC-SB = southeast Alaska to southern Bering Sea, PI-SMI = Pribilof Islands to St. Matthew Island. Assigned category True category OR-SQC NQC-SB PI-SMI OR-SQC NQC-SB PI-SMI 63 83% 10 7% 7 9% 112 84% 5 17% 12 9% 25 83% Table 7 Cross validation results of a three-category discriminant function classification for juvenile Pacific halibut in the northeast Pacific based on parasite data. Numbers of fish assigned to a category and the corresponding percentage of the sample in that category are shown. Correct classi- fications are shown in bold (66% of 62). NQC = northern Queen Charlotte Islands, NI = Nagai Island, UP = Unimak Pass, PI = Nunivak Island (central Bering Sea). True category NQC-NI UP PI Assigned category NQC-NI UP PI 30 91% 6 46% 10 63% 6 46% 1 6% 1 8% 5 31% not lose their southern character. Winter sampling could potentially determine if this is the case. With respect to the Bering Sea, we suggest that the majority of the mixing occurs in the southern Bering Sea because classification breaks down when the southern Bering Sea is included in the northern region. This mix- ing is consistent with larval studies that show that larvae enter the Bering Sea through the Aleutian chain. Those fish may not disperse far into the Bering Sea. Rather, they either remain in the southern Bering Sea or migrate back to the Gulf of Alaska area (Skud [1977] believed that both occurred). A migration may explain why fish tagged in the Bering Sea tend to be recovered at greater distances from the tagging site than those tagged elsewhere (Geernaert, 1996). Migrations of the central and northern Bering Sea group appear to be in a more northerly direction (Skud, Fishery Bulletin 101(1) 1977), which would preclude mixing in the Aleutians and the Gulf of Alaska. Zoogeographic analysis with patterns of prevalence showed that Bering Sea parasites are rarely found outside the Bering Sea (Blaylock et al., 1998b). The patterns identified in our analysis agree only in part with zoogeographic analyses (Blaylock et al., 1998b). The southern boundaries in both studies are in the vicinity of the Queen Charlotte Islands, providing additional support for the existence of a southern group of halibut. However, this analysis, unlike the zoogeographic analyses, indicated no sign of a division in the vicinity of Kodiak Island, sug- gesting that the division near Kodiak Island depends on short-lived species not included in this analysis. The evi- dence for the existence of a northern Bering Sea group is equivocal; it was supported by the clustering of localities by using prevalences and, to some degree, the clustering of individuals, but was not supported by any other analyses (Blaylock etal., 1998b). With respect to juveniles, Skud's (1977) analysis clearly indicates compensatory movement from the Gulf of Alaska and southern Bering Sea to southern areas, and, as such, predicts that juveniles should have more similar parasite faunas among areas. Our data show this similarity, but there are significant caveats. First, our samples of juve- niles came from areas that form a single group in the clas- sification of adults. The sample from the northern Queen Charlottes is near the southern boundary of that group, and the sample from Nunivak Island is near the northern boundary. Samples of juveniles from other areas, particu- larly the southern area, should be examined to help clarify this issue. Second, and maybe more important, in these smaller fish, prevalences and intensities are low and per- haps hinder separation. However, because halibut at this stage are susceptible to bycatch in other fisheries (IPHC, 1996), management should probably consider juveniles a mixed stock to prevent impacts on future halibut popula- tions in distant localities. Overall, our analysis provides a less clear picture than that of Arthur and Albert (1993) for Greenland halibut in the northwest Atlantic. Part of the lack of clarity may be due to our use of the training and test set method rather than the bootstrapping method used by Arthur and Albert, which would increase the likelihood of correctly classify- ing similar fish. Also, Arthur and Albert were dealing with a very different system. Geological and oceanographic conditions around the Gulf of St. Lawrence are quite com- plex and create great potential for the isolation of stocks. The northeast Pacific is more open and has fewer isolating mechanisms than the northwest Atlantic. Further, the system is clinal (Blaylock et al., 1998b) and Pacific hali- but are quite capable of migrating along the entire Pacific coast; therefore, less clear cut divisions are expected. Nev- ertheless, wc successfully identified groups of fish, some with a high degree of accuracy. Skud (1977) suggested that juveniles will, as adults, homo to the areas in which they were spawned, making the existence of reproductive stocks at least possible. Modern molecular methods could address the issue. For example, molecular methods could potentially address the existence of separate stocks in the south and in the northern Bering Sea. The limited molecular studies done to date, however, have not elucidated any indentifiable stock structure be- cause of limited sampling localities, the limited number of loci examined, and the use of juveniles only. Tsuyuki et al. (1969) examined a single serum hemoglobin transferrin lo- cus in halibut from ten sites from Vancouver Island to the Bering Sea and found that only one southeast Alaska local- ity was different. Grant et al. (1984) found no differences between Gulf of Alaska and Bering Sea halibut at five loci but were able to distinguish northeast Pacific halibut from Japanese halibut. However, it is important to note that biochemical and genetic information measures differentia- tion at a different time scale than that reflected in parasite data (Lester et al.. 1988). According to Grant (1984), move- ment of only a few Atlantic herring (Clupea harengus) may be sufficient to obscure true differences between different breeding stocks. Thus, even limited gene flow could obscure any differences in the loci examined. Parasite or tagging information alone, however, can not determine whether or not the groups we identified are reproductive stocks. Therefore, all potential factors that might refine the halibut stock concept should be consid- ered. The parasite data suggest a conservative approach to management that recognizes a mixed stock of juveniles and three potential stocks of adults — one in the south, an- other in the northern Bering Sea, and a third and largest centered in the Gulf of Alaska. Acknowledgments We thank the International Pacific Halibut Commission, Seattle, WA, for coordinating sampling and for financial support. Mark Higgins and John Quintero provided invaluable assistance in the laboratory. Tom McDonald and Dave Whitaker provided technical assistance. We also thank Al Shostak and Jeff Lotz for advice and comments. Literature cited Arthur, J. R., and E. Albert. 1993. Use of parasites for separating stocks of Greenland halibut (Reinhardtius hippoglossoides) in the Canadian northwest Atlantic. Can. J. Fish. Aquat. Sci. 50:2175- 2181. Blaylock, R. B., J. C. Holmes, and L. Margolis. 1998a. The parasites of Pacific halibut iHippoglos- siis stenolepis) in the eastern North Pacific; host-level influences. Can. J. Zool. 76:536-547. Blaylock. R. B.. L. Margolis. and J. C. Holmes. 1998b. Zoogeography of the parasites of Pacific halibut [Hippoglossus ulenolepis) in the northeast Pacific. Can. J. Zool. 76:2262-2273. Boje. J., F. Riget, and M. Koie. 1997. Helminth parasites as biological tags in population studies of Greenland halibut tRciuhariltiiia hippoglossoi- dex (Walbauml). in the north-west Atlantic. ICES J. Mar Sci. 54:886-895. Bush. A. O.. K. D. LafTcrty, J. M. Lotz. and A. W. Shostak. 1997. Parasitology meets ecology on its own terms: Margo- lis et al. revisited. J. Parasitol. 83:575-583. Blaylock et a\. Use of parasites in discriminating stocks of Htppoglossus stenolepis Geemaert, T. 1996. Tagging studies. In Report of assessment and re- search activities, 1995, p. 277-288. International Pacific Halibut Commission, Seattle, WA. Gibson, D. I. 1972. Flounder parasites as biological tags. J. Fish Biol. 4:1-9. Grant, W. S. 1984. Biochemical population genetics of Atlantic herring, Clupea harengus. Copeia 1984:357-364. Grant, W. S., D. J. Teel, T. Kokayashi, and C. Schmitt. 1984. Biochemical population genetics of Pacific halibut {Hippoglossus stenolepis) and comparison with Atlantic halibut (H. hippoglossus). Can. J. Fish. Aquat. Sci. 41: 1083-1088. IPHC (International Pacific Halibut Commission). 1996. Report of assessment and research activities, 1995, p. 121-172. IPHC, Seattle. WA. Khan, R. A., M. Dawe. R. Bowering, and R. K. Misra. 1982. Blood protozoa as an aid for separating stocks of Green- land halibut, /ft'(>i/iard/(Hs hippoglossoides. in the northwest Atlantic. Can. J. Fish. Aquat. Sci. 39:1317-1322. Krzykawski, S. and J. Wierzbicka. 1982. An attempt to determine the systematic position of Greenland halibut. Reinhardtius hippoglossoides (Wal- baum. 1792), from Labrador region and Barents Sea on the basis of morphometric, biologic, and parasitological studies. Acta Ichthyol. Piscat. 22:59-75. Lester, R. J. G. 1990. Reappraisal of the use of parasites for fish stock identification. Aust. J. Mar. Freshwater Res. 41:855-864. Lester, R. J. G., K. B. Sewell, A. Barnes, and K. Evans. 1988. Stock discrimination of orange roughy, Hoploslelhus atlanticus, by parasite analysis. Mar Biol. 99:137-143. Moser. M. 1991. Parasites as biological tags. Parasitol. Today 7:183- 186. Skud, B. E. 1977. Drift, migi-ation, and intermingling of Pacific halibut stocks. International Pacific Halibut Commission, Scien- tific Rep. No. 63, 42 p. IPHC, Seattle, WA. Tsuyuki, H., E. Roberts, and E. A. Best. 1969. Serum transferrin systems and the hemoglobins of the Pacific halibut iHippoglossus stenolepis). J. Fish. Res. Board Can. 26:2351-2362. Wilkinson, L., M. Hill, J. Welna, and G. Birkenbeuel. 1992. SYSTAT for Windows: statistics, version 5 edition, 750 p. SYSTAT Inc., Evanston, IL. Williams, H. H., K. MacKenzie, and A. McCarthy. 1992. Parasites as biological indicators of the population biology, migrations, diet, and phylogenetics of fish. Rev. Fish Biol. Fish. 2:144-176. 10 Abstract— Extensive plankton collec- tions were taken during seven Septem- ber cruises (1990-93) along the inner continental shelf of the northcentral Gulf of Mexico ( GOM ). Despite the high productivity and availability of food during these cruises, significant small- scale spatial variability was found in larval growth rates for both Atlantic bumper (Chloroscombrus chrysurus, Carangidae) and vermilion snapper {Rhomboplites aurorubens, Lutjani- dael. The observed variability in larval growth rates was not correlated with changes in water temperature or asso- ciated with conspicuous hydrographic features and suggested the existence of less-recognizable regions where condi- tions for growth vary. Cruise estimates of mortality coefficients (Z) for larval Atlantic bumper (n=32,241 larvae from six cruises) and vermilion snapper (n = 2581 larvae from four cruises) ranged from 0.20 to 0.37 and 0.19 to 0.29, re- spectively. Even in a subtropical cli- mate like the GOM, where larval-stage durations may be as short as two weeks, observed variability in growth rates, particularly when combined with small changes in mortality rates, can cause order-of-magnitude differences in cumulative larval survival. To what extent the observed differences in growth rates at small spatial scales are fine-scale "noise" that ultimately is smoothed by larger-scale processes is not known. Future research is needed to further characterize the small-scale variability in growth rates of larvae, particularly with regard to microzoo- plankton patchiness and the temporal and spatial pattern of potential preda- tors. Small-scale spatial variability in larval growth rates may in fact be the norm, and understanding the implica- tions of this subtle mosaic may help us to better evaluate our ability to partition the causes of recruitment variability. Small-scale spatial and temporal variability in growth and mortality of fish larvae in the subtropical northcentral Gulf of Mexico: implications for assessing recruitment success Bruce H. Comyns Department of Coastal Sciences College of Marine Sciences The University of Southern fVlississippi 703 East Beach Drive Ocean Springs, Mississippi 39566 E mail address bruce.comyns(g)usm.edu Richard F. Shaw Department of Oceanography and Coastal Sciences School of The Coast and Environment Louisiana State University Baton Rouge, Louisiana 70803 Joanne Lyczkowski-Shultz Southeast Fisheries Science Center National Manne Fishenes Service P,0- Drawer 1207 Pascagoula, Mississippi 39568 Manuscript accepted 1 1 July 2002. Fish. Bull. 101(2):10-21 (2003). For many marine fishes year-class strength undergoes large fluctuations because of the inherent variability in larval, postlarval, and juvenile survi- vorship (Hjort, 1914; Gushing, 1975; Lasker, 1975; Hunter, 1982; Houde, 1987; Goshorn and Epifanio, 1991; Pepin and Myers, 1991). Understanding and quantifying recruitment variability remains one of the greatest challenges in fisheries science today (Fritz et al., 1990; Gushing and Horwood, 1994; Leggett and Deblois, 1994; Mertz and Myers, 1995). Early survival rates are influenced not only by predation pres- sure but also by gi-owth rate which can alter the duration of the larval stage when larvae are exposed to accumula- tive high mortality rates (Houde, 1987; Ghambers and Leggett, 1987; Ander- son, 1988; Bailey and Houde, 1989). Pepin (1991) formalized this concept by depicting the cumulative mortality (C) of a population from stage a to older stage b as the direct function of the instantaneous growth ig{x\) and mor- tality (AfUl) rates such that ■mx] g\x] dx, where .v are factors that influence the vital rates (M andg) such as food avail- ability, temperature, and abundance of potential predators. Many questions remain concerning the causes of recruitment variabil- ity. Reasons for variability include the following: the inherent variability in growth and mortality rates and result- ing survivorship; difficulties in estimat- ing mortality rates with sufficient ac- curacy and precision; and the complex interrelationships among factors that affect survivorship of larvae (Parrish, 1973; Laurence, 1979; Houde, 1987; Beyer, 1989; Pepin, 1991 ). Houde ( 1989) hypothesized that cohort survivorship is more sensitive to small changes in vital rates in high latitude systems than in tropical or subtropical systems because the colder temperatures cause slower growth rates and longer larval stage durations, i.e. up to 100 days. Comyns et al.: Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 11 88" / MISSISSIPPI SOUND • 59 •SS ,/ 9 -eo »58 • 57 > • 56 \. • 55'. u • 54 • 53 • 52 51 •46 / • 50 •49 ,47 •AS BRETON SOUND ^ •AS ,•44 • 36 • 37, • 38 • 39 • 4Q • 41,: 7 *4?- •43 J;^ • 34 • 33 . • 32 • 31 • 30 • 29 • 28 ■^n^ T.rH.^^ j^f4 .-^ tlUkm Figure 1 Station locations (•) of plankton collections in the northcentral Gulf of Mexico, September 1990 to 1993. Pepin (1991) questioned this conclusion because he found no net effect of temperature on postlarval stage-specific mortality rates, although his study was based mainly on interspecific variation in mortality (Francis, 1994). The objectives of our study were to determine if growth rates of Atlantic bumper (Chloroscombrus chiysurus, Carangidae) and vermilion snapper (RhombopUtes auro- riibens, Lutjanidae) varied over small spatial scales in the northcentral Gulf of Mexico (GOM); determine the magni- tude and variability of cruise estimates of larval mortal- ity; and determine the potential influence of variability in these vital rates on cohort survivorship in a region where summer water temperatures approach 30°C and larval stage durations are as short as two to three weeks. Vermilion snapper is the most abundant species of snap- per in the northern GOM (Goodyear and Schirripa'), and Atlantic bumper is the most abundant carangid. Materials and methods Sampling location and shipboard procedures Seven, three-day cruises were conducted in inner-shelf waters of east Louisiana, Mississippi, and Alabama during ' Goodyear, C. P., and M. J. Schirripa. 1991. A biological profile for vermilion snapper with a description of the fishery in the Gulf of Mexico. Unpublished report CRD 87/88-16, 53 p. " South- east Fisheries Science Center, National Marine Fisheries Ser- vice, 75 Virginia Beach Drive, Miami FL 33149. September 1990-93 (Fig. 1). Cruise estimates of larval mortality were determined by using data from all cruises during the four-year period. Specimens used for age and growth analyses were collected during 14-16 September 1991 when larvae of both vermilion snapper and Atlantic bumper were abundant. Larvae were collected with a 1 m x 1.4 m Tucker trawl fitted with a 333-pm mesh nitex net and a mechanical flowmeter. Oblique tows were taken from the surface to within a few meters of the bottom and back to the surface at a speed of approximately two knots (1.0 m/s). Samples were concentrated and stored in 95'7f ethanol. At each sampling location surface, midwater and bottom measure- ments of temperature and salinity were obtained with water-bottle casts. Laboratory procedures Lengths of larvae were measured to the nearest 0.1 mm by using a stereomicroscope (12x or 25x) fitted with an ocular micrometer and the larvae were sorted into 0.5-mm size classes. Measurements were taken from the tip of the snout to the end of the notochord in preflexion larvae (notochord length), and from the tip of the snout to the end of the urostyle or hypural plate (whichever was more distal) in flexion or postflexion larvae (standard length). Larval shrinkage was not accounted for because between- station and between-cruise comparisons of growth rates were made with larvae that were preserved in the same concentration of ethanol and stored for approximately the same length of time. Shrinkage of ethanol-preserved 12 Fishery Bulletin 101(1) lan'ae is not large, e.g. to 7% (Theilacker, 1980; Fowler and Smith, 1983; Kruse and Dalley, 1990). It is unlikely that size-related shrinkage effects would have biased our estimates of growth rate because these estimates were based on larvae in similar size classes. Addition- ally, Theilacker (1980) found that preserving northern anchovy larvae after they had died during net capture caused additional shrinkage, but this shrinkage was at a constant rate that was proportional to fish length. Catches of larvae were standardized to account for sampling effort and expressed as number of larvae under 10 m- of sea sur- face. This method of expressing the abundance of larvae more accurately reflects station differences in abundance than a mean density (number/ni'^) when fish larvae are not homogeneously distributed throughout the water column, as has been shown with other species from this area (Lyc- zkowski-Shultz and Steen, 1991), and when sampling (sta- tion) depths are variable, as they were in our study. Dry weights of larvae were determined by rinsing speci- mens with distilled water, drying for 24 h at 60°C, and weighing to the nearest 0.1 pg. Both sagittal otoliths were removed following rehydration for 12 h. Otoliths were mounted convex side up on a glass microscope slide with a drop of Pro-Texx mounting medium and a cover slip. Oto- lith growth increments were counted in the sagittal plane under oil immersion (12.50x). A total of 140 Atlantic bumper larvae and 119 vermilion snapper larvae were selected for age analyses. Specimens were selected from stations where a wide size range of larvae were collected. Daily otolith increment formation has been validated for larval Atlantic bumper (Leffler and Shaw, 1992). Daily increment formation has not been vali- dated for vermilion snapper; however, otolith increments observed in larval vermilion snapper were very similar in width and spacing to validated daily increments found in red snapper from this region (Szedlmayer, 1998; Lycz- kowski-Shultz and Comyns^). Slopes of age-length regres- sions for larval vermilion snapper (n=ll) and red snapper (n=25) collected during July 1992 in our study area were not significantly different, further indicating that vermil- ion snapper, like red snapper, form daily otolith growth increments. Otolith growth increments were counted by using the sagitta (right or left) that provided the most distinct incre- mental zones. Paired /-test analyses showed no significant difference (P<0.05) in diameters of left and right sagittae in both vermilion snapper (/j = ll) and Atlantic bumper (n=20). Daily increments were counted along the longest axis of the otolith from the core to the outer edge. Otoliths were read once by a single reader, and a random subsam- ple of otoliths from vermilion snapper («=30) and Atlantic bumper (n=30) was read a second time to examine within- Lyczkowski-Shultz, J., and B. H. ("omyns. 1992. Karly Hfi- history of snappers in coastal and sholf waters of tho north- central Gulf of Mexico late summer/fall months, 1983-1989, 12 p. + 9 tables. 17 figures. Technical Report submitted to the National Marine Fisheries Service, Southeast Regional Office, 9721 Executive Center Drive North, St. Petersberg, FL. 33702. reader variability. Otolith increment counts differed by one day for only two of the 30 otoliths during the second reading for both species. Data analysis Age-length and age-weight relationships were described by using the exponential equation L or IV = exp(a -i- bt), where, in its linearized form, L = notochord or standard length in mm; W = dry weight in mg; a = y-intercept; b = slopeof regression line (instantaneous growth rate); and t = age of larvae in days. Values of a and h were calculated from the linearized form of the growth equation after the length or weight data were transformed to their natural logarithms. The instantaneous growth rate (b), i.e. the slope of the log- transformed, age-length or age-weight relationship, is also referred to as the growth coefficient. Caution must be exercised when making dry-weight comparisons because of preservation-induced weight loss. Kristoffersen and Salvanes (1998) found that body weight loss was as high as 37-39% in small ethanol-preserved mesopelagic fishes. Dry weight data were used only to determine whether relative changes in weight tracked trends found in age- length relationships. Analysis of covariance (ANCOVA) was used to determine if differences existed among station estimates of instantaneous growth coefficients (Sokal and Rohlf 1969; SigmaStat, 1995). If differences were found ( a=0.05 ), the simultaneous test procedure ( STP; Sokal and Rohlf 1969) was used as an a posteriori test to determine station differences. Cruise estimates of total larval abundance for each size class (catch curves) were developed for Atlantic bumper and vermilion snapper by summing the abundance esti- mates of each size class under 10 m^ of sea surface from each station. Length-frequency distributions were con- verted to age-frequency distributions by assigning ages to mid-points of the 0.5-mm size classes with the age-length relationship previously described. Age-class abundances were corrected for stage duration by dividing the abun- dance estimate of each age class by their respective dura- tions (Houde, 1977). It is necessary to correct for stage du- rations of age classes if growth rates are nonlinear. Stage durations of age classes were determined by assigning ages based on previously determined growth equations to end-points of the 0.5 mm size classes. This customary method for constructing catch curves relies on the rarely examined assumption that larvae at different sampling locations are growing at similar rates. The high r~ values of the age-length relationships (0.92 for Atlantic bumper; 0.84 for vermilion snapper) that resulted when aged lar- vae from all stations were combined indicated that growth Comyns et al : Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 13 rates over the study area were similar enough to justify use of this technique. Cruise estimates of larval mortality rates for Atlantic bumper and vermilion snapper larvae were estimated from catch curve analyses (e.g. Houde, 1977; Essig and Cole, 1986; Watanabe and Lo, 1988; Deegan. 1990; Co- myns et al., 1991 ). The instantaneous mortality rate (Z) was estimated by the slope of the exponential function relating duration-corrected larval abundance and age (Ricker, 1975): D, = D^ exp i-Zt), where D, = total abundance of larvae at time /; Dy= total abundance of individuals at time 0; Z = instantaneous mortality rate; and t = age of size class in days since spawning. Age and abundance of size classes were fitted to this exponential function with a nonlinear least squares routine, and only the descending limb of the regression was used to estimate mortality rates. To reduce poten- tial biases associated with 1) any trend of increasing variance in the length-at-age distribution with increas- ing age, and 2) net avoidance by larger larvae, only Atlantic bumper and vermilion snapper larvae smaller than 6.1 mm and 6.0 mm, respectively, were used to estimate mortality rates. Kolmogorov-Smirnov two- sample tests showed no significant differences (P<0.05) between size-frequency distributions for day versus night catches within this size range for vermilion snap- per and Atlantic bumper larvae. Results 221 0.2 18 0.16 0.14 0.12 g 0.1 n=19 1=26 „.,, n=22 T f T T ~ _L n=23 I I — I — I — I I — I — 12 23 24 34 37 41 42 44 47 B i 0.7- 0.6- 0.5- 0.4- " n= 8 — 25 " 11 n= n= 13 22 n=19 -I- n J_ n=23 n= -8 10 12 23 24 34 37 41 Station 42 44 47 Figure 2 Growth coefficients (horizontal bars) for Atlantic bumper larvae collected at nine stations in the northcentral Gulf of Mexico, during 14-16 September 1991. Vertical lines repre- sent 95% confidence intervals around the growth coefficients, and numbers above bars depict sample sizes; (A) shows age versus In length growth coefficients and (B) shows age versus In dry-weight growth coefficients. Age and growth Atlantic bumper larvae, which were commonly found throughout the study area, ranged from 2 to 14 days old, 1.4 mm to 8.1 mm in length, and 0.003 mg to 1.446 mg in dry weight. Estimates of age versus length growth coefficients were not similar for all stations (ANCOVA; P=0.001). The STP revealed no overlap in 95% confidence intervals around growth coefficients for larvae collected at station 42 and larvae collected at stations 41, 23, and 47 (Figs. 2A and 3). According to their respective growth equations, Atlantic bumper larvae at station 42 grew at approximately 0.43 mm/d and reached a length of 6 mm in approximately 13.3 days. Larvae collected at adjacent sta- tion 41 grew faster, approximately 0.63 mm/d, and reached a length of 6 mm in 10.4 days. Similarly significant differences in station estimates (n=9) of age-dry-weight growth coefficients were also found (ANCOVA; P=0.01), and growth coefficients for larvae collected at station 42 were significantly different from larvae collected at stations 41 and 23 (STP; Fig. 2B). By 11 days, the estimated dry weight of an Atlantic bum- per larva at station 42 was 0.38 mg, whereas at station 41 larvae gained weight faster and the estimated dry weight of an 11-d-old larva was 0.58 mg. Adjacent stations 41 and 42 were 10 km apart, and water temperatures at these two locations were very similar. Surface temperatures varied by only 0.1°C (28.7°-28.8°C), and surface and midwater temperatures varied by only 0.5°C. Daily surface water temperatures recorded at a weather bouy within the study area showed that temperatures varied by less than 2°C during the 31-d period prior to our study. Significant differences in station (n=7) growth rates of vermilion snapper larvae were also found in our 14-16 September 1991 cruise (ANCOVA; P=0.03). Vermilion snapper larvae ranged from 4 to 16 days old, 2.5 mm to 6.5 mm in length, and 0.014 mg to 0.696 mg in dry weight. Growth coefficients for larvae collected at stations 15 and 25 were significantly different (STP; Figs. 4A and 5). Ac- cording to their respective growth equations, vermilion snapper larvae collected at station 15 reached a length of 5 mm in 10.7 days, whereas larvae collected at station 25 grew more slowly and did not reach a length of 5 mm until 12.6 days. Stations 15 and 25 were located 17 km apart on the inner shelf at water depths of 29-30 m. Surface water temperatures at these stations varied by 2.2°C, and both surface and midwater station temperatures differed by less than 2°C. 14 Fishery Bulletin 101(1) 10 ~ I 6 en £ 4 Larvae from station 41 L = exp(-0.052 + 0.1 68t) = 0.95, n= Larvae from station 42 L = exp(0.147 + 0.1 24t; r^= 0.91, n= 23 n r 7 9 Age (days) 11 13 Figure 3 Age versus length data for Atlantic bumper larvae (/! = 140) collected at nine stations in the northcentral Gulf of Mexico during 14-16 September 1991. High- lighted are the age-length relationships at two adjacent stations where growth rates differed. L = notochord or standard length in mm; t = larval age in days. A 0.151 0.13 11 009 0.07 1 05 n 12 n=12 - T T _ — n =5 n=15 - - - — ~ __ n=23 n=16 11 15 16 T 1 r 24 25 30 31 B 0.5- n=12 45- f)=ii 4 - — — n=8 n =5 35- - - - n= 16 n= 15 ^ ^M n=23 ^ ^^ ^^ ^^ 0.3- L - - - 0.2b- 11 15 16 24 Station 25 30 31 Figure 4 (irowth coofTicients (horizontal bars) for vermilion snapper larvae collected at seven stations in the northcentral Gulf of Mexico during 14-16 September 1991. Vortical lines represent 9.')7( confidence intervals around the growth coefficients, and numbers above bars depict sample sizes. (Al shows age versus In length growth coefficients, and iBl shows age versus In dry- weight growth coefficients. Differences in age versus dry-weight growth coefficients were also significantly different (ANCOVA; P=0.03) and once again stations 15 and 25 (Fig. 4B) were significantly different (STP). Vermilion snapper larvae gained weight faster at station 15 where an 11.0-d-old larva had an esti- mated dry weight of 0.28 mg. At station 25 the estimated dry weight of this same larva was only 0.17 mg. Although vermilion snapper larvae were collected at most of the sta- tions within the study area (Fig. 1), abundances were low at shallow (12-14 m depth) stations immediately south of the Mississippi-Alabama coast, and larvae were never collected at stations within Chandeleur Sound. These sta- tions were very shallow (4-9 m). Although our study did not assess microzooplankton prey availability, macrozooplankton dry-weight estimates varied widely over space and time. At the 33 stations east of Chandeleur Sound where larvae of vermilion snapper and Atlantic bumper used in our study were captured, macro- zooplankton dry-weight estimates at 20 stations exceeded 3g/100 m'*, and at eight of those stations values exceeded 5g/100 m'^. Seven days later only at five of the 33 stations were macrozooplankton dry-weight estimates >3g/100 m' and at no station did estimates exceed 5g/100 m-^. Mortality estimates Atlantic bumper was generally the most abundant spe- cies in plankton collections; 32,241 larvae were collected during six cruises conducted in September of 1990, 1991, and 1993. Mortality rates were not estimated for Atlantic bumper larvae collected during the two cruises conducted in September 1992 because abundances of larvae were very low. When station abundance data were pooled for each of the six cruises, size-frequency distributions gener- Comyns et al,: Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 15 8-1 Larvae from station 15 L = exp(0.368 -H oust) ^/ 6 - ^ °''"'' \^^^^ ? E £ 4- C ^^^^^^ o -1 ^^g^^^^^^^^»°°^ Larvae from station 25 ^^^^^^""^"^"^ L = exp(0,512 -(- 0,087t) 2 - n - r^= 0.91, n=23 U 1 1 1 1 1 4 6 8 10 12 14 Age (mm) Figure 5 Age-length data for vermilion snapper larvae in =89 ) collected at seven stations in the northcentral Gulf of Mexico during 14-15 September 1991. L = body length in mm; t = larval age in days. Highlighted are the age-length relation- ships at two stations where growth rates differed. ally exhibited a similar decrease in abundance with suc- cessively larger size classes; however, the smallest size class (1.1-1.6 mm) was the most abundant in only three of the six cruises (Fig. 6, A, B, and F). This under repre- sentation of the smallest size class in several cruises was likely influenced by several potential factors, including a possible decrease in spawning prior to sampling and patchiness of eggs and newly hatched larvae caused by the aggregation of spawning adults. Cruise-estimates of mortality coefficients, which were derived by pooling data from all stations sampled during a cruise and omitting the smallest size class, ranged from 0.20 to 0.37 (Fig. 6). It is likely that mortality rates varied between stations, but as previously mentioned, an average cruise-estimate of mor- tality was determined to ascertain a realistic level about which the effects of small variations in growth rates could be assessed on the cumulative survival of larvae. Standard errors of Z estimates were low, ranging from 0.02 to 0.05. Size-frequency distributions were derived for vermilion snapper lar\'ae (n=2581) taken during two September 1991 cruises, and single late-September cruises in 1992 and 1993 (Fig. 7) when vermilion snapper larvae were abundant. Mortality estimates could not be estimated for five September cruises during the period 1990-93 because relatively few larvae were collected. Larvae collected dur- ing three of the four cruises when they were abundant showed a steady decrease in abundance of successively larger size classes (Fig. 7, A, C, and D). During the fourth cruise (late September 1991; Fig. 7B), the size-frequency distribution showed a distinct peak in abundance of in- termediate-size larvae (4.0-mm size class). Mortality coef- ficients (Z) from the four cruises ranged from 0.19 to 0.30 and standard errors for the mortality coefficients were relatively low ranging from 0.02 to 0.05. Discussion Plankton collections taken in the northcentral GOM during September showed that growth and mortality rates did vary in time and space for Atlantic bumper and vermilion snapper larvae, and that these differences were great enough to significantly impact the cumulative sur- vival of larvae in a subtropical climate where larval-stage durations are short (i.e. two weeks). Growth and mortal- ity estimates of vermilion snapper larvae were previously unknown. Two previous studies of growth and mortality of Atlantic bumper larvae (Leffler and Shaw, 1992; Sanchez- Ramirez and Flores-Coto, 1998) provided no information on variability in growth rates at small spatial scales and no estimates of mortality during the period when our study was conducted. Highly significant between-station differences in growth rates were observed for both Atlantic bumper and vermilion snapper larvae. The largest difference in age versus length growth coefficients for Atlantic bumper larvae was found at adjacent, inner-shelf stations located approximately 10 km apart. According to growth equations, the faster grow- ing larvae grew to a length of 6 mm 2.9 days sooner than larvae at the adjacent station, and differences in larval weight gain as expressed by dry weight of 11-d-old larvae varied by over 30^?^. Water temperatures at these two sta- tions were extremely similar; surface temperatures varied by only 0.1°C, midwater temperatures varied by 0.4°C, and surface and midwater temperatures varied by 0.5°C. It is likely that a similarly small temperature differential was present during the two-week period prior to this cruise, i.e. throughout the life of larvae used in our study because dai- ly surface water temperatures recorded at a weather bouy within the study area during the previous month showed 16 Fishery Bulletin 101(1) 3000 2500 2000 1500 1000 500 Sep. 7-9 1990 49/52 stations n=5,531 Z=0-37 r^=0.97 SE=0.05 2 1 2.6 3,1 3 6 4 1 4.6 5,1 5 6 700 600 500 400 300 200 100 11 16 2.1 2.6 3.1 3.6 4.1 4.6 5.1 5.6 Sep. 14-16 1991 41/50 Stations n=3,611 Z=0.20 r^=0.82 SE=0.03 1400 1200 1000 800 600 400 f 200 J J Sep. 9-11 1993 33/47 stations n =7,282 Z=0.30 1^=0.93 SE=0.04 -fi^JZL 1 1 1 6 2 1 26 3 1 3.6 4,1 4,6 5.1 5 6 Sep. 14-16 1990 49/54 stations n=9,948 Z=0.30 r=0.98 SE=0.02 ^ wn Wl m^ _ 3 1 36 4 1 46 5 1 5 6 Sep. 21-23 1991 45/51 stations n =4,379 Z=0.28 r^=0.97 SE=0.02 B H PI PI PI - 1.1 16 2.1 2,6 3,1 3.6 4.1 4.6 5 1 56 Sep. 19-21 1993 25/32 stations n = 1,490 Z=0.32 r^=0.90 SE=0.05 11 16 2 1 2 6 3.1 36 4 1 46 5 1 5 6 Size class (mm) Figure 6 Size-frequency distributions of Atlantic bumper larvae collected during six cruises in the northcentral Gulf of Mexico during September 1990. 1991, and 1993. Values of Z, SE, and r'^ refer to mortality curves produced from the duration-corrected, age-frequency distributions (omitting the smallest size class). The fraction listed for each cruise refers to positive (larvae were collected) stations/total stations sampled. Abundance of each size class is pooled estimates of station abundances. that temperatures varied by less than 2°C. Significant dif- ferences in both age versus length and age versus weight relationships were also found for vermilion snapper larvae collected at relatively close stations (i.e. 17 km apart). Wa- ter temperatures at these two stations were similar but differed by as much as 2°C. Faster growing larvae reached a length of 5 mm approximately 2 days sooner than larvae growing in nearby areas. Significant differences were also found in larval weight-gain; dry weight of 11-d-old larvae from different stations varyicd by as much as 657f . The variability in growth rates that we observed was likely caused by station differences in food availability and size-selective mortality, and to a lesser degree by wa- ter temperature. Unfortunately our data did not allow us to determine the individual effects of these factors on observed growth rates. At least for Atlantic bumper, the effects of temperature changes were probably minimal. Larval survival is generally more influenced by factors other than temperature. Morse (19891 found a positive cor- relation between length-dependent mortality and surface water temperature for 26 larval fish taxa and attributed this to increased predator consumption rates (caused by increased metabolic rates) at higher temperatures. He also concluded that increased growth due to increases in temperature alone would generally impart no advantage to reduce larval mortality because of the concomitant increased predatory consumption rates. Increased larv-al- stage duration at cooler temperatures is not necessarily associated with increased cumulative larval mortality be- cause predation rates decline with decreasing temperature (Pepin, 1991; FVancis, 1994). Methot ( 1981 ) concluded that after correcting for the effect of temperature on growth rates, the mean growth rate of larval fish is an indicator of the degree to which lar\'al growth, and presumably sur- vival, is food limited. We acknowledge that size-selective predation, i.e. "cull- ing out" the slowest (or fastest! growing lai-vae, could have produced the differences in size-at-age structure among Comyns et a\: Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 17 i 0) 100 - o CO 50 > B Sep, 14-16 1991 24/36 stations 938 larvae Z=0 25 r^=097 SE=0.03 71 5 55 180 160 140 120 100 80 60 40 20 all II Sep 21-23 1991 29/35 stations 975 larvae Z=0.29 71 r-=0.95 SE=0.05 L II 200 . 180 1 ^ 160 H ^ 120^ ^ 100 ^ y 80 '^"' 60 > 40 i ^ 20 J K 25 lis ^j / / D Sep. 27-29 1992 50 29/37 stations 45 458 larvae Z=0-18 "o ^ pj 35 . < r=0.87 SE=0,03 30- :j > 25 /■ K 20 15 10 1 1/1 , g^ Sep. 19-21 1993 11/32 stations 210 larvae Z=0.25 3 5 4 4.5 5 55 25 3 Size class (mm) Figure 7 Size-frequency distributions of vermilion snapper larvae collected during four cruises conducted in the northcentral Gulf of Mexico during September 1991, 1992. and 1993. Estimates of Z, SE, and r- refer to mortality curves produced from duration-corrected age-frequency distributions. The fraction listed for each cruise refers to positive (larvae were collected) stations/total stations sampled. Abundance of each size class was pooled from estimates of station abundances. stations that we observed. However, predation pressure seems unlikely to have been the primary cause of this variability. If among-station variability in size-selective mortality was largely responsible for the differences in larval growth rates, one would expect the variability in size-at-age at each station to be quite variable and this was not the case. Stations where the effects of size-selec- tive mortality were minimal (or less) should have had both fast and slow growing larvae present; yet coefficients of determination (r^) were >0.90 for age versus length regressions at all stations. Furthermore, there was no correlation between observed growth rates and r- values which would be expected if size selective predation was largely responsible for the variability in growth rates that we observed. Many studies have shown that food availability has a large influence on growth rates of larvae (e.g. Houde and Schekter, 1981; Buckley et al., 1987; Pepin, 1991i and it is likely that station differences in food availability influ- enced our observed differences in larval growth rates. We did not collect the small size-fraction of prey eaten by fish larvae, but our data did reveal extensive spatial and tem- poral variability in the abundance of macrozooplankton. Macrozooplankton biomass at station 42, where relatively slow growth of Atlantic bumper occurred, was 2.6 mg/100 m-' whereas at station 41, where larvae were growing faster, macrozooplankton dry weight (3.9 g/100 m') was bWc higher. When all stations were considered, there was no correlation between macrozooplankton dry weight and growth coefficients of lan'ae, but macrozooplankton biomass was certainly very patchily distributed. For most stations there was at least a 509^ difference in macrozoo- plankton dry weight between one of the adjacent sta- tions. It is equally likely that the smaller size fraction of zooplankton that fish lai-vae eat were also very patchily distributed. In addition, several other studies have shown that primary production in the northern GOM is dynamic and spatially heterogenous (Lohrenz et al., 1990, 1994; Re- dalje et al., 1994), although these studies have focused on regions influenced by discharge from the Mississippi and Atchafalaya rivers. In many studies the significant spatial variability in growth rates of field-caught lai-\'ae cannot be explained by changes in water temperature. These reported differences in growth rates have often been associated with factors such as storm events (Lasker, 1975; Maillet and Checkley, 1991). different geographical locations (Mokness, 1992; Nixon and Jones, 1997; Allman and Grimes, 1998;), or distinct hydrographic features such as tidal fronts (Munk, 1993) and riverine discharge plumes (Govoni et al., 1985; DeVries et al., 1990; Lang et al., 1994). All studies in the GOM that have reported spatial differences in larval 18 Fishery Bulletin 101(1) growth rates have involved comparisons in the vicinity of the Mississippi River discharge plume (Govoni et al., 1985; DeVries et al., 1990; Lang et al., 1994; Allman and Grimes, 1998). The observed variability in larval Atlantic bumper and vermilion snapper growth rates reported in our study was not associated with conspicuous hydrographic fea- tures (e.g. hydrographic convergence zones) and suggests the existence of less-recognizable regions where condi- tions for growth vary. Cruise estimates of mortality were determined to as- certain a realistic level about which the effects of small variations in growth rates on the cumulative survival of larvae could be assessed. In order to do this, data from all stations sampled during a cruise were pooled. This provided the most reliable general estimate of mortality for each cruise despite likely site-specific differences in mortality rates that are extremely difficult to measure. Such pooling of data is not unusual; in fact Morse (1989) suggested that samples should be summed over the larval production cycle. Essig and Cole (1986) estimated mortal- ity rates of larval alewives iAlosa pseudoharengus) by us- ing both converted length-frequency distributions, as we did, and actual age-frequency distributions. They found no statistical difference between the two methods. Pepin and Miller ( 1993 ), however, warned that because variability in observed length-at-age increases with larval age (Cham- bers et al., 1988), analyses that use size in older fish to rep- resent age may yield biased estimates of mortality rates. Yet, Pepin and Miller (1993) observed that their mortality rates, which were estimated by using size as a proxy for age, were consistent with mortality rates reported from other environments and species. Ideally, all fish would be aged, but for our study this was not possible because of the large sample sizes, multiple cruises, and the labor-inten- sive nature of otolith preparation for age determination. Atlantic bumper lai-vae were extremely abundant in = 32,241 for six cruises), and cruise estimates of age-fre- quency distributions showed consistent, well-defined de- scending limbs. Estimates of mortality coefficients (Z) for Atlantic bumper larvae were similar for September cruis- es conducted in the same year. For example, in 1990 the two cruise estimates of Z were 0.37 and 0.30, in 1991 the two Z estimates were 0.20 and 0.28, and in 1993 estimates of Z were 0.30 and 0.32. These mortality rates are similar to estimates reported by Leffier and Shaw (1992) during four September cruises in the same area during 1986-87 (Z=0. 17-0.35) and by Sanchez-Ramirez and Flores-Coto (1998) in the southern Gulf (0.15-0.30). In addition, stan- dard errors of the mortality estimates from our study were low, ranging from 0.02 to 0.05. Cruise estimates of mortality rates for vermilion snap- per were determined during four cruises when larvae were relatively abundant («=2581). The descending limbs of three of the size-frequency distributions uniformly spanned all seven size classes, but during one cruise the middle size class was most abundant and the descending limb of this size-frequency distribution was restricted to four size classes. However, mortality rates were quite simi- lar during all cruises (Z=0.19 to 0.30) and each had a low standard error {SE=0.02 to 0.05). Collections of Atlantic bumper and vermilion snapper larvae were taken when water temperatures ranged from 25° to 30°C, and the mortality coefficients estimated from these collections were similar to those reported for other species under similar temperature regimes. Houde (1989) summarized vital rates of six species of larval fish as reported in seven studies where the mid-points of water temperatures at the time of collection ranged from 26° to 28°C. Most of these studies generated a range of mortality estimates, and the mid-points of the ranges reported in six of these studies varied from 0.21 to 0.38, values that are consistent with the mortality estimates (0.19 to 0.39) that we observed. Our primary reason for estimating mortality rates was to ascertain a realistic level about which small variations could be assessed for potential effects on the cumulative survival of larvae, particularly in conjunction with vari- ability in larval growth rates. Our method assumes a con- stant birth rate, or recruitment rate into the population, and assumes that fish leave the population only through death. There is clearly some expected variability in the de- gree to which these assumptions were met; however, based on the similarity of mortality estimates, not only between cruises but also to previously published estimates, it is concluded that our mortality estimates are biologically meaningful. The well-accepted fisheries paradigm holds that changes in year-class strength are determined by variability in mor- tality during early life stages (Sissenwine, 1984; Houde, 1987; Bailey and Houde, 1989; Gushing and Horwood, 1994). Despite extensive efforts to understand the causes of recruitment variability, significant questions remain because the operant factors are likely to be interrelated parts of the ecosystem dynamics that comprise a multidi- mensional system (Ellersten et al., 1995). For example, it is not the mortality or growth rate alone that determines survival during the early life-stages, but the ratio MIG, the stage-specific mortality rate (Pepin, 1991). Examin- ing previously published information, Houde (1989) found an exponential increase in predicted lai-val-stage dura- tion with decreasing water temperature for 26 species of larval fishes and surmised that when temperature is low, small changes in growth rates can induce large changes in larval-stage duration that may significantly affect the recruitment process. To determine the potential effects that variability in vi- tal rates might have on the cumulative survival of larvae, hypothetical numbers of newly hatched Atlantic bumper were projected to a size of 6 mm under the influence of the growth and mortality rates we observed in the study area (Table 1). According to these vital rates, and a hypo- thetical initial cohort size of 1 x lO*" individuals, 124,930 larvae survive to a length of 6 mm under the scenario of relatively fast growth and low mortality (G=0.61 mm/d; Z=0.20). If the growth rate is slowed (G=0.45 mm/d) and it takes approximately three days longer to reach a length of 6 mm, the number of larvae that sui-vive to this length is reduced by 44*7^. If the slower growing lai-vae are exposed to the higher mortality rate (Z=0.37), cumulative survival of larvae decreases by an order of magnitude, and only Comyns et al : Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 19 Table 1 Hypothetical survival of Atlantic bumper larvae to a size of 6 mm un the study area. der the infl uence of growth and mortality rates observed in Initial number in cohort Instantaneous mortality coefficient (per day) Age of 6-mm larva (d) Number of 6-mm larvae 1 X 10« 0.20 10.4 124,930 1 X 106 0.20 13.3 69,948 1 X 106 0.30 10.4 44,157 1 X 106 0.30 13.3 18,499 1 X 106 0.37 10.4 21,322 1 X 106 0.37 13.3 7292 7292 larvae survive to a length of 6 mm. Houde (1987) published a projection on the mortality of larvae exposed to hypothetical levels of mortality and growth rates, but the theoretical exercise used relatively long larval-stage durations (45-56 days). Results of our study show that even in a subtropical climate where larval stage durations may be as short as two weeks, relatively small changes in observed larval growth rates, particularly when com- bined with small differences in mortality, can have a large impact on cumulative larval survival. To what extent the observed differences in growth rates at small spatial scales are fine-scale "noise" that is ultimately smoothed by larger-scale processes is not known. Future research is needed to further characterize the small-scale variability in growth rates of larvae, particularly with regard to mi- crozooplankton patchiness and the temporal and spatial pattern of potential predators. Small-scale spatial vari- ability in larval growth rates may in fact be the norm, and understanding the implications of this subtle mosaic may help us to better evaluate our ability to partition the causes of recruitment variability. Acknowledgments Collections serving as the basis of this research were sup- ported by the SEAMAP program (Southeast Area Moni- toring and Assessment Program) and the NOAA/NMFS MARFIN program (Marine Fisheries Initiative). Several cruises were also conducted by personnel from National Marine Fisheries Service in Pascagoula, Mississippi. Sorting of these plankton samples was made possible by funding provided by the U.S Fish and Wildlife Service through the Wallop Breaux program. This program is administered in Mississippi by the Department of Marine Resources (DMR) whose personnel must be thanked for providing support. Sorting of the plankton samples was made possible by the efforts of several people, including Mae Blake, Cindy Gavins, Pam Bond, Dianne Scott, Ngoc Bui, and Jean Bennett. We also thank Pam Bond for many contributions, including acting as field party leader during cruises, for larval identifications, much of the data entry and management, and for preparing otoliths. We are also grateful to Chet Rakocinski for providing help with statis- tical analyses. Literature cited Allman, R. J., and C. B. Grimes. 1998. Growth and mortality of little tunny (Euthynnus alletteratus) larvae off the Mississippi River plume and Panama City, Florida. Bull. Mar. Sci. 62:189-197. Anderson, J. T. 1988. A review of size dependent survival during pre-recruit stages of fishes in relation to recruitment. J. Northwest Atl. Fish. Sci. 8:55-66. Bailey, K. M., and E. D. Houde. 1989. Predation on eggs and larvae of marine fishes and the recruitment problem. Adv Mar Biol. 25 (ISBN 0-12- 026125-1): 1-83. Beyer, J. 1989. Recruitment stability and survival: simple size-spe- cific theory with examples from the early life dynamics of marine fish. Dana 7:45-147. Buckley, L. J.. T. A. Halavik, A. S. Smigielski, and G. C. Laurence. 1987. Growth and survival of the larvae of three species of temperate marine fishes reared at discrete prey densities. Trans. Am. Fish. Soc. Symp. 2:82-92. Chambers, R. C, and W. C. Leggett. 1987. Size and age at metamorphosis in marine fishes: an analysis of laboratory-reared winter flounder iPseudopleu- ronectes americanum with a review of variation in other species. Can. J. Fish. Aquat. Sci. 44:1936-1947. Chambers, R. C, W. C. Leggett, and J. A. Brown. 1988. Variation in and among life history traits of labora- tory-reared winter flounder iPseudopleuronectes ameri- canus). Mar. Ecol. Prog. Sen 47: 1-15. Comyns, B. H., J. Lyczkowski-Shultz, D. L.Nieland, and C.A.Wilson. 1991. Reproduction of red drum in the north-central Gulf of Mexico: seasonality and spawnor biomass. In Larval fish recruitment and research in the Americas; proceedings of the 13th annual larval fish Conference, Merida, Mexico, 21-26 May, 1989, p. 17-26. U.S. Dep. Commer. NOAA Tech. Rept. MNSS 95. Gushing, D. H. 1975. Marine ecology and fisheries, 278 p. Cambridge Univ. Press, Cambridge. 20 Fishery Bulletin 101(1) Gushing, D. H., and J. W. Hoi-wood. 1994. The gT'owth and death of fish larvae. J. Plank. Res. 16:291-300. Deegan, L. A. 1990. Effects of estuarine environmental conditions on population dynamics of young-of-the-year gulf menhaden. Mar Ecol. Prog. Ser 68:195-205. De Vries. D. A., C. B. Grimes, K. L. Lang, and D. B. Wliite. 1990. Age and growth of king and Spanish mackerel larvae and juveniles from the Gulf of Mexico and U.S. South Atlantic Bight. Environ. Biol. Fish. 29:135-143. Ellersten, B., P. Possum, P Solemdal, and S. Sundby. 1995. The 'critical period' concept — a century of recruitment research. Mar. Ecol. Prog. Ser 128:306-308. Essig,R. J, and C.F.Cole 1986. Methods of estimating larval fish mortality from daily increments in otoliths. Trans. Am. Fish. Soc. 115:34-40. Fowler, G. M., and S. J. Smith. 1983. Length changes in silver hake (Merluccius bilinearis) larvae: effects of formalin, ethanol and freezing. Can. J. Fish. Aquat. Sci. 40:866-870. Francis, M. P. 1994. Duration of larval and spawning periods in Pagrus auritus (Sparidael determined from otolith daily incre- ments. Environ. Biol. Fish. 39:137-152. Fritz, E. S., L. B. Crowder, and R.C. Francis. 1990. The National Oceanic and Atmospheric Administra- tion plan for recruitment fisheries oceanography research. Fisheries 15:25-31. Goshom, D. M., and C. E. Epifanio. 1991. Development, survival, and growth of larval weakfish at different prey abundances. Trans. Am. Fish. Soc. 120: 69.3-700. Govoni, J. J., A. J. Chester, D. E. Hoss, and P. B. Ortner 1985. An observation of episodic feeding and growth of larval Leiostomus xanthurus in the northern Gulf of Mexico. J. Plank. Res. 7:137-146. Hjort, J. 1914. Fluctuations in the great fisheries of northern Europe. Rapp. P.-V. Reun. Cons. Int. Explor Mer 20:1-13. Houde, E. D. 1977. Abundance and potential yield of the round herring, Etrumeus teres, and aspects of its early life history in the eastern Gulf of Mexico. Fish. Bull. 75:61-89. 1987. Fish early life dynamics and recruitment variability. Trans. Am. Fish. Soc. Symp. 2:17-29. 1989. Comparative growth, mortality, and energetics of marine fish larvae: temperature and implied latitudinal effects. Fish. Bull. 87:471-495. Houde, E. D., and R. C. Schecter 1981. Growth rates, rations and cohort consumption of marine fish larvae in relation to prey concentrations. Rapp. R-V. Reun. Con.s. Int. Kxplor Mer 178:441-453. Hunter, J. R. 1982. Feeding ecology and predation of marine fish larvae. In Marine fish larvae — morphology, ecology, and relation to fisheries (R. Lasker, ed.), p. 34-77. Univ. Washington Press, Seattle, WA. KristolTersen, J. B., and A. G. V. Salvanes. 1998. Effects of formaldehyde and ethanol pre.servation on body and otoliths ui Maurolicus MuvUen and Benthosenui glaciate. Sarsia 83:95-102. Kruse, G. H., and E. L. Dalley 1990. Length changes in capelin, Mallotus villosus (Muller), larvae due to preservation in formalin and anhydrous alcohol. J. Fish. Biol. ,36:619-621. Lang, K. L., C. B. Grimes, and R. F Shaw. 1994. Variations in age and growth of yellowfin tuna larvae, Thunnus albacores, collected about the Mississippi River plume. Environ. Biol. Fish 39:259-270. Lasker, R. 1975. Field criteria for survival of anchovy larvae: the rela- tion between inshore chlorophyll maximum layers and suc- cessful first feeding. Fish. Bull. 73:453-462. Laurence, G. C. 1979. Larval length-weight relations for seven species of northwest Atlantic fishes reared in the laboratory. Fish. Bull. 76:890-895. Leffler, D. L., and R. F. Shaw. 1992. Age validation, growth, and mortality of larval Atlan- tic bumper (Carangidae: Chloroscombrus chrysurus) in the northern Gulf of Me.xico. Fish. Bull. 90:711-719. Leggett, W. C, and E. Deblois. 1994. Recruitment in marine fishes: is it regulated by star- vation and predation in the egg and larval stages? Neth. J. Sea Res. 32:119-134. Lohrenz, S. E., M. J. Dagg, and T. E. Whitledge. 1990. Enhanced primary production at the plume/oceanic interface of the Mississippi River. Cent. Shelf Res. 10: 639-664. Lohrenz, S. E., G. L. Fahnenstiel, and D. G. Redalje. 1994. Spatial and temporal variations of photosynthetic parameters in relation to environmental conditions in coastal waters of the northern Gulf of Mexico. Estuaries 17:779-795. Lyczkowski-Shultz, J., and J. P. Steen Jr 1991. Diel vertical distribution of red drum Sciaenops ocel- latus larvae in the northcentral Gulf of Mexico. Fish. Bull. 89:631-641. Maillet, G., and D. M. Checkley Jr 1991. Storm-related variation in the growth rate of otoliths of larval Atlantic menhaden Brevoortia tyrannus: a time series analysis of biological and physical variables and implications for larva growth and mortality. Mar Ecol. Prog. Ser 79:1-16. Mertz, G., and R. A. Myers. 1995. Estimating the predictability of recruitment. Fish. Bull. 93:657-665. Methot, R. D., Jr 1981. Spatial covariation of daily growth rates of larval northern anchovy, Engraulis mordax. and northern lamp- fish, Stenobrachius leucopsarus. Rapp. P.-V. Reun. Cons. Int. Explon Mer 178:424-431. Morse, W. W. 1989. Catchability, growth, and mortality of lai-val fishes. Fish. Bull. 87:417-446. Mokness, E. 1992. Differences in otolith microstrueture and body growth rate of North Sea herring [Cliipea harengus L.) larvae in the period 1987-1989. ICES J. Mar Sci., 49:223-230. Munk, P 1993. Differential growth of larval sprat Sprattus spratlus across a tidal front in the eastern North Sea. Mar Ecol. Prog. Ser 99:17-27. Nixon, S. W., and C. M. Jones. 1997. Age and growth of larval and juvenile Atlantic croaker, Micropogimias undulatus. from the Middle Atlan- tic Bight and estuarine waters of Virginia. Fish. Bull. 95: 773-784. Parrish, B. B. 1973. Foreward, fish stocks and recruitment. Rapp. P.-V. Reun Cons. Int. Explor Mer 164:1-3. Comyns et al.: Spatial and temporal variability in growth and mortality of fish larvae in the Gulf of Mexico 21 Pepin, P. 1991. Effect of temperature and size on development, mor- tality, and survival rates of the pelagic early life history stages of marine fish. Can. J. Fish. Aquat. Sci. 48:503-518 Pepin, P., and R. A. Myers. 1991. Significance of egg and larval size to recruitment variability of temperate marine fish. Can. J. Fish. Aquat. Sci. 48:1820-1828. Pepin, P, and T.J. Miller. 1993. Potential use and abuse of general empirical models of early life history processes in fish. Can. J. Fish. Aquat. Sci. 50:1343-1345. Redalje, D. G., S. E. Lohrenz, and G. L. Fahnensteil. 1994. The relationship between primary production and the vertical export of particulate organic matter in a river impacted coastal ecosystem. Estuaries 17:829-838. Ricker, W. E. 1975. Computation and interpretation of biological statistics offish populations. Bull. Fish. Res. Board Can. 191:382. Sanchez-Ramirez, M., and C. Flores-Coto. 1998. Growth and mortality of larval Atlantic bumper Chlo- roscombrus chrysurus (Pisces: Carangidae) in the southern Gulf of Mexico. Bull. Mar. Sci. 63: 295-303. SigmaStat. 1995. Statistical software version 2.0 for Windows 95, NT and 3.1. Jandel Scientific, San Rafael, CA. Sissenwine, M. P. 1984. Why do fish populations vary? In Exploitation of marine communities (R. M. May, ed.), p. 59-94. Springer- Verlag, Berlin. Sokal, R. R.,andFJ. Rohlf 1969. Biometry. 776 p. W.H. Freeman and Company, San Francisco, CA. Szedlmayer, S. T. 1998. Comparison of growth rate and formation of otolith increments in age-0 red snapper. J. Fish. Biol. 53:58-65. Theilacker, G. A. 1980. Changes in body measurements of larval northern anchovy, Engraitlis mordax, and other fishes due to han- dling and preservation. Fish. Bull. 78:685-692. Watanabe, Y, and N. C. H. Lo. 1988. Larval production and mortality of Pacific saury, Cololabis saira, in the northwestern Pacific ocean. Fish. Bull.78:601-613. 22 Abstract— Fecundity (F, number of bixHwied eggs ) and egg size were estimated for Hawaiian spiny lobster iPanulirus margirmtus) at Necker Bank, North- western Hawaiian Islands (NWHI), in June 1999, and compared with previous (1978-81, 1991) estimates. Fecundity in 1999 was best described by the power equations F = 7.995 CL ^■""", where CL is carapace length in mm (f-^'=0.900), and F = 5.174 TW ^ ^ss^ ^here TW is tail width in mm (r2=0.889) (both n=40; P< 0.001). Based on a log-linear model ANCOVA, size-specific fecundity in 1999 was 18% greater than in 1991, which in turn was 16% greater than during 1978-81. The additional increase in size- specific fecundity observed in 1999 is interpreted as evidence for flirther com- pensatory response to decreased lobster densities and increased per capita food resources that have resulted either from natural cyclic declines in productivity, high levels of harvest by the commercial lobster trap fishery, or both. The density decline is well-documented by a fivefold decrease in commercial catch-per-trap- haul (CPUE) during the late 1980s to early 1990s and by a similar decrease m research CPUE for all-sized (includ- ing juvenile) P. marginatus through the 1990s. Fecundity increases are consis- tent with decreases in median body size at sexual maturity, first described from comparisons of 1977-81 and 1986-87 specimens and consistently observed thereafter during the 1990s. Egg size covaried with fecundity; in 1999, indi- vidual eggs within broods had a 11% greater mass (15% greater volume) than eggs brooded in 1991. Implications of these obsei-vations are discussed in rela- tion to possible future management mea- sures for a commercial lobster fishery in the NWHI. More generally, our findings argue for the need to routinely reevalu- ate compensatory responses in e.xploited stocks of lobsters and other resources. Temporal changes in population density, fecundity, and egg size of the Hawaiian spiny lobster iPanulirus marginatus) at Necker Bank, Northwestern Hawaiian Islands Edward E. DeMartini Gerard T. DiNardo Happy A. Williams Honolulu Laboratoi7 Southwest Fisheries Science Center National Manne Fishenes Service 2570 Dole Street Honolulu, Hawaii 96822-2396 E-mail address (for E. E DeMartini): Edward DeMartini@nooa gov Manuscript accepted 20 September 2002. Fish. Bull. 101:22-31 (2003). The endemic Hawaiian spiny lobster (Panulirus marginatus) has been the principal target of the Northwestern Hawaiian Island (NWHI) commercial trap fishery since the mid- to late 1970s (Uchida and Tagami, 1984; Polovina, 2000). Landings and exvessel (whole- sale) value have fluctuated greatly over the years, in part because of annual variations in trapping effort and a 1-yr fishery closure in 1993, but have been generally lower during the 1990s because of declines in oceanic produc- tivity and recruitment and increased exploitation (Polovina and Moffitt, 1995; Polovina et al., 1995). The fishery was closed in 2000 because of increasing uncertainty in the population models used to assess stock status. In Decem- ber 2000 President Clinton, through Executive Order (EO) 13178 and later through EO 13196, established the Northwestern Hawaiian Islands Coral Reef Ecosystem Reserve which may prohibit commercial lobster fishing in the NWHI for at least 10 years. Ajinual research surveys of the National Marine Fisheries Service (NMFS), Honolulu Laboratory, have demonstrated a decline (Fig. 1) in spiny lobster density (CPUE, catch-per-trap-haul) at Necker Bank, NWHI, one of the sites at which spiny lobsters have been consistently targeted since about the mid-1970s. Polovina (1989) first described a den- sity-dependent decrease in median body size at sexual maturity and an increase in asymptotic body size for spiny lobster at Necker Bank, based on a contrast between specimens collected during 1977-81 and 1986-87. DeMartini et al. (1993) observed an increase in size-spe- cific fecundity for specimens collected in 1991, used to further characterize the Necker Bank population's status after- exploitation. Compensatory increases in juvenile growth and survival and in- creases in size at maturity as responses to decreased density following increased fishery exploitation have been observed for other spiny lobster stocks (e.g. see Pollock, 1995a, 1995b). In this article, our objectives were to estimate recent (1999) berried female fecundity and egg size for the Hawai- ian spiny lobster at Necker Bank and to relate these to prior, analogous esti- mates for lobsters collected in 1991 and 1978-81, analyzed by DeMartini et al. (1993). We then use the 1999 fecundity estimates and 1999 commercial catch data to characterize recent egg produc- tion by the Necker Bank population. We conclude with a brief discussion of the management implications of com- pensatory reproductive responses by the population. Methods and materials All specimens used in this study were trapped from Necker Bank surround- ing Necker Island (23°34'N, 164°42'W), NWHI, during the species' mid-spring to mid-summer peak period of egg brooding at mid-archipelago latitudes (Uchida and Tagami, 1984). Specimens DeMartini et al.: Population density, fecundity, and egg size of Panulirus marginatus 23 for 1999 were collected on a cruise of the NOAA ship Townsend Cromwell. Details of specimen collection and processing of the 1978-81 and 1991 samples are described by DeMartini et al. (1993). The 1999 samples were collected during 9-22 June 1999 from the bank terrace at a median 27-m depth by using molded plastic ("Fathom Plus") traps baited with 1 kg of mackerel (Scomber Japonicus) and fished with a standard (over- night) soak. Shipboard processing Specimens were processed identically to those collected in June 1991. All specimens were processed alive within minutes of trap retrieval. Both carapace length (CL: defined as the straight line distance between the anterior edge of the supraorbital ridge and the posterior edge of the carapace along the dorsal midline) and tail width (TW: defined as the straight line distance across the abdo- men at the widest spot between the first and second abdominal segments ) of each specimen were measured to 0.1 mm with dial calipers. TW is the present metric of choice for lobster management in the NWHI trap fishery. CL was the metric used to characterize body size in many prior research and management studies of the species, and its measurement was needed for comparison with results of studies made prior to the mid-1980s. Ber- ried (ovigerous) females were scored for egg developmental stage by using a gross visual proxy (brooded eggs noted as either orange or brown in color to the unaided eye). Berried specimens were individually flash-frozen for laboratory evaluation ashore. Laboratory analyses 700 ^ 600 o o 500 300 - 100 - - 1982 1986 1988 1998 2000 B Commercial CPUE ^TV / \ A: / » * I V \ \ \ %s^ 1 1 1 1 1 1982 1984 1986 1988 1990 1992 Year 1994 1996 1998 2000 Fecundity, here defined in the limited sense of a single brooded egg mass (see Chubb, 2000), was estimated for 5-10 females per 5-mm TW class in order to provide at least 40 total specimens spanning the entire size range for analyses. Except for sample sizes, procedures were identical to those used for the 1991 collection. Only females bearing orange egg clusters with embryos lacking visible melanin pigment (early embryonic development) were considered in order to minimize the probability of physical damage, egg loss, and fecundity underestima- tion during capture and handling, which is an apparent problem only for broods of heavily pigmented (brown), late- development eggs with soft capsules (DeMartini, unpubl. data). Frozen specimens were thawed overnight at 3°C. All four pairs of egg-bearing pleopods were then removed from the abdomen, gently blotted (damp-dry) on a paper towel, and weighed individually to 0.1 mg on an electric Figure 1 Time series plots of (A) the Northwestern Hawaiian Islands (NWHI) commercial trap catch and landings of Panulirus marginatus (no. of lobsters x 1000) and effort (no. of trap-hauls x 1000); and (B) total P. marginatus catch-per-trap-haul (CPUE) at Necker Bank, NWHI, during the 1983-99 commercial fishing seasons and as assessed on 1988-1999 lobster research cruises. (Research CPUE data are lacking for years prior to 1988.) Dashed lines framing the research CPUE curve in B represent bootstrapped 95'7f confidence intervals (DiNardo et al.-). microbalance. Eggs were then carefully teased off pleopod setae with jeweler's forceps and stored after being wrapped in cool, damp paper towels to minimize evaporative weight loss. Individual pleopods were then rew-eighed and the weight of each pleopod's egg complement was calculated by difference. Three subsamples of 0.1-0.2 g, each compris- ing about 700-1000 eggs total (about 100 eggs per pleopod, pooled over all 8 pleopods), were next weighed to 0.1 mg, their component eggs counted, and relative fecundity (RF, number of eggs per gram of brooded eggs) was calculated as a simple ratio, with the three subsamples used to calcu- late a mean and standard error of RF. Fecundity (F, defined 24 Fishery Bulletin 101(1) Table 1 (A) Summary catch statistics {Townseiid Cromwell research cruise, Necker Bank, June 19991 and (B-D) fundamental linear-mass interrelationships for body and egg sizes of female Hawaiian spiny lobster (Pa/iulirus marginatus) at Necker Bank, Northwestern Hawaiian Islands. A Female catch statistics Total Berried TW„ TW,, median range median range 834 54.6 350 42.0 50.1 24-72 51.1 38-72 B Relation of tail width to carapace length and vice versa; model: Y - aX + b TW = 0.6087 CL + AAA and CL = 1.5772 TW - 4.00, where TW = tail width in mm, CL = carapace length in mm, and a and h are fitted constants. [H-'=0.963, ?i=825,P<0.0011 C Relation of body weight to carapace length; model: Y = aX'' SW = 0.00090 CL 2 9^=2 [range: 51.9-114.7 mm CL,n= 197,P<0.0011 where BW = total body weight in g, and CL = carapace length in mm, for unbemed females (source: Uchida and Tagami [1984] ). D Relation of egg weight to egg diameter; model: Y = aX'' EW = 0.3985 ED2 2472 where EW = egg weight in 0.001 mg, and ED =egg diameter in mm. lr-'=0.833,;!=40, P<0.001] as the total number of pleopod-brooded eggs) was calcu- lated as the product of mean RF and total weight of the brooded egg mass. Pilot tests indicated that this procedure estimated F with coefficients of variation (CV, SD/mean x 100%) consistently <5'7f. Subsamples of 25 eggs were ran- domly taken from each female's total egg complement and the diameter of individual eggs were measured (random axis) at 500x magnification by using a dissecting micro- scope and an optical micrometer. Average individual egg weight was also independently derived as the ratio of the weight to numbers of eggs present in the parent sample. Statistical analyses Relations of female body size to fecundity and body size to egg size were evaluated for the 1999 samples by using both linear and nonlinear least squares procedures (proc REG, proc NLIN) of PC SAS for Windows v 6.12 (SAS Institute, 1990a, 1990b). Analysis of covariance (ANCOVA; proc GLM; SAS Institute, 1990c; Chubb, 2000) was used to compare size-specific fecundity estimates of Necker Bank P. ryjarginaliis among the three exploitation periods: 1978-81, 1991, and 1999. Subseasonal variation within spawning seasons was controlled by the aforementioned restriction on month of specimen collection, and single collections were assumed to provide accurate character- izations within exploitation periods. Fecundity data used to characterize the 1978-81 and 1991 periods at Necker Bank are listed in Appendix A of DeMartini et al. (1993). Analogous comparisons of size-specific egg sizes were lim- ited to the 1991 and 1999 periods because no data on vari- ance of egg sizes were available for tiic 1978-81 samples (DeMartini et al., 1993). Body-size-fecundity relations were allometric, hence log-linear (see Somers, 1991); natu- ral logarithms were used for ANCOVAs and regressions of log-linear relations. An index of reproductive potential (IRP; Kanciruk and Herrnkind, 1976) was computed for the 1999 specimens in order to determine the size classes of females that con- tributed most to population egg production. The IRP was constructed by using data for female P. marginatus caught by the commercial fishery at Necker Bank during 1999, collected by several contracted fishery observers. Results Fecundity and egg size of lobsters in 1999 Fecundity A number of the female Panulirus marginatus trapped on the research cruise at Necker Bank during June 1999 were berried (Table 1). The estimated fecundity of 40 females, spannmg 54.3 to 105.4 mm CL (39.3-67.4 mm TW, see CL-TW relation; Table 1), ranged more than fivefold, from 109,865 to 590,530 eggs (Appendix A). Fecundity was positively and nonlinearly related to TW (Fig. 2) and CL and best described by the power equations F = a TW' and F = a CL'', respectively, as F= 5.1743 rW''^ 7580^ |7-=0.889| and F= 7.9952 CL-^-^'i' lr2=0.900. both n=40. P<0.001. DeMartini et al.: Population density, fecundity, and egg size of Panulirus marginatus 25 600 r • • / • / 500 y 1 / • (numbers of eggs o o o o ••/ * • 1 LL 200 • y^ ^ • 100 >* • 1 1 1 1 1 1 35 40 45 50 55 60 65 70 Tail width (mm) Figure 2 Scatterplot and fitted power curve describing the relation between fecundity (no. of brooded eggsxlOOO) and tail width (TW, in mm) for Panulirus marginatus collected at Necker Bank. 1999. 0.80 • • • 075 • • • • • E, 1 0.70 •D O) UJ , , • • • • . •• • • • • • 065 . • • • • • • 060 • • t 1 1 1 1 35 40 45 50 55 60 65 70 Tail width (mm) Figure 3 Scatterplot describing the relation between egg diameter (mm) and tail width (TW, in mm) for Panulirus marginatus collected at Necker Bank. 1999. The standard errors of 6 were 0.1787 and 0.1472, respec- tively. A log-linear fit of the F-CL data (LaF=2.5533 LnCL -hi. 3977) was nominally inferior (r-=0.886) to the curvilin- ear fit but was required for the general linear model used in the ANCOVA comparisons that follow below. Fecundity subsamples averaged 1.5 ±0.14(SE)'7f of total brooded egg mass weight. Brooded egg masses weighed an average 51.4 g and ranged from 15.8 to 109.2 g. CVs of the three replicate estimates of RF averaged 1.2 ±0.1V'/(. Mean RF was 5882 ±160(SE) eggs per g of brooded eggs and ranged nearly twofold from 4030 to 7930 eggs per g of eggs among the 40 females. Based on the aforedescribed nonlinear best fit, the fecundity of the median-size (53.8 mm TW, 80.9 mm CD female caught at Necker Bank in 1999 by the com- mercial trap fishery was an estimated 306,400 ±90,200 (95% CI) eggs. Egg size The mean (±SE) diameter of early-stage eggs carried by the 40 berried females was 0.69 ±0.007 mm, with a range of 0.61 to 0.79 mm. Analogous median (25th, 75th percentile) diameters were 0.70 (0.65, 0.72) mm. The corresponding egg weights were 0.17 ±0.005 and 0.18 (0.15, 0.20) mg (range: 0.13-0.25 mg). Individual egg size (diameter: Fig. 3; weight) was unrelated (both P>0.63) to female body size (TW). Individual egg weight was a power function of egg diameter (Table 1). Temporal comparisons of fecundity and egg size Fecundity Size-specific fecundities differed among the three periods, and body-size-adjusted means differed for each period (Table 2, Fig. 4). Size-adjusted mean fecun- dity in 1999 was 18% greater than in 1991, which in turn was 16% greater than during 1978-81. Lobster in 1999 thus exhibited a cumulative 36% increase in size-specific fecundity over that described for lobster collected during 1978-81. The statistical power ( 1 minus /3, where ji is the probability of making a type-II error) to detect an effect size equal in magnitude to the changes observed between 1978-81, 1991, and 1999 was estimated as >97% at a = 0.05. Egg size Brooded eggs on average were about 5% greater in diameter (equivalent to 15% greater volume assuming the volume of a sphere, V=4/3 nr') and were 11%^ heavier in 1999 compared to 1991 (Table 3). The precision of our 26 Fishery Bulletin 101(1) Table 2 (A) ANCOVA and (B) component least squares regression statistics for log-linear (LnV = Ina + 6LnX) relations of fecundity (F) to carapace length {CD for P. margmatus caught at Necker Bank during three periods: 1978-81, 1991, and 1999. Natural logs are used throughout. Under- lines illustrate that the least-square means for each period differ from one another. MSE = mean square error A ANCOVA model: LaF = = LnCL -1- period (r-'= 0.855; root MSE= =0.1843) Factor df MS F P Model 3 6.82 199.6 0.0001 LnCL 1 17.16 502.1 0.0001 Period 2 0.87 25.5 0.0001 LnCL X period (P=0.27 — ns; not included in final model ) Error 105 0.03 Total 108 12.426 > 12.284 > 12.119 1999 > 1991 > 1978-81 B Regression models: LnY = Lna -i- 6LnA' 1978-81 LnF= 2.7994 + 2. 1569 LnCL, r2=0.708, /!=35, P<0.001, SEa=1.0367, SE 6=0.2412 1991 LnF = 1.5859 + 2.4778 LnCL, r2=0.881, n=34, P<0.001, SE a=0.6934, SE 6=0.1607 1 999 LoF = 1 .3977 -i- 2.5533 LnCL, r-'=0.886, /I =40, P<0.001, SE a=0.6452, SE 6=0.1485 measurements was sufficient for the observed change in egg diameter to have had an 87'* chance of being detected at a = 0.05. Individual and population egg production Based on the IRP of Kanciruk and lierrnkind (1976), most egg production by the Necker Bank population of P. margincitus in 1999 was by small adults (<60 mm TW) that now dominate the population (Table 4). Large adults (>60 mm TW), although highly fecund, are now too rare to contribute substantially to total population egg produc- tion (Table 4). Table 3 (A) ANCOVA and (B) component least squares regression statistics for linear relations of egg diameter (ED, in eye- piece units, X 0.020=mm) to carapace length (CL, mm) of P. marginatus caught on research cruises to Necker Bank during 1991 and 1999. See Table 2 caption for additional details. A ANCOVA model: ED = CL + period (r2=0.142; root MSE=2. 193) Factor df MS F P Model 2 27.36 5.69 0.005 CL 1 3.33 0.69 0.41 Period 1 46.85 9.74 0.003 CL X period iP = 0.81 — ns; not included in final model) Error 69 4.81 Total 71 0.691 > 0.658 1999 > 1991 B Regression models: ED = a + b CL 1991 i5:L> = 31.1646 -I- 0.0230 CL, r2=0.021. n=32. P<0.001, SEa=2.1224, SE 6=0.0282 1999 ED = 33.5760 -i- 0.0128 CL, r2=0.006. ?!=40, P<0.001, SEa=2.1768, SE 6=0.0274 Discussion Size-speciFic fecundity and egg size Fecundity The initial 16% increase in body size-specific fecundity between 1978-81 and 1991 occurred while commercial CPUE decreased fivefold. Unlike commercial data, research CPUE data were collected at fixed sta- tions (including juvenile habitat), were uninfluenced by increased catchability (the targeting of larger adult lobster in more productive habitats by commercial fishermen ), and continued to show a decline of similar magnitude during the 1991-99 period when size-specific fecundity increased an additional 18''i (Fig. IB). Thus both observed fecundity responses occurred simultaneously with declining lobster densities. The cumulative 36% increase in size-specific fecundity observed for Necker Bank P. marginatus over a >20-yr period of exploitation is not unreasonable given the evidence for concurrent, compensatory declines in body size at sexual maturity in this population (Polovina, DeMartini et al : Population density, fecundity, and egg size of Panulirus marginatus 27 1989; DeMartini et al.^). Density-dependent changes in somatic growth, survival rates, and body sizes at sexual maturity have been described for numerous other palinurid spe- cies (Pollock, 1995a, 1995b). At least one case study provides further evidence for reproduc- tive compensation. Chittleborough (1979) doc- umented a decreased interval between broods as a response to increased exploitation in the Western Australian rock lobster, P. cygnus. Prior to the present study, the study of De- Martini et al. (1993) was the only published record of changes in size-specific fecundity in a spiny lobster, perhaps attributable to density declines resulting from exploitation, although perhaps only reflecting natural interannual variation independent of fishing (Pollock, 1995b). The fecundity update for 1999 in this ar- ticle further supports DeMartini et al.'s (1993) original interpretation of an increase in size- specific fecundity as a density-dependent re- sponse at lower population densities. Other data on body size at sexual maturity for the period 1988-99, to be reported elsewhere (De- Martini et al.'), extend the temporal pattern of smaller body size at sexual maturity first documented during 1986-87, after 10 years of exploitation, by Polovina (1989). The observed decrease in size at maturity could have been caused by slower growth (Pollock, 1995a) re- sulting from lower levels of oceanic productivity (Polovina et al., 1994, 1995). However, if smaller size at maturity has been a proximal response to decreased rather than in- creased per capita food availability, it is inconsistent with the simultaneous increases in size-specific fecundity and egg size which have occurred. Evidence for changes in the nutritional status of P marginatus at Necker Bank dur- ing 1991-95 is equivocal (Parrish and Martinelli-Liedtke, 1999). Resolution of whether the lower densities of spiny lobsters at Necker Bank have resulted from natural de- clines in productivity, increased fishery exploitation (or both) would require comparative evaluations for lobsters collected from fished as well as unfished control areas at the bank; unfortunately, as of 1999 unfished lobster habi- tat at Necker Bank does not exist. The observed increase in size-specific reproductive output of P. marginatus probably has been a phenotypic response to lower densities and higher per capita food availabilities at Necker Bank. It is unlikely, given the 3-yr generation time of P. marginatus (Uchida and Tagami, 1984) and relatively short (20-1- yr) period over which the responses have occurred, that a genetic, rather than phenotypic, dynamic has been involved. More extensive comparisons of the egg productions of P. marginatus popu- lations among Necker and other NWHI banks differing in natural and fishery-induced densities would be necessary • 1999 ^ O 1991 ® 1978-81 J^ 13 - 73 7 2 0.11 0.09 782 0.01 0.08 242.3 Total 6226 2340 100.00 100.00 100.00 for all-sized females has been observed in two Caribbean species, Panulirus inflatus (Gracia, 1985) and P. argiis (Fonseca-Larios and Briones-Fourzan, 1998). Briones- Fourzan and Contreras-Ortiz ( 1999), however, could detect no difference in egg size among P. guttatus sampled during three consecutive years. Mean egg size declined within the spawning season, but this decline reflected a loss in mass caused by embryonic development within individual eggs rather than the production of smaller eggs later in the breeding season (Briones-Fourzan and Contreras-Ortiz, DeMartini et al.; Population density, fecundity, and egg size of Panulirus marginatus 29 1999). Pollock {1995c) noted that P. guttatiis produces unusually few, but large eggs for a shallow-water tropical palinurid. Using the CL-to-body-weight regression listed in Table 1, we estimated an inverse index of egg size (Pollock, 1997) for Necker Bank P. marginatus in 1999 that was 660 eggs per g total body weight. Such small eggs are typical within the derived lineage of shallow-water, subtropical and tropical members of the genus (Pollock, 1997). We could find no other studies documenting changes in egg size as a response to density fluctuation in palinurid lobsters. Prior to the mid-1990s, information on temporal and size-related patterns of fecundity and egg size were largely restricted to cold- and warm-temperate members of the genus Jasi/s and Panulinif; (Pollock, 1995c, 1997). Per- haps egg size, like size-specific fecundity, is phenotypically labile in tropical reef species of the genus Panulirus for which high and variable predation pressure makes such plastic responses adaptive. More research on size-specific, individual reproductive output is needed for P. marginatus and other tropical reef species of spiny lobsters. It is unknown whether egg size lability in P. marginatus has a genetic or environmental basis. One could perhaps evaluate this for individual females by repetitively mea- suring egg subsamples from successively brooded egg masses of berried tagged and recaptured females. Fixed but differing egg sizes among individual females would be consistent with a genetic basis. On the other hand, changes in the size of eggs produced by the same individual female in successive broods would suggest that environmental factors are involved. Management implications One of our observations has major relevance to the man- agement of P. marginatus in the NWHI lobster fishery. Based on the IRP of Kanciruk and Herrnkind (1976), egg production by the Necker Bank population of P. margin- atus was dominated by the 50-57 mm TW classes in 1999, which together contributed >43'7f to population egg pro- duction (Table 4). Even though each large (>60 mm TWi individual produces a disproportionately great number of eggs, large females are now so poorly represented in the population that they no longer drive population egg production (Table 4). The eggs produced by smaller (50- 57 mm TW) females are more important to the population now than before exploitation. In 1996 a "retain all" size policy was established for the commercial fishery, replac- ing a 50-mm-TW minimum size limit used previously, in part because of the high mortality of discarded lobsters (DiNardo et al., 2002). If a commercial lobster fishery with a minimum size limit were to be reinstated in the NWHI, a minimum size larger than the previous (50 mm TW) should be considered. Our findings on the size distribution of population egg production indicate that smaller adult females, which now produce most of the population's eggs, should be further protected, perhaps by using larger escape vents in traps. Doing so would increase total population egg production and might assist in countering recruitment overfishing (Botsford, 1991; Pollock, 1993). Panulirus mar- ginatus production at Necker Bank historically and pres- ently dominates archipelago-wide production by the spe- cies; this production is supported by empirical catch data (DiNardo et al.'^) as well as modeling of its recruitment dynamics (Polovina et al., 1999). Augmenting egg produc- tion by the Necker Bank population might significantly bolster stock-wide productivity. The body size distribution of egg production by P. marginatus at other NWHI banks is presently unknown, however, and egg production by large females elsewhere possibly could partly offset the deficit in production at Necker. Our observations on the size distribution of egg production at Necker Bank none- theless merit important consideration for setting size limits for spiny lobster management. By necessity we calculated the IRP assuming that all size classes produced the same (single) brood per spawn- ing period because data on size-specific spawning frequen- cy were lacking. We caution that, if females >60 mm TW (whose size-specific egg production is greatest) produce broods more frequently than smaller females (Lipcius, 1985), we have proportionately underestimated the con- tribution of larger females to population egg production. Individual Panulirus marginatus of all sizes likely produce multiple broods per individual spawning season, based on the protracted period during which females are berried (Uchida and Tagami, 1984; Polovina and Moffitt, 1995) and the occasional presence of new, intact (unused) spermatophore plates on spent females (unpubl. data, Honolulu Laboratory, NMFS). (The latter observation in fact suggests that Necker Bank P. marginatus can pro- duce more than one brood per molt [like P. argus; Sutcliffe, 1953].) There are no time-series growth-rate data avail- able with which to evaluate whether females of a given body size might now be producing larger broods at more frequent intervals than previously. If females are now growing faster, it is likely that the rates of both molting and brood production are now greater Accurate estimates of individual spawning frequencies and how these might differ among females of varying body sizes, would be needed to fully describe the compensatory increase in reproduction which has occurred for the Necker Bank population of P. marginatus. Acknowledgments We thank several anonymous fishery observers for collec- tion of invaluable commercial catch data and R. Moffitt, J. Polovina, and an anonymous reviewer for constructive criticisms of the manuscript. - DiNardo, G. T, W. R. Haight, and J. A. Wetherall. 1998. Sta- tus of lobster stocks in tfie Northwestern Hawaiian Islands, 1995-97, and outlook for 1998. Southwest Fish. Sci. Cent. Admin. Rep. H-98-05, 35 p. Honolulu Laboratory, Southwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2.396. 30 Fishery Bulletin 101(1) Literature cited Annala, J. H. 1991. Factors influencing fecundity and population egg production of Jasus species. In Crustacean egg produc- tion, Crustacean issues 7 (A. Wenner and A. Kuris, eds.), p. 301-315. Balkema, Rotterdam. Botsford, L. W. 1991. Crustacean egg production and fisheries manage- ment In Crustacean egg production. Crustacean issues 7 (A. Wenner and A. Kuris, eds.), p. 379-394. Balkema, Rotterdam. Briones-Fourzan, P.. and G. Contreras-Ortiz. 1999. Reproduction of the spiny lobster Panulirus gitttatus (Decapoda: Palinuridae) on the Caribbean coast of Mexico. J. Crust. Biol. 19:171-179. Chittleborough, R. G. 1979. Natural regulation of the population of Panulirus longipes cvgnus George and responses to fishing pressure. Rapp. P-V. Reun. Cons. Perm. Int. Explor. Mer 175:217- 221. Chubb, C. F 2000 Reproductive biology: issues for management. In Spiny lobsters: fisheries and culture, 2"0.05; 1999: x-=^.2A, df=3, P>0.05) and for Kailua Bay (1998: ^-=5.74, df=3, P>0.05; 1999: ;f-=7.14, df=3, P>0.05). Angling (pole and line fishing) accounted for the great- est number of hatchery-reared and wild Pacific threadfin acquired in the reward program but had the lowest CPUE among gear types (Table 5 ). Over both survey years, for wild and hatchery-reared fish combined, angling accounted for 63.0% of the total Pacific threadfin catch, followed by gill- nets ( 19.2% ), thrownets ( 13.9% ), and surround nets (3.9% ). Length of Pacific threadfin caught varied among gear types, between years, and between hatchery-reared and wild Pacific threadfin. For 1998, mean size of Pacific threadfin captured was significantly different among gear types (P=6.378, df=3, 632, P<0.001) and between hatch- ery-reared and wild fish (P=11.833, df=l, 632, P<0.001). Gillnets tended to catch larger fish, and surround nets captured the smallest fish (Tukey multiple comparison test results — gillnets: 255.5 mm > angling: 235.1 mm > thrownet: 214.0 mm > surround nets: 193.7 mm). Mean length for hatchery-reared Pacific threadfin pooled over all gear types was 207.0 mm (SD=28.6) and mean length for wild Pacific threadfin was 242.2 mm (SD=52.7). 38 Fishery Bulletin 101(1) A 1998 N Kahana Bay release site <-, Kaneohe Bay B 1999 Maleakahana (, Kahana Bay release site \ Waimanalo Figure 3 Movement patterns of hatchery-reared Pacific Ihreadfin released in Kahana Bay from the reward fishery program during the lAi 1998, n - 20 and (B) 1999. n = 40 fishing seasons. Friedlander and Zlemann: Impact of hatchery releases on recruitment of Polydactylus sexfilis 39 A 1998 B 1999 N ^\ Kallua Bay release site r^ Lanikai Waimanalo Sandy Beach Maleakahana Kailua Bay release site Figure 4 Movement patterns of hatchery-reared Pacific threadfin released in Kailua Bay from the reward fishery program during the (A) 1998, n = 43 and (B) 1999, n = 57 fishing seasons. Information on sex of Pacific threadfin obtained in the reward program was available only during 1999 because whole fish were not acquired in 1998. The ratio of male to hermaphrodite to female was 56:17:27 for wild Pacific threadfin and 81:15:4 for hatchery-reared fish (Fig. 5). Three hatchery-reared fish that had changed from males 40 Fishery Bulletin 101(1) Table 5 (\atch per unit of effort (CPUE) for Pacific Mean rank computed for Kruskal-Wallis dure. Gear types with the same letter (A, threadfin caught by three different gear types used in rank sum test (//=19.303, df=2, P<0.001). Results of E B) are not significantly different. the recreational-artisanal fishery, unn's multiple comparison proce- Gear type n Total hours Total no. of threadfin No hatchery-reared Mean CPUE SD CPUE Dunn's multiple comparisons Thrownet 16 55.0 166 7 5.49 9.34 A Gillnet 31 83.5 221 20 3.52 4.89 AB Angling 113 592.5 817 56 1.63 1.44 B Grand total 160 731.0 1204 83 2.38 3.98 Comparison of size (mm PL) and sex for h 1999. Table 6 atchery-reared and wild catch Pacific threadfin acquired from the reward fishery in Hatchery-reared Wild Sex Mean Min. Max. n Mean Min. Max. n Males 257.1 Hermaphrodites 267.2 Females 311.3 210 220 247 307 323 356 61 11 3 250.0 275.0 315.6 174 200 249 360 372 380 112 462 179 to females were recovered during 1999; one was released during summer 1996, the other two were released during summer 1997. Mean size for males was 250.0 mm (SD=32.1) for wild fish and 257.1 mm (SD=24.8) for hatchery-reared fish; mean size for hermaphrodites was 275.0 mm (SD=27.5) for wild fish and 267.2 mm (SD=43.3) for hatchery-reared fish; mean size for females was 315.6 mm (SD=39.2) for wild Pacific threadfin and 311.3 mm (SD=57.1) for hatchery-reared threadfin (Table 6). The GSI for hatchery- reared males (.v=0.629, SD=0.728) was not significantly different (7=6524, P=0.073) than the GSI for wild males (.v=0.619, SD=0.842), likely because of the larger size of hatchery-reared males during 1999. Number of hatchery- reared females and hermaphrodites was too low for statis- tical comparisons. Condition factor for hatchery-reared Pacific threadfin during 1999 was not significantly different between re- lease sites (F=0.074, df=l, 79, P=0.786) or among release sizes (F=1.488, df=3, 79, P=0.224). Therefore, condition factors for ail hatchery-reared fish were pooled and com- pared to condition factors for all wild fish recovered in 1999. No significant difference was found in condition fac- tors between the.se two groups ( 7=47733.0, P=0.087). Discussion Cultured Pacific threadfin juveniles released into the ocean survived and recruited successfullv into the recreational Sex Figure 5 Sex ratio for wild and hatchery-reared Pacific threadfin returned for the 1999 reward fishing sea.son. Values are total number offish in each sex category. fishery, accounting for Wf and 8*^; of the catch on the windward side of the island of Oahu in two years ( 1998 and 1999, respectively). Hatchery fish from the 1997 release constituted the majority of the hatchery fish returns to the recreational fishery in 1998 (89.4':^ ) and 1999 (95.9'^; ). Few of the hatchery fish released in years prior to 1997 have been recovered from the recreational fishery. The large Fne(dlander and Ziemann: Impact of hatchery releases on recruitment of Polydactylus sexftlts 41 Comparisons of length of juveniles, males for wild fish from the 1999 reward fishery Table 7 hermaphrodites, and females for Pacific threadfin around Oahu from 1962 to 1968 and program. One standard deviation of mean fork length is shown in parentheses. Sex 1962-68, n = 1651 1999 reward program, n=1105 Mann Whitney T-value Fork length (mm) Percentage of total Fork length (mm) Percentage of total P Juveniles Males Hermaphrodites Females 227(30) 268(29) 317(33) 378(45) 6.4 52.3 17.8 23.5 191(24) 249(35) 275(27) 316(39) 39.7 33.8 10.3 16.2 44864 184831 11952 26200 A A A A O O O O b b b b o o o o impact of a relatively small number of released fish on the recreational fishery shows that hatchery releases of limited numbers of fish have the potential to impact both the number of fish taken in the fishery and the rate at which the fishery can recover. The differences in contribu- tion rates for different release years suggest that natural factors affecting the survival of juveniles, as well as early larval stages, vary between years. Hatchery-reared and released fish collected in the rec- reational fishery showed growth rates, condition factors, and gonadosomal indices similar to those of wild fish, suggesting that hatchery-reared fish are able to adapt to the natural environment and integrate into the wild population. Our data (unpubl.) for wild and hatchery fish collected in nursery habitats showed no significant differ- ences in growth rates. The mean size of hatchery-reared fish collected in 1998 was smaller than in 1999 (over 95% of the fish collected in 1998 and 1999 came from the same releases in 1997). Mean size for hatchery-reared fish in 1999 was not different from the mean size of wild fish for both years, which suggests that size of hatchery fish in 1999 represents the approximate size of 2-3 year-old threadfin and mean age of fish in the recreational fish- ery is also 2-3 years. The size-frequency distributions of hatchery and wild fish in 1998 and 1999 suggest that a significant portion of the wild fish in the fishery is younger than two years. Small hatchery fish at release made a higher relative contribution to the recreational fishery than did the larger size group (but not significantly so, except for fish taken in Kailua Bay in 1999), and the nursery habitat sampling conducted after the 1997 releases showed the same (Leber et al., 1998; Ziemann et al.'-). This pattern is in contrast to that observed for mullet in Hawaii (Leber, 19951 and Pacific threadfin for other years (Ziemann et al.'^) Hatchery fish disperse slowly from the point of release along the windward coast of Oahu. In nursery habitats three months after release (Ziemann et al.-i, hatchery fish represented in excess of 70'7( of the threadfin and they decreased within nine months to 10% or less. Some decrease is due to predation, but some is due to dispersal because in 1998, after 1 year at large, fish were caught in the recreational fishery a mean distance of 11.2 km from the release point, and after two years, mean distance had increased to 15.2 km. Dispersal from the two release sites differed: after one year mean distance for Kahana Bay releases was 14.6 km, whereas mean distance for fish releases in Kailua Bay was 9.6 km. The 1999 reward sample contained 16% females, 44% males and hermaphrodites, and 40% immature fish. The life cycle of Pacific threadfin (protandric hermaphrodites) makes this skewed sex ratio even more problematic be- cause individuals do not become functional females until about 30 cm FL and these larger fish are selectively re- moved from the population by fishing. For protogynous species, size-selective fishing mortality may result in differential loss of larger males (Sadovy, 1996; Beets and Friedlander, 1999). The percentage of juveniles in the catch was high. Mean size of Pacific threadfin in all sexual categories was significantly smaller than that reported by Kanayama* in 1962-68 (Table 7, Fig. 6); further, females constituted 23.5% of the catch in the 1960s, but only 16.2% of the catch in 1999. We demonstrated that cultured Pa- cific threadfin juveniles released in known nursery habi- tats survive and recruit successfully into the recreational fishery 1-2 years later. Our Pacific threadfin data indicate that recruitment of young fish to the population may be jeopardized because there are few mature females left in the population (recruitment overfishing), even with supplementation of hatchery-reared fish. The underlying problem of the threadfin fishery on Oa- hu and the other Hawaiian Islands is primarily an intense local harvest by subsistence and recreational fishermen, as well as habitat loss from coastal and upland develop- ment. Current state regulations, as well as unregulated removal of larger individuals from the population, contrib- ute to the male-biased sex ratios observed in our study. Stock recovery based on natural reproduction will be a long-term process. Implementation of an enhancement progi-am for Pacific threadfin focused on juveniles and perhaps larger females could speed the rate of recovery of the local population. Kanayama, R. 1973. Life history aspects of the moi Po/vrfac- tylus sexfilix in Hawaii, 50 p. State of Hawaii, Department of Lands and Natural Resources, Honolulu, Hawaii. 42 Fishery Bulletin 101(1) 0-25 1 A 1962-68 0.20 0.15 0.10 0.05 Z 0.00 0.25 0.20 0.15 0.10 0.05 I I Juveniles ^y////////i Males ^^^m Hermaphrodites ^^^ Females 0.00 10 30 40 Fork length (cm) Figure 6 Proportion of juveniles, males, hermaphrodites, and females for Pacific thread- fin from I A) 1962 to 1968 and (B) during the 1999 reward fishery program. Acknowledgments The authors acknowledge the contributions to this re- search made by Ken Leber, Peter Craig, Reiji Masuda, Robert Cantrell, Steve Arce, Scott Bloom, Tom Ogawa, Don Dela Pena, Rich Hall, Karl Keller and other members of the slock managenii'nl staffat The Oceanic Institute, and of the culture support provided by Tony Ostrowski and the staff of The Oceanic Institute finfish program. Jim Parrish, Reiji Masuda, Ken Leber, two anonymous reviewers and editors provided valuable .suggestions for the manuscript. This research was supported under NOAA gi-ant NA7(jFY0()59. Literature cited Anderson, R. O., and R .M. Neumann. 1996. Length, weight, and associated structural indices. In Fisheries techniques, 2"'' ed. (B. R. Murphy and D.W. Willis, eds. ), p. 447-482. Am. Fi.sh. Soc, Bethesda, MD. Blankenship, H. L., and K. M. Leber 1995. A responsible approach to marine stock enhancement. Am. Fish. Soc. Symp. 15:165-175. Beets, J., and A. Friedlander 1999. Evaluation of a con.servation strategy: a spawning aggregation closure for grouper in the Virgin Islands. En- viron. Biol. Fish. 55:91-98. Fnedlander and Ziemann: Impact of hatchery releases on recruitment of Polydactylus sexfilis 43 Bleeker, P. 1875. Recherches sur la fauna de Madagascar et de ses dependances d"apres les descouvertes de Francois P. L. Pollen et D. C. van Dam. 4" Parte. Poissons de Madagascar et de rie de la Reunion. Leiden, The Netherlands. Grimes, C. B. 1998. Marine stock enhancement: sound management or techno-arrogance? Fisheries 23(9):18-23. Howell, B. R., E. Moksness, and T. Svasand (eds.). 1999. Stock enhancement and sea ranching, 606 p. Fish- ing News Books. Oxford, England. Hosaka, E. Y. 1990. Shore fishing in Hawaii, 175 p. Petroglyph Press, Ltd., Hilo, HI, Imai. T, H. Takama, and I. Shibata. 1994. Estimates of the total amount of red sea bream caught by recreational party boats in Kanagawa Prefecture. Sai- bai Giken 23:77-83. |In Japanese.] Jefferts, K. B., P. K. Bergman and H. F Fiscus. 1963. A coded-wire identification system for macro-organ- isms. Nature (London) 198:460-462. Kaeriyama, M. 1996. Population dynamics and stock management of hatchery-reared salmons in Japan. Bull. Natl. Res. Inst. Aquacult. Suppl. 2:11-15. Kitada, S., Y. Taga, and H. Kishino. 1992. Effectiveness of a stock enhancement program evalu- ated by a two-stage sampling survey of commercial land- ings. Can. J. Fish. Aquat. Sci. 49:1573-1582. Leber, K. M. 1995. Significance of fish size-at-release on enhancement of striped mullet fisheries in Hawaii. J. World. Aquacult. Soc.,26(2):143-153. Leber, K. M. and S. M. Arce. 1996. Stock enhancement in a commercial mullet, Mugil cephalus L., fishery in Hawaii. Fish. Manage. Ecology 3: 261-278. Leber, K. M, S. M. Arce, D. A. Sterritt, and N. P. Brennan. 1996. Marine stock-enhancement potential in nursery habi- tats of striped mullet, Mugil cephalus. in Hawaii. Fish. Bull. 94:452-471. Leber, K. M., N. P. Brennan , and S. M. Arce. 1998. Recruitment patterns of cultured juvenile Pacific thread- fin, Polydactylus sexfilis (Polynemidae), released along sandy marine shores in Hawaii. Bull. Mar Sci. 62:389^08. Lowell, N. 1971. Some aspects of the life history and spawning of the moi [Polydactylus sexfilis). M.A. thesis, 45 p. Univ. Hawaii, Honolulu, HI. Masuda, R., and K. Tsukamoto. 1998. Stock enhancement in Japan: review and perspective. Bull. Mar Sci. 62:337-358. McEachron, L. W., R. L. Colura, B. W. Bumguardner, and R. Ward. 1998. Survival of stocked red drum in Texas. Bull. Mar Sci. 62:359-368. McEachron, L. W., and K. Daniels. 1995. Red drum in Texas: a success story in partnership and commitment. Fisheries 20:6-8. Myers, R. F 1991. Micronesian reef fishes: a practical guide to the iden- tification of the coral reef fishes of the tropical central and western Pacific, 298 p. Coral Graphics, Barrigada, Guam. Ostrowski, A., T. Iwai, S. Monahan, S. Unger, D. Dagdagan, P. Murawaka, A. Schivell, and C. Pigao. 1996. Nursery production technology for Pacific threadfin {Polydactylus sexfilis). Aquaeulture 139:19-29. Randall. R. E. 1996. Shore fishes of Hawaii, 216 p. Natural Worid Press, Vida, OR. Randall, J. E., G. R. Allen, and R. C. Steene. 1990. Fishes of the Great Barrier Reef and Coral Sea, 507 p. Univ. Hawaii Press, Honolulu, HI. Sadovy, Y. J. 1996. Reproduction of reef fishery species. In Reef fish- eries (N. V. C. Polunin and C. M. Roberts, eds.), p. 15-59. Chapman and Hall, London. Santerre, M. J., G. S. Akiyama, and R. C. May. 1979. Lunar spawning of the threadfin, Polydactylus sexfi- lis, in Hawaii. Fish. Bull. 76:900-904. Santerre, M. J., and R. C. May 1977. Some effects of temperature and salinity on labora- tory-reared eggs and larvae of Polydactylus sexfilis (Pisces: Polynemidae). Aquaeulture 10:341-351. Schramm, H. L., Jr, and R. G. Piper, eds. 1995. Uses and effects of cultured fishes in aquatic ecosys- tems, 608 p. Am. Fish. Soc. Symp. 15. Sokol, R. R., and F J. Rohlf 1981. Biometry, 859 p. WH. Freeman, San Francisco, CA. Szyper, J. P, M. J. Anderson, and N. H. Richman. 1991. Preliminary aquaeulture evaluation of moi {Polydac- tylus sexfilis ). The Progressive Fish-Culturist 53:2025. Tinker, S. W. 1982. Fishes of Hawaii, 532 p. Hawaiian Services, Inc., Honolulu, HI. Titcomb, M. 1972. Native use of fi.sh in Hawaii, 175 p. LTniv. Hawaii Press. Honolulu, HI. 44 Abstract — A general model for yield- per -recruit analysis of rotational (per- iodic) fisheries is developed and ap- plied to the sea scallop (Placopecten magellanicus) fishery of the northwest Atlantic. Rotational fishing slightly increases both yield- and biomass-per- recruit for sea scallops at F^y^. These quantities decline less quickly when fishing mortality is in-creased beyond ^MAX than when fishing is at a constant rate. The improvement in biomass- per-recruit appears to be nearly inde- pendent of the selectivity pattern but increased size-at-entry can reduce or eliminate the yield-per-recruit advan- tage of rotation. Area closures and rota- tional fishing can cause difficulties with the use of standard spatially averaged fishing mortality metrics and reference points. The concept of temporally aver- aged fishing mortality is introduced as one that is more appropriate for seden- tary resources when fishing mortality varies in time and space. Yield- and biomass-per-recmit analysis for rotational fisheries, with an application to the Atlantic sea scallop (Placopecten magellanicus) Deborah R. Hart Northeast Fisheries Science Center 166 Water St Woods Hole, MA 02543 E mail address Deborati HartsSnoaa gov There has been growing interest in using rotational fishing to manage ses- sile or sedentary stocks (e.g. Caddy and Seijo, 1998). Under such a strategy, fish- ing mortality in a given area is varied periodically. Typically, the area is closed for a period of time, then fished, and then closed again. The openings of the different areas are timed so that at least one area is open to fishing each year. This approach has been proposed or is being used for abalone, corals, sea cucumbers, geoduck clams, sea urchins, and several species of scallops (Sluc- zanowski, 1984; Garcia, 1984; Botsford et al., 1993; Caddy, 1993; Heizer, 1993; Campbell et al., 1998; Caddy and Seijo, 1998; Lai and Bradbury, 1998). Recently, area closures have been used to help manage the Atlantic sea scallop (Placopecten magellanicus) fish- ery off the northeastern United States. Three areas on Georges Bank were closed to scallop and groundfish fish- ing in December 1994 to help protect depleted groundfish resources. Subse- quently, there have been substantial in- creases in scallop abundance, biomass, and mean size in these areas; mean scallop biomass in the closed areas, as measured by the Northeast Fisheries Science Center (NEFSC) sea scallop survey, rose from 0.6 kg/tow in 1994 to 15.8 kg/tow in 2000.' During limited Manuscript accepted 20 September 2002. Fish. Bull. 101:44-.57(2003). ' NEFSC (Northeast Fisheries Science Cen- ter). 2001. Report of the 32nd north- east regional stock assessment workshop (32nd SAW). Stock Assessment Review Committee (SARC) consensus summary of assessments. NEFSC Ref Doc. 01-05. 289 p. (Available from NEFSC, 166 Wa- ter St., Woods Hole MA 02M:\.\ openings of these areas to fishing in 1999 and 2000, nearly 5000 metric tons (t) of scallop meats (about 20'* of the total landings during this period) were landed, while biomass levels remained high. In April 1998, two areas in the Mid-Atlantic Bight were closed to scal- lop fishing for three years in order to protect high concentrations of juvenile scallops. Scallop biomass has increased markedly since the closures in these areas as well, from 0.8 kg/tow in 1997 to 9.7 kg/tow in 2000^ and about 3500 t of scallop meats have been landed in these areas in the year since they were reopened in May 2001. These data suggest that temporary or rotational closures can help increase scallop bio- mass and yield. For these reasons, a rotational management system for the U.S. Atlantic sea scallop fishery is cur- rently under consideration. Many common fisheries models may not be appropriate for sessile stocks because these models assume spati- ally uniform fishing mortality (Caddy, 1975). Such a "dynamic pool" assump- tion is strongly violated when a sessile stock is fished rotationally so that a portion of the stock is not fished in a given year. For this reason, many previous analyses of rotational fisher- ies have used either spatially explicit simulations (e.g. Caddy and Seijo, 1998), per- recruit analyses of pulse fishing, where all vulnerable individu- als are removed from an area when the area is fished (e.g. Sluczanowski, 1984), or per-recruit analyses of a single co- hort (e.g. Gribble and Dredge, 1994). Spatially explicit models suffer from then- complexity, making it difficult to extract general principles from model Hart: Yield- and biomass-per-recruit analysis of rotational fisheries 45 simulations. Analysis of pulse fishing, although simple to apply and understand, is not applicable to those situations where only a portion of the available resource is removed when an area is opened periodically to fishing. Per-recruit analysis of a single cohort is not applicable to relatively nonselective multiple age-group fisheries. Botsford et al. (1993) developed a mixed-age rotational yield-per-recruit model for red sea urchins. They showed that rotational fishing for these urchins would increase egg production considerably, while slightly decreasing yield-per-recruit. Recently, Myers et al. (2000) presented a mixed-age per-recruit analysis of a possible rotational fishery strategy for sea scallops. The emphasis of this study was on the effect of putative high levels of indirect (noncatch) fishing mortality on yield-per-recruit, and on a proposed rotational plan that Myers et al. suggested would help ameliorate this effect. The purpose of the present article is to present a general theory for any type of periodic or rotational fishing strat- egy for a mixed-age sessile or sedentary stock. This work generalizes many of the above mentioned studies (in par- ticular, that of Botsford et al., 1993) and does not require an assumption of constant recruitment or specific spatial configuration (or both). This theory is applied to the Atlan- tic sea scallop fishery of Georges Bank. Measures of fishing mortality and overfishing defini- tions are usually based on models where fishing is as- sumed constant in space and time. In rotational fisheries, or in cases where part of a fishing ground has been closed indefinitely to fishing, these assumptions may be seriously violated, especially for stocks that are relatively immobile. Alternative measures of fishing effort and overfishing definitions are presented here that are more appropriate to fisheries of nonmobile stocks where rotational or indefi- nite closures are used. Methods The object of this analysis was to compute the expected yield-per-recruit and biomass-per-recruit of a cohort located in an area where fishing mortality may depend on the year and the variation in fishing mortality is periodic with time. Rotational fishing is usually thought of as a sequence of periodic closures and openings of different areas. The theory described here is more general, and can be applied to any situation where fishing mortality is varied periodically in a given area. Suppose the fully recruited fishing mortality in an area during year k is F^ and that fishing mortality rates vary pe- riodically with period p ( where p is in years ), so that F , = Ff. for all k. Let ^wi; be the mean of F^.F.^ F^^ For sim- plicity, it is assumed that there is one recruitment event and one new cohort per year. However, extension of the theory to multiple cohorts per year is straightforward. There are p possible patterns of fishing mortality expe- rienced by a cohort, depending on the point of the cycle when it enters the fishery. The cohort that enters in the first year will experience fully recruited fishing mortality rates: F F F F F F F (1) during successive years. The next cohort will experience the same fishing mortality rates, but in a different order: F„F„F„...,F^.F,K. (2) and so on. Two special cases are of particular interest: pulse rota- tion and symmetric rotation. Pulse rotation means that F^ = for ^=l,2,...,p-l (the area is closed for p-1 years), then F >0 (the area is pulse fished for one year), and then Ff^=0 for k=p-¥l,p+2,..., 2p-l (the area is closed again), etc. Symmetric rotation, where p is even, means that F/, = for 1 < ^ < p/2, and Ff. = 2F^v^. for p/2 < k < p, i.e. the area is closed for p/2 years and then fished at a constant rate for the next p/2 years. For each of the p patterns of fishing mortality, yield- and biomass-per-recruit can be calculated by using standard per-recruit techniques. Here, a method similar to the "ge- neric per-recruit" model described in Quinn and Deriso (1999) is used (see Appendix). The only unusual aspect is that the mortality terms Z and F^. in Equations 11-13 (see Appendix) depend explicitly on time, i.e. on the year of the rotational cycle. Each of the p cohorts will have differ- ent yield-per-recruit Y^,Y.^...,Y , and biomass-per-recruit ByB.^.-Mp values because the ages at which they experi- ence the fishing mortalities F^,F2,..-,F are different. Define y^VG ^"^^ ^avg ^^ be the means of the p patterns of cohort yield- and biomass-per recruit, respectively. Y^yg is the expected yield of a recruit chosen randomly with re- spect to cohort. In other words, F^vc '^ ^^^ long-term mean yield-per-recruit that can be expected from the rotational fishing strategy. Similarly, Bavg '^ the expected long-term mean biomass-per-recruit. Note that unlike conventional per-recruit theory, yield- and biomass-per-recruit vary with cohort, so that the mean yield- and biomass-per-re- cruit obtained at any point in time may be different from ^AVG and Bavg- Let y^',y2',...,y'P' be yield-per-recruit of the p cohorts, in decreasing order, so that y" is the highest yield-per- recruit out of all the p cohorts and Y'l'' the lowest, y" is an upper bound on the yield-per-recruit that might be obtained with a rotational strategy if, for example, the closure pattern were timed to optimize yield-per-recruit from a large year class. It is important when comparing rotational and con- stant fishing strategies to compare alternatives that have the same long-term survival rates, i.e the same natural mortalities and mean fishing mortality rates. If this is not done, then effects of rotation can be confounded with those due to variations in fishing mortalities. If there are initially N^ fully recruited individuals in an area that are fished at a constant rate F„, then there will be N=N,,exp{-pM-pFj (3) of these individuals remaining alive after p years. If instead, fishing mortality was varied on ap year rotation, so that in each year of the cycle, fishing mortality in an 46 Fishery Bulletin 101(1) Table 1 Estimated life history parameters for Georges Bank sea scallops. Von Bertalanffy growth parameters are from Serchuk et al. ( 1979). Relations of shell height (SH) to meat weight (MW) (see Eq. 7) were obtained by combining the data of Serchuk and Rak' with that of NEFSC (Footnote 2 in the general text). The natural mortality estimate is from Merrill and Posgay (1964). The selectivity pattern is based on the current gear configuration of scallop dredges with 89-mm rings (NEFSC, Footnotes 1 and 2 in the general text). M(/vr) a (In g) b ^mmtmm' ''full <™ni' h^ (mm) d KUyr) L^ ( mm ) (natural (SHMW (SHMW (MinSH (SH for full (cull (discard (growth) ( growth ) mortality) parameter) parameter) selected) selectivity) size) mortality Value 0.3374 152.46 0.1 -11.6038 3.1221 65 75 0.2 Serchuk, F. M, and R, S. Rak. 1983. Biological characteristics of offshore Gulf of Maine scallop populations: size distribution, relations of shell height to meat weight, and relative fecunditv patterns. Reference document 83-07, 42 p. [Available from Northeast Fisheries Science Center, 166 Water St.. Woods Hole, MA 02543.1 area is given by Fj, F.^ of individuals remaining alive after p years would be F , respectively, then the number N'p = No exp -pM-Y^F, (4) In order for the long-term survivorship of the two strate- gies to be equal (i.e., A^ =N^' ), the uniform fishing mortal- ity F|^ must equal the average fishing mortality AVG = it.. P L 1=1 (5) of the rotation plan. Therefore, F^vc. ^^ used to scale all the graphs and per-recruit comparisons. The model described above and in the Appendix was implemented as a Fortran-90 program where the integrals were numerically computed with a time step of 0.01 y. Parameters used in the model are given in Table 1 and represent estimates for growth and mortality of Atlantic sea scallops (Placopecten magellanicus), for which rota- tional management is currently under consideration. Results Yicld-per-recruit curves for no rotation (continuous uni- form fishing), three-year pulse rotation (i.e. the area is closed for two years and fished for one year), six-year pulse rotation, and nine-year pulse rotation are given in Figure 1. Note that the x axis in Figure 1 is the mean fishing mortality rate Fi^vi;- ^^^ they axis is mean (i.e. expected) yield-per-recruit V^vc- averaged over cohorts. Because the mean fishing mortality rate is the same for all points at the same .v coordinate, the three-year rotation has a fish- ing mortality rate during years when fishing occurs (F ) that is throe times as high, and the six-year rotation six times as high, as the constant F i no rotation ) strategy with the same F, AVG- Rotation affects the yield-per-recruit curve for sea scal- lop in three different ways (see Fig. 1). First, rotation modestly increases the maximum mean yield-per-recruit Yf^js^; the maximum mean yield-per-recruit for the nine- year rotation is about 9% greater than without rotation. Second, Fj^l« *i-^- the value of Fj^vc, where Yt^j^x '® °^" tained) increases somewhat under rotation, especially for longer rotation periods. Third, there is less yield penalty in rotational management for exceeding Fiy^a^x. For example, fishing at F = 1 results in a 38% loss of yield if there was no rotation, but only an 8% loss under a six-year pulse rotation. Although 6-yr pulse rotation results in only a 5% increase in yield-per-recruit over no rotation at their re- spective F^y^ values, the advantage of 6-yr pulse rotation atF= 1 is43'7c. Maximum yield-per-recruit for pulse fishing as a func- tion of the rotation period p is shown in Figure 2. The best yield-per-recruit is obtained for long periods of 9 to 10 years. However, this type of strategy would imply that a number of years would pass before any yield would be obtained from most recruits and this strategy would only slightly increase maximum yield-per-recruit over that of steady fishing. Depending on management goals, it might be reasonable to discount future yields, so that the pres- ent value of yield taken t years into the future would be discounted by exp(-&), where 5 is the annual discount rate (assumed 10%/yr). The rotation period that maximizes discounted yield-per-recruit is 6 years (Fig. 2). If prices as a function of meat weight are known, it would also be possible to do a similar analysis to maximize discounted value-per- recruit. Yield isopleths, commonly used to visualize yield-per- recruit analysis (Beverton and Holt, 1957), are given in Figure 3A (yield-per-recruit) and 3B (discounted yield- per-recruit). For rotational analyses, fishing mortality is placed on the .v-axis and rotational period on the y-axis. Note again that for longer rotation times, the decline in yield for fishing mortalities greater than F^,^^ is much less than without rotation. The value of F^^^j^^ and maxi- mum yield-per-recruit increases slightly with longer rota- tion periods. Biomass-per-recruit for no rotation, 3-, 6- and 9-yr pulse rotation strategies is given in Figure 4. Compared to con- stant fishing, rotational fishing gives increased biomass- per-recruit; this increase is most evident for the longer Hart: Yield- and biomass-per-recruit analysis of rotational fisheries 47 No rotation 3 yr rotation 6 yr rotation 9 yr rotation 04 F (/yr) Figure 1 Yield-per-recruit curves for Georges Bank sea scallops with no rotation and with 3- 6-. and 9-yr pulse rotations. rotations and higher fishing mortalities. At F[^i,\x- the increase in biomass-per-recruit is shght, un- less a very long (e.g. 9-yr) rotational period is employed. The performance of rotation can be assessed as a function of selectivity at size. Figure 5A gives maximal yield-per-recruit for a number of pulse rotation strategies and a variety of values of h^^^^, the smallest size selected by the gear; the size of full selectivity, /if^,,, was taken as h^^^^ + 23 mm (consistent with the assumed current gear selec- tivity pattern; see Table 1). Rotation can give sub- stantial yield-per-recruit advantages when the gear selects animals of well below optimal size, especially for longer periods. However, long-pe- riod rotation actually gives less yield-per-recruit than constant fishing for larger values of h^^^. Figure 5B gives a similar plot for biomass-per- recruit, where fully recruited fishing mortality is fixed at F = 0.3 in all cases. Unlike yield-per- recruit, rotational fishing increases biomass-per- recruit regardless of the selectivity pattern, espe- cially when the rotational period is long. Yield-per-recruit from difTerent cohorts under a rotational system can vary considerably. The cohort which recruits into the fishery at about the time of the closure produces the highest yield-per-recruit, whereas the cohort that reaches exploitable size at about the time that the area is opened has the lowest. Figure 6 gives the mean yield-per-recruit together with that of the cohorts with the highest and lowest yield-per-recruit under six- year pulse rotational management (i.e. i^^vc^ ^"'' ^""^ ^*'' • • • • o o o o o 2 4 6 8 10 Rotation period p(yr) Figure 2 Maximum yield-per-recruit (solid circles) and discounted (Wv) yield-per-recruit (open circles) for Georges Bank sea scallops with a pulse rotation of periods between 1 and 1 1 years. respectively). A 319^ increase in maximal yield compared to constant fishing (and 25'7c increase over the average yield- per-recruit under rotation) can be obtained from the cohort whose yield-per-recruit is the highest under rotation. Note that, unlike conventional yield-per-recruit cun'es for sea scallops, yield-per-recruit from this cohort is almost com- pletely insensitive to effort beyond a certain level. 48 Fishery Bulletin 101(1) :!'■ '•(i '■ (i ■■'■'■ (i 10 10 B 2 4 6 8 1,0 Figure 3 Yic'ld-per-rocruit (Ai and discounted ( 10'* ) yield-pcr-rotrult (B) isopletlis for Georges Bank sea scalloi)s with pulse rotation. Note that the v axes represent rotation period. Hart: Yield- and biomass-per-recruit analysis of rotational fisheries 49 250 n \\ \'\ \\ No rotation 200 - V 6 yr rotation 9 yr rotation 150 - 3 J 100 - 50 - X: >^^ ^::^;;;;;:;:rr^^ - 1 1 I 1 0.0 02 0.4 0.6 08 F.J'V') Figure 4 Biomass-per-recruit for Georges Bank sea scallops with no rotation and with 3-, 6-, | and 9-yr pulse rotations. Table 2 Calculated values of F^^^, y^^X' ^maX' ^"""^ discounted iDsc) Yj^j^^^ (in grams, with a 10% discounting factor) for (A) pulse rotation with no incidental fishing mortahty, (B) pulse rotation with 15'7f incidental fishing mortality, and (Cl symmetric rotation with no incidental fishing mortality, p = period of rotation. P ^.iM.v ■'w.u- 5„^v DscY^,^^ P ^MAX ^M.\.\ ^M.\.\ ^^'^ ^ M.W A 1 0.217 17.25 84.0 10.66 B (cont.) 4 0.205 14.81 88.4 9.01 2 0.219 17.27 83.6 10.70 5 0.219 14.99 85.4 9.16 3 0.225 17.38 82.4 10.83 6 0.236 15.20 82.8 9.25 4 0.239 17.57 79.6 11.04 7 0.257 15.43 81.2 9.25 5 0.259 17.84 76.8 11.24 8 0.277 15.63 82.2 9.15 6 0.287 18.17 74.5 11.32 9 0.292 15.75 86.0 8.96 7 0.324 18.47 73.8 11.25 10 0.300 15.8 91.8 8.70 8 0.351 18.71 77.4 11.03 11 0.302 15.75 99.0 8.4 9 0.363 18.84 84.2 10.71 10 0.372 18.82 92.0 10.34 11 0.374 18.69 100.6 9.93 C 2 4 6 0.219 17.27 83.6 10.70 0.225 17.4 82.5 10.82 0.235 17.56 81.0 10.90 B 1 0.192 14.62 91.9 8.79 8 0.244 17.68 81.1 10.85 2 0.193 14.62 91.6 8.80 10 0.248 17.73 83.5 10.68 3 0.197 14.68 90.5 8.88 12 0.253 17.64 86.3 10.44 Yield-per-recruit curves for 6 yr symmetric rotation (i.e. closed for three years and then opened for three years), 10-yr symmetric rotation, 6-yr pulse rotation, and no rotation are given in Figure 7. Symmetric rotation gives yields-per-recruit that lie between that of pulse rotation and constant fishing. Maximum yield-per-recruit. together with the associated B^^x, and maximal discounted yield for pulse rotation, with and without 15'^'i incidental mortal- ity, and for symmetric rotation without incidental mortal- ity, are given for various rotation periods in Table 2. Results from yield-per-recruit runs with incidental fishing mortality (Table 2B) show similar patterns to 50 Fishery Bulletin 101(1) ; 20 No rotation 3 year rotation 6 year rotation 9 year rotation 160 140 120 100 - Figure 5 (A) Yield-per-recruit and (B) biomass-per- recruit for no rotation (solid line), 3-yr (dotted line), 6-yr (dashed line) and 9-yr (dot-dashed line) pulse rotations of sea scallops as a function of minimum selected shell height /i„,|„. Scallops are assumed to be fully recruited to the fishery at shell height /i,,,,^ + 23 mm. results with no incidental fishing mortality (Table 2A). Note that both yields and the value oi F^^.^y^ are reduced if incidental fi.shing mortality exists and that the pen- alty for overfishing without rotation is somewhat higher (about 679f loss in yield-per-recruit for fishing at F=l without rotation compared to 389f without incidental mortality). However, at ^\,,\x• the loss of yield due to inci- dental mortality is about the same for rotational fishing as for steady fishing. Discussion Rotational fishing can generate increased yield- and bio- mass-per-recruit for sea scallops compared to nonrota- tional fishing. The expected increase in maximum vield- per-recruit is modest (<109; ) for a fixed rotational pattern. The over SO'/f gain in yield-per-recruit obtained from cohorts that reached exploitable size near the time of the closure is partially cancelled by the loss of yield-per-recruit Hart: Yield and biomass-per-recruil analysis of rotational fishenes 51 25 -1 20 - 15 - -"''" 3 f" "~ £ 10- > / •"- ^ 5 / - f ^^^ Average cohort Worst yielding cohort — Best yielding cohort 0.2 4 6 8 ''wr. cyo Figure 6 Yield-per-recruit for a 6-yr pulse rotation of Georges Bank sea scallops from an average cohort, and those that benefit the most and the least from rotation. on those cohorts that reached exploitable size at about the time the area was reopened, thereby resulting in only a modest gain in yield-per-recruit. A more substantial gain in maximum yield-per-recruit (up to 30% greater than constant fishing) can be obtained if the closure is timed to optimally exploit an unusually large year class. These results are consistent with several studies that indicate that periodic fishing can often increase yields over con- stant fishing (Botsford, 1981; De Klerk and Gatto, 1981; McCallum, 1988; Clark, 1990; Myers et al., 2000). A second, and perhaps more important, advantage of rotational fishing is that it alleviates the impact of both growth and recruitment overfishing. Growth overfishing (i.e. fishing at a level higher than Fj^^j^) under rotational management induces a substantially smaller reduction in yield-per-recruit than would occur with constant fish- ing. Rotation also increases biomass-per-recruit for sea scallops, especially for levels of F above F^^y^, thereby reducing the impact of possible recruitment overfishing. It might be argued that overfishing should not be occurring in any case. However, even when management measures are taken to eliminate overfishing, it can still occur, for example, if 1) reference points are incorrect because of un- certainty in life history parameters; 2) fishing mortality, or the effect of management measures on fishing mortality, has been underestimated; or 3) there is localized overfish- ing because of spatial variation in fishing intensities or life history parameters (or variation in both), even though when averaged spatially, F,^y(. < F^^^ (Caddy, 1975; Hart, 2001). Rotational fishing can thus be thought of as part of a precautionary strategy. In so much as it may increase maximum yield, rotational management is superior to many other precautionary measures that reduce yield. The only costs of rotational management are the costs of administrating and enforcing such a system, and socioeco- nomic costs from temporary closures of traditional fishing grounds. The latter might be significant if closures force fishermen to make long distance steams to unfamiliar areas. Because the optimal F^vg under rotation is only slightly greater than the nonrotational F.^^^^^. the amount of effort and fleet capacity required to optimize yield-per- recruit under rotation is about the same as that needed under uniform fishing. Rotation also imposes practical constraints on the level of average fishing effort, thereby limiting the extent to which stocks can be overfished. Fishing mortality rates for U.S. sea scallop stocks were estimated as exceeding 1.0/yr during the late 1980s and early 1990s. ^ This would corre- spond under a 6-yr pulse rotation to an unaveraged fish- ing mortality of over F = 6 in the area open to fishing. Such a high fishing morality rate, corresponding to about a 98% exploitation rate for fully recruited scallops, is likely to be impractical for both physical and economic reasons. Thus, F.^y^, in a rotation plan would likely be considerably below the high levels seen in the late 80s and early 90s, even if there was no other restriction on fishing effort other than pulse rotation. Myers et al. (2000) claimed that "near-optimal yields are achieved across a wide range of fishing mortalities" in their rotational scheme. However, much of their analysis was confounded by their use of unaveraged open area fish- ing mortality (=pF^yq) on the x axis of their per- recruit curves. For example, in the case analyzed in Myers et al (2000), where one of p areas would be fished each year, the fishing mortality F applied in the area open to fishing in a 9-vr rotation (i.e. 1/9 of the area would be fished each 52 Fishery Bulletin 101(1) 15 - /"^"^^'^'SS^ 3 g 10 - / 5 - / 6 year pulse rotation 6 year symmetric rotation — - - 10 year symmetric rotation 0.0 0.2 0.4 0.6 0.8 10 f.ufiCyo Figure 7 Comparison of yield-per-recruit for Georges Bank sea scallops with 6-year and 10- year symmetric rotations with that of a 6-yr pulse rotation or no rotation. year and that F = F^yc, ^9' would represent one ninth of the actual effort as the same F applied to the whole area under non-rotational fishing. Use of unaveraged fishing mortality has the effect of stretching the .v axis by a factor of p, thereby making their graphs appear flatter than they actually are. F^vg ^^ representative of not only the true time-averaged fishing mortality but also in many cases would be proportional to spatially averaged fishing effort (as measured by, e.g., hours fished). Myers et al. (2000) also suggested that rotational fishing would help lessen the impact of indirect (incidental) fish- ing mortality on yield-per-recruit. The analysis given in the present study indicates that incidental mortality low- ers yield-per-recruit at /^max about the same amount re- gardless of whether or not rotational fishing is employed. At levels of fishing mortality well above F^i^y^, rotational fishing does appear to modestly decrease the loss of yield- per-recruit due to incidental mortality. This decrease is due to the fact that incidental mortality, by somewhat lowering F^^^, exacerbates the effects of overfishing, whereas rotation alleviates the loss of yield-per-recruit due to overfishing. The effectiveness of sea scallop rotational fishing can be understood by examining fishing mortality at size for vari- ous rotational strategies. F'igure 7 shows fishing mortality as a function of shell height for no rotation and for 3-, 6-, and 9-yr rotations for F^^q = 0.2 (Fig. 8A), and /-^vp = 0.6 (Fig. 8B). Rotational fishing (especially for longer periods) tends to reduce the fishing mortality on small scallops and shift this effort onto larger individuals, thereby in- creasing yield-per-recruit, especially when overfishing is occurring. The periodic peaks in fishing mortality seen in the rotational strategies occur at the sizes where a new cohort begins to be fished (i.e. when the scallops are some integer number of years past their age at 40 mm). In practice, these peaks are likely to be much less pro- nounced because of variations in individual gi'owth rates and settlement times. However, the qualitative pattern of increasing selectivity with size should not be affected by such variations. Sea scallops are an ideal candidate for rotational man- agement, combining fast growth and low natural mortality with a sedentary adult lifestyle. In addition, sea scallops are recruited into the fishery at a size that is well below optimal from a yield-per-recruit perspective. The increase in size-selectivity induced by rotation described above should therefore induce an increase in yield-per-recruit. However, in those fisheries where the size-at-entry to the fishery is much larger, rotation would not be expected to induce gains in maximal yield-per-recruit (see Fig. 5A). On the other hand, it appears that rotation increases biomass-per-recruit regardless of the size-selectivity of the fishery (Fig. 5B). Botsford et al. ( 1993) found that rota- tion increased biomass- but not yield-per-recruit for red sea urchins. These results are consistent with the above discussion because the minimum legal size for landing the urchins was already near-optimal. Although the exact levels of yield- and biomass-per- recruit obtained with or without rotation are sensitive to such factors as natural mortality and growth rates, the relative gains of rotation over constant fishing are much less sensitive to these factors. Rotation will improve yield-per-recruit under a broad range of parameter choices provided that li the ratio of growth to mortality K/M is Hart: Yield- and biomass-per-recruit analysis of rotational fisheries 53 0,30 025 020 15 - 0,10 - 0-05 - 0.00 1.2 1.0 0.8 0.6 0.4 0.2 - 0.0 No rotation Ttiree year rotation Six year rotation Nine year rotation 80 No rotation 3 yr rotation 6 yr rotation 9 year rotation 100 120 140 80 100 120 Stiell tieight (mm) 140 Figure 8 Mean fishing mortality at length for Georges Bank sea scallops with no rotation and with 3-. 6-. and 9-yr pulse rotations for (A) F = 0.2, and (B) f = 0.6. sufficiently high (greater than about 0.5 with the other parameters in the model fixed as given in Table 1), and 2) size-selectivity is suboptimal. Rotation improves biomass- per-recruit under even a wider range of parameters. Allee effects may occur in broadcast spawners such as urchins and scallops. Areas that are closed for several years may allow these animals to form dense aggregations (that would likely be heavily fished if not closed), thereby improving fertilization success (Botsford et al., 1993). Such an effect would mean that rotation could produce greater benefits in fecundity than would be suggested by biomass- or eggs-per-recruit curves. Metapopulation structure might also be considered when designing a rotational strategy. If recruitment is limited by the supply of settling larva, an area that is a source of larva might be fished less than that required to maximize yield-per-recruit in order to increase larval sup- ply (Tuck and Possingham, 1994). The calculations that indicate long optimal rotational periods assume low constant natural mortality, indepen- dent of age or density, based on the study of Merrill and Posgay (1964). There is some evidence that the natural mortality rate of sea scallops may increase with age or size for shell heights greater than about 110 mm (Mac- 54 Fishery Bulletin 101(1) Donald and Thompson, 1986). If this is the case, optimal rotational periods would be shorter than calculated here, although the jdeld-per-recruit formalism would remain valid. More serious problems would be caused if there is density-dependent mortality of adults or if high adult den- sities inhibited recruitment because rotational closures can induce higher densities than would constant fishing. If either of these processes occur, shorter rotation periods would be advisable to minimize this problem. For sea scal- lops, however, observations of areas that have been closed to fishing for a number of years give no indication that such density-dependent processes are occurring (Fig. 2b in Hart 2001, and Table B5-8 in NEFSC^). An extreme case of rotational fishing is true pulse fishing, where all exploitable individuals are removed at periodic intervals (see e.g. Sluczanowski, 1984). Thus, true pulse fishing corresponds to pulse rotation (as defined in the present study) with an infinite fishing mortality. Such pulse fishing is not optimal for sea scallops, as can be seen by the slight decline of yield-per-recruit at high fishing mortalities in Figure 1 because at very high fish- ing mortalities, the partial selectivity of the gear loses its effectiveness and all individuals that are even slightly selected to the gear (i.e. that are even slightly greater than /in,,,,) will be removed. To put it another way, at high fishing mortality rates, the additional (i.e. marginal) catch obtained from a further increase in F will disproportion- ately consist of small animals, thereby reducing yield- per-recruit. Pulse fishing would be optimal if knife-edge selectivity is assumed. For this reason, the assumption of knife-edge selectivity would lead to unrealistic results for cases such as sea scallops, where gear selectivity increases more gradually with size. Proper application of rotational theory therefore requires a careful examination of fishing selectivity with size. Pulse fishing can be related to the classic Faustmann theory of forest rotation (see e.g. Reed, 1986; Clark, 1990). In this theory, if a stand of trees in an area that has last been harvested t years previously has value Vit), then the optimal time to harvest the trees satisfies V'(t) = 8(V(t)-c) + 5 V{t)- exp(&)- 1 (6) where 5 c the discount rate; and the cost of harvesting. In the case of a fishery, Vit) would represent the expected value of the exploitable stock (e.g. those of shell height greater than /!„„„) at time t. (Note that in this context, unlike the original forest application, it is not necessary to assume that all exploitable individuals arc the same age.) In the case of scallops, assuming all scallops command the same price per unit weight, c = 0, and 5 = 0.1, this formula would give an optimal rotation period of about 6.1 years (the optimal period would be moderately longer for realis- tic positive values of c). This value corresponds well to the rotation time of 6 years that optimizes discounted yield- per-recruit (see Fig. 2). However, the yield-per-recruit for 6-yr pulse fishing, V(6), is less than 80% of the maximal yield-per-recruit obtained by fishing uniformly. Again, the reduced yield-per-recruit is due to the fact that pulse fish- ing induces knife-edge selectivity at h^^^. rather than the usual gradual increase in vulnerability to the gear Symmetric rotational strategies appear to give less benefit than does pulse rotation. However, optimal pulse rotation would require high, and possibly impractical, lev- els of effort in an area when it is opened (e.g. F of about 1.7 for a 6-yr pulse rotation). In addition, such a strategy would require that areas be closed most of the time, pos- sibly inducing social-economic disruptions by closing traditional fishing grounds for long periods. Compared to pulse rotation, symmetric rotation requires less concen- trated effort, allows areas to be open half the time, and is less sensitive to the assumption of constant natural mortality. One possible compromise between pulse and symmetric rotation is to close an area for half the time and then gradually increase effort during the opening. For example, an area might be closed for three years and then fished for the next three years at Ffj^^, 2Fjj^^, and ■^^MAX- respectively. (Questions have been raised regarding the appropriate- ness of the use of whole-stock fishing mortalities as tar- gets or reference points for fisheries of sedentary stocks that include rotational or long-term closures (or both) (NEFSC-). The solid line in Figure 9 gives the whole stock (biomass-weighted) fishing mortality (assuming constant recruitment everywhere) for a pulse rotational system consisting of six areas, one of which is fished each year in turn. This whole-stock fishing mortality was obtained by simply dividing the yield-per-recruit for a 6-yr pulse rotation by the corresponding biomass-per-recruit. The x axis is F^vG' which should be proportional to true effort. As can be seen, whole-stock fishing mortality is proportional to effort for low fishing mortalities, but then flattens to a maximum of just under 0.4. A similar situation can happen even if an area is fished uniformly, except that a portion of the area is set aside as an indefinite closure. The dashed line in Figure 9 gives an example for the case when lO"^'; of the area is permanently closed and is allowed to equilibrate to the biomass-per-re- cruit corresponding to zero fishing mortality Whole-stock fishing mortality shows a relationship to the actual fishing effort (the fishing effort in the open area only) in the open areas that is similar to that of rotational fishing. In both cases, closed area biomass dominates the whole-stock fish- ing mortality calculation at high fishing effort. The yield at high fishing effort is essentially derived from incoming recruitment, which is not sensitive to fishing effort for very high effort levels. Therefore, the whole-stock fishing mortality becomes nearly constant when effort is high. ^ NEFSC (Northeast Fisheries Science Center). 1999. Report of the 29th northeast reffional stock assessment workshop (29th SAW). Stock Assessment Review Committee (SARC) consen- sus summary of assessments. NEFSC Ref Doc. 99-14, 347 p. [Available from NEFSC, 166 Water St., Woods Hole, MA 02543.1 Hart; Yield- and biomass-per-recruit analysis of rotational fisheries 55 Figure 9 Whole-stock fishing mortality as a function of efTort (F^vg' ''"' ^^^ scallops with a 6-yr pulse rotation (solid line), and constant fishing with 10% of the area permanently closed (dashed line). The dotted line is the line y = x. The current situation for sea scallops in Georges Bank gives an even more extreme example of this phenomenon. About 80% of the biomass lies in the groundfish areas that have been closed to scallop fishing for most of the time since December 1994. Because these areas will be closed to scalloping in 2002, the whole-stock fishing mortality in this year cannot exceed the F^y^ reference point of 0.24. Therefore, according to the current overfishing definition (the whole-stock F is below F-^j^x'- ^^^ stock cannot be overfished. Nonetheless, the fishing mortality in the open areas may exceed F^i^x- resulting in growth overfishing in these areas. Thus, the stock in the open areas could be overfished from a yield-per-recruit perspective even if the whole-stock F is below F^y^. The opposite situation could also occur If scallops in the groundfish closed areas on Georges Bank were fished more than slightly above the Ff^^j^ = 0.24 reference point, the whole-stock fishing mortality would also be above this ref- erence point and overfishing would be considered to be oc- curring. However, an area that has been closed for a number of years should be fished harder, compared to an area that has never been closed, once the area is reopened in order to maximize yield-per-recruit. Thus, a strategy that would maximize yield-per-recruit might require a whole-stock F that would in some years be higher, and in some years lower, than the conventional overfishing reference point. A whole-stock fishing mortality rate may therefore not be the most appropriate metric for overfishing definitions when some areas are temporarily or permanently closed to fishing. Its value may not be representative of the yield- per-recruit that could be obtained at that level of fishing mortality. Furthermore, when most of the biomass is in closed areas, estimated whole-stock fishing mortality may be more sensitive to variations in recruitment and mea- surement error than to actual changes in effort. As an alternative to a whole-stock fishing mortality metric, the following considerations are suggested for a fishing effort measure that is compatible with yield-per- recruit calculations. (1) Stock from areas that are not fished in a given time period should not be included in the fishing mortality calculation for that time period. In a relatively sedentary stock, the amount of biomass in the closed areas is irrelevant in determining the yield-per- recruit that will be obtained from the stock in the open areas. (2) Time-averaging of fishing mortality in the open areas is required to take into account the previous fish- ing history of the area. An area that has been closed for a number of years needs to be fished harder once opened than an area that has been continuously fished in order to maximize yield-per-recruit. Based on these considerations, the time-averaged fish- ing mortality computed from the open areas only, F.wg' is an appropriate measure of fishing mortality in fisher- ies managed by using rotational or indefinite closures. It is natural to take the averaging period equal to the rotational period p. With this metric. Fy^,^-^ is only slightly sensitive to the rotational period p and completely insensi- tive to the level of closed area biomass. Indeed, even if no closures existed, but fishing effort varied with time, it may still be advisable to employ a time-averaged fishing mor- tality because the previous history of fishing mortalities strongly affects the level of effort required to maximize future yield-per-recruit. If an area has been fished harder than Fm.yx for a number of years, so that the population 56 Fishery Bulletin 101(1) size-structure in this area is smaller than the equilibrium size-structure obtained by fishing at F^^^, then fishing the next year at a level somewhat below i^M^x will improve long term yield-per-recruit. Similarly, if an area has been fished below F^^j^, so that its size structure is larger than what would occur when fishing at a constant rate of Fj^j^^^, then it may be optimal to temporarily fish at a level higher than F,^^. In summary, rotational fishing can improve yield- and biomass-per-recruit for long-lived sedentary stocks such as sea scallops. Rotational management can be part of a precautionary strategy because it can help alleviate the effects of growth and recruitment overfishing. Rotational management will however require a rethinking of conven- tional yield-per-recruit reference points. Acknowledgments I would like to thank T. Kenchington for discussions regarding the equivalence of long-term survivorship under rotational and constant fishing. This paper also benefitted from discussions with and comments from P. Rago, L. Jacobson, S. Murawski, F. Serchuk, A. Applegate, and the reviewers. Literature cited Beverton, R. J. H.. and S. J. Holt. 1957. On the dynamics of exploited fish populations, 533 p. Chapman and Hall, London, United Kingdom. Botsford, L. W. 1981. Optimal fishery policy for size specific, density-depen- dent population models. J. Math. Biol. 12:265-293. Botsford, L. W., J. F. Quinn, S. R. Wing, and J. G. Brittnacher 1993. Rotating spatial harvest of a benthic invertebrate, the red sea urchin, Strongylocentrotus franciscanus. In Pro- ceedings of the international symposium on management strategies for exploited fish populations. Alaska Sea Grant Report AK-SG-93-02, p. 409-428. Alaska Sea Grant Pro- gram. Anchorage, AK. Caddy, J. F. 1973. Underwater observations on tracks of dredges and trawls and some effects of dredging on a scallop ground. J. Fish. Res. Board Can. 30:173-180. 1975. Spatial models for an exploited shellfish population, and its application to the Georges Bank scallop fishery. J. Fish. Res. Board Can. .32:1305-1328. 1993. Background concepts for a rotating harvesting strat- egy with particular reference to the Mediterranean red coral, Corallium rubrum. Mar Fish. Rev. 55:10-18. Caddy,J. F,andJ. C. Soijo. 1998. Application of a spatial model to explore rotating harvest strategies for sedentary species. Can. Spec. Publ. Fish. Aquat. Sci. 125:3,59-365. Campbell, A., R. M. Harbo. and C. M. Hand. 1998. Harvesting and distribution of Pacific geoduck clams, Panopea abrupta. in British Columbia. Can. .Spec. Publ. Fish. Aquat. Sci. 125:349-358. Clark, C.W. 1990. Mathematical bioeconomics. The optimal manage- ment of renewable resources, 2"'' ed., 386 p. Wiley, New York, NY. De Klerk, P, and M. Gatto. 1981. Some remarks on periodic harvesting of a fish popu- lation. Math. Biosci. 56:47-69. Garcia, S. 1984. Modelisation et exploitation rationnelle des stocks de corail precieux: une premiere approche. FAO Fish. Rep. 306:109-121. Gribble, N., and M. Dredge. 1994. Mixed-species yield-per-recruit simulations of the ef- fect of seasonal closure on a Central Queensland coast prawn trawling ground. Can. J. Fish. Aquat. Sci. 51:998- 1011. Hart, D. R. 2001. Individual-based yield-per-recruit analysis, with an application to the Atlantic sea scallop, Placopecten magel- lanicus. Can. J. Fish. Aquat. Sci. 58:2351-2358. Heizer, S. 1993. "Knob cod"-management of the commercial sea cucum- ber fishery in British Columbia. J. Shellfish Res. 12:144- 145. Lai, H., and A. Bradbury. 1998. A modified catch-at-size analysis model for a red sea urchin iStrongylocentrotus franciscanus) population. Can. Spec. Publ. Fish. Aquat. Sci. 125: 85-96. MacDonald, B. A., and R. J. Thompson. 1986. Production, dynamics and energy partitioning in two populations of the giant scallop Placopecten magellanicus (Gmelinl. J. Exp. Mar Biol. Ecol. 101:285-299. McCallum, H. I. 1988. Pulse fishing may be superior to selective fishing. Math. Biosci. 89:177-181. Merrill, A. S., and J. A. Posgay. 1964. Estimating the natural mortality rate of sea scallop. Res. Bull. Int. Comm. N.W. Atlantic Fish. 1:88-106. Myers, R. A., S. D. Fuller, and D. G. Kehler 2000. A fisheries management strategy robust to ignorance: rotational harvest in the presence of indirect fishing mortality Can. J. Fish. Aquat. Sci. 57:2357-2362. Quinn, T J., and R. B. Deriso. 1999. Quantitative fish dynamics, 542 p. Oxford U Press. New York, NY, and Oxford, United Kingdom. Reed, W. J. 1986. Optimal harvesting models in forest management — a survey. Natural Resource Modeling 1:55-79. Serchuk, F M., P W. Wood, J. A. Posgay, and B. E. Brown. 1979. Assessment and status of sea scallop ^Placopecten magellanicus) populations of the northeast coast of the United States. Proc. Natl. Shellfish. Assoc. 69:161-191. Sluczanowski, P. R. 1984. A management oriented model of an abalone fishery whose substocks are subject to pulse fishing. Can. J. Fish. Aquat. Sci. 41:1008-1014. Tuck, G. N., and H. P. Possingham. 1994. Optimal harvesting strategies for a metapopulation. Bull, Math. Biol. .56: 107-127. Hart: Yield and biomass-per-recruit analysis of rotational fisheries 57 Appendix Basic yield-per-recruit model This appendix describes the basic yield-per-recruit model used for a cohort. In this model, recruits start at a specified shell height (or length) /((,. The shell height is converted into a starting age Oq by using a von Bertalanffy growth equation. The shell height at time t is also obtained by using the von Bertalanffy growth cui-ve. The shell height is converted into a meat weight by using a shell-height/ meat weight formula: w = exp(o -(- h ln(/?)), (7) where w and h are in units of grams and millimeters, respectively. Natural mortality occurs at a rate M, assumed for these simulations to be constant for all ages (M=0. 1 ). The fishing mortality rate F(h} on a scallop of shell height /; is given by Fih) = FffJ(h), where Fq is the fully recruited fishing mortality rate and J(h> is the selectivity of the gear. J(h) was taken to be if /; is less than a minimum shell height ^min' 1 'f ' '^ greater than a fully recruited threshold size hr^ii, and interpolated linearly as Jih)-. h-K H, full Aim (8) ^f'min < '' < /'full- Individuals that are caught by the gear but are smaller than a maximum cull size hj, are dis- carded and are subject to a discard mortality d. In these simulations, d is taken to be 0.2 (DuPaulM; however, the results are not very sensitive to the exact value of this parameter All individuals caught at a size greater than hj are assumed to be landed and are included in the total yield. F^(h) denotes the rate at which scallops of shell height h are caught and retained (i.e. not discarded). The possibility has been raised that some scallops may be killed but not captured by the gear (Caddy, 1973; My- ers et al. 2000). Caddy (1973) estimated that 15-20% of the scallops remaining on the bottom in the path of a scallop dredge are killed but not captured by the dredge. Murawski and Serchuk^ estimated that less than 59; of the scallops remaining in the path of the dredge suffered incidental (noncatch) mortality. In order to use the above studies to estimate the relationship between incidental fishing mortality F, and the fully recruited capture fish- ing mortality rate F^^, it is necessary to know the efficiency '* DuPaul, W. D. 2000. Personal commun. Virginia Institute of Marine Science, P.O. Box 1346, Gloucester Point, VA 23062- 1346. ■• Murawski, S. A., and F. M. Serchuk. 1989. Environmental effects of offshore dredge fisheries for bivalves. ICES CM. 1989/K:27. e of the dredge on a fully recruited individual. Denote by ; the fraction of scallops that suffer mortality among those that were in the path of the dredge but that were not caught, so that / is estimated at 0.15-0.2 by Caddy (1973), and less than 0.05 by Murawski and Serchuk.' The ratio R of fully recruited scallops in the path of the dredge that are caught to those killed but not caught is R = el\i(l-e)\. (9) If fully recruited scallops suffer capture fishing mortality at rate F^, then the rate of incidental fishing mortality will be F, = FJR F,J (l-e)/e. (10) If e is taken as 50% (estimated as the average scallop dredge efficiency on Georges Bank^), then F, would be in the range 0.15 F^ to 0.2 F,, according to Caddy (1973) and less than 0.05 Fg according to Murawski and Serchuk. -^ To ascer- tain the effects of incidental fishing mortality on the 5aeld- per-recruit calculations, model runs were performed with no incidental mortality, and also when F| = 0.15 F^; incidental fishing morality was applied to all size groups. Let Z(h) be the total mortality rate at shell height /; (i.e. the sum of natural mortality, and discard, indirect, and landed fishing mortality). Then the fraction of recruits re- maining t years after the beginning of the simulation is S(n = exp Z{T}dT (11) Total yield- and biomass-per-recruit are calculated by the formulas: Y= [s(t)F^{h{t))w(h(t}}dt B= [S(t)w{hit))dt, (12) (13) where a^ = the ending age of the simulation, taken to be 30 -^a^. For convenience in these simulations, a^ is taken to be 2 years; this age is assumed to correspond to a shell height of precisely 40 mm. In the rotational simulations reported in this study, the fully recruited landed fishing mortalities Fjh> (h>hf^i^) are assumed to vary periodically and are given in year k by F^^^, where > is the year that the cohort reaches the starting age Oq. Rago, P. J. 2001. Personal commun. Northeast Fisheries Science Center, 166 Water St., Woods Hole. MA 02543. 58 Abstract— Southern bluefin tuna( SBT) ^Tluinnus maccoyit) growth rates are estimated from tag-return data associ- ated with two time periods, the 1960s and 1980s. The traditional von Ber- talanffy growth model (VBG) and a two-phase VBG model were fitted to the data by maximum likelihood. The traditional VBG model did not provide an adequate representation of growth in SBT, and the two-phase VBG yielded a significantly better fit. The results indicated that significant change oc- curs in the pattern of growth in rela- tion to a VBG curve during the juvenile stages of the SBT life cycle, which may be related to the transition from a tightly schooling fish that spends sub- stantial time in near and surface shore waters to one that is found primarily in more offshore and deeper waters. The results suggest that more complex growth models should be considered for other tunas and for other species that show a marked change in habitat use with age. The likelihood surface for the two-phase VBG model was found to be bimodal and some implications of this are investigated. Significant and substantial differ- ences were found in the growth for fish spawned in the 1960s and in the 1980s, such that after age four there is a difference of about one year in the expected age of a fish of similar length which persists over the size range for which meaningful recapture data are available. This difference may be a density-dependent response as a con- sequence of the marked reduction in the SBT population. Given the key role that estimates of growth have in most stock assessments, the results indicate that there is a need both for the regu- lar monitoring of growth rates and for provisions for changes in gi'owth over time (possibly related to changes in abundance) in the stock assessment models used for SBT and other species. Estimating long-term growth-rate changes of southern bluefin tuna (Thunnus maccoyii) from two periods of tag-return data William S. Hearn CSIRO Marine Research Private Bag No. 5 Wembley, Western Australia 6020 Australia E-mail address, bill.hearnigimanne.csiro au Thomas Polacheck CSIRO Marine Research GPO Box 1538 Hobart, Tasmania 7001 Australia Manuscript accepted 28 May 2002. Fish. Bull 101:58-74(2003). Estimating growth rates has been a major focus of fisheries research throughout the twentieth century, and a large body of literature exists on the topic (e.g. Lee, 1912; Ford, 1933; Wal- ford, 1946; Manzer and Taylor, 1947; Allen, 1966; Yukinawa, 1970; Pitcher and MacDonald, 1973; Kimura, 1980; Fournier et al., 1990). This literature reflects, at least in part, the funda- mental importance of information on growth rates in stock assessments and the subsequent provision of man- agement advice for commercially har- vested fish populations. For example, growth information is required for yield-per-recruit analyses and for the estimation of spawning stock biomass in the estimation of stock-recruitment relationships. In addition, for a number of species, estimates of growth rates have been the primary or only source of information that can be used to esti- mate the age of individual fish and the age distribution of commercial catches (particularly for tropical species and for tunas and billfish). Such informa- tion on age is a critical component required in the analyses and models used to assess and manage these fish stocks (Bayliff 1991; Clay, 1991; Caton. 1991; Wild, 1994; Wild and Hampton, 1994; Polacheck etal.'). Almost all the work on modeling growth has centered on modeling growth rate as a continuous, smooth, monotonically decreasing function of age, and the von Bertalanffy (1938) growth (VBG) equation, and its ex- tensions, have been the most common approach used. In addition, the growth process has frequently been modeled as static. Temporal variations in aver- age growth for fish of the same size, or age, (due, for example, to changes in the physical environment or popu- lation density) are often ignored or considered to be relatively minor (with some notable exceptions — e.g. Le Cren, 1958; Southward, 1967; de Veen, 1976; Toresen, 1990; Ross and Nelson, 1992; Kaeriyama, 1996; Sinclair and Swain, 1996). For the large pelagic tunas and billfishes, the von Bertalanffy growth equation and extensions has been the standard used for modeling growth (Bayliff, 1980). For a variety of tuna species, numerous growth studies have been conducted, and generally a rea- sonable range of parameter values has been estimated (e.g. see the sets of pa- rameter values estimates for the eight scrombrid species in Bayliff, 1980). ' Polacheck, T, A. Preece, A. Bctlchem, and N. Klaer 1998. Treatment of data and model uncertainties in the assessment of southern bluefin tuna stocks. In Fish- ery stock assessment models (F Funk et al", eds.), p. 613-637. Alaska Sea Grant College Program Report AK-SG-98-01. Univ. Alaska, PO. Box 755040, Fairbanks, AK 99775-5040. Hearn and Polachek: Long-term growth rate changes in Thunnus maccoyii 59 Bayliff ( 1988) also investigated regional growth differenc- es in Pacific skipjack and yellowfin tunas. Interpretation as to whether any differences found are merely an artifact of the data collection or procedures used or whether they reflect real temporal or spatial difference has generally not been possible because the basic data (e.g. tagging, hard parts, length-frequency data), data collection procedures, analytic approaches, and the areas and time periods from which the data were collected have varied greatly among studies. For southern bluefin tuna (SBT) (Thunnus maccoyii), extensive juvenile tagging programs were conducted in the 1960s and 1980s, and a large number of returns with measured lengths were recovered. From both periods, some returns were received after times at liberty in excess of 10 years. These two sets of tagging experiments provide the basis for the direct comparison of growth over a time span of 30 years. Also, because of the large number of tags returned in these studies, a more detailed examination of the adequacy of the von Bertalanffy growth equation as a model of the growth process is possible than with many data sets. These tagging data (primarily those from the 1960s) have been used as a basis for a number of analyses of growth rates (Murphy, 1977; Kirkwood. 1983; Hearn, 1986; Hampton, 1991; Lucas-). In the present paper, we present results of the analyses of the growth increment data from these two sets of tagging experiments. We ex- amine these data both in terms of 1) whether SBT growth differed between the tagging periods and 2) whether there was a change in the growth process between adult and juvenile SBT (i.e. whether a more complex model than the simple von Bertalanffy equation is required to provide an adequate description of SBT growth). The results presented here incorporate and build upon the already cited published analyses of these tag-return data, unpublished reports, and discussions of SBT growth in scientific meetings on SBT (e.g. Hearn and Hampton'; Hearn and Polacheck''; Anonymous^). - Lucas, C. 1974. Working paper on southern biuefin tuna pop- ulation dynamics ICCAT ( Intenational Commission for the Con- servation of Atlantic Tunas), SCRS/74/4. Collective Volume of Scientific Papers, vol. HI, p. 110-124. [Available from CSIRO Marine Laboratories, GPO Box 1538, Hobart, Tasmania 7001, Australia.) 3 Hearn, W. S. and J. Hampton. 1990. SBT growth change. Ninth trilateral meeting on SBT, Hobart, Australia, September 1990, SBFWS/90/8. 19 p. (Available from CSIRO and the Com- mission for Conservation of Southern Bluefin Tuna. P.O. Box 37, Deakin West, ACT 2600, Australia.) ■> Hearn, W. S., and T Polacheck. 1993. Estmiating SBT age- at-length relations for the 1960s and 1980/90s. Twelfth trilat- eral meeting on SBT, Hobart, Australia, October 1993. SBFWS/ 93/4, 21 p. [Available from CSIRO and the Commission for Conservation of Southern Bluefin Tuna, P.O. Box 37, Deakin West, ACT 2600, Australia.) ■' Anonymous. 1994. Report of the southern bluefin tuna trilat- eral workshop; Hobart, Australia, January/February 1994, 161 p. [Available from CSIRO and the Commission for the Conser- vation of Southern Bluefin Tuna, P.O. Box 37, Deakin West, ACT 2600, Australia. [ Background: the SBT stock and fishery SBT is a highly-migratory species that begins to spawn at about 10-12 years of age in waters south of Java during the southern summer, mainly from September to April (Farley and Davis, 1998). During the first year of life they tend to be transported south by the tropical Leeuwin Current to inshore waters between Perth and Esperance, Western Australia. From ages 1 to 4 years, they appear to mainly inhabit, at least in the summer months, the waters off the Great Australian Bight, southern New South Wales (NSW) and eastern Tasmania. Many move to oceanic waters during the winter months and apparently progressively so as they age. By five years of age almost all have migrated to oceanic waters between 30° and 50°S at all longitudes, but mostly in the Eastern Hemisphere. Substantial surface fisheries operated off the south coast of Western Australia from 1969 to the mid 1980s, off the south coast of NSW from 1963 to the early 1980s, and off South Australia from 1964 to the present. Since 1959 a major Japanese longline fishery has operated in oceanic waters between 30° and 50°S, mainly from the mid-Atlan- tic and westwards to a few degrees west of New Zealand. Materials and methods Tagging programs Description Large numbers of tagged fish were released by CSIRO staff in the period from 1959 to 1968 and again in the period from 1980 to 1984. The releases from these two periods are used in our present study. Most of the tagged fish were initially caught with pole-and-Iine gear with barbless hooks, although a relatively small number were caught with troll lines. After a fish was hooked, it was hauled aboard the vessel and placed on a measuring board (in the 1960s) or a vinyl cradle (in the 1980s), and its nose to caudal fork length was measured. The fish was then tagged by an operator who inserted a 12-cm plastic spaghetti dart tag into the fish about 4 cm to the rear of the second dorsal fin on either side of the fish and rer- eleased it into the water within about 30 seconds. After 1963 almost all fish were double tagged. The tag numbers and length of each fish were recorded, together with loca- tion and date of release. This information was later trans- ferred to a computer database. Tagging operations in both the 1960s and 1980s were concentrated in the nearshore, surface-water fisheries bordering the central and western southern coast of Aus- tralia and the southern coast of NSW. In the 1980s no tags were released from the NSW coast area because this com- ponent of the fishery had collapsed, and surface schools of juvenile SBT could no longer be found (Caton, 1991). The South Australian tagging took place in the Great Australian Bight or in the adjacent shelf waters generally between longitudes 128° and 136°E. Releases in Western Australia occurred in the Albany (between longitudes 112° and 119°E) and Esperance (between longitudes 119° and 125°E) areas. There were 33,309 juvenile SBT tagged by 60 Fishery Bulletin 101(1) CSIRO personnel during the 1960s (1959 to 1968) and 10,743 during the early 1980s (1980 to 1984). Of these fish, 1972 and 4280, respectively, were later recaptured. On recapture, fishermen recovered the tags and re- corded the fish's length (if measured), location, and date. The tags with the recorded information were returned to the scientific staff at CSIRO, who then provided a reward. Most of the recapture lengths were measured by fisher- men or factory staff but about 31% were measured by scientists. Those measured by scientists cannot be con- sidered a representative sample. In particular, all of the measurements for longer-term recaptured fish come from fishermen aboard Japanese longline vessels. In addition to length, longliners often reported the dressed weight and sometimes the whole weight, or both, of recaptured fish. In the 1960s Australian fishermen seldom reported any weight measurements, but in the 1980s they commonly reported the whole weight of recaptured fish. Data selection The tagging experiments were conducted mainly within a narrow range of months at each site; therefore returns within a few months would be most strongly influenced by the seasonal differences found in SBT growth (Hearn, 1986; Burgess et al., 1991; Leigh and Hearn 2000). A nine-month period at liberty coincides with a low frequency in the times at liberty for the experi- ments; therefore we excluded data from analyses with less than 270 days at liberty. We also excluded data for which fish were tagged by fishermen, or when the recovery length, year, or month were reported by the tag finder to be unknown or uncertain. Previously reported weight-length relationships (Wara- shina and Hisada, 1970; Hampton, 1986; Robins^) were used to identify and screen out dubious recapture data. The details of the screening procedures are documented by Hearn" and Anonymous.-^ Longline recaptures were excluded if the expected weight of a recaptured fish for its reported length was either less than 2/3 of the reported weight or greater than 1.5 times the reported weight. Some of the major inconsistencies were thought to be due to measuring the length of a fish without its tail or with- out its head (Lucas^). For surface fish in the 1980s, a high proportion of the weight-length data for recaptured fish from four vessels was inconsistent with the weight-length relationships noted above. All tag-return data from these four vessels were excluded. Another 2.5% of the 1980 data were excluded because of highly unlikely values for the ra- tio of the reported weight to length of the recaptured fish. '" Robin.s, J. P. 1963. Synopsis of biolofjical data on .southern bluefin tuna, Thiinnus ihvnnus maccoyn (Ca.stlcnaui 1872. FAO Fisheries Report 6(2), p. 562-587. [Available from CSIRO Marine Laboratories, GPO Box 1538, Hobart, Tasmania 7(X)1, Australia.) " Hearn, W. S. 1982. Fish tagging: data processing, editing and storage. In CSIRO data base for .southern bluefin tuna {Than- nus maccoyii (Castlenau)) (J. Majkowski, ed.l. p. 8-9. CSIR(J Marine Laboratories, Rep. 142. lAvailablc from CSIRO Marine Laboratories, GPO Box 1538, Hobart, Tasmania 7001, Australia.] For the screening methods used, no assumption was made about the underlying growth curve, and these meth- ods were designed so that they would not induce a bias into the results. The selection process yielded data sets that were sufficiently large for valid analyses, being 730 and 1450 for the 1960s and 1980s data sets, respectively. Note that for other tuna species the selection process used in our study (particularly the deletion of recaptured fish with short times-at-liberty) may cause problems because of smaller data sets (e.g. skipjack and yellowfin tunas in Bayliff, 1988). Experimental assumptions The use of the tag-return in- crement data for estimating growth rates requires the fol- lowing assumptions about the tagging protocols and data collection procedures: 1 Tagging does not retard growth. 2 The tagged fish are uniquely and correctly recorded at release and recapture. 3 The lengths of fish are measured without bias at re- lease and recapture. 4 A wide range of fish sizes are represented, in recap- tures at least. 5 There are no significant size-selection processes for fish within similar age ranges. With respect to tagging effects, Hampton (1986) and Hearn (1986) have shown that there can be a significant weight loss of 7-12% for tagged fish in the first month after release. However, tagged fish recover this weight loss within a year at liberty, and there is no apparent difference between tagged and untagged fish after this time (Hearn, 1986). (There is little information available on weight loss of tagged fish at liberty between one month and one year) In terms of length, Hearn and Hampton' could not detect a reduction of growth from growth increment residuals in the tag-return data even within the first 30 days after release. Limited data from the effect of handling and tag- ging fish in commercial farm pens indicated no retarda- tion in growth in length after 150 days. These farm fish did show a loss in weight when first caged, but the weight was regained over a period of a few months (Anonymous''); therefore we do not think that tagging had any substan- tial effect on the growth rate of tagged fished in our study. With respect to the other assumptions, all fish were tagged with uniquely numbered tags. During tagging operations, tags were arranged in blocks of sequential numbers to avoid confusion and the misrecording of tag numbers. Return of the physical tag was required for fishermen to obtain the reward, and the double tagging of almost all fish since 1963 has allowed cross verification of tagging numbers, which allows little scope for error in the record- ing of tag numbers. Approximately 23*7^ of the length mea- surements for the selected recaptured fish were measured by scientists. Mainly due to the deletion of short-term recaptured fish (i.e. < 270 days), this is less than that for all data (31%). For the fishermen-measured lengths, there was no reason to suspect any consistent bias, and comparison of the residuals for fishermen- and scientist- Hearn and Polachek; Long-term growth rate changes in Thunnus maccoyii 61 measured returns in the fitted models below did not indi- cate any systematic pattern. The recaptured fish used in our study ranged in size from 60 to 175 cm, although the number of fish in the larger size ranges was relatively small — less than 2% were larger than 140 cm. (The con- sequences of the small number of fish in the large-size category are discussed below. ) Within both the surface and longline fishery, a range of sizes and age classes is har- vested within a single operation. No indication exists that within the size range encompassed by a cohort at a given age, there existed significant gear or fishery size selectiv- ity. Overall, the above basic assumptions seem reasonable in modeling growth from these SBT tagging data. (1991). In this model, fish grow according to one model (or parameter set) up to a certain length and according to another thereafter. In our analyses, we assumed that fish have VBG throughout their lives but grow according to one set of VBG parameters (L_j and k-^) up to length L* and according to a second set (L_.^ and k.-,) at larger sizes, the two-phase VBG model. Thus, the predicted length as a function of time for this model is L , l-e-*>"-'°' for tt* Analytical methods Models Two basic models were used to analyze growth information from the tag-return data. The first was the simple VBG model: where /, =L__(l-e-*"-'"'), (1) = the length that fish grow to asymptotically; = the length of a fish at age (or time) t; = the exponential rate at which the growth rate slows; and = the hypothetical age (or time) when a fish is of length zero. where t* = the predicted time for a fish to reach L*. Note that t* can be solved for in terms of four of the parameters of the model ltQ,k^, L^j, and L*): where k. (o = h + -r^°s Mogfl ^ ^~: ^«i (4) and 1,= the length of a fish at the time of tagging, t^ When applied to tag-return data, this equation can be used to predict the growth increment as a function of the length at release and the time at liberty: As with this simple VBG model. Equation 3 can be solved to predict the growth increment as a function of the release length and the time of liberty 5t = t.2-t{. SI = {L_ -l)(l-e -kSl) (2) where SI = the growth increment; St = the time at liberty; and / = the length of release. Note, in this study we simplified the growth model by not accounting for seasonal growth. However, data on recap- tured fish with short times at liberty were specifically deleted to ensure that our results were robust after this simplification. Preliminary analyses of the tag-return data suggested that a simple and time invariant von Bertalanffy growth model may not provide an adequate description of the growth rate for SBT. These preliminary analyses suggested S = (L^,i-/,,)(l-e"*''') i{t2(l-e"''^'*) if/,>^*. (5) It should be noted that in some of the analyses considered below, the estimate of L^j did not converge (i.e. the esti- mate for L. , was essentially infinite). In such cases, the estimated growth rate is linear, with growth rate /?,, and for the first phase we replaced the von Bertalanffy growth function with a simple linear one: « = /?,*, 1 Growth rates in the 1960s and the 1980s were not equal; 2 There were systematic deviations from a VBG curve, possibly corresponding to different growth processes or models for adults and juveniles. Consequently, in the present study, we considered a more complex model than the simple VBG and conducted sepa- rate, as well as combined, analyses of the tag data from the two periods. The more complex model selected was the two-phase growth model developed by Bayliff et al. and t* = t^-(L*-l,^)/Ry (6) Model-fitting procedure A large body of literature exists on statistical approaches for estimating growth from tag- return data (e.g. Fabens, 1965; Sainsbury, 1980; Kirkwood and Somers, 1984; Francis, 1988; James, 1991; Hampton, 1991; Wang et al., 1995). The most appropriate approach depends on the error structure assumed for the model. We followed the maximum-likelihood approach and general error structure described by Hampton (1991). The mea- sured growth increment offish "/" is 62 Fishery Bulletin 101(1) SI, = E\a] + e, + e„ (7) where f, is due to measurement error in the observed growth increment (i.e. the combined effect of any errors in measuring the lengths at the time of release and recap- ture) and e, is due to process or model error The latter may be a function of /j, St, SI, and the model parameters. For the measurement error component, we allowed for different variances, depending upon whether the recap- tured fish was measured by an independent and scientifi- cally trained individual or by a fisherman. Scientifically trained individuals (i.e. scientists) included fishery observ- ers, port samplers, and CSIRO staff We assumed that the measurement error was normally distributed, with mean zero and variance o^-, where .v is one of /"or s for recaptured fish measured by fishermen or scientists. The choice of the functional form for the process error in growth models is a complex issue. One approach has been to consider that process error stems from variability among individuals in the expected value of the growth parameters (e.g. Sainsbury, 1980; Hampton, 1991; Wang et al., 1995). This approach in the case of the two-phase VBG model would result in many potential structures for the process error component because there could be indi- vidual variability in the expected value of any single or possible combination of parameters (of which there are 25 combinations). There is little theoretical basis for deciding which of these 25 combinations to use. As an alternative, we selected a more empirical approach. A function that increases with longer times at liberty seemed appropriate, and was also consistent with preliminary analyses. We ex- plored both linear and quadratic functional relationships between the times at liberty and the process error com- ponent. The quadratic term was found to be insignificant, and therefore we chose to report only results for a simple linear functional relationship, namely 0;;^St. Hence the cor- responding variance of the expected gi'owth increment of fish ( is V( ^^ ) = a^- + a,~ St^. It should be noted that without independent data on measurement error any constant component in process error would be totally confounded with the measurement error term in the model. Therefore, a^ should be considered as a combined measurement and process error term. Both a^ and (T„, were estimated empiri- cally by maximum likelihood tag increment data. Assuming a Gaussian error distribution, the likelihood function is =ni--,.^.p(.it«!j (8) The estimates of the parameters are found by minimizing ■ln(L) = ^^ ,„,p.„.,„.it«a! (9) D.E. Shaw, CSIRO Div Maths, and Stats.), which uses the Nelder and Mead (1965) method. Model selection The estimation of the full two-phase VBG models across both tagging periods contains 16 parameters (five model parameters plus three variance parameters for each time period). We examined a variety of alternative hypotheses to test whether the number of parameters could be reduced by eliminating some or equating them. For the model parameters, we considered whether the L„ or k terms were equal either between time periods or between the first and second phases within a time period. We also considered the simple VBG model, for which L* doesn't exist. For the L '■ parameter, we considered whether the esti- mates were different between the two time periods. We also examined models in which the value of L* was de- termined by assuming that the expected growth rate for a fish of length L* was equal for both growth phases (i.e. by assuming that the changes in growth rates as a function of length is a continuous function). Under this assumption (10) The minimum value was obtained for all models by using the minimizing subroutine MINIMD (programmed by This model is referred to as the continuous rate two- phase model in the "Results" section. However, this model is not smooth because it has a discontinuity in the deriva- tive of the growth rate atL*. For the variance parameters, we considered whether any of them could be eliminated and also whether a^ = Of. We used the log-ratio test and AIC criterion (Akaike, 1974) to identify the most parsimo- nious model. Results Best fits to the 1960s and 1980s data Table 1 contains the maximum likelihood solutions for various assumptions when fitting growth models to either the 1960s or 1980s tag-return data separately. Using the AIC criteria, we found the best-fit model for the 1960s tag- return data was one with linear growth in the first phase and with the change between the two phases at approxi- mately 74 cm (row 1, Table lA). The fit to this model com- pared with all other parameter combinations yielded both the lowest AIC and negative log-likelihood values. The fit, however, was only marginally better then the fit (row 3, Table lA) to the two-phase VBG curve with common k parameters (e.g. where the difference in the negative log- likelihood values is 1.21). Except for the first phase growth parameters, the estimates for the other parameters are nearly identical between these two models. This similar- ity reflects the fact that growth is nearly linear over the initial part of a VBG curve. Thus, by having a relatively high L. I (271 cm), essentially similar gj-owth rates can be achieved up through the 74 cm size range when ^j = ^-2, as compared with linear growth in the first phase. It should Hearn and Polachek: Long-term growth rate changes in Thunnus maccoyii 63 Table 1 (A) Estimation of SBT growth parameters, and tests, from 1960s tag-return data with time at liberty of at least 270 days; (Bl Estimation of SBT growth parameters, and tests, from 1980s tag-return data with time at liberty of at least 270 days, "na" = not applicable because this is the normal von Bertalanffy curve, i.e. with only one phase. Common parameters Number of parameters ^.., *, L^2 k2 L* a^ ^/ cr„, -Log likelihood AIC A none 6 22.23' 0.0000 210.90 0.1063 74.24 0.000 3.122 2.992 2055.53 4123.07- ^«l=^»2 6 212.73 0.1451 212.73 0.1044 75.75 0.000 3.134 2.999 2057.39 4126.77 6 271.35 0.1060 211.35 0.1060 74.71 0.000 3,130 2.997 2056.74 4125.49 5 172.67 0.1723 172.67 0.1723 na 2.201 3.782 2.855 2099.04 4208.08 continuous rate 6 114.14 0.4289 205.45 0.1128 81.55 0.000 3.203 3.021 2065.89 4143.78 «fm=0 7 760.47 0.03425 191.33 0.1330 70.00 3.478 5.258 0.000 2088.19 4190.39 0=0f 6 22.20 0.0000 209.65 0.1085 74.04 2.301 2.301 3.180 2068.33 4148.66 a=af=0 6 454.10 0.05660 214.26 0.1033 74.70 0.000 0.000 3.752 2071.57 4155.15 B none 8 226.70 0.1649 182.52 0.1841 84.90 2.305 4.501 3.018 4509.99 9035.97 L^,-L^, 7 183.09 0.2276 183.09 0.1832 85.65 2.266 4.497 3.031 4510.44 9034.88 «1=«2 i.l=i.2 h -h 7 210.24 0.1841 182.61 0.1841 84.99 2.276 4.492 3.0.30 4510.02 9034.03* 5 156.45 0.2884 156.45 0.2884 na 1.626 4.209 3.405 4530.88 9071.76 continuous rate 7 141.07 0.3590 182.25 0.1842 97.70 2.148 4.473 3.085 4514.54 9043.09 f^„,=o 6 206.71 0.1883 180.82 0.1883 84.85 4.149 5.920 0.000 4526.06 9064.11 ci=af 6 210.74 0.1858 182.23 0.1858 85.36 3.948 3.948 3.183 4529.25 9070.49 a=a^O 5 209.46 0.1875 181.49 0.1875 85.30 0.000 0.000 4.737 4545.25 9100.49 ' Here A, is zero,i.e. the growth rate is constant in the first phase; therefore we give the estimate of the growth - The least AIC value for estimates from the 1960s data. Na= not applicable because this is the normal von phase. ' The least AIC value for estimates from the 1980s data. rate instead ofL.j. Bertalanffy curve, i.e. w th only one also be noted that the two-phase VBG model with common L^ (row 2, Table lA). was very similar to the common k parameterization, reflecting the high correlation between L„ and k in the VBG models. For the 1960s, the continu- ous two-phase VBG model was rejected, P< 0.005 (row 5, Table lA). For the 1980s data, the best-fit model based on the AIC values was for the two-phase VBG model with a common value for the k parameter in both phases (row .3, Table IB). The estimate of the size at which the change between the two phases occurs was 85 cm (compared to the estimate of 74 cm for the 1960s data). As with the results of the 1960s data, the common-/^ model, common- L model, and the full two-phase model yielded very similar values for both the likelihood and parameter estimates in the second phase, but not for those in the first phase. This similarity reflects the high correlation between k and L in the VBG model, so that over the limited size range below L* nearly identical growth rates can be achieved in the common-^ model by decreasing the value of L^j. For the 1980s (as with the 1960s), the continuous two-phase VBG model (7 parameters), was rejected, P< 0.005 (row 5, Table IB). For both the 1960s and 1980s data, the two-phase model provided a substantially and significantly better fit to the tag return data than a simple VBG model. This can be seen in Table 1 (A and B) by comparing the negative log-likelihood and AIC values for the simple VBG model (row 4) with any that include a two-phase component, particularly the continuous rate two-phase VBG model. We also fitted a smooth Richards' (1959) growth model (a generalization of the VBG model) to the data, which fitted better than the simple VBG model, but worse than the two-phase VBG models. Note, however, that the log-ratio test and AIC criterion may not be fully applicable for testing the differences between the simple and two-phase VBG models because the simple VBG model can arise in more than one way as a submodel of the two-phase model (e.g. with common L . and k parameters or from L* equaling zero or infinity) (Davies, 1977, 1987). Nevertheless, the large magnitude of the differences in the log-likelihood values indicates a significance difference. For the 1960s data, it should be noted that the scientist measurement error (crj was estimated to be essentially zero when it was included as an explicit term in several of the models. In these cases, we refitted the models exclud- ing this parameter. Common sense dictates that measure- ment errors would not be zero. The most informative data 64 Fishery Bulletin 101(1) Table 2 Comparison of SBT growth parameter estimates for the 1960s and 1980s, between absolute maximum likelihood and local maxima | likelihood. Common Number of -Log parameters parameters ^-1 ^ L„., k. L* -30 ( 1 t r 1 1 1 1 E D 2 4 6 8 10 12 14 c Time at liberty (yr) O) c ^ o 30 - 3 D "D U> Q <> ^ O OJ % o tr 20 10 ,\:->t^^^i.' .> o o o o 00 ^wsp^j^^^yfj o o o o -10 oo ^ o o o -20 o -30 1 1 1 1 1 1 I 60 80 100 120 140 160 180 Expected recapture lengtti (cm) Figure 1 (continued) Hearn and Polachek: Long-term growth rate changes in Thunnus maccoyii 67 2100 A 2095 - /' 2090 - / 2085 1 / 2080 \ / 2075 \ / 2070 \ 1 2065 \ /"\ _,.,--''' 2060 - ' 2055 - ■o o o J= OJ S" _l 1 2050 5 4550 1 1 1 1 1 1 1 1 1 60 70 B 80 90 100 110 120 130 140 4545 - 4540 - 4535 - 4530 - ~-N, 4525 - 4520 - / —■'' 4515 4510 ^ ^^s -•'''' 4505 - 4500 5 1 1 1 1 1 1 1 1 1 60 70 80 90 100 110 120 130 140 L- (cm) Figure 2 Negative log-likelihood value as a function ofL'; (Al for the 1960s; (Bl for the 1980s. 68 Fishery Bulletin 101(1) 350 300 250 200 150 100 50 rr 60 350 300 250 200 150 100 50 B 60 70 70 80 80 90 100 90 100 /.• (cm) 110 110 120 130 120 130 Figure 3 F"requoncy distribution for the estimates of L* from lOOO bootstrap simulation results: (A) for the 1960s; iBiforthe 1980s. Hearn and Polachek; Long-term growth rate changes in Thunnus maccoyii 69 Table 3 Southern bluefin tuna growth parameters and tests, from jointly analyzing the 1960s and 1980s data. Common Number of -Log parameters parameters ^-, ^ l^.. k.^ L* 0, ""t fJ,„ likeHhood AIC none (60) 13 22.23' 0.0000 210.90 0.1063 74.24 0.000 3.122 2.992 (80) 210.24 0.1841 182.61 0.1841 84.99 2.273 4.496 3.030 6565.55 13157.10 CT,„60=CT„,80 12 22.23 0.0000 210.98 0.1062 74.25 0.000 3.115 2.998 210.31 0.1840 182.67 0.1840 85.00 2.309 4.514 2.998 6565.56 13155.11 L60,,,,=Z.80.,, 12 213.29 0.1462 210.49 0.1069 75.03 0.000 3.133 3.001 213.29 0.1804 184.20 0.1804 85.00 2.323 4.518 3.001 6567.43 13158.85 L60„i=L60_j, =Z,80„, U 211.41 0.1480 211.41 0.1060 75.08 0.000 3.128 3.004 211.41 0.1827 183.24 0.1827 85.00 2.315 4.515 3.004 6567.44 131.56.87 /t60,=A'80| 12 185.24 0.1804 210.46 0.1069 75.11 0.000 3.119 3.008 213.26 0.1804 184.21 0.1804 85.01 2.325 4.510 3.008 6568.07 13160,15 L60.,,=L80„2 11 22.20 0.0000 195.25 0.1232 73.74 0.000 3.120 3.010 233.55 0.1584 195.25 0.1584 85.22 2.302 4.516 3.010 6568.89 13159.78 /f60,=*802 11 23.85 0.0000 186.76 0.1374 70.48 0.000 3.097 3.044 258.76 0.1374 207.04 0.1374 86.23 2.262 4.495 3.044 6575.52 13173.03 L*60=Z,*80 12 110.51 0.4707 196.71 0.1269 85.45 0.000 3.196 3.014 211.13 0.1827 182.97 0.1827 85.45 2.306 4.505 3.014 6574,09 13172.19 CT,60=a^80 12 22.22 0.0000 210.59 0.1072 74.10 1.602 3.023 3.068 209.66 0.1842 182.49 0.1842 84.94 1.602 4.472 3.068 6570.46 13164.92 0^60=0)80 11 22.28 0.0000 210.43 0.1064 74.23 0.000 4.174 2.975 210.40 0.1844 182.62 0.1844 85.02 2.394 4.174 2.975 6577.96 13177.91 ' Here k^ is zero. i.e. the growth rate is constant in the first phase; therefore we give the estimate of the growth rate insteac ofZ..,. SBT growth rates in the 1960s are estimated to be less than those in the 1980s up to 144 cm (Fig, 4), Comparison of the 1960s and 1980s expected growth curves over time for a 55-cm fish are presented in Figure 5. In making this comparison, we assumed that a 55-cm fish is approximately one year of age (Anonymous"') and that size at age one did not change between the 1960s and 1980s, as supported by length-frequency data from these two time periods (Leigh and Hearn, 2000; Anonymous''). Thus, Figure 5 can also be considered as an estimate of the expected length-at-age curve. Figure 5 indicates that the overall expected growth was significantly faster in the 1980s than in the 1960s, so that a fish of 55 cm or age 1 would take approximately four years in the 1960s to achieve the same length that would have been achieved in three years in the 1980s. A feature of the best-fitted estimated growth param- eters is that the expected growth curves intersect at -170 cm, so that after age 13 a fish from the 1960s is estimated to be larger than a fish from the 1980s. This crossover is driven primarily by the difference in the estimates of L .;. The standard log-likelihood test indicates a low probabil- ity, P=0.01, that L,^., for the 1960s and 1980s are the same. However, the analyses of the bootstrap estimates of L ,., indicate that the estimates are bimodal, reflecting the bi- modal distribution of L*. Random sampling from the boot- strap distributions for L.2 showed that in G.V/c of cases the 1960s L„,2 estimate was less than the 1980s estimate. For a two-sided test at the 5'7c significant level, at least 2.5'7f (and at most 97.5%) of the bootstrap samples would have been expected to have the 1960s L ., less than that of the 1980s to justify the hypothesis that the two L^., are equal. Thus, based on the bootstrap results, the hypothesis of equality cannot be rejected. Most of the 6.1% of cases are associated with the 1960s L,,., less than 180 cm, which are in turn associated with the upper mode of L* in Figure 3A, i,e, near L* = 91 cm. It is worth noting that only three recapture lengths were greater than 170 cm. There are, therefore, very minimal data for estimating growth rates beyond 170 cm and for precisely estimating L^.,. Discussion The results in this study indicate that the traditional VBG model does not provide an adequate representation of growth in SBT. There appears to be a significant change in the pattern of growth in relation to a VBG curve during the juvenile stages of the SBT life cycle. This, in turn, may be related to the transition from a tightly schooling fish that spends substantial time in near and surface shore waters 70 Fishery Bulletin 101(1) 40 r 35 - 30 -"■•-... IT 25 >s "'*.^ E o 0) 2 20 - 1 'irn'" 1 you b ^ 1980's 5 o 5 15 - ^^ """"■■■■-.....^ 10 ^^-^::.i,^^ 5 n 1 1 1 1 1 1 1 1 U 40 60 80 100 120 140 160 180 200 Length (cm) Figure 4 Comparison of the 1960s and 1980s best-fit estimates of southern bluefin tuna growth rates as a function of length. to one that is found primarily in more offshore and deeper waters. In this regard, recent information from archival tags indicates that SBT between 80 and 90 cm (about two to three years old) commonly migrate during winter months to offshore oceanic waters in the Indian Ocean and the Tasman Sea and begin to feed at substantial depth (Gunn and Block, 2001). In contrast, catches and samples off Albany, Western Australia, show that many SBT less than 70 cm stay in nearshore Australian waters during winter (Hynd, 1965; Murphy"^; and release data analyzed in this study). Thus, the growth changes estimated to be near L* = 80 cm may correspond to a marked change in the SBT behavior during these winter months. The von Bertalanffy growth equation and its modifica- tions have been the standard for modeling tuna growth. The life history dynamics for most tuna species (e.g. north Pacific bluefin, albacore, bigeve, and yellowfin tuna) have a bimodal component analogous to that of SBT. Thus, juveniles are frequently found in densely packed surface schools, whereas at larger sizes individuals are rarely found near the surface and appear not to occur in densely " Murphy, G.I. 1979. Southern bluefin tuna. Au.st. CSIRO Div. Fish. Oceanogr. Fishery Situation Report 1, 10 p. (Available from CSIRO Marine Research, CPO Hox 1538 Hobart 7000, Australia. I packed schools (although there is little direct information on schooling for these larger fish). Moreover, mature tuna expend considerable energy in the spawning process, and in some cases swim thousands of kilometers and incur considerable weight losses during spawning (Warashina and Hisada, 1970). Bayliff et al. (1991) also found that growth models with a rate discontinuity at a certain size provided a better fit to Pacific northern bluefin tuna tag- return data than a simple continuous growth model. The extent to which this may be a general phenomenon in tuna or other fish species with marked changes in habitat use with age is not clear However, the results from our study and those of Bayliff et al. ( 1991 ) suggest that a growth rate with a discontinuity at a certain size may be more common than existing modeling of growth may indicate. Complex growth models, which deviated from a simple continuous growth curve, have generally not been considered, and the available data, in many cases, may not have sufficient power to be able to statistically identify more complex growth processes if they exist. Although the complex two-stage growth model used in our study clearly provided a substantial and significant improvement in fit to the growth-increment data, the mod- el itself presents problems in terms of the biological inter- pretation of the parameter estimates for L*. The bimodal nature ol' the likelihood function means that the size and Hearn and Polachek: Long-term growth rate changes in Thunnus maccoyli 71 Figure 5 Comparison of the 1960s and 1980s best-fit estimates for the expected length of southern bluefin tuna as a function of age, assuming that the expected length of an age-1 fish is 55 cm. age where the change in growth occurs is not well defined. This, in turn, confounds the evaluation of the plausibility of different specific possible biological hypotheses underly- ing the change. Moreover, although the change in growth patterns may be quite rapid, a large discontinuity in the growth rates at a specific length seems unrealistic. The continuous two-stage VBG model did not fit the 1960s and 1980s data as well as the discontinuous two-stage VBG models. However, the two-phase VBG models fitted the data better than the simple von Bertalanffy growth curve (Table 1, A and B) and its generalization — the simple Richards' ( 1959) curve (senior author, unpubl. results). From both the statistical estimation and biological per- spective, we think there is scope for the development of more appropriate complex growth models. In this context, there is also need for the development of estimation proce- dures for these complex models that can take into account alternative error structures that allow for individual variability in the growth rate parameters (e.g. Sainsbury, 1980: James, 1991; Wang et al., 1995 ). In the joint analysis of the 1960s and 1980s data, (7„, was the only parameter found not to be significantly different between the two data sets. However, caution is warranted in any comparison and interpretation of growth curves determining parameter values because of the well-known negative correlation between k and L. of the VBG growth model and the bimodal nature of the likelihood surface, as already noted. In particular, the differences in the esti- mates of L^2 should not be taken as strong evidence that the asymptotic growth of SBT decreased or that there was a crossover in the growth rates. These complex growth changes are difficult to explain from a biological perspec- tive and, as noted above, the bootstrap results indicate that the hypothesis that the L^g parameters are equal cannot be rejected. Moreover, we would note that there is a paucity of tag return data for larger fish. A total of only seven tags were recovered from fish with lengths exceed- ing 165 cm and only three for fish with lengths in excess of 170 cm. Fitting VBG models does not provide reliable es- timates of growth when extrapolated beyond the range of the data because of the large negative correlation between k and L ^. We, therefore, do not think that the current data are sufficient to determine whether, in fact, L^^ differed between the 1960s and 1980s. One of the primary applications of the estimated SBT growth curves is to provide estimates of the age distribu- tion of commercial catch in stock assessments based on catch-at-age analyses (e.g. Anonymous'*). The predicted growth curves (assuming that an age-1 fish is 55 cm) indicate that the estimated ages of 165-cm fish have di- verged by about a year for the curve based on a common L ,,, compared with those for which L„2 is allowed to differ 72 Fishery Bulletin 101(1) between the 1960s and 1980s. For smaller sizes, the diver- gence is substantially less (e.g. for fish 140 cm or less the divergence is less than three months). In terms of using the growth rate data to estimate ages from lengths, these results indicate that for the older reproducing fish the re- sults will be highly sensitive to assumptions about L^2- The results from these tagging studies clearly show that growth rates for SBT hatched in the 1980s had increased in relation to those cohorts hatched in the 1960s. The in- crease in growth rates is substantial, so that a fish, on av- erage, would have been expected to take four years to grow from 55 cm to 111 cm in the 1960s, but only three years to do so in the 1980s. In other words, after age 4 there is a difference of about one year in the expected age of a fish of similar length, and this difference persisted over the size range for which meaningful recapture data were available. The change in growth and its magnitude are consistent with the analyses in Leigh and Hearn (2000) of the modes in length-frequency distributions of juvenile fish captured in the Australian surface fishery. The underlying causes of the change in SBT growth rates are unknown. They could be associated with changes in environmental conditions, population size, or a combination of the two. The change in SBT growth rates between the 1960s and 1980s is associ- ated with very substantial declines in both the adult and juvenile components of the SBT stock (Polacheck et al.'; Anonymous^). There is an increasing number of examples in which growth rates have been reported to be inversely corre- lated to fish population numbers because of intraspecific competition. For example, Le Cren (1958) documented an increase in the growth rate of perch after a planned reduc- tion of a lake population. In a converse case, Kaeriyama (1996) reported a decline in the growth rate of Japanese chum salmon following a many-fold increase in its popula- tion size because of a most successful hatchery enhance- ment scheme. Other accounts are published in Southward (1967), de Veen (1976), Toresen (1990), Ross and Nelson (1992), and Sinclair and Swain (1996). However, the re- ports are mainly on species for which direct aging data are reliable and regularly collected over a lengthy period, or the fish are hatchery reared. The hypothesis that the increase in SBT growth rates was the result of the marked reduction in SBT papulation size would seem plausible, given the similar associations that have been found in a number of fisheries phenom- ena. As discussed in Leigh and Hearn (2000), changes in juvenile SBT growth rates based on analyses of length- frequency data are also consistent with the change having a density-dependent component. In this regard, it is worth noting that preliminary analyses of tag return data from the 1990s indicate that growth rates in the 1990s were similar to those in the 1980s (Polacheck and Preece"^). Thus, these preliminary results are also consistent with the change in growth being a density-dependent response as both juvenile and adult SBT abundances remained at low levels during this period (e.g. Anonymous^; Polacheck and Preece'°). Large uncertainty exists about possible recovery of the SBT stock in the near future (e.g. Anony- mous^), but continued monitoring of SBT gi'owth may provide one indicator of stock recovery. To simplify our investigation we did not consider sea- sonal growth. We avoided possible bias, due to seasonal growth, by analyzing data only from fish with times at liberty more than or equal to 270 days. This restriction provided an efficient mechanism to focus on the long-term growth process and was effective because the resultant sets were large. Large numbers of recaptured fish with reliable information and times at liberty more than 9 months seem rare for other tunas, in which case the added complication of accounting for possible seasonable growth would be required to ensure the robustness of the results. The analyses in this paper represent the first docu- mented examples of substantial temporal changes in growth rates that persisted for an extended portion of the life span in a large pelagic tuna resource. For tuna stocks in general, estimates of growth rates play a major role in stock assessments and in the subsequent management advice derived from these assessments. Acknowledgments We thank the many crew and scientific staff who par- ticipated in the 1959-84 SBT tagging operations. We are especially grateful for Australian and Japanese fishermen who returned tags with information on recapture length. The 1983-84 tagging program was financially supported by an Australian Government grant. Literature cited Akaike, H. 1974. A new look at the statistical model identification. Institute of Electrical and Electronic Engineers Transac- tions on Automatic Control, AC-19, p. 716-723. IEEE Control Systems Society, New York, NY. Allen, K. R. 1966. A method of fitting growth curves of the von Berta- lanfl'y type to observed data. J. Fish. Res. Board Can. 23: 163-179. Bayliff.W. H 1980. Synopsis of biological data on eight species of scorn- brids. Inter-Am. Trop. Tuna Comm., Spec. Rep. 2 (W. H. Bayliff, ed. ), 530 p. lATTC, San Diego. CA. 1988. Growth of skipjack, Katsuwonus pelamis. and yellow- fin, Thiinniis alhaccires. tunas in the eastern Pacific Ocean, as estimated from tagging data. Inter-Am. Trop. Tuna Comm. Bull. 19(41:311-385. ■' Anonymous. 1998. Report of the 1998 Scientific Committee meeting 3-6 August 1998, Tokyo, Japan. lAvailable from the Commission for the ('onservalion of Southern Hluefin Tuna, PO Box 37, Deakin West, ACT 2600, Australia.] '" Polacheck, T., and A. Preece. 1998. Preliminary comparisons of the growth rates of southern bluefin tuna in the 1990s with tho.sc in the 1960s and 1980s. Tenth SBT recruitment moni- toring workshop, 14-17"' September 1998, Hobart. Australia. RMWS/9H/5, 11 p. lAvailable from CSIRO and the Commis- sion for the Conservation of Southern Bluefin Tuna, P.O. Box 37, Deakin West, ACT 2600, Australia.] Hearn and Polachek: Long-term growth rate changes In Thunnus maccoyii 73 1991. Status of northern bluefin tuna in the Pacific Ocean. Inter-Am. Trop. Tuna Comm., Spec. Rep. 7:29-88. Baylifi; W. H., I. Ishizuka, and R. B Deriso. 1991. Growth, movement, and attrition of northern bluefin tuna. Thunnus thynnus, in the Pacific Ocean, a.s determined by tagging. Inter-Am. Trop. Tuna Comm. Bull. 20(1): 1-94. Burgess, D., A. Caton. J. Gunn, W. Hearn, T Murray, and C. Proctor. 1991. Aging and growth of juveniles and adults. In Review of aspects of southern bluefin tuna: biology, population and fisheries (A. E. Caton, ed.), p. 210-224. Inter-Am. Trop. Tuna Comm., Spec. Rep. 7. Caton, A. E. 1991. Review of aspects of southern bluefin tuna: biology, population and fisheries. Inter-Am. Trop. Tuna Comm., Spec. Rep. 7:181-357. Clay D. 1991. Atlantic bluefin tuna (Thunnus thynnus thynnus (L.)l: a review. Inter-Am. Trop. Tuna Comm., Spec. Rep. 7:89-179. Davies, R. B. 1977. Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrics 64:247- 254. 1987. Hypothesis testing when a parameter is present only under the alternative. Biometrics 74: 33-43. de Veen, J. F. 1976. On changes in some biological parameters in the North Sea sole (Solea solea L.). J. Cons. Int. Explor. Mer 37:60-90. Fabcns, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Farley, J. H., and T L. O. Davis. 1998. Reproductive dynamics of southern bluefin tuna, Thunnus maccoyii. Fish. Bull. 96: 223-236. Ford, E. 1933. An account of the herring investigations conducted at Plymouth during the years from 1924 to 1933. J. Mar Biol. Assoc. U.K. 19:305-384. Fournier, D. A., J. R. Sibert, J. Majkowski, and J. Hampton. 1990. MULTIFAN a likelihood method for estimating growth parameters and age composition from multiple length frequency data sets illustrated using data from southern bluefin tuna (Thunnus maccoyii). Can. J. Fish. Aquat.Sci. 47:301-317. Franci-s, R. I. C. C. 1988. Maximum likeHhood estimation of growth and growth variability from tagging data. NZ J. Mar Fresh- water Res. 22:42-51. Gunn.J. S. andB.A. Block. 2001. Advances in acoustic, archival and satellite tagging of tunas. In Tuna — physiological ecology and evolution (B. A. Block and E. D. Stevens, eds. ), p. 167-224. Academic Press, New York, NY. Hampton, J. 1986. Effect of tagging on the condition of southern bluefin tuna, Thunnus maccoyii, (Castlenaul. Aust. J. Mar Fresh- water Res. 37:699-705. 1991. Estimation of southern bluefin tuna Thunnus mac- coyii growth parameters from tagging data, using von Ber- talanffy models incorporating individual variation. Fish. Bull. 89:577-590. Hearn, W. S. 1986. Mathematical methods for evaluating marine fish- eries. Ph.D. diss., 195 p. Univ. New South Wales, Sydney, New South Wales. Hynd, J. S. 1965. Southern bluefun tuna populations in south-west Australia. Au.st. J. Mar Freshwater Res. 16: 25-32. James, I. R. 1991. Estimation of von Bertalanffy growth curve param- eters from recapture data. Biometrics 47: 1519-1530. Kaeriyama, M. 1996. Population dynamics and stock management of hatchery-reared salmons in Japan. Bull. Natl. Res. Inst. Aquacult., Suppl. 2:11-15. Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-776. Kirkwood, G. P 1983. Estimation of von Bertalanffy growth curve param- eters using both length increment and age-length data. Can. J. Fish. Aquat. Sci. 40:1405-1411. Kirkwood, G. P., and I. F. Somers. 1984. Growth of two species of tiger prawn, Penaeus esculentus and P. semisulcatus, in the western Gulf of Carpentaria. Aust. J. Mar Freshwater Res. 35:703-712. Le Cren, E. D. 1958. Observations on the growth of perch (perca fluviatilis L.) over twenty two years with special reference to the effects of temperature and changes in population density. J.Anim.Ecol. 27:287-334. Lee, R. M. 1912. An investigation into the methods of growth determi- nation in fishes. Publ. Circ. Cons. Explor. Mer 63:34. Leigh, G. M., and W. S. Hearn. 2000. Changes in growth of juvenile southern bluefin tuna (Thunnus maccoyii): an analysis of length-frequency data from the Australian fishery. Mar Freshwater Res. 51:143- 154. Manzer, J. I., and F H. C. Taylor 1947. The rate of growth of lemon sole in the Strait of Georgia. Fish. Res. Board Can. Prog. Rep. Pac. Coast Stns. 7224-27. Murphy, G. I. 1977. New understanding of southern bluefin tuna. Aust. Fish. 36(11:2-6. Nelder, J. A., and R. Mead. 1965. A simplex method for functional minimization. Com- put. J. 7:308-313. Pitcher, T J., and P D. M. MacDonald. 1973. Two models for seasonal growth in fishes. J. Appl. Ecol. 10:559-606. Richards. F J. 1959. A flexible growth function for empirical use. J. Exp. Bot. 10:290-300. Ross, M. R., and G. A. Nelson. 1992. Influences of stock abundance and bottom-water temperature on growth dynamics of haddock and yellow- tail flounder on Georges Bank. Trans. Am. Fish. Soc. 121: 578-587. Sainsbury, K. J. 1980. Effect of individual variability on the von Bertalanffy growth equation. Can. J. Fish. Aquat. Sci. 37:241-247. Sinclair, A. F, and D. P. Swain. 1996. Comment: Spatial implications of a temperature-based growth model for Atlantic cod Gadus morhua off the eastern coast of Canada. Can, J. Fish. Aquat. Sci. 53:2909-2911. Southward, G. M. 1967. Growth of Pacific halibut. Rep. Int. Pac. Halibut Comm. 43, 40 p. 74 Fishery Bulletin 101(1) Toresen, R. 1990. Long-term changes in growth of Norwegian spring- spawning herring. J. Cons. Int. Explor. Mer 47:48-56. von Bertalanfly, L. 1938. A quantitative theory of organic growth. (Inquiries on growth laws. II). Hum. Biol., 10:181-213. Walford. L.A. 1946. A new graphic method of describing the growth of animals. Biol. Bull. (Woods Hole, Mass.) 90:141-147. Wang. Y. G., M. R. Thomas, and I. F. Somers. 1995. A maximum likelihood approach for estimating growth from tag-recapture data. Can. J. Fish. Aquat. Sci. 52:252-259. Warashina, I., and K. Hisada. 1970. Spawning activity and discoloration of meat and loss of weight in the southern bluefin tuna. Bull. Far Seas Fish. Res. Lab. (Shimizu) 3:147-166. |In Japanese with English abstract.) Wild, A. 1994. A review of the biology and fisheries for yellowfin tuna, Thunnus albacares, in the eastern Pacific Ocean. FAO Fish. Tech. Pap. 336/2:52-107. Wild, A., and J. Hampton. 1994. A review of the biology and fisheries for skipjack tuna, Katsuwonus pelamis, in the Pacific Ocean. FAO Fish. Tech. Pap. 336/2:1-51. Yukinawa, M. 1970. Age and growth of southern bluefin tuna, Thunnus maccoyii (Castlenau) by use of scale. Bull. Far Seas Fish. Res. Lab. (Shimizu) 3:229-257. 75 Abstract— The life history of the At- lantic sharpnose shark {Rhizoprion- odon terraenouae) was described from 1093 specimens collected from Virginia to northern Florida between April 1997 and March 1999. Longitudinally sectioned vertebral centra were used to age each specimen, and the period- icity of circuli deposition was verified through marginal increment analysis and focus-to-increment frequency dis- tributions. Rhizoprionodon terraenovae reached a maximum size of 828 mm precaudal length (PCD and a maxi- mum age of 11-1- years. Mean back-cal- culated lengths-at-age ranged from 445 mm PCL at age one to 785 mm PCL at age ten for females, and 448 mm PCL at age one to 747 mm PCL at age nine for males. Observed length- at-age data (estimated to 0.1 year) yielded the following von Bertalanffy parameters estimates: L„= 749 mm PCL (SE=4.60), K = 0.49 (SE=0.020), and tg=-0.94 (SE=0.046) for females; and Z,„ = 745 mm PCL (SE = 5.93), ft:=0.50 (SE=0.024l, and ?o = ^0.91 (SE = 0.052) for males. Sexual maturity was reached at age three and 611 mm PCL for females, and age three and 615 mm PCL for males. Rhizoprionodon terrae- novae reproduced annually and had a gestation period of approximately 11 months. Litter size ranged from one to eight (mean=3.85l embyros, and in- creased with female PCL. Life history of the Atlantic sharpnose shark {Rhizoprionodon terraenovae) (Richardson, 1836) off the southeastern United States Joshua K. Loefer South Carolina Department of Natural Resources Manne Resources Research Institute 217 Fort Johnson Road P,0, Box 12559 Charleston, South Carolina 29412-2559 E mail address loefer|(S'mrddnr state scus George R. Sedberry South Carolina Department of Natural Resources Marine Resources Research Institute 217 Fort Johnson Road PO Box 12559 Charleston, South Carolina 29412-2559 Manuscript accepted 22 July 2002. Fish. Bull. 101:75-88 (2003). The Atlantic sharpnose shark (Rhizo- prionodon terraenovae) is a small carcharhinid that inhabits the coastal waters of the western North Atlantic from the Bay of Fundy to the Yucatan (Castro, 1983). It is the most common small coastal species off the southeast- ern U.S. coast and the Gulf of Mexico ( Branstetter, 1990). This species is frequently encountered by a variety of commercial fishing gear, including bottom longline, gill net, bandit reel (used by the snapper-grouper fishery), and shrimp trawl. Rhizoprionodon ter- raenovae is also a common catch in the recreational hook-and-line fishery. The age and growth of this spe- cies has been described in the Gulf of Mexico by Parsons (1981, 1983a, 1985) and Branstetter (1981, 1986, 1987a). Although those studies provided sig- nificant information on the age and growth of R. terraenovae, data were collected from 1979 to 1984, a time in which fishing pressure on the R. terrae- novae population was probably not as high as at present (Cortes, 1995). The previous studies dealt with fishes only from the northern Gulf of Mexico, and therefore may not represent the entire stock, although the stock structure for R. terraenovae in the northwestern Atlantic remains unclear (Heist et al., 1996). No published age and growth studies exist for specimens collected from the southeastern U.S. Atlantic coast. The reproductive biology of this species has been studied in both the Gulf of Mexico and off the southeast- ern U.S. coast (Parsons, 1983b; Cas- tro, 1988, 1993; Castro and Wourms, 1993), but the lack of concurrent age and growth data off the southeastern United States limits the utility of these data for fishery management. Considering the importance of accu- rate and timely age, growth, and repro- ductive information to fishery manage- ment, this study had two objectives: to describe age, growth, and reproduction in the southeastern U.S. population of R. terraenovae; and to compare these data to those of previous studies on the same species in the Gulf of Mexico. Materials and methods Rhizoprionodon terraenovae (/!=1093) were collected throughout the year in coastal waters from April 1997 through March 1999. Collection sites ranged from Chesapeake Bay, Virginia, to Port Canaveral, Florida (Fig. 1). The majority of specimens were collected off the coast of South Carolina. A vari- ety of sampling gears were employed for sample collection: bottom longline (47'7f of specimens), otter trawl (22%), port-sampling of commercial fishing 76 Fishery Bulletin 101(1) -38° •34" 26' 84° 80° 100 76° 200 300 400 km Figure 1 Sample collection sites and distribution by area (roughly equivalent to state borders) for R. terraenovae collected during this study, 1997-99. (D) represents locations where one or more R. terraenovae were captured. (A) = 13 males (694-793 mm PCL); (B) = 52 females (215-786 mm PCD, 51 males (200-765mm PCL); (C) = 497 females (197-813 mm PCL), 441 males (225-828 mm PCL); (D) = 8 females (302-763 mm PCL), 7 males (320-658 mm PCL I; (E) = 6 females (335-738 mm PCL), 16 males (271-720 mm PCL). vessels (16%), rod and reel (12'7( ), gill net (39^ ), and other miscellaneous gear types (2%). Following capture, the sex of each specimen was deter- mined and the specimen was weighed (to the nearest 0.1 kg), evaluated for sexual maturity, and its body length was measured. Four body length measurements (to the nearest mm) were taken from each individual: precaudal length (PCL, measured from the tip of the snout to the anterior termination of the precaudal pit), fork length (FL), natu- ral total length (NTL, measured with tail in a "natural" swimming position (Parsons, 1985]), and total length (TL, measured with dorsal portion of tail bent parallel to the body axis). Unless otherwise noted, precaudal lengths are used throughout this study. Regression relationshiijs of TL, NTL, and FL on PCL were derived to facilitate com- parison with other studies. The claspers of males were measured from the clasper tip to the anterior termination of the vent. The siphon sac was measured from the base of the clasper fin (where the sac originates) to the anterior termination of the sac. The condition of the seminal vesicles was also recorded. Male maturity was indicated by calcification of the claspers and the presence of a fully formed siphon sac (Clark and von Schmidt, 1965; Parsons, 1983b). Gonadosomatic indices (GSIs, Parsons, 1983b) were calculated for male sharks with the formula GSl = gonad weight (g)/hody tveight (g) x 100. Loefer and Sedberry: Life history of Rhizophonodon terraenovae off the southeastern United States 77 The ovaries and uteri of females were examined macro- scopically for indicators of maturity, such as yolking eggs, embryos, or placental scars. Vitellogenic oocytes were eas- ily identified by their bright yellow coloration in contrast to the pale white coloration of nonvitellogenic oocytes. If vitellogenic oocytes were present, the diameter of all vitel- logenic oocytes in the ovary was measured (to the nearest 0.1 mm) with dial calipers. If maturing oocytes were not present, the most differentiated nonvitellogenic oocytes (which were noticeably larger that the rest of the oocytes in the ovary) were measured. Any embryos were removed from the uteri, counted, their sex determined, and mea- sured (TL). Female maturity was determined by the presence of embryos, umbilical scars in the uterus from previous pregnancy, or the presence of large vitellogenic oocytes (greater than 15 mm diameter) nearing ovulation (Parsons, 1983b). A segment of the vertebral column extending from the cervical region (dorsal to the branchial chamber) to the origin of the first dorsal fin was removed from each specimen and frozen. Vertebrae from the cervical portion of the spinal column were used for aging because of the shallow concavity of the intermedalia and the size simi- larity between adjacent centra in this region. The shal- low concavity of the vertebrae facilitated processing and measurement during aging (Branstetter and McEachran, 1986). Age determination was attempted on 890 of the 1093 specimens collected during the study. Vertebrae selected for aging were separated from the frozen seg- ment, defrosted, and soaked in 5% sodium hypochlorite for 5-30 min (depending on size) and were removed from the solution as soon as all excess connective tissue had been dissolved. A longitudinal section approximately 500 pm thick was cut from the center of each vertebrae with a Mark- V watering saw and allowed to air-dry for at least 24 h. Dried sections were then attached to glass slides with Accu-mount 60 mounting medium and hand polished with wet 600-grit sandpaper to a thickness of approximately 350 pm. Several staining or ring elucidation techniques (e.g. Parsons, 1983a; Branstetter, 1986; Brown and Gru- ber, 1988; Hoenig and Brown, 1988) failed to significantly increase increment visibility; therefore all aging was per- formed with unstained vertebral sections. Vertebral sections were read on a dissecting microscope with transmitted light and a polarizing filter at 20x mag- nification. Increment radii and marginal increments were measured through the center of the corpus calcareum (Fig. 2) with OPTIMAS image analysis software (Media Cyber- netics, 1999). Precaudal length was regressed on centrum radius (CR) for males and females to test for an isometric relationship. The increments observed in vertebral sections were narrow circuli similar to those described by Simpfen- dorfer (1993), as opposed to the gi-owth bands described by Branstetter (1987a). All increment counts were made without knowledge of the size, sex, or collection date of the specimen. The primary reader (senior author) counted increments on all samples twice; each reading was sepa- rated by at least two months. Increment counts that were not in agreement were counted a third time. If the third Figure 2 Diagrammatic representation of a vertebral sec- tion; bm = birth mark, c = circuli, cc = corpus calcareum, cr = line of centrum radii and annuli measurements, f = focus, i = intermedalia. count did not agree with one of the first two counts, the specimen was excluded from the analysis. The secondary reader (coauthor) counted increments from all specimens not eliminated by the primary reader's analysis. Between- reader disagi'eements were re-examined by both observ- ers simultaneously. All specimens for which a consensus could not be reached were discarded. The index of average percentage error (lAPE; Beamish and Fournier, 1981) was used to estimate precision between the final readings of the primary reader and the initial readings of the second- ary reader The annual periodicity of increment formation was verified through marginal increment analysis and focus- to-increment frequency distributions. Absolute marginal increment distances were converted to "relative" marginal increments by dividing the distance between the last in- crement and the edge of the centrum by the width of the last fully formed growth band (Skomal, 1990; Natanson, et al., 1995). This conversion compensated for differences in growth rates between age classes. Back-calculated lengths at previous ages were esti- mated from vertebral measurements by using a modified Fraser-Lee equation proposed by Campana ( 1990): L„ = L,, + |(C„ - C,.) (L,. - L„)/{C,, - Cf,)], where L^ = length at age; L_ = length at capture; C^ = centrum radius from focus to increment a; and C = centrum radius at capture. 78 Fishery Bulletin 101(1) Lq and Cq are biologically derived intercepts that repre- sent the fish length and centrum radius, respectively, at which the proportionality between fish length and centrum growth are initiated. For the purposes of this study, mean body length and centrum radius at birth were used as the biologically derived constants (Sminkey and Musick, 1995). The observed age-class data were used to estimate "actual ages" to 0.1 year These were calculated by the number of circuli present plus growth since the formation of the last circulus. All specimens were given a 1 June birth date, which approximates the middle of the pupping season. This process corrected for growth since the last in- crement, preventing the potential overestimation of size- at-age that might result from analyzing the data by year class alone. All three types of length-at-age data (observed age class, observed actual age, and back-calculated age) were fitted to the von Bertalanffy growth equation (VBGE; von Bertalanffy, 1938): L, = LJ1 -Kit- 'o'), where L, = length at age t; L^ = asymptotic length; K = growth coefficient; and Tq = theoretical age at zero length. Each of the three types were analyzed for sexes combined, as well as for each sex separately. The parameters for the VBGE were estimated through a stepwise Gauss-Newton iterative fitting process computed by JMP statistical analysis software (Anonymous, 1998). Results The sharpnose shark was abundant throughout the year in coastal waters within the sampling area. The ratio of males to females in the overall sample was not signifi- cantly different from a 1:1 ratio (chi-square test, « = 1091. a=0.65, v=l,;^2=1.39,P=0.24). Linear regression of TL, NTL, and FL on PCL resulted in the following equations: TL = 29.804 + 1.279PCL NTL = 31.678 + 1.254PCL FL= 11.249 -H 1.075PCL (;i = 1009, ;-=0.99,P<0.0001); (n=493,r2=0.99,P<0.0001); (n = 1083, r2=0.99, P<0.0001). Reproduction and maturity Size-at-maturity estimates were based on observations of 526 males and 564 females. The smallest fully mature male was 600 mm PCL, and the largest immature male was 615 mm PCL. All males greater than 615 mm PCL and 36'" '" -ales from 600 to 615 mm PCL were fully mature. The onset and completion of maturity in male R. tcrraenorae were demonstrated by the onset of develop- ment in the claspers and siphon sac (Fig. 3). Males began to mature at 500 mm PCL. The maturation of claspers and siphon sac reached completion approximately one year later, at 600 to 615 mm PCL. The smallest maturing female was 509 mm PCL and con- tained one maturing oocyte five mm in diameter The second smallest maturing female was 529 mm PCL. The smallest gravid female was 591 mm PCL. The largest immature fe- male, based on lack of embryos or uterine scarring, was 611 mm PCL. Females from 591 to 611 mm PCL were either gravid (63%) or contained large (>10 mm diameter) matur- ing oocytes and were close to their first ovulation (377^ ). All females greater than 611 mm PCL were mature. Mean GSI and mean ovarian egg diameter (MOD) both demonstrated prominent peaks during the calendar year Male GSI values were highest in April and high values were also present in March and May (Fig. 4). However, the seminal vesicles remained turgid and full of semen for some time following the seasonal testicular degeneration which began in May. Female MOD values were highest in May and June. An increase in standard error along with a drop in mean value for the month of June (Fig. 4) demon- strated that ovulation began at that time. The extremely low MOD in July indicated the completion of ovulation. Litter sizes ranged from one to eight, and generally in- creased with female PCL (Fig. 5). Mean litter size was 3.85 embryos, and significantly more embryos were found in the left uterus (mean=2.19) than in the right (mean=1.65; chi-square test, «=558, a=0.05, v=4. ^2=62.62, P<0.0001). Nonlinear regression of litter size on female PCL resulted in the following equation («=278, /■'-=0.51, P<0.0001): Litter size = -11.07 -i- 0.021 PCL + 1.37 X lO-^lPCL- 710.9)2 Rhizoprionodon terraenovae were born at approximate- ly 212 mm PCL. The smallest free-swimming neonate was 190 mm PCL, and the largest full-term embryo was 242 mm PCL. Most pupping occurred from mid-May to early June. However, a small number of neonates appeared as early as mid-April. Consequently, mean embryo total length was at a minimum in July and at a maximum in June (Fig. 6). The sexes of uterine embryos were not sig- nificantly different from the expected 1:1 ratio (chi-square test, ?!=844, a=0.05, v=l, ;f-=0.076, P=0.78). Age and growth Separate linear regressions of PCL on centrum radius (CR) for males and females were not significantly different (ANCOVA, P=0.065l and were therefore combined (Fig. 7) to yield the following formula: PCL = 61.80 -I- 124.48Ci? (r2=0.963, n=812, P<0.0001). The regression line slightly overestimated centrum radius for the largest individuals (>700 mm PCL) of both sexes. Data transformation, as well as nonlinear regression, failed to increase the r^ value, and only the largest speci- mens were affected. Nonlinear regression of total body weight on length was significantly different between males and females (AN- COVA after log-transformation, P<0.001), and resulted in the following equations: Loefer and Sedberry: Life history of Rhtzopnonodon terraenovae off tfie southieastern United States 79 O 300 250 - 200 - 150 100 - 50 - itu - 120 • ••• • °«9 100 80 60 40 20 • n n = 441 2«^°^ n = 290 200 300 400 500 600 700 800 900 Precaudal lengtti (mm) Figure 3 Relation of clasper and siphon sac length to precaudal length for male R. terraenovae; (•) represents individuals with uncalcified claspers; (O) represents individuals with fully calcified claspers. Females: W, = e(-is.62)pcL(3M) (/•^'=0.99, P<0.0001, /i=458); Males; w = e(-i8.i8.pcL'2 96i (?-^'=0.99, P<0.0001. n=454). where W, = total body weight. Aging was attempted on 890 specimens, 812 of which were aged without elimination. Agreement between the first and second counts conducted by the primary reader was 66%, with 91% within one increment, and 99% within two. Those sections that showed disagreement between the first and second reading (n=303) were counted a third time, and 96% agreed with one of the first two readings. The remaining 4% (12 specimens) were excluded from the analysis. Agreement between readers was 72%, with 95% within one increment and 99%. within two. Vertebrae for which counts did not agree between readers (246 out of 878) were re-examined by both readers simultaneously. A concurrent age could not be reached on 66 vertebrae, which were eliminated from the study. The lAPE between the final readings of the primary reader and the initial readings of the secondary reader was 7.4%. Size-frequency distributions of the discarded individuals (data not shown) closely matched those of the raw data set and did not indi- cate the elimination of a large number of individuals from any age class during the aging process. Mean relative marginal increments for age classes 1+ through 7+ combined demonstrated a minimum in July (Fig. 8). The O-i- age class was excluded from this analysis to ensure that growth from the birth mark did not affect the results. Frequency distributions of focus-to-increment measurements for ages 0+ through 7+ demonstrated single modes for all annuli in each age class for both males and females (Fig. 9). Most R. terraenovae were found to have an increment in the intermedalia and an associated change in the angle of the corpus calcareum. which is similar to the birth mark 80 Fishery Bulletin 101(1) described by other authors (e.g. Casey et al., 1985; Branstetter, 1987b; Simpfendorfer, 1993). There were 239 young of the year R. terraenovae collected during this study, 88 of which contained no discernible birth mark. All young of the year lacking a birth mark were captured in June and July (Fig. 10), whereas all young of the year captured from August through April had a birth mark. Both marked and unmarked centra were noted in July and showed a readily apparent trend; individuals with a birth mark were sig- nificantly larger than those without a birth- mark (/-test, df=96, ^=-7.138, P<0.0001). Back-calculated lengths-at-age were sim- ilar to observed lengths-at-age in all cases, although observed values were slightly higher for all age classes (Table 1). There was no evidence of Lee s phenomenon in the older age classes. Back-calculated size at the birth mark overestimated size at birth as determined by observations of neonates and full-term embryos. The VBGE estimates calculated by age class, actual age, and back-calculated age demonstrated little variation either within or among data types (Table 2). The VBGE parameters from all data types corresponded well with known life his- tory parameters for size at birth and maximum size. Unless otherwise noted, all comparisons throughout the remainder of this study were based on VBGE es- timates derived from the "actual age" data type. Discussion Reproduction Length-at-maturity estimates for male R. terraeno- vae were similar among the three published studies. Parsons ( 1983b) estimated male maturity at -610 to 653 mm PCL (lengths from other studies were con- verted to PCL by using the formulae derived from the current study) and Branstetter (1987a) estimated the same at 600 mm PCL. We determined that males reach full maturity at -600 to 615 mm PCL. The three studies failed to agree on length at maturity for female R. terraenovae. Branstetter (1987a) and Parsons (1983b) approximated the size of females at maturity at 660 mm PCL and from 650 to 690 mm PCL, respectively. We found, however, that females mature at a smaller size, from 590 to 610 mm PCL. The reproductive seasonality of R. terraenovae in our study appeared to lack synchrony; males reached their reproductive peak in April and females in May and June. Mature males dissected in late May and June had vis- ibly atrophied testes compared to those collected in April and early May. However, their seminal vesicles were still highly swollen and contained large amounts of semen. This condition indicated that male R. terraenovae were 20 - r 1.2 18 - 1 16- £ 14 - ^ b^ o ■04 1 X 3 a ■0.2 » 2 - n - Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Figure 4 Mean gonadosomatic index and mean ovarian egg diameter by month for female R. terraenovae . Open circles indicate females (n=275). closed circles indicate males l/i=214). Error bars represent mean ± one standard error 8l 5 T 7 - i « 6- Q. =) 18 /^ ■'■ 36/^* 57 ^ C 10 'V^ ^^ J 20 40 ^L^^ S 3- 1 - 2 •^6^ 'h -^ "o -^s- ■h ^s 'h Female precaudal length (mm) Figure 5 Mean litter size on female size class. Solid line represents best-fit quadratic equation. Numbers indicate sample size for each data point. Flrror bars represent mean ± one standard error. still capable of mating during May and June, when female MOD values were highest. Therefore, the mating season of fl. terraenovae off the southeastern U.S. coast appeared to last from mid May to early July. Simpfendorfer (1992) noted a similar misalignment of peaks in reproductive seasonality between the sexes in R. taylori. The largest litters noted in our study contained eight pups (/?=4l. This increases the maximum litter size re- ported for R terraenovae in the northwestern Atlantic Loefer and Sedberry: Life history of Rhizoprionodon terraenovae off the southeastern United States 81 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Figure 6 Mean total length by month for embryos collected from April 1997 through March 1999. Numbers indicate sample size for each data point. Error bars represent mean + one standard error 1200 1 1000- PCL = 61 80+ 124 48CR ^^ {mm) CO o o " iifirf^ ^ 600 ^^0^' ra a> -400 - -ii^^\ Q. _^JgPP^ X POO - ^^^ 00 10 2 30 40 50 60 70 80 90 Centrum radius (mm) Figure 7 Linear regression of precaudal length I mm) on centrum radius (mm i for R Icr- | racnovae; (X) represents males, (Ol represents females. (Parsons, 1983b; Castro and Wourms, 1993). Early reports of up to 12 pups in sharpnose sharks collected from Cuban waters (Bigelow and Schroeder, 1948) were likely the re- sult of misidentification (Castro and Wourms, 1993). Age and growth The PCL-CR regression line slightly overestimated cen- trum radius for large individuals. This trend has also been noted in large female Carcharhiniis obscurus (Natanson et al, 1995) and appears to result from a change in the slope of the linear relationship as growth becomes asymptotic near the maximum length of the species. This phenomenon was deemed to have a minimal effect on the linear regres- sion formula used in this study. Although the linear rela- tionship appears to undergo an immediate change in slope at about 700 mm PCL. there are not enough data following this change (that is. the animal does not increase substan- 82 Fishery Bulletin 101(1) Jan Feb War Apr May Jun Jul Aug Sep Oct Nov Dec Figure 8 Mean relative marginal increment (mm) by month for age classes 1+ through 7+. Numbers indicate sample size for each data point. Error bars represent mean + one standard error. Table 1 Mean, minimum, and maximum length s-at-age (mm) and statistics for observed actual and back-calculated ages (O-lO-i- years). O-i- 1 + 2+ 3-^ 4-1- 5-1- 6-1- 7-1- 8+ 9-1- lO-i- Females Back-calculated mean 249 452 556 619 665 698 722 740 754 775 777.9 minimum 189 307 422 521 563 600 627 673 711 754 maximum 301 573 646 742 764 795 795 800 785 804 SD 19 32 35 36 35 34 29 28 22 19 n 379 305 273 232 186 137 90 42 13 5 1 Observed mean 320 513 629 676 700 717 741 755 762 788 787.0 minimum 197 391 469 606 615 345 663 688 726 764 maximum 465 624 707 780 765 805 812 810 796 813 SD 63 51 49 33 30 66 26 31 23 20 n 123 32 42 46 50 47 48 29 8 4 1 Males Back-calculated mean 247 452 564 634 675 695 708 717 728 715 minimum 191 310 372 519 582 625 651 690 706 maximum 317 553 681 760 778 809 753 764 743 SD 21 36 45 40 38 32 24 21 16 n 337 260 225 191 159 102 49 15 4 1 Observed mean 323 509 600 676 716 722 722 732 743 720 minimum 200 340 387 578 623 653 661 699 729 riKixinnini 466 602 730 777 796 828 763 773 757 SD 63 59 69 46 39 34 24 21 14 n 116 35 34 32 57 53 35 10 3 1 Loefer and Sedberry: Life history of Rhtzopnonodon lerraenovae off tfie southeastern United States 83 Females Age 0+ n =74 Age 1 + n =30 80 Males 70 60 OSOi A AgeO+ n =76 40 H / \ 30 / \ 20 / \ 10 / \ Age 1 + n =34 Age 4+ n =47 Age 4 + n =41 Focus to increment distance (mm) Figure 9 Focus to increment distance (mm) frequency distributions for males and females age 1+ to 7+. The first distribution represents the birth mark in all cases, subsequent distributions represent (from left to right! measurements to the first, second, third, fourth, fiflh, sixth, and seventh increments, respectively. 84 Fishery Bulletin 101(1) 40- 35- O 30- E E 25. w ra 20. E ^ 15. O e 8 o 8 o 8 o if » = o 1 i 1 Q. 05. n n Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 10 Centrum radius of age 0+ R. terraenovae, by month. (♦) represents individuals without a birth mark, (O) represents individuals with a birth mark. Table 2 Von Bertalanffy growth parameters oi Rhizoprionodon terrraenovae from the southeastern coast of the United States. Von Berta- lanfTy growth parameters from previous studies in the northern Gulf of Mexico are included for comparison. Von Bertanalanffy growth parameters L^ SE SE SE Sex (mm PCD K ^0 ofZ.„ ofK oU, n Data type Study Females 752 0.52 -1.07 5.33 0.025 0.052 433 age class current Males 746 0.53 -1.07 6.83 0.030 0.059 379 Sexes combined 750 0.52 -1.07 4.23 0.019 0.039 812 Females 749 0.49 -0.94 4.60 0.020 0.046 433 estimated actual age current Males 745 0.50 -0.91 5.93 0.024 0.052 379 Sexes combined 748 0.50 -0.92 3.65 0.015 0.034 812 Females 738 0.46 -0.90 2.64 0.006 0.015 1856 back-calculation current Males 726 0.53 -0.79 3.14 0.009 0.016 1447 Sexes combined 732 0.49 -0.85 2.02 0.006 0.011 3303 Sexes combined 820 0.36 -0.99 — — — 20 estimated actual age Branstetter 11987a) Males 709 0.39 to 0.53 -2.01 — — — 15 age class Parsons(1985) tially in length following the .shift) to reliably fit a second regression line. The back-calculation equation used in our study does not employ the linear regression in its calcula- tions and was minimally affected by the negative bias that this phenomenon had on the slope of the regression. Marginal increment analysis in the present study in- dicated that growth increments form in summer. This finding is contrary to that of earlier studies on R. ter- raenovae, which indicated winter deposition (Parsons, 1985; Branstetter and McEachran, 1986; Branstetter 1987a). However, other species in this genus have been shown to deposit increments during the summer months. Simpfendorfer (1993) demonstrated summer (February) increment deposition in R. taylori in Australian waters. He cited stress during the breeding season as a possible cause because hepatosomatic index and condition factor in both sexes were low during the mating season, an indica- tion of probable stress. Furthermore, growth increments in elasmobranchs may reflect periods of slow calcareous accretion that have been compressed by increased growth (Gelsleichter, 1998). This pattern of deposition may result in increments from periods ol' slow growtii not becoming Loefer and Sedberry: Life history of Rhizopnonodon terraenovae off tfie soutfieastern United States 85 900 - 800 - 700 - E 600- g' 500- ™ 400 3 to m 300 dI 200 100 - ^gsi"^^^ "^ - Branstetter (1987b) Parsons (1985) Present study 01 23456789 10 11 12 Age (years) Figure 11 A comparison of growth curves for R. terraenovae from the present study with those from previous works by other authors. Growth curve from the present study is based on estimated actual ages for both sexes (parameter values are presented in Table 2). visible for some time after their actual formation until enough new tissue has grown distally to provide the compression and contrast necessary for reliable identifica- tion. In other words, the increments observed in our study first became visible in July, but may have actually formed one to several months earlier. It should be noted that the methods of vertebrae processing and examination followed during our study were more similar to those of Simpfen- dorfer (1993) than to those of Parsons (1985) or Branstet- ter (1987a). These methods may have contributed to the close similarity found in both the physical appearance (i.e. that of "check marks" as opposed to pairs of growth bands) and temporal deposition of increments between our study and that of Simpfendorfer ( 1993). We found young of the year R. ter-racnovae with and without a birth mark. This is unusual in that most studies that have documented the presence of a birth mark have found one present in all specimens examined (e.g. Casey et al., 1985; Branstetter, 1987b; Simpfendorfer, 1993). Simp- fendorfer ( 1993) suggested that the "birth" mark in R. tay- lori was probably laid down sometime after birth because he observed the same overestimation of size at birth by back-calculations noted previously in our study. No tempo- ral estimation of the lag between birth and the formation of a birth mark has been published. The young-of-the-year R. terraenovae examined during our study demonstrated a distinct temporal transition from the lack of a birth mark to the presence of a birth mark iFig. 10). The data sug- gest that the birth mark is not actually laid down at birth in June, but approximately one month later in July. This time lag may explain the overestimation of size at birth by back-calculation. It is possible that the mechanism for the formation of the birth mark lies in the switch from embryonic to normal somatic growth, which may not occur immediately following parturition. The von Bertalanffy growth parameters derived for our study demonstrated differences from those derived for previous studies (Fig. 11). Parsons (1985) estimated an L^ of 709 mm PCL, and Branstetter (1987a) 820 mm PCl" L„ for our study was 745 mm PCL for males, and 749 mm PCL for females. The ^q value produced by Parsons was low at -2.01 yr, whereas the values produced by Branstetter (-0.99 yr) and our study (-0.90 yr for males and -0.94 yr for females) agreed well with the known gestation period of approximately 11 months. Parsons (1985) estimated K by several methods, resulting in values ranging from 0.39 to 0.53. The higher values agreed well with the estimates of our study (0.49 for females and 0.50 for males). Brans- tetters ( 1987a) estimate of K was 0.36, lower than that of the current study. Yearly growth rate estimates by Parsons (1983b) and Branstetter (1987a) revealed an increase of 133 to 211 mm PCL during the first year of life. 94 mm during the second year, 55 mm during the third year, and 16 to 32 mm growth after maturity. We found similar, though slightly higher growth rates: 198 to 202 mm PCL during the first year, 100 to 108 mm during the second, 63 to 69 mm dur- ing the third, and from to 46 mm thereafter. Parsons (1985) determined age at maturity by three methods: extrapolation of growth rates to size at maturity, the VBGE, and Holdens method (Holden, 1974). The esti- mates produced by these methods ranged from 2.0 to 3.5 for males, and 2.4 to 3.9 for females. Branstetter (1987a) compared his von Bertalanffy-derived estimates to those of Parsons (1985), and found his results in general agree- ment with Parsons" higher estimates. Branstetter (1987a) 86 Fishery Bulletin 101(1) thus concluded that males mature in three years and fe- males in four. In our study, males reached full maturity at 2.4 to 2.6 years of age, making them functionally mature at the third breeding season following birth. Females were found to mature at 2.2 to 2.5 years, which would also re- sult in full maturity just prior to the third postnatal breed- ing season. Although it was noted in both previously cited studies that males matured six months to one year earlier than females, no such discrepancy in age at maturity be- tween the sexes was apparent in our study. Differences between studies The differences between this and previous studies on R. terraenovae are likely a combination of many contributing factors. These studies were conducted in different regions at separate times and may reflect clinal or temporal differ- ences (or both) between Gulf of Mexico and northwestern Atlantic R. tej-raenovae populations. However, there are other contributing factors that must be considered as well, most notably differences in data collection and analysis techniques. Parsons' ( 1985) growth curves were based on males and were grouped into age classes (not assigned actual ages). His von Bertalanffy parameters were then derived by us- ing the Ford and Walford plot method (Parsons, 1985), re- quiring the use of mean lengths of each age class. This age class grouping does not take into account growth since the deposition of the last increment, and may therefore bias the Ford and Walford plot by pulling the data to a faster asymptote (Branstetter and McEachran, 1986; Branstet- ter, 1987a;). This bias produced a low L, (706 mm PCD and /f,(-2.01 years) in Parsons' estimates (Branstetter and McEachran, 1986; Branstetter, 1987a). This phenomenon was not evident in VEGE estimates based on age classes in our study, which were very similar to estimates based on actual ages (Table 2), and was probably due to the fact that iterative fitting of age data to the VBGE by computer software (an option unavailable to Parsons at the time of his study) is less sensitive to unaddressed growth than the graphically based Ford and Walford plot method. Although the aging technique used by Branstetter (1987a) was similar to that of our study (counts on lon- gitudinal sections of cervical centra). Parsons' (1985) aging technique took ring counts from the face of centra that had been removed from a more posterior region of the vertebral column than the region chosen in our study. It has been stated by several authors (Branstetter and McEachran, 1986; Martin and Cailliet, 1988; Kusher et al.. 1992) that increment counts made from sections of verte- bral centra are generally preferable to those taken from the face of unsectioned centra. Sectioned centra allow for better documentation of the increment structure near the edge because the increments become narrower and more difficult to delineate with increasing age (Branstetter and McEachran, 1986; Martin and Cailliet, 1988; Kusher et al., 1992). This distinction is critical when the potential consequences of age underestimation (including overesti- mation of K', growth rate, and maximum sustainable yield) are considered. Based on comparison of our work to that of previous studies (Branstetter, 1981, 1987a; Parsons, 1983a, 1983b, 1985), there may be differences between the Gulf of Mexico and southeastern U.S. Atlantic populations of Atlantic sharpnose sharks. The question then becomes whether these differences are clinal or temporal in nature. Clinal variation, for instance, may explain the differences noted in size and age at maturity in female/?, terrae/iovae. Simpfendorfer ( 1993 ) noted differences in size at maturity between populations of R. taylori in Australia, as did Par- sons (1993) and Carlson et al. (1999) between populations of Sphyrna tiburo and Carcharhinus acronotus. respec- tively, off the Gulf coast of Florida. However, the extended time frame between the current and previous studies ( 15 to 20 years), also opens the possibility that the differences are related to a temporal change in population structure of the species across the entire Gulf and Western Atlantic region. In the earlier studies, data were collected during a time when fishing pressure (both directed and indirected) on R. terraenovae was lower than at present, and fisheries were shown to have dramatic effects on shark populations in less time (Anonymous'). The differences noted between the studies may thus be a manifestation of temporal changes in population structure of the species as a whole over the last two decades. A more current study on Gulf of Mexico R. terraenovae is needed to properly address these potential population differences. Conclusion Small shark species such as R. terraenovae tend to show rapid growth in the first few years of life and a dramati- cally slower growth rate once maturity is reached. This aspect of their growth complicates age estimation by ver- tebral increments because the most recent marks in older specimens are so closely spaced that accurate counting and measurement become problematic. The overlapping of increments in these older specimens or the lack of iden- tifiable increment formation altogether due to asymptotic growth may lead to an underestimation of ages in large adults. Althhough the maximum age demonstrated in our study was ll-i- years, the actual life span ofR. terraenovae may be longer The life history parameter estimates that have been pre- sented in our study are based on one of the largest short- term samples collected for any study of elasmobranch life history to date. The most significant aspect of this study is the documentation of differences in size and age at matu- rity between female R. terraenovae in the Gulf of Mexico and females off the southeastern U.S. coast. A difference in age of maturity of one year in an animal with a relatively short life span, such as R. terraenovae, can have a dramatic effect on the outcome of population models (see Cortes, 1995). Although the documentation of age at maturity dif- ferences by different researchers may be highly susceptible Anonymous. 1993. Fishery management plan for sharks of the Atlantic Ocean, 167 p. U.S. Dep. Commerce., NOAA, NMFS, Silver Spring, MD 20910. Loefer and Sedberry: Life history of Rhizopnonodon terraenovae off tfie soutfieastern United States 87 to analytical bias during the aging process, the documenta- tion of differences in size at maturity is unmistakable. Acknowledgments This project was funded by the University of Charleston (South Carolina), The Marine Resources Monitoring, Assessment and Prediction program (MARMAP), the Cooperative Atlantic States Shark Pupping and Nursery Survey (COASTSPAN), and the W. F. Pate Memorial fund. The authors would like to thank the following: MARMAP personnel; the Southeast Area Monitoring and Assess- ment Program (SEAMAP) personnel; Glenn Ulrich and the crew of the RV Anita: Dean Grubbs, Jack Musick, and the crew of the RV Bay Eagle: Bill Roumillat and the SCDNR Inshore Fisheries Project; Reese Hair and the crew of the FV Malachi HI: Steve Johnson and the crew of the FV Miss Gina: and Gary Mckillop and the crew of the FV Boiizai, for assistance in sample collection. Pat- rick Harris. Charles Wenner, Steven Branstetter, Antony Harold, Glenn Parsons. Enric Cortes, Jim Gelsleichter. Colin Simpfendorfer, and Gregor Cailliet offered advice and constructive criticism. Literature cited Anonymous. 1998. JMP: Statistics and graphics guide, .593 p. SAS Institute, Inc., Cary, NC. Beamish. R. J., and D. A. Fournier. 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. 38:982-98.3. Bigelow, H. B., and W. C. Schroeder 1948. Sharks. In Fishes of the western North Atlantic, part one, vol. 1 (J. Tee-Van, C. M. Breder, S. F. Hildebrand, A. E. Parr and W. C. Schroeder, eds. ). p. 59-546. Mem. Sears Found. Mar Res., Yale Univ. Branstetter, S. 1981. Biological notes on the sharks of the North Central Gulf of Mexico. Contrib. Mar Sci. 24: 13-34. 1986. Biological parameters of the sharks of the Northwest- ern Gulf of Mexico in relation to their potential as a com- mercial fishery resource. Ph.D. diss., 138 p. Texas A&M University. TX. 1987a. Age and growth validation of newborn sharks held in laboratory aquaria, with comments on the life history of the Atlantic sharpnose shark. Rhizopnonodon terraenovae. Copeia 1987(21:291-300. 1987b. Age and growth estimates for blacktip. Carcliarhi- niis limbatus. and spinner, C. brevipinna. sharks from the northwestern Gulf of Mexico. Copeia 1987(4):964-974. 1990. Early life-history implications of selected carcha- rhinid and lamnoid sharks of the northwest Atlantic. //; Elasmobranchs as living resources: advances in the biol- ogy, ecology, systematics, and the status of the fisheries (H. L. Pratt Jr, S. H. Gruber, and T. Taniuchi. eds.), p. 17-28. U.S. Dep. Commer. NOAA Tech. Rep. NMFS 90. Branstetter, S., and J. D. McEachran. 1986. Age and gi-owth of four carcharhinid sharks common to the Gulf of Mexico: a summary paper. In Indo-Pacific fish biology: proceedings of the second international confer- ence on Indo-Pacific fishes (T. Uyeno, R. Arai, T. Taniuchi, and K. Matsuura, (ds.), p. 361-371. Ichthyological Soc. of Japan, Tokyo. Brown, C. A., and S. H. Gruber 1988. Age assessment of the lemon shark, Negapriun bre- virostns, using tetracycline validated vertebral centra. Copeia 1988(31:747-753. Campana, S. E. 1990. How reliable are growth back-calculations based on otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227. Carlson, J. K., E. Cortes, and A. G. Johnson. 1999. Age and growth of the blacknose shark. Charchannus acronotus in the Eastern Gulf of Mexico. Copeia 1999(3 1: 684-691. Casey J G., H. L. Pratt Jr, and C. E. Stillwell. 1985. Age and gi-owth of the sandbar shark {Carcharhiniis plumbeus) from the western North Atlantic. Can. J. Fish. Aquat. Sci. 42:963-975. Castro, J. I. 1983. The sharks of North American waters, 180 p. Texas A&M Univ Press, College Station, TX. 1988. Investigations in the reproductive biology of sharks. Ph.D. diss., 110 p. Clemson Umv, SC. 1993. The shark nursery of Bulls Bay, South Carolina, with a review of the shark nurseries of the southeastern coast of the United States. Environ. Biol. Fishes 38:37-48. Castro, J. I., and J. P. Wourms. 1993. Reproduction, placentation, and embryonic develop- ment of the Atlantic sharpnose shark. Rhizopnonodon terraenovae. J. Morph. 218:257-280. Clark, E., and K. von Schmidt. 1965. Sharks of the central gulf coast of Florida. Bull. Mar Sci. 15:13-83. Cortes, E. 1995. Demographic analysis of the Atlantic sharpnose shark, Rhizoprionodon terraenovae, in the Gulf of Mexico. Fish. Bull. 93:57-66. Gelsleichter, J. 1998. Vertebral cartilage of the clearnose skate. Raja egtan- teria: development, structure, ageing, and hormonal regu- lation of growth. Ph.D diss., 215 p. College of William and Mary, Gloucester Point, VA. Heist, E. J., J. A. Musick, and J. E. Graves. 1996. Mitochondrial DNA diversity and divergence among sharpnose sharks, Rhizopnonodon terraenovae. from the Gulf of Mexico and Mid-Atlantic Bight. Fish. Bull. 94:664- 668. Hoenig, J. M., and C. A. Brown. 1988. A simple technique for staining growth bands in elas- mobranch vertebrae. Bull. Mar Sci. 42:334-337. Holden, M. J 1974. Problems in the rational exploitation of elasmobranch populations and some suggested solutions. /:; Sea fisher- ies research (F. R. Harden Jones, ed.), p. 117-137. Elek (Scientific books), London. Kusher, D. I., S. E. Smith, and G. M. Cailliet. 1992. Validated age and growth of the leopard shark, Tiia- kis semifasciata, with comments on reproduction. Envi- ron. Biol. Fishes 35:187-203. Martin, L. K., and G. M. Cailliet. 1988. Age and growth determination of the bat ray, Myli- obatis californica, in central California. Copeia 1988(3): 762-773. Media Cybernetics 1999. Optimas image analysis software, version 6.51. Me- dia Cybernetics, L. P., Silver Spring, MD. 88 Fishery Bulletin 101(1) Natanson, L. J., J. G. Casey, and N. E. Kohler. 1995. Age and growth estimates for the dusky shark, Car- charltinus ohscurus, in the western North Atlantic Ocean. Fish. Bull. 93:116-126. Parsons, G. R. 1981. Growth and reproduction of the Atlantic sharpnose shark, Rhizoprionodon terraenovae (Richardson). M.S. thesis, 71 p. Univ. of South Alabama, Mobile, AL. 1983a. An examination of the vertebral rings of the Atlantic sharpnose shark. Rhizoprionodon terraenovae. Northeast Gulf Sci. 6:63-66. 1983b. The reproductive biology of the Atlantic sharpnose shark, Rhizoprionodon terraenovae (Richardson). Fish. Bull. 81:61-73. 1985. Growth and age estimation of the Atlantic sharpnose shark, Rhizoprionodon terraenovae: a comparison of techniques. Copeia 1985(1 ):80-85. 1993. Geographic variation in reproduction between two pop- ulations of the bonnethead shark, Sphyrna tiburo. Envi- ron. Biol. Fishes 38:25-35. Simpfendorfer, C. A. 1992. Reproductive strategy of the Australian sharpnose shark, Rhizoprionodon taylori (Elasmobranchii: Carcharh- inidae), from Cleveland Bay, Northern Queensland. Aust. J. Mar. Freshwater Res. 43:67-75. 1993. Age and growth of the Australian sharpnose shark, Rhizoprionodon taylori, from North Queensland, Australia. Environ. Biol. Fishes 36:233-241. Skomal, G. B. 1990. Age and growth of the blue shark, Prionace glauca, in the North Atlantic. M.S. thesis, 82 p. Univ. of Rhode Island, Kingston, RI. Sminkey, T. R., and J. A. Musick. 1995. Age and growth of the sandbar shark, Carcharhinus plumbeus, before and after population depletion. Copeia 1995(4):871-883. von Bertalanffy, L. 1938. A quantitative theory of organic growth (inquiries on growth laws II). Hum. Biol. 10:181-213. 89 Abstract— We present a method to integrate environmental time series into stock assessment models and to test the significance of correlations between population processes and the environmental time series. Param- eters that relate the environmental time series to population processes are included in the stock assessment model, and likelihood ratio tests are used to determine if the parameters improve the fit to the data significantly. Two approaches are considered to integrate the environmental relation- ship. In the environmental model, the population dynamics process (e.g. recruitment) is proportional to the environmental variable, whereas in the environmental model with pro- cess error it is proportional to the environmental variable, but the model allows an additional temporal varia- tion (process error) constrained by a log-normal distribution. The methods are tested by using simulation analy- sis and compared to the traditional method of correlating model estimates with environmental variables out- side the estimation procedure. In the traditional method, the estimates of recruitment were provided by a model that allowed the recruitment only to have a temporal variation constrained by a log-normal distribution. We illus- trate the methods by applying them to test the statistical significance of the correlation between sea-surface tem- perature (SST) and recruitment to the snapper (Pagrus auratus) stock in the Hauraki Gulf-Bay of Plenty, New Zea- land. Simulation analyses indicated that the integrated approach with additional process error is superior to the traditional method of correlating model estimates with environmental variables outside the estimation pro- cedure. The results suggest that, for the snapper stock, recruitment is posi- tively correlated with SST at the time of spawning. A general framework for integrating environmental time series into stock assessment models: model description, simulation testing, and example Mark N. Maunder George M. Watters Inter-Amencan Tropical Tuna Commission Scnpps Institution of Oceanography 8604 La Jolla Shores Drive La Jolla, California 92037-1508 E-mail address (for M N Maunder) mmaunderfgiiattcorg Manuscript accepted 20 September 2002. Fish. Bull. 101:89-99 (2003). Identifying a clear relationship between an environmental variable and pro- cesses in the dynamics of the population (recruitment, natural mortality, growth) or the fishery (catchability) would al- low improved estimation and prediction of model parameters and derived quan- tities. It is well known that the environ- ment plays a large role in the population dynamics and catchability offish stocks. Many researchers (Green, 1967; Joseph and Miller, 1989; Hinton and Nakano, 1996; Lehodey et al., 1997; Shepherd et al., 1984) have identified correlations between population processes and envi- ronmental factors, and others (Hunter, 1983; Bertignac et al., 1998; Lehodey et al., 1998) have suggested hypotheses for the underlying causes of these cor- relations. Incorporation of environmen- tal time series into stock assessment models may provide additional informa- tion to help estimate model parameters, particularly when fishing observations (catch, effort, length-frequencies) are missing. For the management of fish stocks, it can be advantageous to be able to predict future catch rates and popula- tion sizes. Because there is often a delay due to the propagation of the recruit- ment signal in the population structure or because statistical and numerical models can provide predictions for some environmental variables (e.g. tempera- ture) (or for both reasons), the relation- ship can be used to predict future catch rates or population sizes. Statistical catch-at-age analysis (e.g. Fournier and Archibald, 1982; De- riso et al, 1985; Methot, 1990) is more appropriate than cohort analysis (vir- tual population analysis) to include relationships between an environ- mental variable and processes in the dynamics of the population. In cohort analysis, if there are missing data, they are simply extrapolated without any statistical methods, which may cause bias in the parameter estimates. Also, the potential correlation with an envi- ronmental series is calculated outside of the estimation procedure, producing several disadvantages, including the loss of information and the difficulty of propagating uncertainty (Maunder, 1998a, 2001a, 2001b). However, in sta- tistical catch-at-age analysis, there are robust statistical methods (maximum likelihood, with all the parameters esti- mated together by obtaining the best fit between predicted and observed data) that allow inclusion of multiple data sets and the integration of the environ- mental series into the stock assessment model. These methods automatically al- low for missing data and provide confi- dence inteivals, and the hypotheses can be easily incorporated and tested. The methods used to integrate the environmental series into the stock as- sessment model can be applied to differ- ent processes in the population, but are illustrated here with the case of recruit- ment. Recruitment is the fundamental process in the population dynamic that is responsible for the fluctuations of the stock size. Many studies (e.g. Francis, 1993) show that environmental vari- ables affect the recruitment. In statisti- cal catch-at-age analysis, recruitment combines an average value with an an- nual deviate, constrained by using a 90 Fishery Bulletin 101(1) distributional assumption (e.g. Maunder and Starr, 2001). This constraint allows the estimation when there is no in- formation (i.e. missing data). Traditional methods that re- late recruitment to environmental factors use correlation analysis of an environmental time series with estimates of recruitment from a stock assessment model. For example, cohort analysis is first used to generate a time series of re- cruitment. Then the time series of recruitment is regressed against sea-surface temperature (SST). This two-step pro- cedure has a number of disadvantages (Maunder, 1998a, 2001a, 2001b), including the loss of information and the difficulty of propagating uncertainty. We introduce a method suggested by Maunder (1998a; see Maunder and Starr, 2001) that incorporates environ- mental time series into stock assessment models and tests the significance of the correlation between the population processes and the environmental time series. We test the model with simulated data and compare the results to correlating model estimates with environmental vari- ables outside the estimation procedure. We illustrate this method with an application that investigates the correla- tion between SST and recruitment within the context of a statistical catch-at-age analyses used to assess snapper (Pagrus auratus) in the Hauraki Gulf-Bay of Plenty, New Zealand (Maunder and Starr, 2001). Materials and methods Integrating environmental indices into stock assessment models Parameters that relate the environmental time series to population processes were included in the statistical catch- at-age stock assessment model. We added additional struc- ture to the stock assessment model for each parameter of the stock assessment model (X) that was hypothesized to 1) have temporal variation, 2) be correlated with an envi- ronmental time series, and 3) have sufficient information in the data to be estimated for multiple time periods. This structure included a mean value for the stock assessment model parameter (^i), temporal deviations in the stock assessment model parameter (f,), a parameter that relates the environmental series to the stock assessment model parameter (/3), and a scaling parameter (a) that ensures that ji is the mean value for the stock assessment model parameter over the time period used in the model. X, = // exp (a + lili + e,). (1) where t = time, and /,= the value of the environmental time series at time t. The parameter a ensures that ^ is equal to the mean over the whole time period (Gilbert'; see Maunder and Starr, 2001). Therefore, a removes the log normal bias and bias caused by an unnormalized environmental time series and is defined as a = In (2) where n is the number of time periods. The additional structure requires that a set of param- eters (f,) that are constrained by a distributional assump- tion and two free parameters (|j, /3) be estimated. The distributional assumption (referred to as a "prior" in the following description and represented by the negative logarithm of the prior probability, see Eq. 3) is a prior on the degree of temporal variation in the stock assessment model parameter The default assumption is a normal distribution (assuming that the stock assessment model parameter is lognormally distributed) with mean zero and given standard deviation. Information about this distribu- tion can be obtained from estimates for similar species (e.g. Myers et al., 1995). The prior -In Prior (f |cr) = V (£^ 2ct- (3) ' Gilbert, D. J. 1999. Personal cdiiirmin. National Institute of Water and Atmospheric Research Limited. P.O. Box 14-901, Wellington, New Zealand. keeps the temporal deviations close to zero if there is no information in the data to the contrary. It is important to note that the prior is also needed to avoid making /3 a redundant parameter The parameters ^ and j5 and the set of parameters £, are estimated simultaneously with the other parameters of the stock assessment model, and the negative logarithm of the prior is added to the negative log-likelihood function of the stock assessment model. The parameter estimates are really the mode of the posterior distribution, but we treat them in a likelihood context. The influence of the environ- mental time series can be removed from the analysis by fixing /3 at zero. Therefore, likelihood ratio tests can be used to determine if the /3 parameter significantly improves the fit to the data. If the addition of /3 reduces the total negative log likelihood by more than about 1.92 units (/-, ^^^y qc^ ), then the additional parameter significantly improves the fit to the data at the 0.05 level, and there is a statistically sig- nificant correlation between the population process and the environmental time series. Similar tests can be performed to test the significance of the set of temporal deviation parameters, f,, by taking into account the number of ad- ditional parameters. Hilborn and Mangel ( 1997 ) provided a simple description of the likelihood ratio test. (The Akaike information criterion or the Bayes information criterion could also be used. ) Therefore, by fixing, or not, /J or f at zero we can define three types of statistical models; 1 Traditional model /J is (ixi'd at zero, the parameter set f, is estimated, and a significant relationship is determined by testing if the correlation coefficient between f, and the environ- mental time series is significantly different from zero. Maunder and Watters: Integrating environmental time series into stock assessment models 91 2 Environmental model /3 is estimated, each value in the parameter set f, is fixed at zero, and a significant relationship is deter- mined by testing if /i = 0, using a likelihood ratio test. 3 Environmental model with process error Both p and e, are estimated and a significant relation- ship is determined by testing if /3 = 0, with a likelihood ratio test. Simulation testing Simulation analysis was carried out to test the perfor- mance of the integi'ated approach and to compare this approach to the traditional model. A simple age-structured model (Appendix I) was set up to simulate a population for 20 years, starting from an unexploited population and gen- erating catch, effort, and catch-at-age data. The simulated recruitment was generated as having a component based on an environmental time series and a random compo- nent. Each component was given the same variance (0.6- ). The environmental time series was randomly generated with /3 = 1 for each simulation. The standard deviation of the observation error in the CPUE index, o^^p^,^, was set at 0.6, and the sample size of the catch-at-age data was set to 50. The same age-structured model was then fitted to the data to estimate the model parameters. The three models defined in the previous section (traditional model, environmental model, and environmental model with pro- cess error) were tested with the simulated data. In addi- tion to the parameters outlined in the description of the three models, average recruitment, the catchability coef- ficient, and the standard deviation of the fit to the CPUE data were also estimated. We also used a model that had constant recruitment to provide likelihood values to use in testing the significance of the environmental model. The simulation analysis was repeated 500 times for four scenarios: 1) using catch-at-age data for all years, 2) using catch-at-age data for the first 10 years, 3) using catch-at- age data for the last 10 years, 4) using catch-at-age data for all years, but using j3 = when generating the simulat- ed data. Scenario 4 was used to investigate the probability of type-I error of the models when used in combination with the statistical tests. For each set of simulated data and each model, we determine how often a significant rela- tionship between the logarithm of annual recruitment and the environmental time series is detected, the estimate of the slope of the relationship between the logarithm of an- nual recruitment and the environmental time series, the estimates of average recruitment, and the depletion level (ratio of current to unexploited biomass). We also calcu- late minimum-width 95% confidence intervals for average recruitment, using the likelihood profile method for the simulated data sets with catch-at-age data for all years. Application: relating recruitment in the Hauraki Gulf-Bay of Plenty snapper stock to SST Recruitment to the Hauraki Gulf snapper iPagrua auratus I stock is correlated with temperature (Paul, 1976). The abundance of l-i- snapper in the Hauraki Gulf estimated by trawl surveys has been shown to have a positive corre- lation with SST (Francis, 1993) and air temperature (Gil- bert, 1994) around or just after the time of spawning in the previous year This relationship has also been shown with catch-at-age analysis to continue to hold as snapper enter the fishery at ages 4 and older (Maunder and Starr, 1998). We applied the integrated approach described in this study in combination with the age-structured statistical catch-at-age model described in Maunder and Starr (2001 ) to the Hauraki Gulf-Bay of Plenty snapper stock. The model was fitted to catch-at-age data and biomass estimates. The biomass estimates were available for 1985 and 1994 and were obtained from analysis of tagging data. The majority of the catch-at-age data were available from 1990 to 1997, but there were some catch-at-age data of dubious quality, small sample size, and high variability for 1970 to 1973. The annual recruitment at age 1 was estimated for the time pe- riod of the model ( 1970-98) and also for 18 age classes (ages 2 to 19) that comprised the initial conditions in 1970. Results Simulation analysis For all four sets of simulated data the environmental model had the highest probability of detecting a relation- ship between recruitment and the environmental time series (Table 1). This model had a very high probability of detecting a relationship even when there was no relation- ship in the simulated data (Table ID). This indicates that the likelihood ratio test is not appropriate for the environ- mental model (see Appendix III). For all data sets, except that with only catch-at-age data in the first 10 years, the traditional model had a higher probability of detecting a relationship between recruitment and the environmental time series than did the environmental model with process error The environmental model with process error had a lower probability of detecting a true relationship than the traditional model, but also had a slightly lower probability of type-I error (the probability of incorrectly accepting a nonexistent relationship) than the traditional model. The probability of detecting a relationship was reduced as the number of catch-at-age data sets was reduced. The environmental model with process error did not show any bias in the estimate of the slope of the relation- ship between the logarithm of annual recruitment and the environmental time series, /3 (Table 1 ). For this model, the variation in the estimates of /3 increased when fewer years with catch-at-age data were available. The environmental model showed a small negative bias and slightly more er- ror in the estimates of p. The traditional model showed a large negative bias in the estimate of the slope of the relationship between the logarithm of annual recruitment and the environmental time series, and this bias increased as the amount of catch-at-age data was reduced. The tra- ditional method also had larger error, which increased as less catch-at-age data were available. The errors in the estimates of average recruitment and ^ajr^B,, increased slightly with less catch-at-age data 92 Fishery Bulletin 101(1) Table 1 (A) Results from the simulation analysis in which all the catch-at-age data were used. 9c = percentage of data sets that produced a | significant relationship between the environmental time series and recruitment; fi = average (average absolute relative error) of the estimates of the slope of the relationship between the environmental time series and recruitment; R„ = average (average absolute relative error) of the estimates of the average recruitment; B^^^^/Bg = average error (average absolute relative error) in the estimate of the ratio of current to unexploited biomass. p was set to 1 when generating the simulated data. (B) Results from the simulation analysis using only the first 10 years of catch-at-age data. /3 was set to 1 when generating the simulated data (C) Results from the simulation analysis using only the last 10 years of catch-at-age data, fi was set to 1 when generating the simulated data. (D) Results from the simulation analysis using all the catch-at-age data, but setting /3 = when generating the simu lated data. (Abso- lute, rather than relative, error was used for j3. ) EMwPE = environmental model with process error % P Ro BcJBg A Traditional 92 0.86(0.25) 996(0.05) -0.01 (0.15) Environmental 99 0.95(0.25) 1,024(0.10) 0.03(0.24) EMwPE 83 1.01(0.23) 988 (0.05) -0.04(0.14) B Traditional 57 0.45(0.55) 1.011(0.07) -0.03(0.17) Environmental 95 0.94(0.33) 1.078(0.17) 0.01(0.27) EMwPE 60 1.05 (0.30) 1,004 (0.07) -0.03 (0.15) C Traditional 81 0.68 (0.36) 1.022(0.09) 0.04(0.21) Environmental 97 0.95(0.32) 1,035(0.10) 0.04(0.25) EMwPE 74 1.00(0.26) 1.007(0.08) -0.02(0.19) D Traditional 3 -0.01 (0.20) 1.005(0.04) 0.01(0.10) Environmental 61 -0.02 (0.22) 1.038(0.09) 0.07(0.21) EMwPE 2 -0.01(0.21) 1,003(0.04) 0.01 (0.10) Table 2 Results related to the confidence intervals for average recruitment from age data. EMwPE = Environmental model with process error the simulation analys is obtained by using all the catch-at- Lower bound average Upper bound average Lower bound SD Upper bound SD True is within True is below True is above Traditional Environmental EMwPE 888 963 887 1,120 1,117 1,118 59 94 56 132 176 120 0.91 0.44 0.92 0.03 0.32 0.01 0.08 0.24 0.07 (Table 1). The errors in these estimates were shghtly greater for the environmental model (more bias and larg- er absolute error) than for the environmental model with process error and traditional model. The confidence intervals on R^ were, on average, greater for the traditional and environmental model with process error than for to the environmental model (Table 2), which greatly underestimated the width of the confidence inter- vals. However, the confidence intervals for the traditional and environmental model with process error showed the true value falling below the confidence interval less often than it fell above it. As expected, an environmental relationship was more difficult to correctly detect with the traditional model in situations with missing data (e.g. when catch-at-age data were missing in the last few years of the series), and, as stated above, using the environmental model is inappro- priate because it has a high probability of detecting a sig- nificant relationship when none exists. The environmental model also has a tendency to under estimate the width of the confidence interval for /?„ and the true value frequently falls outside of thi.s confidence interval. Therefore, the envi- ronmental model with process error is the model of choice. Application: relating recruitment in the Hauraki Gulf-Bay of Plenty snapper stock to SST The environmental model with process error has the lowest negative log-likelihood, but this model has many more parameters than the environmental model (Table 3). Maunder and Walters: Integrating environmental time series Into stock assessment models 93 Table 3 Results from applying the three methods (traditional, environmental, and environmental with process error) to the snapper appli- cation. Constant = recruitment is constant each year and equal to the average recruitment. EMwPE = environmental model with process error, n/a = not applicable. Average recruitment Constant Traditional Environmental EMwPE 13,315 (11,381-15,381) 11,406 (8,500-14,603) 13,530 (11,527-15,569) 12,029 (9,147-15,328) Number of p -ln( Likelihood) parameters n/a 482.5 3 0.20 466.8 50 0.48 473.5 4 0.55 464.4 51 Using the likelihood ratio test at the 0.05 level and com- pensating for the difference in the number of parameters being estimated in each model, we consider the environ- mental model to be the model of choice. If the likelihood ratio test were used, the environmental model with pro- cess error would be chosen over the traditional model, indicating a statistically significant correlation between SST and recruitment. Due to the weaknesses of the envi- ronmental model discussed above, we concentrated on the results of the traditional model and the environmental model with process error. The time series of estimated recruitments from the traditional model showed very little annual variation in recruitment for the first half of the time series and for the last few years of the time series (Fig. lA). This indicates that there is very little information in the data (catch-at- age) about annual recruitment for these time periods and that the prior on the recruitment residuals constrains the estimated recruitment to be close to the average recruit- ment. This result is consistent with the catch-at-age data, which, ignoring the inconsistent data from the 1970s, started in 1990. The greatest age in the catch-at-age data that had individual information was age 19; therefore the 1971 cohort is the first for which there is information. However, at the current exploitation rates, very few snap- per live to be more than 10 years of age, so that there is very little information about cohort size for any of the co- horts produced during the 1970s. The environmental model with process error indicated high variation in recruitment for the whole time period (Fig. IB). This is due to the formulation of the recruitment submodel, for which the annual anomalies are anomalies from the temperature-recruitment relationship; if there is no information in the data about recruitment for a par- ticular year, the recruitment will follow the temperature- recruitment relationship. The correlation of the estimated recruitment from the traditional model with SST had a low r-square (0.26), but it was statistically significant at the 0.05 level when a two-tailed test was used. In addition, the slope of the rela- tionship between recruitment and SST was much less for the traditional model than for to the environmental model with process error (Table 3). The estimates of recruit- ment from the traditional model included a large number of estimates that were close to the mean because there 1950 1960 1970 1980 1990 2000 B 5 ,0 ,5 ,0 ,5 ,0 1950 1960 1970 1980 Year 1990 2000 Figure 1 Annual estimates of relative recruitment strength at age 1 for the Hauraki Gulf-Bay of Plenty snap- per stock from the traditional (A) and environ- mental model with process error (Bl models was no information in the data about these recruitments. Therefore, it was inappropriate to use these recruitments to correlate with SST and, if used, they would result in a poor fit. However, a significant correlation, as obtained in this application, suggests that the correlation is probably stronger than apparent from the analysis, which should give confidence that a relationship exists and provide an incentive to apply the integrated models. The environmental model with process error did not show a statistically significant improvement over the environmental model because, ignoring the 1970s data, the catch-at-age data were available only for the last part of the time period. The recruitment anomalies were esti- mated for the whole time period, as well as for the initial conditions. Many of these recruitment anomalies had very 94 Fishery Bulletin 101(1) little information associated with them and therefore did not add anything to the estimation procedure. However, they do add additional parameters, which reduce the pos- sibility of accepting the model when using the likelihood ratio test. If the recruitment anomalies were estimated for only a limited number of years, it is likely that the envi- ronmental model with process error would be a statistical- ly significant improvement over the environmental model. Statistical tests could be carried out to determine which annual recruitment anomalies should be estimated, but this would be very time consuming. Reducing the number of annual recruitment anomalies may also cause an un- derestimation of the confidence intervals. For the snapper example, removing the anomalies for the initial conditions may be a good compromise. Discussion We have developed a general framework for integrating environmental time series into stock assessment models that appears to perform better than traditional methods. The method is flexible and it can be used to model many dif- ferent functional relationships between population or fish- ing processes and environmental time series and to include multiple environmental time series for any population model parameter (see Appendix II). Furthermore, it can be used with any statistical stock assessment model. The method can be used to test whether an environmental time series describes temporal variation in model parameters. The traditional model, which estimates annual recruit- ment within the stock assessment model and subsequently correlates the recruitment with the environmental series outside the stock assessment model, performs poorly. It has a reasonable probability of detecting a relationship be- tween recruitment and the environmental series, but this probability decreases rapidly as the number of years with missing catch-at-age data sets increases. The probability of incorrectly detecting a relationship when one is not pres- ent is low. This method has reasonable confidence-interval coverage for average recruitment and little bias or variance in the estimates of model parameters. The factor causing the poor performance of the traditional model is the large bias in the estimate of the slope of the relationship between recruitment and the environmental time series, which in- creases as the number of years with missing catch-at-age data increases. The bias occurs because the traditional model has a penalty on the absolute size of the annual recruitment deviations. This penalty constrains an annual recruitment anomaly to be close to the mean recruitment when there is little or no information about the recruit- ment in that year. Therefore, when the logarithm of the annual recruitment is correlated with the environmental time series, the estimated slope of the relationship is biased downward. Even in situations for which there is sufficient information for every recruitment anomaly, there will be a small tradeoflin the size of the anomaly, which reduces the contribution of the penalty to the objective function and the likelihood from the catch-at-age data. Unfortunately, if the penalty on the annual recruitment anomalies is removed. the estimation process can become unstable, particularly in data-poor situations for which the bias is greater. The amount of time that is required by the estimation algorithm also increases if the penalty is removed. When the penalty on the size of the recruitment anomalies is removed, the bias in estimates of the slope of the relationship between recruitment and the environmental time series is reduced when using all the catch-at-age data, but the variance in the estimates is greatly increased. In addition, when removing the penalty there was a large positive bias when using only the last 10 years of catch-at-age data and a large negative bias when using only the first 10 years of catch-at-age data. It is not known what results would be obtained if cohort analysis, which does not use a constraint on the annual recruitment anomalies, is used instead of the statistical catch-at-age analysis. It should be remembered that cohort analysis cannot be used or assumptions that are unlikely to be satisfied will have to be made when catch-at-age data are missing for some years. The environmental model, which has a deterministic relationship between recruitment and the environmental time series that is integrated into the stock assessment model, also performs poorly. This method has poor confi- dence interval coverage for average recruitment because the size of the confidence intervals are gi'eatly underes- timated. The method has larger bias and variance in the estimates of model parameters compared to the other two methods. There is a small negative bias in the estimate of the slope of the relationship between recruitment and the environmental time series. The environmental model has a very high probability of detecting a relationship between re- cruitment and the environmental series, and this probabil- ity only decreases slightly as the number of missing years of catch-at-age data sets increases. However, this model has a very large probability of incorrectly detecting a relationship when one is not present. Therefore, when using the environ- mental model, the likelihood ratio test should not be used to determine if there is a significant relationship between recruitment and an environmental time series. The value used to compare to the x- statistic in the likelihood ratio test for the environmental model is highly correlated with the catch-at-age sample size; therefore simulation analysis is needed to find the appropriate /- statistic for the given sample size (see Appendix III). This is also important for calculating confidence intervals that are also based on the X~ statistic and is the reason for the poor coverage for i?^. The environmental model with process error, which has a relationship between recruitment and the environmen- tal time series that is integrated into the stock assessment model with additional process error, performs well. This model has a reasonable probability of detecting a relation- ship between recruitment and the environmental series, but this probability is lower than those of the other two models, and decreases as the amount of data is reduced. It has a low probability of incorrectly detecting a relation- ship when one is not iircsent. These probabilities could be im[)r()ved by using simulation analysis to find the appro- priate ;f- statistic (see Appendix III). This method has rea- sonable confidence interval coverage for average recruit- ment and has little bias or variance in the estimates of Maunder and Walters: Integrating environmental time series into stock assessment models 95 model parameters. There is very little bias in the estimate of the slope of the relationship between recruitment and the environmental time series. For the environmental model with process error, when there is little or no information in the data to estimate the recruitment for that year, the penalty on the annual re- cruitment anomalies causes recruitment to be estimated close to the recruitment predicted by the relationship between recruitment and the environmental time series. Therefore, if there is a relationship between recruitment and the environmental time series, this model should pro- vide better estimates because additional information is included in the estimation procedure. This model has the favorable property that if there is no relationship between recruitment and the environmental time series, the model estimates ji to be small, eliminating any influence of the relationship between recruitment and the environmental time series, and still estimates the annual recruitment anomalies to represent the variation in annual recruit- ment. The likelihood ratio test can be used to detect a rela- tionship between recruitment and the environmental time series, and if a relationship does not exist, the results with P fixed at zero can be used. However, including [i in the es- timation procedure, even when there was no relationship between recruitment and the environmental time series, did not increase the error in the parameter estimates in relation to the model with j3 fixed at zero (see the results for the traditional model. Table ID). The method we describe can be used to integrate en- vironmental time series for parameters of the stock as- sessment model other than recruitment. The influence of the environment on catchability of the fish would be an obvious choice because there are numerous publications on the topic. For example. Green (1967) suggested that ther- mocline data would improve estimation of tuna abundance from catch and effort data, by allowing for the differentia- tion between changes in tuna abundance and catchability due to vertical distribution of tunas influenced by tempera- ture. We have used a method similar to the method that is presented in the present study to incorporate SST into the purse-seine catchability parameters for yellowfin and big- eye tuna (Maunder and Watters, 2001; Watters and Maun- der, 2001). Maunder (2001a) presented a general method to integrate the standardization of CPUE data into stock assessment models, including the integration of environ- mental variables. Growth rates have been observed to have temporal variation, and this variation has been correlated with environmental factors. Several authors have pre- sented growth curves that include temperature data (e.g. Mallet et al., 1999). Movement is another process that may be influenced by the environment. Lehodey et al. (1997) showed that spatial shifts in the western Pacific skipjack tuna population are linked to the movement of a large pool of warm water and that the movements of this large pool are related to El Nino-Southern Oscillation events. Once a correlation between the environmental time se- ries and the population process has been determined, this relationship can be used to improve the predictive abil- ity of the model. For example, if a relationship between SST at the time of spawning and recruitment has been determined, and the age at recruitment to the fishery is 3 years, recruitment to the fishery can be estimated 3 years in advance. One should be cautious about assuming that these relationships are valid and will continue to hold into the future, however Hilborn and Walters ( 1992) cautioned about using environmental data because there are many environmental indices that one can try, and if the data set has a few large and a few small observations, it is likely that one of the environmental data sets will correlate with the data. Myers (1998) reviewed a number of published cor- relations between recruitment and environmental factors and found that few of the correlations held when retested at later dates. Maunder and Starr ( 1998) also advised cau- tion because they found that a strong cohort may not enter the fishery when expected because of variations in growth rates. We have found that, when applying this method to the bigeye tuna data, there is an inconsistency in the pre- 1997 data and the data for 1997 and 1998 caused by much stronger than expected year classes entering the fishery in 1997 and 1998. There is also difficulty in deciding on the management strategy if environmental regime shifts are influencing the productivity of the stock (Maunder, 1998b). An advantage of the integrated approach, particularly the environmental model with process error, is that it more fully describes the uncertainty in the relationship be- tween the population process and the environmental time series, and therefore this uncertainty can be included in any management advice based on the relationship. Conclusions Integrating environmental relationships in a statistical stock assessment model is an improvement over the tra- ditional statistical model when there are large gaps in the data. However, it is important to include process error to avoid the high probability of detecting spurious correla- tions seen in the environmental model when using the like- lihood ratio test. Therefore, the environmental model with process error is the model of choice because 1 ) there is no bias in the estimates, 2) when there is no relationship with the environmental series, it is equivalent to the traditional model, 3) when such a relationship exists, the recruitment estimates are improved, particularly if there are important gaps in the data, 4) it may be used for prediction, and 5) uncertainty about the relationship can be modeled. Acknowledgments We thank Dave Fournier for advice on using AD Model Builder and related software, and Bill Bayliff, Rick Deriso, Shelton Harley, and Ransom Myers for commenting on the manuscript. Literature cited Bertignac, M., P. Lehodey, and J. Hampton. 1998. A spatial population dynamics simulation model of 96 Fishery Bulletin 101(1) tropical tunas using a habitat index based on environmen- tal parameters. Fish. Oceanogr. 7:326-334. Beverton. R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish. Invest. Ser. II, Mar. Fish. G.B. Minist. Agric. Fish. Food 19, 533 p. Deriso, R. B., T. J. Quinn II. and P. R. Neal. 1985. Catch-age analysis with auxiliary information. Can. J. Fish. Aquat. Sci. 42:815-824. Forsbergh, E. D. 1989. The influence of some environmental variables on the apparent abundance of skipjack tuna, Katsuwonus pela- mia. in the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm. Bull. 19:433-569. Foumier, D., and C. P. Archibald. 1982. A general theory for analyzing catch at age data. Can. J. Fish. Aquat. Sci. 39: 1 195-1 207. Francis, M. P. 1993. Does water temperature determine year class strength in New Zealand snapper (Pa.grusa!/ra;;iS,Sparidae)? Fish. Oceanogr. 2:65-72. Gilbert, D. J. 1994. A total catch history model for SNA 1. New Zealand Fisheries Research Document 94/24. 16 p. NIWA (Na- tional Institute of Water and Atmospheric Research), Well- ington, New Zealand. Granger, C. W. J. 1993. Forecasting in economics. /;; Time series prediction: forecasting the future and understanding the past (A. E. Weigend and N. A. Gershenfeld, eds.), p. 529-538. SFI Studies in the Sciences of Complexity, Proc. Vol. XV. Addi- son-Wesley, Boston, MA. Green, R. E. 1967. Relationship of the thermocline to success of purse seining for tuna. Trans. Am. Fish. Soc. 96:126-130. Hilbom, R., and M. Mangel. 1997. The ecological detective: confronting models with data, 315 p. Princeton Univ. Press, Princeton, NJ. Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dy- namics and uncertainty, 570 p. Chapman and Hall, New York, NY. Hinton, M. G., and H. Nakano. 1996. Standardizing catch and effort statistics using physi- ological, ecological, or behavioral constraints and environ- mental data, with an application to blue marlin iMakaira nigricans) catch and effort data from Japanese longline fisheries in the Pacific. Inter-Am. Trop. Tuna Comm. Bull. 21:171-200. Hunter, J. R. 1983. On the determinants of stock abundance. In From year to year (W. S. Wooster, ed.), p. 11-16. Washington Sea Grant Program, Univ. Wash., Seattle, WA. Joseph, J., and F. R. Miller 1989. EI Nirio and the surface fishery for tunas in the east- ern Pacific. Japan. Soc. Fish. Ocean., Bull. 53:77-80. Lehodey, P.. J. Andre, M. Bertignac, J. Hampton, A. Stones, C. Menkes, L. Memory, and N. Grima. 1998. Predicting skipjack tuna forage distributions in the equatorial Pacific using a coupled dynamical bio-geochemi- cal model. Fish. Oceanogr 7:317-325. Lehodey, P., M. Bertignac, J. Hampton. A, Lewis, and J. Picaut. 1997. El Nino Southern Oscillation and tuna in the western Pacific. Nature 389:715-718. Mallet, J. P, S. Charles, H. Persat, and P Auger 1999. Growth modellingin accordance with daily water temp- erature in European grayling iThymallus thymallus L.). Can. J. Fish. Aquat. Sci. 56:994-1000. Maunder, M. N. 1998a. Integration of tagging and population dynamics models in fisheries stock assessment. Ph.D. diss., 306 p. Univ. Washington, Seattle, WA. 1998b. Problems with using an environmental based re- cruitment index: examples from a New Zealand snapper iPagrus auratus) assessment. In Fishery stock assess- ment models (F. Funk, T J. Quinn II, J. Heifetz, J. N. lanelli, J. E. Powers, J. J. Schweigert, P. J. Sullivan, and C. I. Zhang, eds.), p. 679-692. Alaska Sea Grant College Program Report No. AK-SG-98-01, Univ Alaska, Fairbanks, AK. 2001a. A general framework for integrating the standard- ization of catch-per-unit-effort into stock assessment models. Can. J. Fish. Aquat. Sci. 58:795-803. 2001b. Integrated tagging and catch-at-age analysis ( ITCAAN). In Spatial processes and management offish populations (G. H. Kruse, N. Bez, A. Booth, M. W Dom, S. Hills, R. N. Lipcius, D. Pelletier, C. Roy S. J. Smith, and D. Witherell, eds. ), p. 123- 146. Alaska Sea Grant College Program Report AK-SG-01- 02. Univ. Alaska, Fairbanks, AK. Maunder, M. N., and P. J. Starr. 1998. Validating the Hauraki Gulf snapper pre-recruit trawl surveys and temperature recruitment relationship using catch at age analysis with auxiliary information. New Zealand Fisheries Research Document 98/15, 23 p. NIWA (National Institute of Water and Atmospheric Research), Wellington, New Zealand. Maunder, M. N., and P. J. Starr. 2001. Bayesian assessment of the SNAl snapper (Pagnis aui'atus I stock on the northeast coast of New Zealand. N.Z. J. Mar. and Freshwater Res. 35:87-110. Maunder, M. N., and G. M. Watters. 2001. Status of yellowfin tuna in the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm. Stock Assess. Rep. 1:5-86. Methot, R. D. 1990. Synthesis model: an adaptable framework for analy- sis of diverse stock assessment data. Inter North Pacif Fish. Comm. Bull. 50:2.59-277. Myers, R. A. 1998. When do environment-recruitment correlations work? Reviews in fish biology and fisheries 8:285-305. Myers, R. A., J. Bridson, and N. J. Barrowman. 1995. Summary of worldwide spawner and recruitment data. Can. Tech. Rep. Fish. Aquat. Sci. 2020, iv -i- 327 p. Paul, L. J. 1976. A study on age, gi'owth, and population structure of the snapper, Chrysophrus auratuf (Forsteri. in the Hau- raki Gulf New Zealand. New Zealand Ministry of Fisher- ies and Agriculture, Fish. Bull. 13, 62 p. Ricker, W. E. 1954. Stock and recuitment. J. Fish. Res. Board Can. 11: 559-623. Shepherd, J. G., J. G. Pope, and R. D. Cousens. 1984. Variations in fish stocks and hypotheses concerning their links with climate. Rapp. P.-V. Reun., Cons. Int. Explor Mer 185:255-267. Watters, G. M., and M. N. Maunder 2001. Status of bigeye tuna in the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm. Stock Assess. Rep. 1:109- 210. Maunder and Walters: Integrating environmental time series into stock assessment models 97 Appendix I: description of simulator and estimator The following is a description of the model equations used for the data simulator and for the estimator. The model is run from an unexploited state at the start of the fishery for 20 years. The model includes 10 age classes, where the 10th age class is a plus group. Dynamics Af,,i = floexp(j8/,+ff +a) a = In ^exp(ff +/}/,,) (I.l) (1.2) A^,.. = (^,-i,o-i(l- ",-A-i')'^""" ' f0Ta3.84. Now, con- sider a simple example where Pj = W— and .V, = nexpijil^ + t)). with the penalty - In Prior (f | ct) = V 2a' and CT is a constant. Consider two models: 1) f ^ = and 2) estimate f. For model 1, as A' increases x~ increases in proportion to N, as explained above, because the penalty term is constant. However, for model 2, as N gets large, the Maunder and Walters: Integrating environmental time series into stocl< assessment models 99 relative size of the penalty compared to -InLp gets smaller and therefore the estimates of e^ change so that p, gets closer to p,. Therefore, for model 2, x^ does not increase proportionally with A^. An appropriate test for the environmental model would be to produce sets of random environmental indices that have the same variance and auto-correlation as the actual environmental index to determine the appropriate value of X- that would give the desired type-I error This test would overcome the sample size effect. The method could also be used to refine the test for the environmental model with process error. 100 Abstract— \\V fxamiiu'd the spatial and temporal distribution, abundance, and growth of young-of-the-year (YOY) Atlantic croaker (Micropogonias undu- latus) in Delaware Bay, one of the northernmost estuaries in which they consistently occur along the east coast of the United States. Sampling in Del- aware Bay and in tidal creeks in salt marshes adjacent to the bay with otter trawls, plankton nets and weirs, between April and November 1996-99, collected approximately 85,000 YOY. Ingress of each year class into the bay and tidal creeks consistently occurred in the fall, and the first few YOY ap- peared in August. Larvae as small as 2-3 mm TL were collected in Septem- ber and October 1996. Epibenthic indi- viduals <25 mm TL were present each fall and again during spring of each year, but not in 1996 when low water temperatures in January and Febru- ary apparently caused widespread mortality, resulting in their absence the following spring and summer. In 1998 and 1999, a second size class of smaller YOY entered the bay and tidal creeks in June. When YOY survived the winter, there was no evidence of growth until after April. Then the YOY grew rapidly through the summer in all hab- itats (0.8-1.4 mm/d from May through August). In the bay, they were most abundant from June to August over mud sediments in oligohaline waters. They were present in both subtidal and intertidal creeks in the marshes where they were most abundant from April to June in the mesohaline portion of the lower bay. The larger YOY began egressing out of the marshes in late summer, and the entire year class left the tidal creeks at lengths of 100-200 mm TL by October or November when the next year class was ingressing. These patterns of seasonal distribu- tion and abundance in Delaware Bay and the adjacent marshes are similar to those observed in more southern estuaries along the east coast; however, growth is faster — in keeping with that in other northern estuaries. Seasonal distribution, abundance, and growth of young-of-the-year Atlantic croaker (Micropogonias undufatus) in Delaware Bay and adjacent marshes* Michael J. Miller David M. Nemerson Kenneth W. Able Marine Field Station Institute of Marine and Coastal Sciences Rutgers University 800 c/o 132 Great Bay Boulevard Tuckerton, New Jersey 08087-2004 Email address (for K. W. Able, contact author) ablea'imcsrutgers edu Manuscript accepted 20 August 2002. Fish. Bull. 101(1):100-115(2003). Atlantic croaker (Micr-opogonias undu- latus) is a commercial and sport fishery species that inhabits demersal habitats in estuarine, coastal, and continental shelf systems along the Atlantic coast of North America and in the Gulf of Mexico (Joseph, 1972). They spawn primarily over the continental shelf during a protracted spawning season that, based on the presence of larvae along the Atlantic coast, may extend from early July through March (Lewis and Judy, 1983; Cowan and Birdsong, 1985; Warlen and Burke, 1990; Hettler et al., 1997; Nixon and Jones, 1997; Able and Fahay, 1998). The exact loca- tion of spawning in the Middle Atlantic Bight (MAB) may be related to the areal extent of favorable warm bottom waters for spawning (Norcross and Austin, 1988) and may sometimes occur within or close to the mouth of Chesa- peake Bay (Barbieri et al., 1994b; Reiss and McConaugha, 1999). The larvae and postlarvae have been observed to be more abundant in deeper layers of water that may facilitate transport into and retention within estuarine nursery areas (Weinstein et al., 1980; Norcross, 1991). The young-of-the-year (YOY) usually begin to enter estuaries and tidal creeks along the Atlantic coast in September, or occasionally August, and they are often common components of the fish fauna in tidal creeks and estuaries until fall of the next year from New Jer.sey southward along the Atlantic coast and in the (lull'of Mexico (Chao and Musick, 1977; Knudsen and Herke, 1978; Weinstein, 1979; Currin et al., 1984; Ross, 1988; Able and Fahay, 1998). In some years there is a second pulse of small YOY that arrives in estuaries along the Atlantic coast in the spring or summer; this pulse may be the offspring from later spawning (Chao and Musick, 1977; Ross, 1988). In general, the YOY use estuarine habitats with salinities ranging from almost pure freshwater to seawater (Miglareseetal., 1982). Although YOY Atlantic croaker are present in some Atlantic coast estua- rine habitats during the winter (Ha- ven, 1957; Bearden, 1964; Dahlberg, 1972; Chao and Musick, 1977; Shenker and Dean, 1979; Bozeman and Dean, 1980; Able and Fahay, 1998), they ap- pear to experience winter mortality in the MAB in years with unusually cold winters (Massman and Pachcco, 1960; Joseph, 1972; Chao and Musick, 1977; Wojcik, 1978). Recent laboratory studies have found that YOY Atlantic croaker do not survive in sustained water temperatures of 3°C or lower I Lankford and Targett, 2001 ); therefore extended periods of low winter water temperatures may have drastic effects on their overwinter survival in some estuaries. Previous studies have indicated that YOY Atlantic croaker reach about 107-187 mm TL after their first year Contribution 2002-22 from the Institute of Marine and Coastal Sciences, Rutgers University, New Brunswick, NJ 08901. Miller et al.: Distribution, abundance, and growth of Micropogonias undulatus 101 of growth in estuaries along the Atlantic coast and 102- 250 mm TL in the Gulf of Mexico (Knudsen and Herke, 1978), but only a few studies have reported seasonal growth rates (Hansen. 1969; Knudsen and Herke, 1978). Length-frequency based growth rate estimates for the first year of growth for YOY along the Atlantic coast have ranged from 0.32 to 0.41 nim/d ( Knudsen and Herke, 1978). However, these were based on the entire year, including the larval and early juvenile period when analysis of otolith daily growth rings indicates much slower growth rates of 0.18-0.41 mm/d during the fall and winter months (Nixon and Jones, 1997). Length-frequency data from estuarine nursery areas clearly indicate that most growth occurs during the spring and summer months (Haven, 1957; Chao and Musick, 1977; Ross, 1988; Able and Fahay 1998). Despite various studies of YOY Atlantic croaker in some areas of the Atlantic and Gulf coasts, there is relatively little known of their early life history near the northern part of their range in the MAB and this is especially true for Delaware Bay. Our four-year study used extensive collections in Delaware Bay and in adjacent tidal marsh creeks to describe the timing of Atlantic croaker ingress, their seasonal abundance and size, growth rates, and the timing of their egress out of the marshes. Methods Study sites Delaware Bay is the estuary of the Delaware River and encompasses about 1878 km^ of open water along the southern edge of New Jersey and the northern edge of Delaware (Fig. 1). It has a relatively deep area (10-30 m) in the middle of the lower bay, bordered by narrow shoals and flanked by extensive tidal flats and salt marshes, which contain an additional 85.5 km- of open water in tidal creeks bordered by approximately 640 km- of marsh- plain area. Depending on the amount of river discharge, salinities range from 30-31%f at the mouth of the bay, to 1-10'^f in the lower Delaware River (Table 1; Cronin et al., 1962; Garvine et al., 1992). Ichthyoplankton survey Catch data from an ichthyoplankton survey (Table 2) was used to analyze the distribution, abundance, and size of larval Atlantic croaker in Delaware Bay and the lower Delaware River from April to October 1996. Sampling was performed during daylight hours once a month in April and October and twice a month from May to September Each sampling period included one tow at 70 randomly selected stations distributed among eight designated sam- pling zones (Fig. 1). Samples were collected with a 1-m diameter plankton net (0.5-mm mesh) deployed with a depressor in single stepwise oblique tows from the surface to the bottom. Tows were made at a speed of 1.4-1.9 knots for four to six minutes in the direction of the tidal flow. Up to 50 individuals were measured to the nearest millimeter total length (TL) from each sample. DELAWARE 4 15 km Scale Figure 1 Locations of salt marsh tidal creek sampling sites (1996-99) and designated sampling zones (1-8) in the Delaware Bay and in the lower Delaware River for the ichthyoplankton (April-October 1996) survey and the otter trawl (April-October 1996-98) survey. The heavier line across the bay indicates the boundary between the upper (5-8) and lower ( 1-4) sampling zones and marsh sites. Otter trawl survey We used catch data from a three-year otter trawl survey (Table 2) to analyze the distribution, abundance, and growth of settled YOY Atlantic croaker in Delaware Bay and the lower Delaware River Sampling was performed during daylight hours twice a month from April to Octo- ber in 1996 and once a month in 1997 and 1998, at 40 stations divided up among the same eight sampling zones of the ichthyoplankton survey (Fig. 1). Station locations were selected by using a stratified random sampling design from a grid of 1002 stations, excluding the stations over the deepest water near the mouth of the bay in zone 1. There were eight stations sampled each month in zone 3, six in zone 4, and four in all the other zones. Trawling was done with 4.9-m otter trawls (6 mm stretched codend mesh), made against the prevailing direction of the tide at a speed of 1.8 m/sec for 10 minutes. Up to 100 individu- als were measured from each sample. For presentation 102 Fishery Bulletin 101(1) Table 1 Physical characteristics of marsh and adj Figure 1 for locations of individual sites. icent bay study sites located along the New Jersey shore of Delaware Bay, 1996-99. See Surface temp. Average surface Surface salinity Average surface Average surface Marsh site range (°C) temp. CO range (%f ) salinity (%c) dissolved oxygen (mg/L) Upper hay Mill Creek 5.0-29.0 19.5 0-8.4 2.8 7.4 Mad Horse Creek 0-31.0 19.2 0.7-23.0 9.1 6.3 Browns Run 7.0-3L3 20.5 0.2-14.0 7.0 5.4 Bay 7.0-28.0 20.2 1.5-17.8 10.2 6.3 Lower bay Commercial Township 8.0-30.0 19.8 4.5-22.9 17.0 6.8 Upper Moores Beach 2.0-30.0 18.7 10.0-23.8 17.2 6.2 Lower Moores Beach 5.0-29.0 19.0 4.0-25.0 18.9 7.0 Dennis Township 6.0-32.0 20.4 6.2-24.7 17.0 5.7 Bay 6.0-29.1 19.6 11.1-24.7 17.8 7.3 Table 2 Yearly catch per unit of effort (CPUE= number offish per tow or weir set) for the different types of gear in the marsh creeks or in Delaware Bay and the total number of Atlantic croaker collected by each type of gear. RUMFS = Rutgers University Marine Field | Station; EEP = Public Service Enterprise Group Estuary Enhancement Program. Total number Total number 1996 CPUE 1997 CPUE 1998 CPUE 1999 CPUE of tows/sets offish Source Marsh creeks Otter trawl (creeks) 2.4 3.8 19.1 3.9 4,654 36,295 RUMFS Otter trawl (bay) 164.6 15.3 29.3 49.5 336 12.755 RUMFS Weir 1.1 41.6 46.8 98.7 443 20,714 RUMFS Delaware Bay Otter trawl 4.6 2.8 19.7 — 1,438 13,497 EEP 1-m plankton net 1.7 — — — 957 1,638 EEP Total fish 8,671 12,350 43,942 19,936 7,828 84,899 and statistical analysis of some aspects of the data of the ichthyoplankton and otter trawl surveys in the bay, the upper four zones were combined into an upper bay region and the lower four combined into a lower bay region (Fig. 1 ). Marsh creek survey Tidal creek samphng was carried out at six salt marsh sites on the New Jer.sey side of Delaware Bay (Fig. 1, Table 1). Dennis Township, Commercial Township, and Moores Beach will be referred to collectively as the lower bay sites, and Browns Run, Mill Creek, and Mad Horse Creek will be referred to as the upper bay sites. The average depth of the trawling stations ( ].',i-2.6 m) and Secchi depth values (0.3-0.4 ml were similar at all sites. The upper bay sites in the mostly oligohaline region of the bay had average salinities of 2.8-9. 1'/Jr and the lower bay sites were in the mesohaline region with average salinities of 17.0-18.9%p (Table 1). We sampled each of the marshes (Fig. 1) monthly from April through November 1996-99 (Table 2). Small inter- tidal marsh creeks were sampled with weirs set at high tide and hauled at low tide, approximately six hours later Each weir (2.0 m x 1.5 m x 1.5 m, with 5.0 m x 1.5 m wings, 6.0-mm stretched mesh) consisted of a funnel-shaped net stretched across the channel with wings extended back onto the marsh surface from each end of the net. In cases when the creek did not drain completely the area in ("runt of the weir was seined into the weir Trawling in larger intertidal to subtidal marsh creeks took place aroimd high tide and consisted of four replicate two-minute tows per station, made against the current with a 4.9-m otter trawl (6-mm stretched codend mesh) towed at a constant engine RFM of 2500. Trawling station locations at each site were designed to sample fishes along Miller et al : Distribution, abundance, and growth of Micropogonias undulatus 103 the mouth to upper creek gradients (see Able et al., 2000; 2001; Able et al.M. Thus, at each of the marshes there were six trawling locations. These locations included two large subtidal creeks and two smaller creeks with lower sub- tidal and upper subtidal or intertidal sections in each of the latter. Additional trawling locations were established in the bay immediately outside the mouth of the large creek at the Dennis Township, Moores Beach, Commer- cial Township, and Mad Horse Creek study sites (Fig. 1). The fish collected at these bay stations were used in the length-frequency figures for the bay (exclusive of Novem- ber when there was no trawl survey sampling in the bay) and for the growth calculations, but not in the catch-per- unit-of-effort (CPUE) calculations for the bay. Atlantic croaker collected in each weir set and in each trawl were enumerated, and up to 50 individuals per weir set and 20 per trawl were measured to the nearest millimeter total length. Abundances (CPUE, number of fish per trawl) were compared between the upper and lower bay sites in Delaware Bay, among the six different marsh sites, and among years, by using the nonparametric Mann-Whitney f/-test, or the Kruskal-Wallis ANOVA of ranks for mul- tiple comparisons, and when differences were found, the Dunnis test was used (criteria for significance; P<0.05) to make pair-wise comparisons. Physical variables were measured at the end of each weir and otter trawl sample in the marshes and in Dela- ware Bay (Table 1). Temperature, salinity, and dissolved oxygen concentrations were measured with a hand-held salinity, temperature, and oxygen meter (YSI Model 85), by lowering the probe into the water and recording surface values. Water transparency was measured by lowering a Secchi disc into the water column until it was no longer visible and recording the corresponding depth in 0.1-m increments. Growth J 1996 Ichthyoplankton Survey rT September and October kV CPUE of Atlantic croaker ^•V larvae and postlarvae n= 1,635 ^:^ Upper Bay Sampling \ ^Y Zones 5 - 8 o *\ \ • '^ \^^ •°'S». • -^^ CPUE o • 1-5 \ o ° «o<^o» 2 oo" o • 6- 15 • > 15 \«o « « / \°^° / o o ( ^ 1 "b ^-^ Lower Bay Sampling \ ^ Zones 1-4 \ _^ Figure 2 Catch per unit of effort (CPUE) of larval and postlar\'al Atlantic croaker (Micropogonias undulatus) collected in the ichthyoplank- ton survey in September and October 1996. Growth rates for YOY Atlantic croaker were calculated for samples collected during the late spring through fall in the upper and lower regions of the bay in 1997 and 1998 and in the upper and lower bay marsh sites during 1997. 1998, and 1999. We compared growth using the progres- sion of the monthly median lengths in each area by com- puting the change in the median length of a cohort over a time period divided by the number of days in the period. This method was based on the following assumptions: 1 ) no new (small) recruits join the population during the calculation interval, and 2) no (large) individuals leave the population over the calculation interval. To best meet these assumptions, median growth rates were calculated by using the monthly length data from May to July when there was a minimum of movement offish between differ- ent areas, and then for longer-term monthly comparisons, from May to August, September, and October when fish were moving out of the marshes into the bay. The smaller- Able, K. W., D . M. Nemerson, and T. M. Grothues. In review. Evaluating salt marsh restoration in Delaware Bay: continued analysis offish response at former salt hay farms. size cohort present in the bay in June and July 1998 and in the marshes in 1999 was excluded from the growth cal- culations for those years. The linearity of the progression of median lengths was tested by using linear regression, and the resulting lines were compared between the upper and lower bay Results Distribution, abundance, and size during fall ingress and settlement Atlantic croaker lan'ae were collected only in the late summer and fall during the ichthyoplankton survey in Delaware Bay in 1996 ( Figs. 2 and 3 ). A few individuals were first collected in August (/z=3, CPUE=0.02 fish/tow), and then large numbers of larvae were collected through September («=639, CPUE=3.6) and October (n=996, CPUE=9.0), but they were absent from April to July The overall September-October CPUE was 9.0 fish/tow (range; 104 Fishery Bulletin 101(1) 10 I I 150 August n = 3 c: Upper bay Lower bay > f September n = 494 October n = 905 10 15 20 25 Total length (mm) 30 ^35 Figure 3 Length-frequency distributions of larval and postlarval Atlantic croaker (Micropogonias undulatus) collected in the upper and lower regions of Delaware Bay during the ichthyoplankton sur\'ey in 1996^ 0-56) in the upper bay zones 5-8 and 4.9 (range: 0-36) in the lower bay zones 1-4 (Figs. 1 and 2), and these CPUE values were significantly different between zones (P=0.03). At least one individual was collected in each of the eight zones in both September and October; the highest two- month combined CPUE occurred in the uppermost zone 8 (CPUE=13.4), followed by zone 5 (CPUE=9.6), and the lowest occurred in zone 3 (CPUE=0.1). Larvae were 4-10 mm during August (all in zone 2i, predominantly 2-24 mm in September, and 5-28 mm in October (Fig. 3) — the small- est individuals being caught in the lower bay. Benthic YOY Atlantic croaker of a variety of sizes first appeared in substantial numbers in September in the otter trawl surveys in both the bay (Fig. 4) and marshes (Fig. 5) at lengths >5 mm, and with modes of 15-30 mm for the primary cohort. Exceptions occurred in the bay in 1997, when they were not collected by the trawl survey until October and when they were not collected during September at two of the three upper bay marsh sites each year. The CPUE of benthic YOY Atlantic croaker was usually highest during October in the lower bay marshes (Fig. 6). This pattern of abundance is illustrated by the much high- er four-year overall CPUE of recently ingressed YOY, espe- cially at Dennis Township, Commercial Township, and Up- per Moores Beach (Fig. 7). The combined four-year CPUFI values were significantly different (/'<0.001 ) at each of the six sites, and the CPUE values at the Dennis Township site were significantly greater than at all the sites except for Commercial Township. Similarly, Commercial Town- ship was different from all sites except Dennis Township and Upper Moores Beach, and Upper Moores Beach also was different from Browns Run in the upper bay. Recently settled YOY Atlantic croaker were also caught in the weirs in small intertidal marsh creeks during Sep- tember, October, and November in all three years; the ma- jority were collected at the Dennis Township marsh in the lower bay (Fig. 8). The monthly CPUE (fish per set) in the weirs at Dennis Township was greatest in October 1997 and November 1999 (the weirs were not in place until October 1996) and the largest total number was collected during 1999. The combined four-year CPUE values for 1996-99 at each of the six sites were significantly different (P<0.001), and the CPUE values at the Dennis Township site were significantly greater than those at all the sites, except Commercial Township. The monthly CPUE values during ingress in the bay also were highest in October, but in contrast to the marsh sites were usually higher in the upper part of the bay (Fig. 6). The combined CPUE values for September and October were significantly different between the upper and lower bay regions in 1998 (P<0.001) and 1997 (P=0.048), but not in 1996 (P=0.51). The combined CPUE values for Septem- ber and October for each year ( 1996-98) were significantly different among years (P<0.001) and were different be- tween 1996 and 1997, and between 1996 and 1998. A second, smaller cohort of YOY Atlantic croaker ap- peared in the bay in -June and July 1998 and in the tidal Miller et al : Distribution, abundance, and growth q\ Micropogonias undulatus 105 1996 1997 1998 25 50 75 100 125 150 175 200 225 250 275 300 Total length (mm) Figure 4 Length-frequency distributions of log (« + l) transformed numbers of Atlantic croaker ^Micropogonias undulatus) collected by otter trawl between April and November in Delaware Bay from 1996 to 1998. None were collected in the bay from April to August of 1996. creeks in June 1999 (Fig 4). In the bay these were as small as 15 mm in June 1998 and had a mode of 26-30 mm. They were even more abundant in the bay during July 1998 and had a mode of 41-45 mm. Individuals of this cohort were collected at five of the eight zones in the bay during June and July but were rare in subsequent months (Fig. 4). A smaller size cohort of YOY (;!=69 fish) also appeared in the marshes (Fig. 5) and in the associated bay stations in June 1999. Distribution, abundance, and habitat use during summer residency Young-of-the-year Atlantic croaker were abundant in Del- aware Bay and in the adjacent marsh creeks from April through the fall egress of each year, except in 1996, when trawling in both the marshes and the bay caught no YOY until 26 individuals (115-200 mm) were collected in the bay in September and October (Fig. 4). In contrast, the 106 Fishery Bulletin 101(1) 1996 1997 1998 1999 25 50 75 100 125 150 175 200 225 250 275 300 Total length (mm) Figure 5 Lciiftth-frcquency distributions of log l/i + ll transformed numbers of Atlantic croaker iMkropoffonias undulatus) collected by otter trawl (April to November) in salt marsh creeks along the northern shore of Delaware Bay from 199(t to 1999. None were collected in the marshes from April to August of 1996. marsh creek surveys found YOY in both the larfje and small creeks from April to September during 1997, 1998, and 1999. Typically in the years after 1996, the CPUE was greatest from April to June at the lower bay marsh sites and then decreased after July to an almost total absence offish toward the end of the fall egress out of the marshes in November (Fig. 6). The overall CPUE at each marsh site for all three year classes combined (April to November in 1997, 1998 and 19991 was highest at Dennis Township and Commercial Township and at Upper Moores Beach in the lower bay and lowest at Lower Moores Beach and at the tipper hay sites iF'ig. 7). As a result, the monthly Miller et al : Distribution, abundance, and growth o\ MIcropogonlas undulatus 107 Upper bay zones Lower bay zones 35 30 25 20 15 ^ 10 5 1996-1997 Year Class 1996 n = 2.733 1997 n = 357 8 9 10 4 5 6 7 8 9 10 60 uj 40 Zl Q. " 20 \ 1997 Year - 1998 1998 ^'^^^ l. = 6,117 1997 T n = 578 i / J"}J 89 10 456789 10 1998-1999 / ^ Year Class 1998 n = 3.612 8 9 10 Month Upper bay marshes ' Lower bay marshes 1996- 1997 Year Class 8 9 10 11 4 5 6 7 8 9 10 11 1999-2000 Year Class 1999 i V n = 3.468 8 9 10 11 Month Figure 6 Monthly average catch per unit of effort (CPUE ±SEM) of young-of-the-year Atlantic croaker (Micropogonms undulatus) collected by otter trawl during the fall ingress and from April to November for three year classes in the upper and lower regions of Delaware Bay (left panels: CPUE=fisli/10 min. tow) and of four year classes in tidal creeks at the upper and lower bay marsh sites I right panels: CPUE=fish/2 min. tow). No data are presented from April to July of 1996 because Atlantic croaker were absent from the bav and marshes until October CPUE at the lower bay marshes was consistently more than twice as high as that in the upper bay (Fig. 6). The combined four-year CPUE values for YOY caught during April-October 1997-99, at each of the six sites were sig- nificantly different (P<0.001), and the CPUE values at the Dennis Township site were significantly greater than at each of the other sites. The CPUE values at Commercial Township and Upper Moores Beach in the lower bay also were greater than those at all the upper bay sites. Young-of-the-year Atlantic croaker also used small in- tertidal creeks in the marshes from April to August, where they were collected in weirs. They were most abun- dant at the Dennis Township site in the lower bay in all three years where they were present from May to July in 1997 and from April to August in 1998, 1999 (Fig. 8). The monthly CPUE (fish per set) in the weirs at Dennis Town- ship was greatest in June of both 1997 and 1998. Com- pared to the total catch in the weirs at Dennis Township in all three years (n=3994), far fewer were caught in the weirs at the other sites in the lower bay during all three years (« = 152) and fewer still at the sites in the upper bay (;)=9i. The CPUE of YOY Atlantic croaker in Delaware 108 Fishery Bulletin 101(1) 25 September - November 1996-1999 20 n = 36,295 15 ^ 10 1 - CPUE O en l.ll ^. 50 April - November 40 1997- 1999 J n = 24,440 30 - 1 20 1 10 J l-.A DT LMB UMB CT BR MHC MC Lower Bay Sites Upper Bay Sites Figure 7 Combined overall average catch per unit of effort (CPUE= fish/2 min. tow |±SEM]) of young-of-the-year Atlantic croaker {Micropogonias undulatus) collected by otter trawl at the Dennis Township (DT), Lower Moores Beach (LMB), Upper Moores Beach (UMB), Commercial Town- ship (CT), Browns Run (BR), Mad Horse Creek (MHC) and Mill Creek (MC) marsh sites during the fall ingi'ess of four years 11996-99) and of post-ingress croaker (April to November 1997-99). Bay was higher in the upper bay, which has muii sedi- ments in most areas, and was much higher in 1998 than in 1997. The monthly CPUE in the upper bay peaked in July or August, but in the lower bay it peaked in October of both years (Fig. 6). The combined CPUE values in the upper and lower bay zones were different between the two regions in both 1997 and 1998 (P<0.001) and the catches within each region were different between the two years (P<0.001). During the summer most YOY were collected in areas of Delaware Bay that had muddy sediments (Fig. 9). In the upper bay zones 7 and 8, which likely have mostly pure mud sediments, YOY were collected at 829f of the stations. In contrast, they were absent in the deeper, large central area of the lower bay that has predominantly sandy and gravelly sediments. However, in the shallow portion of the lower bay, sandy mud, muddy sand, and gravelly mud sediments appear to be distributed on both sides of the bay, and YOY were almost exclusively collected over or near these substrates from April to August, Growth Although YOY Atlantic croaker showed rapid growth during the summer, there was no evidence of growth during the winter. The median growth rates for YOY 10000 ' 1997 1000 100 10 1 0,1 1997 I — I n = 803 All 1996-97Year class 1997-98 Year class n = 4,001 H -H 10000 I 1998 1000 1 100 1 n = 3,167 10 a. o 1 1 10000 - 1999 1000 \ 100 10 1 \l 1997-98Year class 1998-99 Year class n= 1,544 ^JJIJ 1998-1998Year class 1999-2000 Year class n = 24 n = 9,378 0.1 i_Ll APR MAY JUN JUL AUG SEP OCT NOV Montti Figure 8 Catch per unit of effort (CPUE, fish/set) of four year classes of young-of-the-year Atlantic croaker (Micropogonias undulatus) caught each year in the two monthly weir sets across small intertidal creeks at the Dennis Township marsh site in the lower bay during the fall ingress (white bars) and during spring and summer of the following year (black bars). Atlantic croaker calculated for two- to five-month periods beginning in May were fast and ranged from 0.5 to 1,.5 mm/d (Table 3), They were slightly higher in the bay than in the marshes (avg,=l,2 mm/d in the bay and 0,9 mm/d in the marshes) and were lowest in 1998 when Atlantic croaker were most abundant. The growth rates dropped off in the marshes when calculated from May to September or October (Table 3), The lowest early summer growth rates occurred in the tidal creeks in the lower bay in 1998 when the CPUE was the highest. The gi-owth rates in the upper and lower regions of Delaware Bay were similar in each year, but as in the lower bay marshes, the values were lower in 1998 when YOY were much more abundant. Linear regressions of the median lengths used to calculate these growth rates showed that median length was strongly cor- related to date (P=0.02-0.001) and illustrated the slightly slower growth rates at the lower bay sites in both 1997 and 1998 (Fig. 10). These pairs of regression lines were not significantly different (ANCOVA) for upper and lower bay regions of either the marsh sites in 1997 (P=0.1), or in the bay in 1998 (P=0.6), except in 1998, when a lower growth rate was indicated at the lower bay marsh sites (P=0.03). In 1999, a similar linear progi-ession of median lengths was observed in the lower bay marshes (r-=0.98), but sample sizes were too small in the upper bay for growth-rate cal- culations. Although there was no sampling in the winter, the length-frequency distributions indicated that most fish collected in April in the bay and marshes were the same Miller et al : Distribution, abundance, and growth o\ Miaopogonias undulatus 109 Delaware Bay Bottom Trawl Survey April-August 1997 and 1998 CPUE of Atlantic croaker n = 5,178 B 250 mm caught in October 1997 and 1998. The baywide trawling survey did not provide samples in November, so it was impossible to determine the timing of egress of the remaining YOY out of the bay, but very few age-1 fish were present in the bay or marshes by spring of the next year. Discussion Ingress and settlement Young-of-the-year Atlantic croaker ingress into bay and marsh nursery areas starting in the fall of each year in Delaware Bay and in other estuaries along the Atlantic coast. The majority appeared in September, October, and November during our study and in previous collections in Delaware Bay (Able and Fahay, 1998), Chesapeake Bay (Haven, 1957; Chao and Musick, 1977), and North Carolina (Ross, 1988). However, the fall ingress of this cohort was not evident until October in Georgia (Dahlberg, 1972) and December in South Carolina iBearden, 1964). The sudden appearance of significant numbers of larger fish (50-75 mm) in September 1996 in both the bay and marshes and to some extent in the marshes in 1999 sug- gests that individuals that experienced different growth rates or came from different spawning events sometimes occurred simultaneously in Delaware Bay. 110 Fishery Bulletin 101(1) Table 3 Estimated daily growth rates of young-of-the-year ^tlantic croaker (Micropogonias undulat us ) based on the monthly progression of median lengths in the upper and lower regions of Delaware Bay in 1997 and 1998 (see Fig. 1) and in tidal creeks in the marsh sites adjacent to the upper and lower bay in 1997, 1998. and 1999. Growth rate ca culations were made for periods of two to five months, 1 with each period starting in May. Calculations for locations with sample sizes <10 fish were excluded. Average Median length Median growth Habitat Year Location in bay collection date Sample size (mmTL) rate (mm/d) Delaware Bay 1997 Upper bay 17 May 25 33 — 7 July 71 103 1.37 5 Aug 88 143 1.38 2 Sep 16 183 1.39 2 Oct 10 190 1.14 Lower bay 26 May 52 43 — 16 Aug 64 154 1.35 20 Sep 14 187 1.23 4 Oct 96 218 1.34 1998 Upper Bay 13 May 275 41 — 15 Jul 873 121 1.27 10 Aug 433 144 1.16 6 Sep 45 177 1.17 Lower Bay 15 May 326 57 — 19 Jul 220 140 1.28 16 Aug 490 164 1.15 10 Sep 131 182 1.06 Marsh Creeks 1997 Upper Bay 21 May 17 46 — 17 Jul 13 115 1.21 20 Aug 16 154 1.19 14 Oct 33 120 0.51 Lower Bay 26 May 505 52 — 23 Jul 334 110 1.00 26 Aug 84 132 0.87 21 Sep 99 161 0.92 19 Oct 13 150 0.67 1998 Upper Bay 8 May 142 41 — 8 Jul 348 115 1.21 5 Aug 41 143 1.15 2 Sep 23 153 0.96 Lower Bay 13 May 467 50 — 14 Jul 851 100 0.81 11 Aug 174 125 0.83 9 Sep 126 142 0.77 1999 Lower Bay 18 May 148 49 — 18 Jul 128 114 1.07 15 Aug 12 151 1.15 According to the sizes of individuals captured by plank- ton net in the water column, versus those collected by otter trawl on the bottom, settlement may occur over a broad size range, i.e. approximately 10-40 mm TL. Scale formation in Atlantic croaker begins at 14-16 mm SL and is completed at 31-38 mm SL during this time (Bridges, 1971) and is an indicator of transformation between larval and juvenile stages. Alternatively, collection of overlap- ping sizes in water column and bottom samples may imply frequent vertical movements as could occur during tidal stream transport (see Weinstein et al., 1980, for recent ex- amples). These movements would provide an appropriate mechanism for small YOY to reach the bay and the lower Delaware River as has been suggested for larval Atlantic croaker in Chesapeake Bay (Norcross, 19911. The length-freiiuency data from our study and from pre- vious studies along the Atlantic coast indicate that a sec- ond, less-abundant cohort of YOY Atlantic croaker often Miller et a!.: Distribution, abundance, and growth o\ Micropogonias undulatus 111 enter nursery areas during the spring and sum- mer This second cohort (between 10 and 45 mm) was observed in June 1998 and 1999 during our study and in May or August (20-30 mm) in the York River of the Chesapeake Bay (Chao and Mu- sick, 1977). Similarly, a second mode was usually apparent from April through August during three years in North Carolina creeks and bays (Ross, 1988), and in May in Georgia (Dahlberg, 1972). In South Carolina, a second smaller cohort began appearing in March and subsequently became the dominant mode in June and July (Bearden, 1964). In addition, the larger-size individuals that have appeared during the fall months simultaneously with the ingressing fall cohort during our study and in the Chesapeake Bay (Haven, 1957; Chao and Musick, 1977), may be individuals of this late-arriving second cohort that did not enter the Chesapeake and Delaware bays until fall. These late arrivals to nursery areas in the Chesapeake and Delaware bays may be individu- als that were spawned close to or south of Cape Hatteras in late winter because there is no evi- dence of spawning in late winter or spring north of Cape Hatteras in the MAB. Atlantic croaker larvae were caught only from August to Janu- ary 1977-1987 over the continental shelf in the MAB and while entering estuaries in central New Jersey (Able and Fahay, 1998), or from November to February in coastal Virginia (Cowan and Bird- song, 1985). In contrast, just south of Cape Hat- teras, larvae as small as 5.2 mm SL were present in collections made from October through mid- April within and offshore of the Newport River estuary in North Carolina in both 1972-73 and 1973-74 (Lewis and Judy, 1983). Small larvae also were collected in the same estuary from November through mid-April in 1985-1986 (Warlen and Burke, 1990) and 1991-92 (Hettler et al., 1997) and in the Cape Fear es- tuary from mid-March to Mid-April in 1978 (Weinstein et al., 1980). Together, these studies indicate that late winter spawning occurs and suggests that it takes place south of Cape Hatteras. Analysis of otolith microstructure of lai^val and juvenile Atlantic croaker from the MAB indicates that later spawned larvae and juveniles have slower growth rates (Warlen, 1982; Nixon and Jones, 1997), which may account for the much smaller size of the later-arriving cohort when it enters the Chesapeake and Delaware bays during the late spring and early summer. Ross (1988) suggested that there may be two groups of Atlantic croaker that overlap and mix in North Carolina. The first group, occurring from North Carolina southward through the northern Gulf of Mexico, with a tendency to- ward high mortality, lower longevity, early maturation, re- sults from winter spawning (White and Chittenden, 1977; Barger, 1985) and mostly spring recruitment to estuaries. The second group ranges from North Carolina to about New Jersey and may exhibit lower mortality, higher longevity, greater size at age, late summer-fall spawning, mostly fall recruitment, and greater size at maturity (Wallace, 1940; • Upper bay sites ' ' Lower bay sites May Jun Jul Aug May Jun Jul Aug Date Figure 10 Linear regressions and goodness-of-fit measures of the monthly median total lengths of young-of-the-year Atlantic croaker (Micropo- gonias undulatus) caught at the upper bay (open circles) and lower bay (black circles) marsh sites and regions of Delaware Bay from May to August in 1997 and 1998 (see Table 3). The coefficient of determination is shown in the upper left for the upper bay regression lines and in the lower right for the lower bay. There is no regression for the lower bay because sample sizes in this region during June and July of 1997 were too small. Morse, 1980; Barbieri et al., 1994a). However, the group of larger, older Atlantic croaker observed by Ross (1988) apparently has been absent in Chesapeake Bay in recent years (Barbieri et al., 1994b). Lankford et al. (1999) did not find statistically significant genetic differences between fall- spawned YOY Atlantic croaker from north of Cape Hatteras and spring-spawned YOY from south of Cape Hatteras, but YOY from the Gulf of Mexico were genetically discrete from those from the Atlantic coast. This lack of marked genetic differences north and south of Cape Hatteras is not sur- prising if there is southward migration of adults from the MAB during winter as has been suggested (Haven, 1957). Although, spawning and recruitment to nursery areas does appear to occur later in the South Atlantic Bight (Bearden, 1964) and in the Gulf of Mexico (Pearson, 1929; Suttkus, 1955; Hansen, 1969), more research is needed to determine if there are significant biological differences between adults in these two areas and if the late arriving YOY in the north originate from spawning at or south of Cape Hatteras. Habitat use Young-of-the-year Atlantic croaker in this study used the entire range of marsh creek habitats, i.e. small intertidal 112 Fishery Bulletin 101(1) creeks and large subtidal creeks within the study area. Intensive tag and recapture studies in marsh creeks at the Dennis Township site found that YOY were resident for periods of up to 78 days from July through October 1998 (Miller and Able, 2002). As a result, our interpretations of habitat use and gi-owth may be representative for much of the summer and fall in Delaware Bay marsh creeks. In the deeper water of the bay, YOY were collected throughout the whole range of salinities but were most abundant over the predominantly pure mud sediments in the lower Delaware River and over areas with mud sedi- ments elsewhere in the lower bay. This pattern is evident elsewhere because YOY have been reported to be most abundant over soft mud sediments in Apalachicola Bay in the Gulf of Mexico (Kobylinski and Sheridan, 1979). As in Delaware Bay, YOY have been collected over the full range of salinities in South Carolina (Bearden, 1964; Miglarese et al.. 19821 and Georgia (Dahlberg, 1972). However, labo- ratory experiments suggest that lower salinities are meta- bolically less costly for YOY (Moser and Gerry, 1989; Pe- terson et al., 1999) and that in some areas of Chesapeake Bay, YOY are most abundant in regions with low salinities (<18'7„) (Haven, 1957). Habitat use and survival in the winter may vary be- tween estuaries. Young-of-the-year Atlantic croaker appear to overwinter in estuaries in the Gulf of Mexico (Pearson, 1929; Suttkus, 1955; Hansen, 1969; Knudsen and Herke, 1978) and in the South Atlantic Bight (Bearden, 1964; Dahlberg, 1972; Bozeman and Dean, 1980), but in the MAB there is probably significant overwinter mortality in years with particularly cold winters. The YOY appear to over- winter in some estuarine habitats in the York River region of Chesapeake Bay in most years (Haven, 1957; Chao and Musick, 1977) and in deeper areas of the bay (Welsh and Breder, 1923), but in some years YOY have been observed to experience winter mortality based on their subsequent disappearance after a cold period (Massman and Pacheco, 1960) and on direct observations of mass mortalities and collections of dead YOY in years with unusually cold winters (Joseph, 1972; Chao and Musick, 1977, Wojcik, 1978). Further, analysis of long-term recruitment indices for Atlantic croaker from 1979 to 1993 indicates that the YOY of this species may have experienced winter mortal- ity due to low water temperatures in 30'7f of the years in Chesapeake Bay and 74*7^ of the years in Delaware Bay I Lankford and fargett, 2001 ). Overwintering mortality apparently occurred in Dela- ware Bay in 1996 when water temperatures in the region dropped below 3°C and remained below 4°C for an extend- ed period of time. The NOAA Buoy 4409, located in the ocean just south of the mouth of Delaware Bay, recorded water temperatures at about 2-4°C for 18 days during January and February 1996, which is at or below the ap- proximate survival temperature of 3°C determined in lab- oratory experiments (Lankford and Targett, 20011. This apparently resulted in a total absence of YOY throughout the bay and in marsh creeks during the spring and sum- mer, which is not surprising because temperatures in the estuary were likely cooler than in the ocean. In contrast, during the winters preceding the relatively high catch years of 1997 and 1998, water temperatures at the same location never dropped below 4.4°C during the winter of 1996-97 or below 5.6°C during 1997-98. Growth Growth rates that we calculated in both upper and lower regions ofthe Delaware Bay (two years) and in the marshes (three years) ranged from about 0.8 to 1.4 mm/d from May to July. The strong linear correlation between median length and date suggested that the average growth rates were relatively constant during the summer from May to August before egress from marshes. Seasonal growth rates of YOY Atlantic croaker in other estuaries along the Atlan- tic coast may be similar to those in Delaware Bay, but the way in which they were calculated influences the values. Knudsen and Herke ( 1978) reviewed the apparent growth rates of YOY Atlantic croaker from a variety of sources but presented growth rates only for the entire first year of growth, which were all less than 0.5 mm/d for studies along the Atlantic coast and in the Gulf of Mexico. How- ever, these estimates included both larval and overwinter- ing periods; therefore they probably underestimated the growth rates during the summer when growth rates are highest. Monthly modal progression in published lengths indicate relatively fast growth rates during the summer in estuarine areas south of Delaware Bay Our calculation of modal progression in lengths from May to July in vari- ous parts ofthe York and Pamunkey rivers of Chesapeake Bay suggested growth rates of approximately 1.3 and 0.7 mm/d in 1952 and 1953, respectively (Haven, 1957) and of 0.9 mm/d in 1972 (Chao and Musick, 1977). Similarly cal- culated values for May to July for fish from shallow creeks in North Carolina indicated growth rates of 0.6, 0.8, and 0.9 mm/d in 1979, 1980, and 1981. but the 1979 estimate is likely to be an underestimate because many of the larger fish appeared to be moving into deeper habitats during that time period (Ross, 1988). In our study, the growth rates remained relatively high when calculated through October (1.1-1.3 mm/d) in the bay but dropped off in the marshes (0.5-0.7 mm/d), potentially reflecting the egress of larger YOY out ofthe marshes into the bay. Data from the Gulf of Mexico suggest slower growth rates of YOY Atlantic croaker in some areas but egress of larger fish out ofthe sampling area may also bias these estimates. Hansen (1969) used length-frequency data to determine growth rate estimates of 0.3 mm/d from January through August in the Pensacola Estuary on the Florida gulf coast in both 1964 and 1965 but noted the highest growth rates were in July (0.6 mm/d). Knudsen and Herke (1978) es- timated gi-owth of YOY in a semi-impounded marsh in Louisiana using recaptured individuals sprayed with fluorescent pigment during winter and spring and found rates of 0.4-0.5 mm/d for fish marked in late January and early February and recaptured into March. Rates for those marked mid-F'ebruary to late March and recaptured into May were 0.8-0.92 mm/d. A previous study at the same location, using the same techniques, estimated that fish marked from December to March and recaptured into June grew at about 0.47 mm/d (Arnold! et al., 1974 ). The monthly Miller et al.: Distribution, abundance, and growth oi Micropogonias undulatus 113 length-frequency data from Lake Pontchartrain, Louisiana (Suttkus, 1955), indicated a constant but slow growth rate of 0.3 rrtm/d from February to September 1954, and no increase in growth rate during the summer As a result of the above, it appears that growth rates may be faster, and thus countergradient in more northern populations, as suggested for Menidia (Conover and Present, 1990), but care should be taken in interpreting growth rates from the literature, especially those based on modal progression. Egress Young-of-the-year Atlantic croaker have a regular pattern of egress out of tidal creeks and estuaries in the MAB during the late summer and fall after reaching lengths of about 100-250 mm. As we observed in the Delaware Bay system, the majority left the marsh creeks from August to October at lengths <200 mm. The larger indi- viduals appeared to leave the marshes first, as has been observed elsewhere (Haven, 1957; Yakupzack et al., 1977), and almost all had left by November. However, the CPUE increased in Delaware Bay in October of both years, and this may have been caused by fish egressing out of the marshes into the bay. Large individuals remained in Delaware Bay longer than in the marshes and substantial numbers of fish 150-300 mm were present in the bay in September and October This finding suggests that egress from the tidal creeks caused the disappearance of Atlantic croaker there, and not gear avoidance, because large fish continued to be caught in the bay. The exact timing of egress of the majority of Atlantic croaker out of the bay is unclear due to lack of sampling throughout the bay after October However, previous collections in Delaware Bay have shown no evidence of any individual >100 mm from November to March (Able and Fahay, 1998), suggesting that egress out of the bay is finished by November in some years. The same pattern of egress out of nursery habitats in the fall has been observed in Chesapeake Bay (Haven, 1957), but in some years there were substantial numbers offish present into November (Chao and Musick, 1977). Very few of each year class reappear in collections during the spring and summer of the next year in either Chesa- peake or Delaware bays (Haven, 1957; Chao and Musick, 1977; Able and Fahay, 1998) and therefore the fate of these individuals is unknown. Fall egress also occurs out of estuaries in the South At- lantic Bight and the Gulf of Mexico, but in contrast to the Chesapeake and Delaware bays, more Atlantic croaker ap- pear to either remain through the winter or re-enter these habitats in some areas in late winter or early spring. In North Carolina, egress out of tidal creeks was mostly com- pleted by November, but this same year class was present again as age-1 fish in the bays in March, April, and May when sampling resumed (Ross, 1988). A similar pattern of egress from estuaries was observed in South Carolina, but the reappearance of age-1 fish in February was even more prominent and they continued to be collected until fall (Bearden, 1964). In the Gulf of Mexico, some age-1 fish have been observed to remain in estuarine habitats for an additional year in Lake Pontchartrain, Louisiana (Suttkus, 1955), or reappear from January to April after leaving the study area in December in coastal Texas (Pearson, 1929). In summary, this study presents the first compre- hensive examination of YOY Atlantic croaker seasonal- ity and habitat use in Delaware Bay and the adjacent marshes. Although patterns of habitat use and season- ality are similar along the east coast, some divergence from the seasonal patterns in Delaware Bay are evident in estuaries in the South Atlantic Bight and the Gulf of Mexico. Growth estimates appear to be the most di- vergent of any characteristics examined — faster growth rates occurring in the more northern estuaries such as Delaware Bay. Acknowledgments Numerous individuals from the Rutgers University Marine Field Station participated in the field sampling or helped with data analysis. We would particularly like to thank Ralph Bush, Bertrand Lemasson, Steven Teo, and James Chitty. John Balletto and Ken Strait provided background information and logistical support. Jonathan Sharp provided data on sediments in Delaware Bay. Financial support was provided by the Estuary Enhance- ment Program of Public Service Enterprise Group. Literature cited Able, K. W., and M. P. Fahay. 1998. The first year in the life of estuarine fishes in the Middle Atlantic Bight, 342 p. Rutgers University Press, New Brunswick. NJ. Able, K.W., D. Nemerson, R. Bush, and P. Light. 2001. Spatial variation in Delaware Bay (U.S.A.) marsh creek fish assemblages. Estuaries 24(31:441-452. Able, K.W., D. M. Nemerson, P. R. Light, and R. O. Bush. 2000. Initial response of fishes to marsh restoration at a former salt hay farm bordering Delaware Bay. In Concepts and controversies in tidal marsh ecology (M. P. Weinstein and D. A. Kreeger eds. ), p. 749-773. Kluwer Academic Publishers, The Netherlands. Arnoldi, D. C, W. H. Herke, and E. J. Clairain Jr. 1974. Estimate of growth rate and length of stay in a marsh nursery of juvenile Atlantic croaker, Micropogonias undu- latus (Linnaeus I, "sandblasted" with fluorescent pigments. Gulf Caribb. Fish. Inst. 26:158-172. Barbieri, L. R., M. E. Chittenden. Jr, and C. M. Jones. 1994a. Age, growth, and mortality of Atlantic croaker, Micro- pogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics. Fish. Bull. 92:1-12. Barbieri, L. R., M. E. Chittenden Jr., and S. K. Lowcrre-Barbieri. 1994b. Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropogonias undulatus. in the Chesapeake Bay and adjacent coastal waters. Fish. Bull. 92:671-685. Barger, L. Y. 1985. Age and growth of Atlantic croakers in the northern Gulf of Mexico, based on otolith sections. Trans. Am. Fish. Soc. 114:847-850. Bearden, C. M. 1964. Distribution and abundance of Atlantic croaker. Mi- 114 Fishery Bulletin 101(1) cropogonias undiilatiix. in South Carolina. Contrib. Boars Bluff Lab., South Carolina 40:1-23. Bozoman, E. L.. Jr., and ,]. M. Dean. 1980. The abundance of estuarine lar\'al and juvenile fish in a South Carolina intertidal creek. Estuaries 3:89-97. Bridges, D.W. 1971. The pattern of scale development in juvenile Atlan- tic croaker (Micropogonias undulatus). Copeia 1971(2): 331-332. Chao. L. N., and J. A. Musick. 1977. Life history, feeding habits, and functional morphol- ogy of juvenile sciaenid fishes in the York River Estuary, Virginia. Fish. Bull. 75:657-702. Conover, D. O., and T. M. C. Present. 1990. Countergradient variation in growth rate: compensa- tion for length of the growing season among Atlantic silver- sides from different latitudes. Oceologia 83:316-324. Cowan, J. H., and R. S. Birdsong. 1985. Seasonal occurrence of larval and juvenile fishes in a Virginia Atlantic coast estuary with emphasis on drums (Family Sciaenidae). Estuaries 8:48-59. Cronin, L. E., J. C. Daiber. and E. M. Hulbert. 1962. Quantitative seasonal aspects of zooplankton in the Delaware River estuary. Chesapeake Sci. 3(2):63-93. Currin, B. M., J. P. Reed, and J. M. Miller 1984. Growth, production, food consumption, and mortality of juvenile spot and croaker: a comparison of tidal and non- tidal nursery areas. Estuaries 7:451-459. Dahlberg, M. D. 1972. An ecological study of Georgia coastal fishes. Fish. Bull. 70:323-353. Garvine, R. W., R. K. McCarthy, and K.-C. Wong. 1992. The axial salinity distribution in the Delaware Estu- ary and its weak response to river discharge. Estuar Coast. Shelf Sci. 35:157-165. Hansen, D. J. 1969. Food, growth, migration, reproduction, and abundance of pinfish, Lagodnn rliomhuidcs, and Atlantic croaker. Mi- cropogonias undulatus. near Pensacola, Florida, 1963-65. Fish. Bull. 68:135-146. Haven, D. S. 1957. Distribution, growth, and availability of juvenile croaker, Micropogonias undulatus. in Virginia. Ecology 38:88-97. 1959. Migration of the croaker, Micropogonias undulatus. Copeia 1959:25-30. Hettler, W. F. D. S. Peters, D. R. Colby, and E. H. Laban. 1997. Daily variability in abundance of larval fishes inside Beaufort inlet. Fish. Bull. 95:477-493. Joseph, E. B. 1972. The status of the sciaenid stocks of the middle Atlan- tic coast. Chesapeake Sci. 13:87-100. Knudsen, E. E., and W. H. Herkc 1978. Growth rate of marked juvenile Atlantic croaker.s, Mu-ropogonias undulatus, and length of stay in a coastal marsh nursery in southwest Louisiana. Trans. Am. Fish. Soc. 107:12-20. Kobylinski, G. J., and P. F. Sheridan. 1979. Distribution, abundance, feeding and long-term Huc- tuations of spot, Leiastomus xanthurus, and croaker, Micro- pogonias undulatus, in Apalachicola Bay, Florida, 1972- 1977. Contrib. Mar Sci. 22:149-161. Lankford, T E., Jr, and T. E. Targett. 2001. Low-temperature tolerance of age-O Atlantic croak ers: recruitment implications for U.S. Mid-Atlantic estu- aries. Trans. Am. Fish. Soc. 130:236-249. Lankford. T E. Jr.. T E. Targett, and P M. Gaffney 1999. Mitochondrial DNA analysis of population structure in the Atlantic croaker, Micropogonias undulatus, (Perci- formes: Sciaenidae). Fish. Bull. 97:884-890. Lewis, R. M., and M. H. Judy. 1983. The occurrence of spot, Leioslomus xanthurus, and At- lantic croaker, Micropogonias undulatus, larvae in On- slow Bay and Newport River Estuary, North Carolina. Fish. Bull. 81:405-412. Massman, W. H., and A. L. Pacheco. 1960. Disappearance of young Atlantic croakers from the York River, Virginia. Trans. Am. Fish. Soc. 89: 154-159. Miglarese, J. V, C. W. McMillan, and M. H. Sealy Jr 1982. Seasonal abundance of Atlantic croaker (Micropogo- nias undulatus) in relation to bottom salinity and tempera- ture in South Carolina estuaries. Estuaries 5:216-223. Miller, M. J. and K.W. Able. 2002. Movements and growth of tagged young-of-the-year Atlantic croaker, Micropogonias undulatus, in restored and reference marsh creeks in Delaware Bay. J. Exp. Mar. Biol. Ecol. 267:15-38. Morse, W. W. 1980. Maturity, spawning and fecundity of Atlantic croaker, Micropogonias undulatus, occurring north of Cape Hat- teras. North Carolina. Fish. Bull. 78:190-195. Moser, M. L., and L. R. Gerry. 1989. Differential effects of salinity changes on two estua- rine fishes, Le!ostomus.va;i//i!/r!;s and Micropogonias undu- latus. Estuaries 12:35-41. Nixon. S. W., and C. M. Jones. 1997. Age and gi-owth of lai-val and juvenile Atlantic croak- er, Micropogonias undulatus, from the middle Atlantic Bight and estuarine waters of Virginia. Fish. Bull. 95:773- 784. Norcross B. L. 1991. Estuarine recruitment mechanisms of larval Atlantic croakers. Trans. Am. Fish. Soc. 120:673-683. Norcross, B. L., and H. M. Austin. 1988. Middle Atlantic Bight meridional wind component effect on bottom water temperatures and spawning distri- bution of Atlantic croaker. Cont. Shelf Res. 8:69-88. Pearson, J. C. 1929. Natural history and consei-vation of the redfish and other commercial sciaenids on the Texas coast. Bull. L'. S. Bur Fish. 44:129-214 Peterson, M. S., B. H. Comyns, C. F Rakocinski, and G. L. Fulling. 1999. Does salinity affect somatic growth in early juvenile Atlantic croaker, Micropogonias undulatus (L.)? J. Exper. Mar Biol. Ecol. 238:199-207. Reiss. S. R.. and J. R. McConaugha. 1999. Cross-frontal transport and distribution of ich- thyoplankton associated with Chesapeake Bay plume dynamics. Cont. Shelf Res. 19:151-170. Ross, S. W. 1988. Age, gi'owth, and mortality of Atlantic croaker in North Carolina, with comments on population dynamics. Trans. Am. Fish. Soc. 117:461-473. Shenker, J. M.. and J. M. Dean. 1979. The utilization of an intertidal salt marsh creek by larval and juvenile fishes: abundance, diversity and tempo- ral variation Estuaries 2:154-163. Miller et al : Distribution, abundance, and growth of Micropogonias undu/atus 115 Suttkus, R. D. 1955. Seasonal movements and growth of the Atlantic croak- er [Micropogonias undulatus) along the east Louisiana coast. Gulf Caribb. Fish. Inst. 7:151-158. Wallace, D. H. 1940. Sexual development of the croaker, Micropogonias iinclulatus, and distribution of early stages in Chesapeake Bay. Trans. Am. Fish. Soe. 70:475-482. Warlen, S. M. 1982. Age and growth of larvae and spawning time of At- lantic croaker in North Carolina. Proc. Annu. Conf , S.E. Assoc. Fish Wild. Agencies 34:204-214. Warlen, S. M., and J. S. Burke. 1990. Immigration of larvae of fall/winter spawning marine fishes into a North Carolina estuary. Estuaries 13:453- 461. Weinstein, M. P. 1979. Shallow marsh habitats as primary nurseries for fishes and shellfish. Cape Fear River, North Carolina. Fish. Bull. 77:339-357. Weinstein, M. P., S. L. Weiss, R. G. Hodson, and L. R. Gerry. 1980. Retention of three taxa of postlarval fishes in an intensively flushed tidal estuary, Cape Fear, North Carolina. Fish. Bull. 78:419-434. Welsh, W. W., and C. M. Breder 1923. Contributions to the life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141- 201. White, M. L., and M. E. Chittenden Jr 1977. Age determination, reproduction, and population dy- namics of the Atlantic croaker, Micropogonias undulatus. Fish. Bull. 75:109-123. Wojcik, F. J. 1978. Temperature-induced croaker mortality. Coast. Oce- anogi'. Climatol. News 1:5. Yakupzack, P M., W. H. Herke, and W. G. Perry 1977. Emigration of juvenile Atlantic croaker, Micropogo- nias undulatus, from a semi-impounded marsh in south- western Louisiana. Trans. Am. Fish. Soc. 106:538-544. 116 Abstract — Goldband snapper {Pristi- pomoides multidens) coWccted (Tom com- mercial trap and line fishermen off the Kimberley coast of northwestern Aust- raHa were aged by examination of sec- tioned otoliths ( sagittae I. A total of 3833 P. multidens, 80-701 mm fork length (98-805 mm total length), were exam- ined from commercial catches from 1995 to 1999. The oldest fish was estimated to be age 30+ years. Validation of age estimates was achieved with marginal increment analysis. The opaque and translucent zones were each formed once per year and are considered va- lid annual growth increments (the translucent zone was formed once per year between January and May). A strong link between water tempera- ture and translucent zone formation was evident in P. multidens. The von Bertalanffy growth function was used to describe growth from length-at-age data derived from sectioned otoliths. No significant differences in length- at-age were found between sexes and growth parameters were L^, = 598 mm, K = 0.187/yr, t„ = -0.173 (r2=0.76). Re- gression models of estimated age as a function of otolith and fish measure- ments indicated a significant relation- ship between estimated age and otolith weight (r^=0.94). Total instantaneous mortality (Z) estimates generated from catch-at-age data off! multidens from the northern demersal scalefish fishery INDSF) were 0.65 for 1995-96, 0.87 for 1996-97, and 0.71 for 1997-98. Esti- mates of the annual instantaneous rate of natural mortality (M) were 0.10-0. 14. The NDSF population of P. multidens is considered to be exploited above opti- mum levels on the basis of these mortal- ity estimates. The protracted longevity, moderately slow growth and low natural mortality rates of P. multidens predis- poses this species as one vulnerable to overfishing, thus cautious management strategies will be required. Furthermore, capture of P. multidens from depths of 60 meters or greater results in a high mor- tality of fish because the physoclistous ruptures causing internal hemorrhaging and hence there is a low probability of survival of any fish returned to the sea. Thus traditional harvest strategies in- volving size limits will bo inappropriate for these fish. Conversely, hai"vest strate- gies that include appropriately targeted spatial fishery closures may provide a useful additional means of preserving the spawning stock biomass of these fish and protect against recruitment over- fishing. Manuscript accepted 21 August 2002. F'ish. Bull: 116-128 12003). Age validation, growth, mortality, and additional population parameters of the goldband snapper iPristipomoides multidens) off the Kimberley coast of northwestern Australia Stephen J. Newman lain J. Dunk Western Australian Marine Researcti Laboratories Department ol Fishenes Government of Western Australia P.O Box 20 North Beach, Western Australia 6920, Australia E-mail address (for S J Newman) snewmanm'fish wagovau The goldband snapper iPristipomoides multidens. Day), known also as gold- banded jobfish. Day's jobfish and large- scaled jobfish, is widely distributed throughout the tropical Indo-Pacific Ocean region from Samoa in the Central Pacific to the Red Sea in the Western Indian Ocean and from southern Japan south to Australia (Allen, 1985). Along Western Australia, P multidens is found as far south as Cape Pasley (34°S) and is landed in commercial quantities from the Ningaloo Reef area (23°30'S) north- wards (Kailola et al., 1993; Newman, unpubl. data). They inhabit hard bot- tom areas and areas of vertical relief and large epibenthos from depths of 60 to at least 245 m and are concentrated in depths from 80 to 150 m (Allen, 1985; Newman and Williams, 1996). Pristipomoides multidens is a com- mercially important species throughout much of its range, forming an important part of the landed catch in both arti- sanal and developed fisheries (Dalzell and Preston, 1992; Newman, 2001). In Western Australia this highly valued resource is marketed whole, usually fresh on ice, and transported by road from regional ports to markets in most state capital cities. It is occasionally ex- ported. In the Kimberley region, within the northern demersal scalefish fishery (NDSF), P. multidens has composed on average 37.7% of the landed catch from 1995 to 1999 (contributing on average 255 metric tons (t)/year). In terms of val- ue to fishermen, it is second only to the red emperor snapper (Lutjanus sehcw). Information on the biology of P. mul- tidens is limited. The juvenile habitats of P. multidens have not been identi- fied, although Newman (unpubl. data) obtained juveniles from uniform sedi- mentary habitat with no relief. In pre- vious studies, several age determina- tion techniques were used to determine the age of P multidens but there were limited attempts at age validation (Ed- wards, 1985; Mohsin and Ambak, 1996; Richards'). The accurate determina- tion of fish age is the key to estimat- ing growth rates and mortality. Errors in determining fish age can result in ambiguous demographic parameters and provide misleading impressions of the production potential of fish stocks (Newman et al., 2000a). There is a lack of reliable information on the longevity, growth parameters, mortality rates, and population characteristics of P. multidens, despite its ecological and commercial importance. This work represents the first com- prehensive study of age, growth, and mortality of a population of P. mul- tidens based on age estimates from sectioned otoliths and contributes to the management of these stocks. The objectives of this study were to validate aging and to provide age, growth, mor- tality and population characteristics of P. multidens from the Kimberley region of Western Australia that are based on age estimates from sectioned otoliths. • Richards, A. H. 1987. Aspects of the bi- ology of some deep water bottomfish in Pa- pua New (Guinea with special reference to Pristipomoides multidens (Day). Report 87-01, 31 p. Fisheries Research and Sur- veys Branch, Department of Primary In- dustry, Port Moresby, Papua New Guinea. Newman and Dunk: Age validation, growth, mortality, and additional population parameters of Pristipomoides multidens 117 122* 123' 124- 125' ly 128" MWW ^?Derby i^l\ { ) \ ,->" •proome f' T A Western Australia A Figure 1 Location of the northern demersal scalefish fishery (NDSF) off the Kimberley coast of northwestern Australia showing the 50-m, 100-m, and '200-m depth contours. The NDSF is bounded in the west by the 120''E latitude line, to the north by the boundary of the Australian fishing zone (AFZ), and to the east by the border with the Northern Territory. Fishing primarily occurs in depths of 80-140 m. A further objective was to investigate the relationship between estimated age and the measurements of both oto- lith and fish dimensions to assess the applicability of these measurements in predicting the age of this species. Materials and methods Commercial landings of P. multidens from the NDSF off the Kimberley coast of Western Australia were sampled from 1995 to 1999. Samples were acquired opportunisti- cally from July 1995 to December 1996, whereas samples obtained from January 1997 to December 1999 were col- lected on a monthly basis among all vessels in the fleet. All specimens were captured with fish traps at depths of 60 to 200 m from 12°-20°S latitude (Fig. 1). Additional specimens were attained from research vessel cruises with fish traps. All fish were measured to the nearest mm total length (TL), fork length (FL) and standard length (SL), weighed to the nearest g total weight (TW) and cleaned weight (CW), and where possible, sex was determined by examination of the gonads. Cleaned weight is defined as the TW after re- moval of the gills and viscera. Length measurements were used to derive conversion equations with linear regression models [TL=a + h(FL), FL=a + b tTL). FL=a + 6 (SL) and SL=a + b (FL)\. Length-weight models The relationships between FL and both TW and CW were described by the power function where W = weight (TW or CW, g); and L = FL(mm). These relationships were fitted to log-transformed data and the parameters were back-transformed (with correc- tion for bias) to the above form. 118 Fishery Bulletin 101(1) Analysis of covariance («=0.05) was used to determine if there were significant differences in the weights-at-length (FL) relationships between sexes for P multidens. Length and weight data were transformed to natural logarithms to satisfy assumptions of normality and homogeneity. Multiple comparisons were performed with Tukey's hon- estly significant difference (HSD) test. Trends in mean length and weight offish over time were assessed by using analysis of variance (a=0.05). Otolith preparation and analysis Otolith removal, measurement, and preparation followed the procedures and protocols described in Newman et al. (1996), Newman et al. (2000b), and Newman and Dunk (2002). All age estimates were based on the analysis of thin transverse sections of otoliths. These thin sections were examined under a dissecting microscope at 10-30x magnification with reflected light on a black background. The otoliths from eight juvenile P. multidens (80-140 mm FL) were examined for daily bands with a different technique. One sagitta per fish was embedded in epoxy resin and a thick transverse section (>500 pm) was cut. The section was then ground and polished from each side to a level near the core (perpendicular to the long axis of the otolith) by hand with ebony paper (1000 grade) and lapping film (9 and 3 pm). A polished thin transverse sec- tion approximately 100 pm thick was produced. The sec- tion was then examined with a compound microscope. Age validation Marginal increment analysis, routinely used to validate fish age, relies on the assumption that if a translucent zone is laid down once a year, there should be a clear pattern of peri- odic growth on the edge of the otolith during the year. Mar- ginal increment analysis is appropriate only if all fish in the population lay down the translucent zone at the same time. Thus, an annulus consists of a single opaque and a single translucent cycle within a r2-month period. The opaque zone is believed to form during periods of slow growth. Marginal increment analysis usually implies measure- ment of marginal growth and hence many researchers have measured the width of the edge of the otolith sec- tion over an annual cycle. This measurement approach has an advantage in that it should be possible to plot growth of the edge over time to validate that only a single translucent mark is laid down each year. However, in P. multidens, it can be difficult to determine a consistent location to measure on the otolith because of the inherent variability of their otoliths; hence this technique was not used in the present study Edge type analysis was adopted for the marginal incre- ment analysis of P. multidens and edge types were clas- sified according to Pearson (1996) as either translucent, narrow opaque (opaque area less than half of the previous opaque zone), or wide opaque (opaque area greater than half of the previous opaque zone). Sectioned otoliths offish of all ages were examined under a dissecting microscope with reflected light on a black background. Age determination Because the peak spawning period off! multidens occurs in late March, all fish were assigned a birth date of 1 April to assure proper year-class identification. Ages were assigned from counts of annual growth increments con- sisting of alternating opaque and translucent rings from sectioned otoliths (opaque rings were counted). Annual growth increments were counted in the ventral lobe of the otolith from the primordium to the proximal surface, as close as was practicable to the ventral margin of the sulcus acousticus. Annual growth increments were counted with- out reference to fish length or date of capture. Each otolith section was examined on four separate occasions. When the counts differed, otolith sections were re-examined. In most cases that required resolution, the fourth and final count was used for analysis of age and growth because by this time considerable experience had been gained in the interpretation of the otolith structure. Otoliths with structural irregularities (such as unusual calcification, deterioration of the ventral lobe, or poorly defined annual growth increments) were considered indecipherable and were excluded from analysis offish age. Counts were compared and the precision of age esti- mates were calculated with the average percent error (APE) of Beamish and Fournier (1981). Greater precision is achieved as the APE is minimized. The relationship be- tween fish length (FL) and age and otolith dimensions was assessed with linear regression techniques. Timing of translucent zone formation in P. multidens and mean sea surface temperatures (SST was assumed to reflect the temperature at depth) were compared by scal- ing values from the two data sets. The scaling process al- lowed direct comparison of each series and any time lags of one in relation to the other. Using the scaling score = 1 - ((maximum data value - data value) -^ range), where (in the month of November) mean SST for the month was 29°C, the maximum for the year (data set) was 29.7 and the range of the data values was 3.7, we calculated the scaled SST = 1 - ((29.7-29) -=■ 3.7) = 0.81; in addition the scaled '7c frequency of otoliths with translucent edge types = 1 - ((67-20) -^ 67) = 0.30. Growth and mortality models The von Bertalanffy growth function (VBGF) was fitted to estimates of length-at-age with nonlinear least squares esti- mation procedures. The VBGF is defined by the equation L, = L„[l- exp[-K{t - 1^)]], where Lf = mean length offish of age t; L , = asymptotic mean length; /f = is a rate constant that determines the rate at which Lf approaches Lj, t = age of the fish; and ^11 = the hypothetical age at which the mean length is zero if it had always grown in a manner described by the VBGF. Newman and Dunk: Age validation, growth, mortality, and additional population parameters of Piistipomoides multidens 119 Table 1 Length-weight relationships for P. multidens off the Kimberley coast of northwestern Australia parameters a and b of the relationship W = aL'', the sample size (n ), and the regression r- value in mm and the weight is total weight |TW] or cleaned weight ICW] in g). . Estimates were lengths used are obtained for the fork length [FL| Group a h n r2 P. multidens (all fish-TW) 2.483 X lO"^ 2.9501 3680 0.983 P. multidens (all fish— CW) 2.356 X lO"^ 2.9425 3073 0.983 P. multidens (male— TW) 2.156 X 10^^ 2.9737 1963 0.985 P multidens (female— TW) 2.825 X 10^^ 2.9281 1671 0.987 The von Bertalanffy growth curves for both sexes were compared with the likelihood ratio test of Cerrato (1990). Estimates of the instantaneous rate of total mortal- ity (Z) were obtained from catch-at-age data from the NDSF. Annual catch in weight was converted to annual catch in numbers-at-age by the use of age-frequency data standardized by fishing effort to obtain catch-per-age class. Catch in weight was converted to catch in numbers based on the mean weight of P. multidens observed in the sampled catch each year. Mortality estimates were then derived between successive years by obtaining the natural logarithm of the catch per age class (e.g. age 7) in year t and subtracting the natural logarithm of the catch per age class (e.g. age 8) in year ^ + 1 for all fully recruited age classes. Mean total Z was then calculated across all fully recruited age classes. Instantaneous natural mortal- ity rates (M) were derived by using the general regression equation of Hoenig (1983) for fish, where logg Z = 1.46 - 1.01 logg t,fi(^jf- itf^nx=^^^ maximum age in years). The Hoenig equation has been shown to provide a reasonable approximation of M in tropical demersal fishes (Hart and Russ, 1996; Newman et al., 1996; 2000b). The annual percentage removal was estimated by annual percentage = [F/Z (l-e-^)! x 100%. Exploitation rates (E) were derived from the estimates of Z and F as defined by the equation E = F/Z iF=the instantaneous rate of fishing mortality derived from the relationship F=Z-M). Reference points for target (optimal) and limit fishing mortality rates (F , and F,,„„,) were calculated for P. multidens by using the estimate of natural mortality (M), because F,^, = 0.5 M (Walters, in press) and F,,„,„ = 2/3 M (Patterson, 1992). Results A total of 3833 P. multidens (ranging in size from 80 to 701 mm FL [10.6-5770 gTW] ) were examined for age analysis. Of the fish collected, 2063 were males ranging from 245 to 671 mm FL and from 296 to 5195 g TW, and 1751 were females ranging from 284 to 701 mm FL and from 450 to 5770 g TW. Length conversion equations were derived for total length: TL = (1.12xFL) -t- 21.84 (n=2137, 7--^=0.995); fork length: FL = (0.89xrL) - 16.61 (/!=2137, r2=0.995); FL = (1.12xSL) -I- 6.44 (;?=2148, /-=0.992); and standard length: SL = (0.89xFL) - 2.14 (n=2148, r-=0.992). Length-weight models Length-weight relationships were calculated separately for males, females, and for both sexes combined (Table 1). The relationship between TW and FL is presented in Figure 2. ANCOVA of TW-at-FL and CW-at-FL were both significantly different between sexes (TW: F=42.56; df: 1, 3234; P<0.001; CW: F=94.29; df: 1, 2652; P<0.001); males were larger than females. The length-frequency distribu- tion for male and female P. multidens is shown in Figure 3. Temporal trends were evident in the mean length and weight of P. multidens over time. Mean FL was signifi- cantly different among years from 1995 to 1999 (ANOVA; F=31.29; df: 1, 4193, P<0.001), with (1995=1996=1997) > (1998=1999). Mean TW was also significantly different among years from 1995 to 1999 (ANOVA: F=89.33; df: 1, 3295, P<0.001), with 1995 > 1996 > 1997 > (1998=1999). Age validation Otoliths displayed alternating opaque and translucent zones. A consistent annual trend was evident; the trans- lucent zone was laid down from January to May and the opaque zone formed from June to December. The trend in thin opaque zone formation in June and July was replicated in both 1997 and 1998. Figure 4 clearly demonstrates that the opaque and translucent zones are laid down once a year and represent valid annual growth increments. Because the marginal increment analysis involved random sampling across all age classes in the sampled population, the validation of annual growth increments can be expected to hold across all age classes. In addition, the formation of the translucent zone in the sagittal otoliths of P. multidens and the annual cycle of sea surface temperatures in the Kimberley region of northwestern Australia were found to be closely related (Fig 5). Otolith structure, analysis, and functionality The sagittae of P multidens are somewhat laterally com- pressed, elliptical structures. The distal surface is concave and the rostrum and postrostrum are somewhat pointed. The sagittae are characterized by variable growth reticu- lations along the dorsal edge from the postrostrum to 120 Fishery Bulletin 101(1) 6()(K) . / / = 4.S y ^ 5000 ^X - 4000 0) g M' S 3000 o 1- j^F 2000 ^^F 1000 ^^^° ° Male JV^^^^ * Female 100 200 300 400 500 600 700 Fork length (mm) Figure 2 Relationship between fork length and total weight for P. multidens off the Kimberlev coast of northwestern Austraha. SO (I 40 SO - 120 - Male ;i = 22SI -B»- 220 2WI 300 340 3S0 420 4(iO 500 540 5SII h20 660 7110 Length class (10mm) Figure 3 Length-frequency distribution (10-mm longth classes) of male and female P. inulti dens sampled for age determination. the antiro.struni and along the ventral edge from the postrostrum to the rostrum. A curved sulcus crosses the proximal surface longitudinally, and the depth of the sulcal groove increase.^ with (ish age. The precision of otolith readings of P. multidens was relatively high lAl^E of lOA'/i). Given the variability en- countered among otoliths, this APE reflects a moderately high level of precision among otolith readings and indi- cates that the aging protocol adopted is replicable. Otolith length and breadth were useful predictors offish length in P. mullulcns. accounting for more than 17' i of the variability (Table 2). In contrast, otolith weight and, in par- ticular, height were poor predictors offish length (Table 2). Otolith weight was the best predictor offish age for f! multi- Newman and Dunk: Age validation, growth, mortality, and additional population parameters of Pristipomoides multidens 121 yo _ 80 -•- Wide Opaque. -O- Translueent, —tr- Thin Opaque % Frequency o o = o ' '/^^ /'^'°^"n '•-. / 30 :n 10 J J A .S O N D J F M A M J ,1 A IWonths (1997-1998) Figure 4 Marginal increment analysis of P multidens sagittal otoliths using an edge type class- ification system (edge types: wide opaque, thin opaque, and translucent). 1.(1 -A- Scaled SST 2^'°/'°^'*^ •-Q- .Sealed Translucent a^^/ \". 0.8 1 V units / * V Scaling 4- 1 p \ °\ / ^ °\ .O 0.2 ■•V / •■■° ^ 0.0 b--^' \ J J A S O N D J I- M A M J J A Months (1997-1998) Figure 5 Comparison of the translucent zone formation on the sagittal otoliths of P miiltidoia with sea surface temperatures in the Kimberley region of northwestern Australia. Values are standardized to take into account any time lags and to allow comparison | of each time series. dens, accounting for Q'iA'yc of the variability in age (Table 2, Fig. 6). Otolith height was also a useful predictor of fish age, accounting for 88'X of the variability in age. In con- trast, otolith length and breadth were poor predictors of age for P. multidens (Table 2). Growth and mortality models The von Bertalanffy growth curve was fitted to FL-at-age for all P. multidens (Fig. 7), and separately for each sex (Table 3>. Growth in FL of P. multidens is relatively fast to 122 Fishery Bulletin 101(1) M) o 27 o Male g „ i Female o o Unknown o8 ^ ^ 24 u 21 S IS 01 1 s * '^^^^^m^ < '"IrJS^^^^A 12 ^.^^^P^^ 9 oa»^^^^ 6 ^^^^^^^^ 3 #■ 0° 0,0 0,1 0,2 0,3 (1.4 0.5 0.6 0.7 O.S (1.4 1.0 1.1 1.2 Otolith weight (g) Figure 6 Relationsh ip between otolith weight and age of P. mtiltidens estimated from sectioned otoliths. Table 2 Comparisons among regression form v = a ses, fish length (FLl a (SEl of the estimate otolith dimensions and length and age of P multiden^. The predictive equations are of the simple linear + bx (OW=otolith weight; OL=otolith length; 0B= otolith breadth; OH=otolith height). For regression analy- nd age were used as the dependent variables (all regressions were significant at P<0.001).The standard error s a measure of the dispersion of the observed values about the regression line. Dependent variable Independent variable Sample size Equation r2 SEof estimate FL OW 2590 FL = (315.37 X OW) + 339.46 0.68 37.57 FL OL 2493 FL = (33.47 X OL) - 111.43 0.77 32.08 FL OB 3745 FL = (50.43x 05) -124.36 0.83 28.48 FL OH 3988 FL = (118.39x0//)+ 175.14 0.53 48.36 Age OW 2408 Age = (21.86 xOWl- 0.68 0.94 0.94 Age OL 2305 A?e = (1.77 xOD- 21.68 0.58 2.63 Age OB 3469 A?e = (2.40 X OB) -19.14 0.60 2.49 Age OH 3652 A^c = (8.52 xO//)- 12.81 0.88 1.36 age 9 but is much reduceci in age cohorts beyond 9 years of age. Parameters of the VBGF are li.sted in Table 3. FL-at- age of P. multich'ns was not significantly different between sexes (log-likelihood=0.9836, test statistic=1.001, P>0.05; no significant differences were found among parameters of the VBGF; see also Fig. 7). Generalized VBGFs of P. multi- dens from previous studies were compared to that derived from our study (Fig. 8). The maximum observed age of P. multidcns in the Kim- berley region was 30 years. Given that the P. multidcns resource in the Kimberley region has been exploited lor over 20 years, it is possible that in an unfished population the longevity of P. multidcns may be closer to 40 years. These two estimates of maximum age in P. multidens were applied to the Hoenig (19831 equation in order to derive an estimate of M. Consequently, M was considered to be in the range of 0. 104-0. 139, representing an annual survi- vorship of 87-90'^f for an unfished population. This range of M estimates for P. multidcns is similar to that observed for other long-lived lutjanid species in the Indo-Pacific region (Newman et al., 1996, Newman et al., 2000a; New- man and Dunk, 2002). The longevity of female and male P. multidens was somewhat similar at 27 and 30 years, respectively. The age structures of P. multidens in the commercial catch differed among years. The 1995 sample had a peak in year Newman and Dunk: Age validation, growth, mortality, and additional population parameters of Pristlpomoides multidens 123 700 bOO . .* ^ oo°". ^^^^^^^Pfe'cp.- f*°^' ° SOO E 400 ^^^pp- ^°-- © 300 ^^ o "" 200 / 1 o Male 100 f '■ '■=""" 2 4 6 S 10 12 14 16 IS 20 22 24 26 2X 30 32 Age (years) Figure 7 Length-at-age and von Bertalanffy growth curve for P. multidens off the Kimberley coast of northwestern Australia. class 5 and relatively strong age classes 6, 8, and 10, but abundance per age class declined rapidly to age 20, after which few fish were found to be older (Fig. 9). The 1996 and 1997 samples were somewhat similar. In 1996 rela- tively strong year classes were present from age 5 through to age 11, and abundance per age class declined rapidly to age 26 (Fig. 9). One year later, the 1997 sample had relatively strong year classes present from age 6 through to age 12 (Fig. 9), providing further evidence of the annual formation of growth increments. The 1998 sample had relatively strong 6, 7, and 8 age classes, and abundance per age class declined rapidly to age 24 (Fig. 9). The 1999 sample was similar to the 1998 sample with relatively strong 6, 7, and 8 age classes, and abundance per age class declined rapidly to age 20 ( Fig. 9 ). Age classes 9 through 12 were somewhat eroded in the 1998 and 1999 samples in comparison to the 1996 and 1997 samples. In all years, abundance per age class declined rapidly to age 20, and fish older than 20 years were not well represented in the catch over the five years of catch sampling. In most years there was a strong mode of age-6 individuals present and this mode may reflect the age at full recruitment to the sampling gear (fish traps). Pristipomoides multidens less than age 6 were in gen- eral not fully recruited to the sampled population and were therefore excluded from the mortality estimates derived from catch-at-age data. The year-specific total an- nual rate of mortality, Z, of P. multidens in the NDSF, was 0.65 for 1995-96 (fish aged 6-21 years), 0.87 for 1996-97 (fish aged 6-21 years), and 0.71 for 1997-98 (fish aged 6-21 years), representing an annual percentage removal of approximately 38%, 49%, and 41%, respectively, for each Table 3 Growth parameters derived from the von Bertalanffy growth function and population characteristics of P. mul- tidens off the Kimberley coast of northwestern Australia (r!=sample size , FL is in mm. and age (t) is in years). Parameters Male Female Total n 1879 1600 3479 i„ 594.49 603.23 598.08 K 0.1868 0.1867 0.1873 'o -0.3601 0.0018 -0.1730 r^ 0.7394 0.7875 0.7630 n 2281 1916 4573 '' ^mean 501.5 493.5 495.1 '' ^m in 245 284 80 '' '^max 671 701 701 n 1872 1597 3833 hnean 10.24 9.54 9.73 hnin 3 4 0.35 ^max 30 27 30 year (Table 4). In addition, exploitation rates were 0.79, 0.84, and 0.80, respectively. The optimum fishing mortality rate, F^^,, for P. multidens was estimated to be 0.052-0.069, and the limit reference point, F,,„,„, was estimated to be 0.069-0.092 (see Table 4). These results indicate that only approximately 6% of the 124 Fishery Bulletin 101(1) 675 600 .525 E E. 450 c 0) .^75 .WO 225 150 75 Briiuard el al. 1 14X4) growth curve Richards ( 1987) growth curve Edwards ( 1985) growth curve This study 10 12 14 16 18 20 Age (years) 24 26 28 30 .^2 Figure 8 Comparison of generalized von BertalanfTy growth cun'es for P. multidens from pre- vious studies with that derived from this studv. Table 4 Summary of total mortality (2) estimates for P. multidens derived from catch-at-age data based on ages determined from sectioned otoliths. Estimates of fishing mortality If) are derived by subtraction because Z = F + M and are compared to estimates of optimum fishing mortality rates. Year Z P F,. F,.,..., 1995-96 0.649 1996-97 0.869 1997-98 0.710 0.510-0.545 0.052-0.069 0.069-0.092 0.730-0.765 0.052-0.069 0.069-0.092 0.571-0.606 0.052-0.069 0.069-0.092 available stock of P. multidens can be harvested on an an- nual basis in a sustainable manner and that annual har- vest rates should not e.xceed 107( of the average stock size. Discussion Sagittal otoliths were determined to be valid structures for age determination in P. multidens. The edge-type clas- sification system of three edge types used in this study is capable of indicating whether the opaque zone has just been formed or whether a new translucent zone is ready to form. The use of marginal increment analysis (MIA) of individuals of all ages exhibits a clear trend and demon- strates conclusively that annual growth increments are formed once per year Annual growth increments were most conspicuous in the ventral lobe of the sagittal oto- liths. However, we observed that experience is a critical iactor in increasing the agreement and hence [precision of successive counts of annual growth increments in P. multidens. The spring-summer peak in opaque zone formation ob- served in our study is in accordance with the peak in opaque zone formation identified by Fowler (1995) and Beckman and Wilson ( 1995) for tropical fishes. The translucent zone (the period of fast growth in the otoliths) is formed in the summer months (January to May ) and the opaque zone ( the slow growth period) is formed in the winter-spring months (June to December). Translucent zones are relatively thin. Declining sea-surface temperature (which was assumed to reflect water temperature change at depth) was associated with the onset of opaque zone formation in the otoliths of P. multidens. Furthermore, reproduction is unlikely to play a significant role in the timing of translucent zone forma- tion in P. multidens because spawning occurs primarily in the March -April period. These results indicate that water temperature, doubtless in association with other factors, provides a stimulus that Influences the endolymph fluid Newman and Dunk; Age validation, growth, mortality, and additional population parameters of Phstipomoides multidens 125 chemistry of these fish, culminating in the formation of an- nual growth increments. Female and male fish older than 20 years of age were uncommon in the landed catch. Fish of both sexes between 5 and 12 years of age were common in the landed catch. The maximum age of P. multidens observed in our study was much greater than that recorded previously. Richards^ reported a maximum age of 14 years in Papua New Guinea from counts of daily rings on otoliths, whereas Brouard et al. (1984) recorded a maximum age of only 8 years in Vanuatu with a similar method. Edwards (198.5) analyzed vertebrae and scales of this species in the Timor Sea and re- ported a maximum age of 14 years. In contrast, Mohsin and Ambak (1996) estimated a maximum age of only 5 years from the east coast of peninsular Malaysia with length- frequency analysis. Variation in the longevity estimates of earlier works is related to the aging methods used and their biases. For example, growth increments in vertebrae are often difficult to detect despite the presence of numerous discontinuities in bone growth (Marriott and Cappo, 2000). Alternatively earlier longevity estimates may have been drawn from sample populations biased by gear selectivity or from populations with varying degrees of exploitation. Otolith weight was a good predictor of age in P. multidens, accounting for 94% of the variability in age. The strong lin- ear relationship between otolith weight and fish age from a very large sample size implies that otolith weight may be used as a proxy for age. The coefficient of determination of the regression model is affected by the degree of colinearity of the independent variables. The high r^ value observed in our study provides the basis for a first-order age approxi- mation. Thus, the potential exists for an age-otolith-weight key to be derived for P. multidens, as for an age-length key, whereby the age composition of the landed catch in future years may be obtained by weighing large numbers of oto- liths. However, the accuracy and precision of adopting this monitoring strategy remains to be tested. The fit of the regression model for the otolith weight- age relationship was much more precise than the fit of the fork-length-age relationship as described by the von Ber- talanffy growth model. Considerable variation in length was observed within most age groups for both sexes. The large variation in length at a given age makes it difficult to accurately determine the age of P multidens from length data alone. For example, fish ranging in length from 450 to 550 mm FL may vary in age from 5 to 30 years. This vari- ability may explain the very low estimate of maximum age obtained by Mohsin and Ambak ( 1996), which was derived with length-frequency analysis. Growth was most rapid through age 9 for both sexes. From age 9 onwards somatic growth slows with increasing age. The estimation of growth parameters is dependent upon adequate sampling across the length range of any species. The fish sampled in our study ranged in length from 80 to 701 mm FL, covering most of the length range of P multidens. Therefore, it is unlikely that the growth parameters of P. multidens are biased because of inad- equate sampling across the length range. Despite methodological differences in age estimation, the estimates of K derived from the studies of Richards' 1995 n = 328 IS 20 22 24 26 28 30 32 1996 n=984 Hn-n^ 4 fi S :() 22 24 2f> 28 .^0 32 1997 n.702 „nnn^ s 1(1 i; 14 199S 11=1126 OQDii 14 16 IK 211 24 26 2S M) ^2 1999 n=573 24 26 2s Ml M Age (years) Figure 9 Age-frequency distributioii.s ol P. multidens in the northern demersal scalefish fishery from ISO,') to 1999. 126 Fishery Bulletin 101(1) (A'=0.188), Ralston (1987; Ar=0.188), and Edwards (1985; /i'=0.219) were somewhat similar to that observed in this study (A'=0.187). However, the asymptotic lengths report- ed by Richards' and Ralston (1987) were larger than our estimates, which were again similar to those of Edwards, 1985). In contrast, the estimates of A' derived by Brouard et al. (1984), Brouard and Grandperrin (1985), and Mohsin and Ambak ( 1996), which ranged from 0.28 to 0.50, provid- ed overestimates of the growth potential off! multidens as observed in our study. Clearly, methodological differences in age estimation have the potential to unduly influence growth parameter estimation and may provide misleading impressions of the production potential of these fishes. The similarity of growth in length-at-age between sexes indicates that there is little trade off in energetic invest- ment into reproductive activity after sexual maturity at the expense of somatic growth as there is for some Lutja- nus species (Newman et al., 1996, 2000b). Information on energy partitioning in the Pristipomoicles is not known. However, females with a large body size would be repro- ductively "fitter" if they could accommodate a large mass of hydrated eggs prior to spawning, especially in a mul- tiple male, multiple female spawning system. The long life span of P. multidens and other lutjanid species (Loubens, 1980; Newman et al., 1996, 2000a; New- man and Dunk, 2002; Rocha-Olivares, 1998) may be an evolutionary adaptation that supports iteroparity. Many demersal reef fish are highly fecund, but egg and larval survivorship is low; therefore, spawning over numerous years may be necessary to maintain stable populations. In addition, numerous years of reproductive output may also be required to contend with environmental variability (e.g. the incidence of cyclones. El Nino-La Nina cycling), which may substantially influence recruitment success. Extended periods of high exploitation results in decreases in the spawning stock biomass and constriction of the age structure of fish populations, and thus diminishes the number of effective spawnings. Any reduction in the number of effective spawnings may result in a decrease in ecological fitness and hence limit the adaptive capacity of the species to combat environmental or anthropogenic induced stress. Variation in life expectancy due to fishing pressure has the potential to bias estimates of M used in our study. To account for any Af associated difference, a range of M es- timates have been considered in our study. Pristipomoides multidens were fully recruited to the commercial fishery in the NDSF by age 6. Catch-at-age data showed relative- ly consistent estimates of Z among years from 1995-96 through to 1997-98 and a relatively broad age structure in the landed catch. Fishery management implications Throughout much of its range P. multidens composes a significant proportion of the demersal catch of tropical multispecios fisheries. Within these multispccies fisheries P. multidens is taken as part of the directed targ(>t catch or as a part of the retained catch. In fish trawl-based fish- eries, P. multidens can be harvested at all stages of their life history from juvenile to adult, making them especially vulnerable to overexploitation. In contrast, fisheries that use trap and line methods of capture (using bait to attract fish) only have the capacity to harvest fish in the subadult- to-adult phase of their life history. Hence, the method of capture and harvest strategy adopted has the capacity to influence the sustainable exploitation of the P. multidens resource. Of particular relevance to fishery managers is the ca- pacity that fish trawl-based fisheries have in being capa- ble of continuing to function and to be economically viable (driven by the more productive, lower value species) while populations of higher valued species such as P. multidens become depleted. Thus, careful monitoring of the P. multi- dens resource will be required, particularly in trawl-based fisheries. Fishery managers need to be responsive to the intrinsic vulnerability off! multidens to overhai-vesting as a corollary of its life history characteristics. Furthermore, fish such as P. multidens, which have low rates of natural mortality, low growth potential, extended longevity, ma- ture relatively late in life and are either dead or moribund as a consequence of internal hemorrhaging when the physoclistous is ruptured during capture, are likely to be particularly sensitive to exploitation pressure. The appar- ent low survival rate for released fish in the fishing depths of the NDSF fieet indicates that the traditional use of legal minimum sizes to increase survival to spawning sizes and hence increase overall yields is not a practical option. Populations off! multidens have a low productive capac- ity and hence are vulnerable to overfishing as a conse- quence of slow gi-owth, extended longevity, late maturity, and low rates of natural mortality The demersal fish re- sources of the NDSF, of which P. multidens is a significant part, is currently being managed with an innovative total allowable effort system that allocates individually trans- ferable effort units equitably to each licensee. However the highly mobile, efficient, and wide-ranging capacity of the NDSF fleet may require more complex management ar- rangements to maintain future breeding stock levels. The incorporation of appropriately targeted spatial or temporal (or both spatial and temporal) closures within the existing effort management framework is likely to provide an addi- tional useful and robust mechanism to maintain spawning stock biomass and protect against recruitment overfishing. In the wider Indo-Pacific region, fishery managers should consider han'est strategies of low frequency or low intensi- ty in conjunction with targeted spatial or temporal closures to protect the spawning stock biomass of these fishes. Harvest strategies such as setting fishing mortality at or hear natural mortality (F=A/) were often prescribed prior to the 1990s (GuUand, 1970). Recently, the adop- tion of harvest strategies such as setting F = F„ , were thought to be quite conset-vative, but usually resulted in F = M harvest strategies (Walters, in press). Following the meta-analysis of Myers et al. (1999), who examined stock-recruitment cui-ve slopes expressed as maximum reproductive rates per spawner at low spawner biomass, Walters (in press) has reported that optimal fishing mor- tality rates are substantially lower than natural mortality rates for most species and stocks. Furthermore, Patterson Newman and Dunk: Age validation, growth, mortality, and additional population parameters of Phstipomoides multidens 127 (1992) reported that fishing mortahty rates above 2/3 M are often associated with stock decHnes, whereas fishing mortahty rates below this level have resulted in stock re- covery. Consequently, exploitation rates for long-lived reef fishes need to be very conservative. The declines evident in the length and weight of fish in the landed catch over the duration of our study support the finding of the high levels of F. These data support the estimates of the annual percentage removals that indicate that the NDSF population of P. multidens is currently exploited above optimum levels. The age structure of the P. multiderjs stock within the NDSF currently consists of close to 30 age classes (ages 2 to 30 years). Therefore, de- pletion of the spawning stock biomass of these fishes will result in long population recovery times and the economic loss associated with recovering and rebuilding these fish- eries may persist longer A minimum of 30 years would be required for the fished population to recover in terms of both virgin spawner biomass and age structure. The results of our study provide the basis for a more detailed age-structured stock assessment for this species. Acknowledgments The authors gratefully acknowledge funding from the Fisheries Research and Development Corporation (FRDC) for this project. This work was undertaken as part of FRDC Project 97/136. The comments and suggestions of Rod Lenanton and Jim Penn and three anonymous reviewers contributed gi'eatly to this manuscript. Logis- tical support was provided by the Department of Fisher- ies, Government of Western Australia. The authors are thankful to the fishermen of the NDSF for the provision of samples and to the fish wholesalers of Perth (Attadale Seafoods Pty Ltd., Kailis Bros., New West Foods [W.A.I Pty. Ltd., Festival Fish Wholesalers) and Broome (Fortes- cue Seafoods) for access to specimens from northwestern Australia. Jerry Jenke provided invaluable support in all areas, Richard Steckis was responsible for maintaining the databases used for this project, and Peta Williamson assisted with the development of the figures. Literature cited Allen, G. R. 1985. FAO species catalogue. Snappers of the world. An annotated and illustrated catalogue of lutjanid species known to date. FAO Fisheries Synopsis 125, vol. 6, 208 p. FAO, Rome. Beamish, R. J., and D. A. Fournier. 1981. A method for comparing the precision of a set of age determmations. Can. J. Fish. Aquat. Sci. 38:982-983. Beckman, D. W., and C. A. Wilson. 1995. Seasonal timing of opaque zone formation in fish otoliths. In Recent developments in fish otolith research (D. H. Secor, J. M, Dean, and S. E. Campana, eds.), p. 27-44. Univ. South Carolina Press, Columbia, SC. Brouard, F., and R. Grandperrin. 1985. Deep-bottom fishes of the outer reef slope in Vanuatu. South Pacific Commission 17th regional technical meeting on fisheries (Noumea, New Caledonia, 5-19 August, 1985). SPC/Fisheries 17/WP.12, 127 p. [Original in French.] Brouard, F, R. Grandperrin, M. Kulbicki, and J. Rivaton. 1984. Note on observations of daily rings on otoliths of deepwater snappers. ICLARM (International Centre for Living Aquatic Resources Management) Translations 3, 8 p. ICLARM, Manila, Philippines. Cerrato, R. M. 1990. Interpretable statistical tests for growth comparisons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-1426. Dalzell, P., and G. L. Preston. 1992. Deep reef slope fishery resources of the South Pacific. A summary and analysis of the dropline fishing survey data generated by the activities of the SPC Fisheries Progiamme between 1974 and 1988. Inshore Fisheries Research Project Technical Document 2, 299 p. South Pacific Commission, Noumea, New Caledonia. Edwards, R. C. C. 1985. Growth rates of Lutjanidae (snappers) in tropical Australian waters. J. Fish. Biol. 26:1-4. Fowler A. J. 1995. Annulus formation in otoliths of coral reef fish — a review. In Recent developments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds.), p. 45-63. Univ. South Carolina Press, Columbia, SC. Gulland, J.A. 1970. The fish resources of the ocean FAO Fisheries Tech- nical Paper 97, 425 p. Hart, A. M.. and G. R. Russ. 1996. Response of herbivorous fish to crown of thorns star- fish AcanthaMer planci outbreaks. III. Age, growth, mor- tality and maturity indices of Acanthurus nigrofuscus. Mar Ecol. Prog. Ser 136:25-35. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortality rates. Fish. Bull. 82:898-902. Kailola, R J., M. J. Williams, R C. Stewart, R. E. Reichelt, A. McNee, and C. Grieve. 1993. Australian fisheries resources, 422 p. Bureau of Resource Sciences, Department of Primary Industries and Energy, and the Fisheries Research and Development Cor- poration, Canberra, Australia. Loubens, G. 1980. Biologic de quelques especes de poissons du lagon Neo-Caledonian. III. Croissance. Cahiers de I'lndo-Paci- fique 2:101-153. Marriott, R., and M. Cappo. 2000. Comparative precision and bias of five different ageing methods for the large tropical snapper Luljanus johnii. Asian Fish. Sci. 13:149-160. Mohsin, A. K. M., and M. A. Ambak. 1996. Marine fishes and fisheries of Malaysia and neigh- bouring countries, 744 p. Universiti Pertanian Malaysia Press, Serdang, Selangor Darul Ehsam, Malaysia. Myers, R. A., K. G. Bowen, and N. J. Barrowman. 1999. Maximum reproductive rate offish at low population sizes. Can. J. Fish. Aquat. Sci. 56:2404-24 19. Newman, S. J. 2001. Northern demersal scalefish interim managed fishery status report. In Slate of the Fisheries Report 1999-2000 I J. W. Penn, ed.), p. 61-64. Fisheries Western Australia, Perth, Western Australia. Newman, S. J., and I. J. Dunk. 2002. Growth, age validation, mortality, and other popula- 128 Fishery Bulletin 101(1) tion characteristics of the red emperor snapper, Lutjanus sebae (Cuvier, 1828). off the Kimberley coast of North- western Austraha. Estuar Coast. Shelf Sci. 55 (1):67- 80. Newman, S. J., and D. McB. Williams. 1996. Variation in reef associated assemblages of the Lut- janidae and Lethrinidae at different distances offshore in the central Great Barrier Reef Environ. Biol. Fishes 46: 123-128. Newman. S. J.. M. Cappo, and D.McB. Williams. 2000a. Age. growth, mortality rates and corresponding yield estimates using otoliths of the tropical red snappers, Lut- janus erythropterus, L. malabaricus and L. sebae, from the central Great Barrier Reef Fish. Res. 48 (1):1-14. 2000b. Age, growth and mortality of the stripey, Lutjanus carponotatus (Richardson) and the brown-stripe snapper, L. vitta (Quoy and Gaimard) from the central Great Bar- rier Reef Australia. Fish. Res. 48 (3):263-275. Newman, S. J.. D. McB. Williams, and G. R. Russ. 1996. Age validation, growth and mortality rates of the trop- ical snappers (Pisces: Lutjanidae), Lutjanus adetii (Castel- nau, 1873) and L. quinquelineatus (Bloch, 1790) from the central Great Barrier Reef, Australia. Mar Freshwater Res. 47(4):575-584. Patterson, K. 1992. Fisheries for small pelagic species: an empirical ap- proach to management targets. Rev. Fish Biol. Fish. 2(4): 321-338. Pearson, D. E. 1996. Timing of hyaline-zone formation as related to sex, location, and year of capture in otoliths of widow rockfish, Sebastesentomelas. Fish. Bull. 94 (11:190-197. Ralston, S. 1987. Mortality rates of snappers and groupers, /h Tropi- cal snappers and groupers: biology and fisheries manage- ment (J. J. Polovina, and S. Ralston, eds.), p. 375^04. Westview Press, Boulder, CO. Rocha-Olivares, A. 1998. Age, growth, mortality and population characteristics of the Pacific red snapper, Lutjanus peru. off the southeast coast of Baja California, Mexico. Fish. Bull. 96:562-574. Walters, C. J. In press. Stock assessment needs for sustainable fisheries management. Bull. Mar. Sci. 129 Abstract — An assessment of the total biomass of shortbelly rockfish t.Sebastes jordani) off the central California coast is presented that is based on a spatially extensive but temporally restricted ich- thyoplankton survey conducted during the 1991 spawning season. Contempor- aneous samples of adults were obtained by trawl sampling in the study region. Daily larval production (7.56 x 10"^ lar- vae/d) and the larval mortality rate (Z=0.11/d) during the cruise were es- timated from a larval "catch curve." wherein the logarithm of total age-spe- cific larval abundance was regressed against larval age. For this analysis, lar- val age compositions at each of the 150 sample sites were determined by examination of otolith microstructure from subsampled larvae (n=2203), which were weighted by the polygonal Sette-Ahlstrom area surrounding each station. Female population weight-spe- cific fecundity was estimated through a life table analysis that incorporated sex-specific differences in adult g^rowth rate, female maturity, fecundity, and natural mortality (M). The resulting statistic (102.17 larvae/g) was insensi- tive to errors in estimating M and to the pattern of recruitment. Together, the two analyses indicated that a total biomass equal to 1366 metric tons (t)/d of age-l+ shortbelly rockfish (sexes combined) was needed to account for the observed level of spawning output during the cruise. Given the long-term seasonal distribution of spawning ac- tivity in the study area, as elucidated from a retrospective examination of California Cooperative Oceanic Fisher- ies Investigation (CalCOFI) ichthyo- plankton samples from 1952 to 1984, the "daily" total biomass was expanded to an annual total of 67,392 t. An attempt to account for all sources of error in the derivation of this estimate was made by application of the delta- method, which yielded a coefficient of variation of 19%. The relatively high precision of this larval production method, and the rapidity with which an absolute biomass estimate can be obtained, establishes that, for some species of rockfish {Sehastes spp.), it is an attractive alternative to traditional age-structured stock assessments. An approach to estimating rockfish biomass based on larval production, with application to Sebastes jordani* Stephen Ralston James R. Bence Maxwell B. Eldridge William H. Lenarz Southwest Fisheries Science Center National Marine FIshenes Service 1 10 Shaffer Road Santa Cruz, California 95060 E mail address (for S Ralston, contact auttior); Steve RalstoniS)noaa gov Manuscript accepted 20 September 2002. Fish. Bull. 101:129-146 (2003). Shortbelly rockfish (Sebastes jordani) is an underutilized species that is dis- tributed from Vancouver Island to northern Baja California (Eschmeyer, 198.3), although it is especially abun- dant along the central California coast. Based on a swept-area bottom trawl survey of demersal rockfish, Gunder- son and Sample ( 1980) estimated there were 24,000 metric tons (t) of shortbelly rockfish in the Monterey International North Pacific Fishery Commission (INPFC) area (35°30'N-40°30'N). This biomass estimate was far greater than that of any other species of rockfish in any area and, significantly, it did not include the midwater portion of the stock. Hydroacoustic estimates of short- belly rockfish biomass in the shelf-slope area between Ascension Canyon and the Farallon Islands (37°00'-38°00'N), a distance of only 110 km, have ranged from 153,000 to 295,000 t (Nunnely'). Although at present there is no di- rected fishery for this species (Low, 1991), much is known of its biology. Ear- ly work by Phillips (1964) provided ba- sic information about the length-weight relationship, growth as estimated from scale annuli, spawning seasonality (i.e. parturition), fecundity, maturity, and the food habits of shortbelly rockfish. Lenarz (1980) later studied shortbelly rockfish growth using ages from whole otoliths and provided preliminary cal- culations of the effect of fishing on the stock. He also demonstrated marked spatial variation in age and length composition along both latitudinal and depth gradients. Growth was re-esti- mated by Pearson et al. (1991) using ages determined from broken and burnt otoliths. From the hydroacoustic biomass estimates cited above and an estimated range for the natural mortal- ity rate (0.20-0.35 yr), they concluded that the maximum sustainable yield (MSY) of the shortbelly rockfish stock in the Ascension Canyon-Farallon Is- lands area was 13,400-23,500 t. Shortbelly rockfish is one of the few Sebastes spp. that can be readily identi- fied at all life history stages. Descrip- tions of the early life stages of short- belly rockfish, from preflexion larvae through the pelagic juvenile stage, were provided by Moser et al. ( 19771. Extend- ing that work, MacGregor (1986) pro- vided a summary of the spatiotemporal distributions of shortbelly rockfish larvae taken in California Coopera- tive Oceanic Fisheries Investigation (CalCOFI) cruises conducted in five different years. His results showed that 99.1"^^ of all shortbelly rockfish lar\'ae (-4-10 mm) were captured within 90 km of shore and that 65. 4"^* were sampled during the month of Febru- ary. Moreover, a strong peak in larval abundance (i.e. 34.7Cf of the coastwide total) was concentrated in the vicin- * Contribution 111 of the Santa Cruz Labor- atory, Southwest Fisheries Science Center, National Marine Fisheries Service, Santa Cruz, CA 95060. ' Nunnely, E. 1989. Personal commun. Alaska Fisheries Science Center, 7600 Sand Pomt Way N.E., Bin C 15700. Seattle, WA 98115-0070. 130 Fishery Bulletin 101(1) ity of Pioneer Canyon (CalCOFI line 63). Later research by Laidig et al. (1991) resulted in the development of a detailed growth model for young-of-the-year shortbelly rockfish, from extrusion through the late pelagic juvenile stage (-180 d), and verified the feasibility of a daily ag- ing protocol by validating a one-to-one correspondence between counts of daily increments and elapsed time in days. More recent work by Ralston et al. (1996) showed that larval shortbelly rockfish can be accurately aged by using optical microscopy. For the year 2000 the Pacific Fishery Management Council (PFMC) revised the shortbelly rockfish acceptable biological catch (ABC) downwards from 23,500 to 13,900 fyr (PFMC-). The new ABC is based on the low end of the estimated MSY range presented in Pearson et al. (1991); it was reduced due to a probable natural decline in stand- ing stock during the 1990s arising from poor ocean condi- tions (MacCall, 1996). The original range, however, was derived by using quite variable data from unpublished hy- droacoustic surveys and Pearson et al. (1991) considered it a strictly preliminary estimate. Given that the biomass of shortbelly rockfish along the central California coast was once thought to be very large (Gunderson and Sample, 1980; MacGregor, 1986), and that the species is still proba- bly the single largest rockfish contribution available to the west coast groundfish fishery, data are needed to estimate the size of the stock more precisely than results available from these previous investigations. The goal of this study was to develop an analytical ap- proach to estimate the total biomass of shortbelly rockfish in the region of Pioneer and Ascension Canyons and to gather field data to evaluate the method. Successful appli- cation of the method to shortbelly rockfish would provide to the PFMC information useful for management. On a more fundamental level, it would also assist in developing fishery-independent survey techniques capable of assess- ing other, more highly exploited, species of rockfish. The assessment approach The basic premise of egg and larval surveys is that it is easier to estimate the absolute abundance of ichthyo- plankton than it is to estimate that of adults (Saville, 1964; Gunderson, 1993). This is especially true when the spatial distribution of adults exhibits some type of size- or age-specific pattern. Shortbelly rockfish is one such spe- cies (Lenarz, 1980) and obtaining a representative sample of the adult population is challenging (Lenarz and Adams, 1980). Conversely, early life history stages (i.e. eggs and preflexion larvae) can be sampled effectively with stan- dard plankton nets (Smith and Richardson, 1977). Due to the direct coupling between egg [iroduction and spawning biomass, mediated through population weight-specific ■^ PFMC (Pacific Fishery Management Council). 1999. Status of the Pacific coast proundfish fi.shcry through 1999 and recom- mended acceptable biological catches for 2000, 44 p. + .54 tables and 5 figs. Pacific Fishery Management Council. 21.30 SW Fifth Ave.. Portland, OR 97201. fecundity(^[eggs/gl), egg and larval surveys have proven successful for estimating spawning biomass in many applications (e.g. Houde, 1977; Parker, 1980; Richardson, 1981; Lasker, 1985; Armstrong et al., 1988; Hunter et al, 1993). Members of the scorpionfish genus Sebastes are distinc- tive because they are primitive viviparous livebearers (Wourms, 1991), resulting in parturition of advanced yolksac larvae (Bowers, 1992). This reproductive strategy lends itself to a larval production stock assessment be- cause the age of all spawning products can be accurately determined from otolith microstructure (Laidig et al., 1991; Ralston et al., 1996). In contrast, in egg surveys, egg age is back-calculated to the time of spawning by 1 ) defining a series of developmental stages, 2) estimating the relationship between stage-specific developmental rates and temperature, 3) assigning a thermal history to each egg, and 4) determining the time required to account for embryo development from spawning to the obsei-ved stage (see for example Lo, 1985; Moser and Ahlstrom, 1985). Because a distinctive extrusion check forms on the otoliths of Sebastes larvae at the time of parturition (Ralston et al., 1996), a rockfish lai-val production estimate does not require information on temperature-dependent developmental rates and the ambient thermal history of egg samples. In the approach presented in the present study, a spa- tially extensive but temporally restricted ichthyoplankton survey was conducted. The age composition of larvae in each plankton tow was determined by subsampling the catch, aging the subsample, and expanding the subsample age composition back to that of the tow total. The total age-specific abundance of larvae in the study region was calculated by weighting the larval catch-at-age in each plankton sample by the polygonal area around it (see Sette and Ahlstrom, 1948). Characterization of the declin- ing trend in total larval abundance at age with an expo- nential mortality model allowed estimation of the produc- tion rate of day-0 larvae and the larval mortality rate at the time the ichthyoplankton cruise was conducted. Information on adult reproduction was obtained from contemporaneous data collected during two separate cruises conducted during the spawning season. In particu- lar, the following functional relationships were estimated from the data collected: 1 ) weight as a function of total length, 2) sex-specific weight at age (i.e. male and female von Bertalanffy growth equations), 3) fecundity as a func- tion of total weight, 4) maturity as a function of age, and 5) the population sex ratio. From these relationships, a life table was constructed, based on an estimate of natu- ral mortality (/yr), that yielded estimates of population weight-specific fecundity (O). As defined here, population weight-specific fecundity includes the biomass contribu- tions to total population size from males and immature females. Together, these estimates (daily larval production and population weight-specific fecundity) can be used to cal- culate the "daily" total biomass of fish in the population required to produce the obsei-ved abundance of lai^vae. The long-term mean seasonal distribution of shortbelly Ralston et al : An approach to estimating rockfish biomass from larval production 131 rockfish larvae and spawning activity was then determined by a retrospective analysis of all CalCOFI ichthyoplankton samples collected in the vicinity of Pioneer Canyon. Based upon the timing of our lar- val survey with respect to the long-term mean distribution of spawning activity, total "annual" biomass was calculated by expansion. Finally, the robustness of the total biomass estimate was evaluated through a simulation study and sensitiv- ity analyses. Methods Trawl sampling of adults 38- 3 37- lOO-fathom contour 124 Specimens of adult shortbelly rockfish were collected during February-March 1991 by the FV New Janet Ann and the RV Novodrutsk (Table 1). Shortbelly rockfish aggregations in and around Ascension and Pioneer Canyons (Fig. 1) were targeted from acoustic sui-veys. A total of 28 trawls (12 bottom and 16 midwater) were con- ducted over bottom depths ranging from 115 to 384 m, at an average net depth of 126 m (range: 18-210 m). The bottom trawl codend mesh size was 3.8 cm and the midwater trawl mesh was 5.1 cm. Dura- tion of the trawls ranged from 3 to 73 min (x=26 min). All landings were either fully weighed or subsampled, depending on the size of the catch. Landings were highly variable, ranging from 23 to 36,300 kg. Two subsamples were taken from each trawl landing. The first was used to determine the overall age, size, sex, and maturity composition of the catch; it was obtained by randomly selecting and examining 100-300 individuals from each trawl, depending on the availability of both time and fish. For these specimens, total lengths (TL) were measured to the nearest mm. gonads were examined to determine sex and to assign gi-oss maturity stages (see below), and otoliths were removed for later age determina- tion by the break-and-burn method (Pearson et al., 1991). A second smaller subsample of 25-50 specimens was also taken to estimate length-weight and fecundity relation- ships. For each of these fish.TL was measured, weight was determined to the nearest mg, the otoliths were extracted, and late vitellogenic ovaries were dissected from females and fixed in Gilsons fluid. We rated females in terms of maturity based on gross gonadal condition. A summary of the scale we used is the following: 1.0 = immature; 2.0-2.9 = vitellogenic oocytes (yolk deposition with associated oocyte and ovary enlarge- ment); 3.0-3.9 = fertilized eggs (embryos to hatched lar- vae); 4.0 = spent; and 5.0 = reorganization and recovery. Fecundity of female shortbelly rockfish was estimated gravimetrically from vitellogenic ovaries after 2-4 months Pioneer Canyon + + Ascension Canyon -I- + bongo station * adult trawl station * CalCOFI station 63.55 123 122 Longitude Figure 1 Map of the central California study region showing bongo net and adult trawl sampling locations. The annual spawning season was estimated by the long-term seasonal distribution of shortbelly rockfish larvae at CalCOFI station 63.55. Table 1 Summary infor nation of trawl collections for ad ults Trawl locations are shown as closed circles in Figure 1 NJA = RV 1 New Janet Ann NOV = RVNoi'odrutsk. Date Vessel Trawl type Bottom Midwater 14 Feb 1991 NJA 3 1 15 Feb 1991 NJA 2 1 22 Feb 1991 NJA 4 1 23 Feb 1991 NJA 3 1 15 Mar 1991 NOV 3 16 Mar 1991 NOV 3 17 Mar 1991 NOV 2 18 Mar 1991 NOV 2 19 Mar 1991 NOV 2 fixation with periodic stirring. Entire fixed ovaries from each female were blotted dry and weighed to the nearest 1.0 mg. Duplicate subsamples of both ovaries were weighed 132 Fishery Bulletin 101(1) to the nearest 0.1 mg and their egg contents counted with the aid of a dissecting microscope. The mean number of eggs per gram from each fish was then expanded to the total ovary weight to estimate annual fecundity (total eggs per individual female). To estimate population weight-specific fecundity, first the length and weight data were fitted to the power func- tion and the bias-corrected regression equation was used to estimate the weight of every fish that was aged. Next, for each sex, growth equations were obtained by fitting the weight and age "data" to the von Bertalanffy growth model ( Ricker, 1975 ). Maturity was quantified by fitting the logis- tic equation to the proportion of females that were mature within 5-mm-TL intervals, and fecundity was estimated by fitting fecundity and female weight data to the power function, with appropriate bias-correction. These various functional relationships were then combined in a life table analysis to determine the expected biomass per female recruit and the expected lifetime lai-val production per fe- male recruit. The ratio of these quantities is the estimated equilibrium cohort weight-specific fecundity (i.e. Oflarvae/ gl ) of female shortbelly rockfish. Given an estimate of total age-0 larvae (N,,), the female biomass responsible for the observed larval production can be estimated as A^q^"'. Finally, the combined sex biomass can be determined by expanding female biomass using weight-based cohort sex ratio estimates from the life table analysis. Ichthyoplankton sampling The primary set of ichthyoplankton samples used in our study was obtained by using bongo nets during a cruise of the NOAA RV David Starr Jordan (DSJ-9102) conducted in the winter of 1991. Sampling began at 1500 h on 8 Feb- ruary and ended at 0230 h on 15 February. During that 6V2-day period 150 stations were occupied in the region bounded from lat. 36°.30'N to lat. 38°00'N and offshore to a maximum distance of 130 km (Fig. 1). The study area included Pioneer and Ascension Canyons — two features in the continental slope known to harbor large numbers of adults (Lenarz, 1980; MacGregor, 1986; Chess et al., 1988). At these sites the sampling density was increased. Field and laboratory processing of the bongo net samples followed prescribed CalCOFI guidelines (i.e. Kramer et al., 1972; Smith and Richardson, 1977), with minor modi- fication. For example, at every fifth sampling station, the bongo frame was deployed with 333-pm and 505-)im mesh nets to determine the extent of extrusion of small larvae in the standard 505-|im mesh (Lenarz, 1972; Somerton and Kobayashi, 1989). After the nets were washed down, samples from both mesh sizes were preserved in 80% EtOH. At all other stations, the net frame was deployed with two 505-pm mesh nets; one sample was preserved in 80'/^ EtOH (to allow later age determinations from larval otoliths), and the other was preserved in lO'/f buffered for- malin. In addition, because Sehastes larvae were believed to occur only in the upper mixed layer (Ahlstrom, 1959), the maximum amount of wire deployed was 200 m, result- ing in a maximum depth fished ctiual to 140 m. Following splitting, sorting, identification, and enumeration of the larvae in the laboratory, abundance was expressed as the number of shortbelly rockfish per 10 m^ of sea surface. To estimate the age composition of the larval popula- tion, the sorted shortbelly rockfish larvae from each of the 150 EtOH-preserved bongo hauls were randomly subsam- pled for otolith microstructure examination. To determine the size of an age subsample (N^), based upon the total number of larvae occurring in a haul (A^^, ), we applied the following rule: 1) for A^,, less than or equal to 10, N^ = N/^, 2); for A^^, greater than 10 but less than or equal to 410, N^ = 10-1- 0.10 [AT; -10]; and 3) for A^;, greater than 410, N^ = 5o! Otoliths were extracted from each specimen in the haul subsample and individual ages determined by methods outlined in Laidig et al. (1991) and Ralston et al. (1996). The age composition of the larvae in each bongo sample was then estimated by expanding the percent age-fre- quency obtained from the subsample to the haul total (N/,). The estimated numbers-at-age of larvae in each haul were standardized to the number per 10 m^ of sea surface irij.^ for age T and haul i) by application of standard haul fac- tors (Kramer et al., 1972; Smith and Richardson, 1977). We expanded the n-j.^ to the entire survey area by using the method of Sette and Ahlstrom ( 1948). The Sette-Ahl- strom estimate is calculated by the following equation (see Kendall and Picquelle, 1990): ^T=Y,^^'^T" where for each of k hauls A, = the area that haul ( repre- sents (units of 10 m'-); and Nj, = the total abundance of lar- vae of age T in the entire survey area. The area for a haul (A, ) is defined as the area circumscribed by a polygon containing all points in space closer to a haul's location than to the location of any other haul. With this definition, we were able to write a simple computer program to calculate Sette-Alilstrom weights by dividing the study area into a fine grid and assigning each grid point to a haul. Note that this definition and procedure for obtaining Sette-Ahlstrom areas is equivalent to construct- ing polygons manually using perpendicular bisectors and measuring their areas (Sette and Ahlstrom, 1948). The mean daily larval production rate during the cruise was estimated as the bias-corrected antilogarithm of the y-intercept of the ordinary least-squares linear regression of log.lA^J against larval age. Moreover, the regression slope provides an estimate of the total instantaneous mor- tality rate of the larvae (Z |/d|). This calculation implicitly assumes that the age distribution of larvae was stationary throughout the 6V2-day period of the cruise. To determine if shortbelly rockfish larvae occur deeper in the water column than 140 m, a series of 1-m^ multiple- opening-closing-net with environmental sensing system (MOCNESS) tows was conducted aboard the RV David Starr Jordan (cruise l)SJ-9203i during the 1992 spawn- ing season. At that time, 21 tows were made in the area of Ralston et al : An approach to estimating rockfish biomass from larval production 133 Pioneer and Ascension Canyons from 1800 h on 21 Febru- ary to 0600 h on 23 February. Where bottom depth permit- ted, discrete depth samples were gathered using 505-pm mesh nets, sampling obliquely in the 0-40, 40-80, 80-120, 120-160, 160-200, 200-300, and 300-400 m depth inter- vals. Sampling was arrayed along seven onshore-offshore transects, each composed of three tows conducted at dif- ferent bottom depths, i.e. mid-continental shelf (110 m), the shelf-break (183 m), and well off the shelf (550 m). All samples were preserved in EtOH and after sorting, iden- tifying, and enumerating the larvae in the laboratory, we expressed abundances as the number of shortbelly rock- fish larvae per 1000 m'^ water sampled. Spawning seasonality At its inception, this assessment was intended to be an application of the fecundity reduction method described by Lo et al. (1992, 1993). However, samples from the Febru- ary and March adult trawl surveys showed that a higher proportion of females had completed spawning in the ear- lier cruise in comparison with the later cruise, an indica- tion that sampling was not representative during one or both of the cruises. Consequently, the fecundity reduction method was abandoned and an alternative approach was devised. Instead, we estimated the seasonal distribution of spawning activity based on the temporal distribution of preflexion shortbelly rockfish larvae in samples col- lected as part of the CalCOFI program from December to April from 1952 to 1984 (see Ahlstrom et al, 1978). We then used this seasonal spawning pattern to expand our estimate of the daily spawning biomass from the short period represented by our 1991 plankton samples to the entire year. To estimate the seasonal spawning distribution, we first identified the appropriate samples from CALCOFI station 63.55 in the vicinity of Pioneer Canyon by using results from MacGregor ( 1986) as a guide. Plankton samples at this location (Fig. 1) were re-examined and the total number of preflexion shortbelly rockfish larvae were enumerated from Sebastes subsets. These samples amounted to 41 plankton tows (bongo and ring nets) taken in 21 different years. Next, we calculated the mean density of preflexion lar- vae for each month, assigned these densities to the mid- point day of the month, and used nonlinear least-squares regression to fit the normal curve to approximate the sea- sonal spawning pattern, i.e. N,(t)- ¥ ^t-iij' where N „, = the estimated density of preflexion larvae on calendar day t (for December t is nega- tive); /J = the expected value of the seasonal distribu- tion of preflexion larvae; the standard deviation i and f = a "nuisance" scaling constant. a = the standard deviation of the distribution; and To determine the seasonal distribution of age-0 larval production that generates the seasonal distribution of preflexion larvae, we assumed that the preflexion larval period has a duration of 15 days (Laidig et al., 1991) and that larvae experience the estimated preflexion mortality rate (see above). As with preflexion larvae, we approxi- mated the seasonal distribution of age-0 larval production with a normal curve, with mean;/f, and standard deviation a^. Given particular values for/^„ and a^^ we calculated the corresponding relative numbers of age-0 larvae produced during each day of the spawning season, and the integrat- ed seasonal distribution of preflexion larvae, i.e. N,(t). ^£iVo«- i)e- We started the estimation with trial values of //q and a^ and then recursively adjusted the parameter estimates until the mean and standard deviation of the inferred sea- sonal distribution of preflexion larvae (A'',,,) converged on the empirical estimates of u and a . Lastly, based on the timing of the larval survey within the long-term mean spawning distribution, total biomass was calculated by simple expansion. Specifically, the midpoint of the 1991 bongo survey was 11 February (i.e. calendar day 42). Consequently we calculated A, which is the proportion of annual spawning under the age-0 larval production curve that occurs from day 41.5 to 42.5. Total population biomass was then estimated by multiplying the estimate of "daily" biomass by Ilk. This calculation im- plicitly assumes that the seasonal progression of spawn- ing has been stable over years. Therefore, the sensitivity of the biomass estimate to a violation of this assumption was evaluated by profiling over a range of values for the mean date of spawning (;/,,), which has a substantial effect on A. Precision of the biomass estimate The determination of total biomass requires the estimation of numerous statistical relationships, each with its own parameter set. The results of fitting these functions were then combined algebraically to produce the final biomass estimate. We calculated the precision of the final biomass estimate by using the delta method (Sober, 1982. p. 8), i.e. v\g(e)\ = Y^v\e,\\ — \ +2^^cov[0„0,]^ he, de, where g(0) = the algebraic combination of functions used to produce the final biomass estimates; and 6 = the full set of estimated parameters. Application of this method requires estimates of variances for each parameter, covariances among parameters, and partial derivatives of estimated biomass with respect to each parameter (dB/dO). Partial derivatives of the final biomass estimate with respect to the parameters were cal- culated numerically by using central differencing by per- 134 Fishery Bulletin 101(1) turbing each parameter ±1% and calculating the resulting effect on biomass. In some instances, variance estimates (i.e. squared stan- dard errors) and covariances were extracted directly from computer output produced by the SAS (1987) procedures PROC REG and PROC NLIN. In the case of linear regres- sions, these are the usual estimates of these quantities (e.g. Draper and Smith, 1981), whereas for nonlinear re- gression, these are asymptotic variances and covariances (e.g. Seber and Wild, 1992). In addition, following the as- sumptions of normal-based regression, the residual error parameter estimates (ct^,,,^,) used in bias adjustments are independent of other regression parameters (i.e. covari- ances are zero) and n&l^_.Ja^^^ has a chi-square distribu- tion with k degrees of freedom, where k is the degrees of freedom associated with (Tj,^^^. (e.g. for a linear regression k=n-2). This chi-square distributed random variable, i.e. no:. 2k, was then used to estimate the variance of the mean square errors (Larsen and Marx, 1981 ) var(o--.,. x2k. 60 5.5 O) 2 en 5.0 K. 0.2843 0.00917 1.73% -1.58% '0.9 -0.78 0.11026 -0.91% 0.96% Male von Bertalanffy growth WL... 209.9 3.0872 -0.67% 0.67% H', = W„„(l-exp[-/f„(r-«o,o.)l)'' K„ 0.2432 0.00984 -2.05% 2.01% to.cr -1.48 0.15183 1.55% -1.54% Adult survival iog,(no' 3.194 0.30662 0.00% 0.00% log,,(A'j.) = log^,(no)-M(r) M 0.2616 0.02656 -1.46% 1.53% Fecundity at weight log,( 0) 3.8155 0.1732 18.91% -15.90% log^,(F) = \ogJ0) + 5(log,,[W1) 8 ' 1.1416 0.0366 19.88% -16.61% MSE 0.2972 0.01827 0.92% -0.91% Maturity at length V 2.888 0.0237 0.00% 0.00% S= 1 + v/|l+exp|-p(rL-TL')ll P 0.6046 0.10318 0.00% 0.00% TV 135.05 0.85515 -0.15% 0.00% Larval production and survival log,(Afo) -0.3775 0.1737 -15.97% 19.01% log^(Nj.) = log^iN g)-zm Z 0.1107 0.01157 0.70% -0.67% MSE 0,1960 0.05659 -2.79% 2.87% Spawning seasonality "V 5.2 xlO" 2.36 X 103 0.00% 0.00% Nj.t) = (H'l tjpVIjr) exp[- (/ - //^ )V2]/*, tt)Q = the .v-intercept of the growth curve (yr); and j3 = the allometric growth parameter estimated from the regression of weight on length (Ricker, 1975). The /3 parameter is often set equal to 3.0, implying isomet- ric growth, although /3 was fixed at 2.980 in this applica- tion (see above). Likewise, 535 adult male fish were aged, their weights estimated from measurements of TL, and the data fitted to the weight-based von Bertalanffy growth model. Regression results for both sexes are presented in Figure 3 and Table 2. A simple exponential mortality model is used to describe the observed pattern of shortbelly rockfish abundance with age. The model is of the form where Wq = female weight (g) W__Q = the asymptotic mean weight of females at a hypothetically infinite age (g); Kq - the instantaneous growth coefficient specific to females (/yr); T = age (yr); N.,. loe ~MT where N-j. = the number of individuals alive at age T(yr); Pq = the extrapolated number at age 7^=0; and M = the instantaneous rate of natural mortality (/yr). 136 Fishery Bulletin 101(1) 300 - 250 - 200 - 150 - S ° ° § S LaJr^rS-a ^ QMS-^gB e e B-'^o ° ° 8 ° ° o §1^ ° S e e J^ ° o ° o 100 - 50 - 1/ |/i o ° Females I : o 1 i^ Weigh Ol , 1 , , , , o o o o 200 -_ 150 - 100 - o o o ° ° o 8 llJTi « o o^ ° o/ « o 50 - n ^/ :/■ o ° o Males ; 8 8 o : ? D 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 . 1 1 1 1 1 . 1 I 1 1 5 10 15 20 25 Age (yr) Figure 3 Weight-based von Bertalanffy growth curves fitted to shortbelly rockfish data collected by trawhng during February-March 1991. In this application a term for fishing mortality is assumeci to be unnecessary because there is no commercial or recre- ational harv'est of shortbelly rockfish. The model was fitted by a weighted linear regression of log^.lA^j,) on T, where the statistical weights were derived from expansions of the aged subsamples to the full trawl catches. Results showed a declining trend in abundance with age (Table 2, Fig. 4), and an estimated adult natural mortality rate of 0.26/yr Like the weight-length relationship, fecundity is typi- cally related to fish size with the power function (Bagenal and Braum, 1968 1, which is linearized by logarithmic transformation, i.e. F = 0W'\ \og^,(F) = log,(0) + Sx log,,(W), where /■' = individual annual Iccuiulilv i larvae/female); log., I W = female weight (g); and and 5 = fitted parameters. In our study fecundity estimates were gathered from 531 females taken during the trawl surveys and the transformed data were fitted by simple regression (Fig. 5, Table 2). There was considerable residual variability in the fecun- dity at weight relationship (r2=0.65; a%^^. =0.29715) and, in this instance, the addition of a bias-correction term had a noticeable effect on back-transformed predictions of fecundity at weight on the arithmetic scale. Although the distribution of regi'ession residuals deviated significantly from normal (P=0.0001i, due in large part to negative skewness, the overall fit was deemed adequate. Predic- tions of fecundity at weight were ~25-35'^( less than the results presented in Lenarz ( 1980), although his equation is based on only the 10 data points provided in I'hillips 11964). Ralston et al.: An approach to estimating rockfish biomass from larval production 137 3 - 2 - u c cr 1.0 T 0.0 r c y a -1.0 T -2.0 -_ a -3.0 ^ -4.0 - 5 ' I ' ' ' I I ' 10 15 Age (yr) 20 '"I 25 Figure 4 Weighted linear regression of log^.-transformed shortbelly rockfish abundance on age (statistical weights based on expansions of aged subsamples to full trawl catches; all samples collected by trawling during February-March 1991). There was no evidence that the sex ratio of the fish sampled during the 1991 spawning season (n = 1121 from February and March cruises combined) varied with age. A two-way test for independence of age and sex yielded X^=27.55, df=23, P=0.23. Moreover, there was no evidence that the overall population sex ratio was other than 50:50 (A^^->=586. N.y=535: ;f-=2.32, df=l, P=0.14). From these re- sults, we concluded that females and males both enter the population at approximately the same rate and thereafter they experience a similar natural mortality rate. The maturity stage data were used to establish a ma- turity schedule for female shortbelly rockfish collected during the 1991 winter spawning season. Fish were first stratified by size class (5-mm-TL intervals), and for each size group, the mean coded ovarian stage was computed. The data were then fitted by nonlinear regression (SAS, 1987) to a logistic model of the form S=l+- v l + e -p'TL-TL't where S = the mean coded maturity stage; TL = total length (mm); and V, p, and TL' = fitted parameters. Results show (Fig. 6, Table 2) an abrupt change in ovarian condition at a length of 135 mm. At the time of sampling (February-March), virtually all fish above that size were gestating or had already released their larvae, whereas fish smaller than that cutoff size were almost exclusively immature. All females were therefore assumed to be reproductively mature if TL was greater than 135 mm. The proportion of fish in each age class that exceeded 135 mm TL was used to define an empirical age-based 12.0 - 11.0 - o (fecundity) o b i 8.0^ 7.0 - o <^ Pi n b.U -j . 1 i 1 1 . . 1 . 1 1 . . . 1 1 1 1 1 1 1 1 , 1 I 1 r I T 1 3.0 3.5 4.0 4.5 5.0 55 60 lege (total weight [gm]) Figure 5 Fit of ordinary least-squares regression to log^-trans- formed fecundity and weight data from shortbelly rockfish sampled by trawling during February-March 1991. 100 150 200 Total length (mm) 250 Figure 6 Fit of the logistic maturity function to coded ovarian devel- opmental stage data. Circles are means, which are brack- eted by ±1.0 standard deviation (all samples collected by trawling during February-March. 1991 ). maturity ogive. Results indicated that 7.9^; of l-yr-old fish spawned, whereas 99. 0"^/ of 2-yr-old females reproduced. When coupled with some type of recruitment model, the four functional relationships given in Figures 3-6 can be used to estimate population weight-specific fecundity and a weight-based population sex ratio. These latter two vari- ables are presented in Table 3 as part of a life table projec- tion for shortbelly rockfish. In the table, age is increment- ed discretely in one year steps to a maximum life span of 30 yr, which extends well beyond the maximum observed age of 22 years (Pearson et al., 1991). All calculations were 138 Fishery Bulletin 101(1) Table 3 Life table for shortbelly rockfish (Sebastes joi ruary-March). ■dani) based upon samples of ad ults obtained during the 1991 spawning season (Feb- Age (yr) 9or o- numbers 9Wt (g) 9 Cohort Wt(g) Fecundity (larvae/9) Weight-specific fecundity (no. of larvae/g of female) Proportion mature Cohort fecundity (larvae) o-Wt (gl cr cohort Wt(g) 1 1.0000 15.9 15.85 1235 77.90 0.079 98 19.9 19.89 2 0.7698 41.0 31.54 3651 89.11 0.990 2783 39.6 30.50 3 0.5926 71.5 42.37 6893 96.42 1.000 4085 62.0 36.72 4 0.4562 102.4 46.72 10.390 101.45 1.000 4740 84.4 38.50 5 0.3512 130.8 45.93 13,736 105.03 1.000 4824 105.3 36.99 6 0.2704 155.3 41.98 16,710 107.61 1.000 4518 124.0 33.52 7 0.2081 175.6 36.55 19,227 109.50 1.000 4002 140.0 29.14 8 0.1602 192.0 30.76 21,291 110.89 1.000 3411 153.6 24.60 9 0.1233 205.0 25.28 22,943 111.93 1.000 2830 164.7 20.32 10 0.0950 215.1 20.43 24,245 112.70 1.000 2302 173.9 16.51 11 0.0731 223.0 16.30 25,258 113.27 1.000 1846 181.3 13.25 12 0.0563 229.0 12.89 26,040 113.70 1.000 1465 187.2 10.54 13 0.0433 233.6 10.12 26,640 114.02 1.000 1154 192.0 8.32 14 0.0333 237.2 7.91 27,098 114.26 1.000 904 195.8 6.53 15 0.0257 239.8 6.16 27,446 114.44 1.000 705 198.8 5.10 16 0.0198 241.8 4.78 27,710 114.58 1.000 548 201.1 3.98 17 0.0152 243.4 3.70 27,910 114.68 1.000 425 203.0 3.09 18 0.0117 244.5 2.86 28,061 114.76 1.000 329 204.5 2.39 19 0.0090 245.4 2.21 28,175 114.82 1.000 254 205.7 1.85 20 0.0069 246.1 1.71 28,261 114.86 1.000 196 206.6 1.43 21 0.0053 246.5 1.32 28,326 114.89 1.000 151 207.3 1.11 22 0.0041 246.9 1.02 28,375 114.92 1.000 117 207.9 0.85 23 0.0032 247.2 0.78 28,412 114.94 1.000 90 208.3 0.66 24 0.0024 247.4 0.60 28,440 114.95 1.000 69 208.7 0.51 25 0.0019 247.6 0.46 28,461 114.96 1.000 53 208.9 0.39 26 0.0014 247.7 0.36 28,476 114.97 1.000 41 209.1 0.30 27 0.0011 247.8 0.28 28,488 114.97 1.000 32 209.3 0.23 28 0.0009 247.8 0.21 28,497 114.98 1.000 24 209.4 0.18 29 0.0007 247.9 0.16 28,504 114.98 1.000 19 209.6 0.14 30 0.0005 247.9 0.13 28,509 114.98 1.000 14 209.6 0.11 Tola 1 411.37 42.028 347.65 based on a discrete time origin that was centered in the spawning season and it was assumed that samples were obtained at that time. For simpHcity, constant annual recruitment to the pop- ulation at age 1 is assumed, although this particular assumption is later relaxed and its specific effect on es- timates of population weight-specific fecundity is evalu- ated. To start the simulated population, recruitment was arbitrarily set equal to N, = 1.0/yr for both females and males. Then, given female weight at age (Fig. 3) and female numbers at age (Fig. 4), one can calculate age-specific female cohort biomass as the product of numbers and individual weights (Table 3), which when summed over all ages yields the total equilibrium female biomass (411.37 g/female recruit). Likewise, age-spe- cific cohort fecundity is calculated a.s female numbers at age, multiplied by estimates of individual female fe- cundity (Fig. 5), multiplied by the proportion of females that are mature (Fig. 6), which when summed over all ages classes yields the expected lifetime larval produc- tion of a cohort (42,028 lai^vae). The ratio of these two quantities (0=102.17 larvae/g of female) estimates the population weight-specific fecundity of female shortbelly rockfish. This population statistic can be compared with individual age-specific estimates presented in Table 3. These weight-specific fecundities range from 77.90 lar\'ae/g offemale for the mature 1-yr-old females, to 114.98 larvae/g of female for the oldest fish. It is also revealing to compare predicted weight-specific fecundity at age from the life table analysis (Table 3i with the observed data calculated directly from individual fish (Fig. 7). Results show a great deal of variability in the observed data, which is consistent with the extensive residual variability in fecundity (see Fig. 5, Table 2). Life table i)re(iictions were generally similar to, but slightly higher than, the observed data. This slight difference is Ralston et al ; An approach to estimating rockfish biomass from larval production 139 175 150 125 100 75 i 50 25 t iU ^ life table prediction o observed ' I ' ' 10 Age (yr) 15 20 Figure 7 Weight-specific fecundity and its dependency on age. Observed values are means of measured females, brack- eted by ±1.0 standard error and ±1.0 standard deviation (all samples collected by trawl during February-March, 1991). The solid line, which is not fitted to the "observed" points, is the predicted relationship from the life table analysis presented in Table 3. due to the bias correction that occurs when the fecundity relationship, which was fitted on the log-scale, is back- transformed to the arithmetic scale. Irrespective of the type of data, however, it is evident that weight-specific fecundity is only weakly dependent on age. This finding implies that changing the equilibrium age structure of the population, as mediated through an alteration in the natural mortality rate (M), will have little effect on the population's weight-specific fecundity. The assumption that recruitment is constant and uni- form does not appear to appreciably bias the estimate of population weight-specific fecundity derived from the life table analysis (102.17 lai-vae/g of female). Population weight-specific fecundity was also estimated in a Monte Carlo life table simulation that used a more realistic lognormal recruitment model (Fogarty, 1993). Annual recruitments in that simulation were determined as N^ = exp(;/ -I- aX). where N^ is the number of recruits at age 1, // = log^,|10,0001, a = 0.921, and X is a standard normal deviate (i.e. X~Af[0,l]). This level of variability in lognor- mal recruitment is comparable to that observed in the widow rockfish fishery (Bence et al., 1993: Hightower and Lenarz*), where 20-fold differences in recruitment have been observed in a 10-yr time period. Results of the simu- lation showed that fluctuating, lognormal recruitment can Hightower, J. E., and W.H.Lenarz, 1990. Status of the widow rockfish fishery in 1990. In Status of the Pacific coast ground- fish fishery through 1990 and recommended acceptable biologi- cal catches for 1991. Stock assessment and fishery evaluation, appendix vol. 2, 48 p. Pacific Fishery Management Council, Portland, OR. 800 -3 700 ^ 600 -. >, 500 H (J § 400 - CT : '^ 300 -. 200 100 ^ p(91.3 < larvae/g < 106.4) = 0.90 > X = 101.06 n = 5000 -.nnn I T I I I T I j yV t VV 'i " i " i " i " | " i " i " i " i " i " i " i " i " i " |' I I I ' r I I 70 80 90 100 110 Population weigtit-specjfic fecundity (larvae/g of female) Figure 8 Results of Monte Carlo simulation evaluating the effect of recruitment stochasticity on calculations of shortbelly rockfish population weight-specific fecundity. produce values of population weight-specific fecundity that range from 71 to 110 larvae/g of female, depending on the exact sequence of year classes and their resulting affect on age structure (Fig. 8). Even so, the mean of the sample distribution (.v = 101.06 larvae/g of female, n=5000) did not differ significantly from the life table calculation that had no recruitment variability. In addition, 90'^f of all the lognormal recruitment realizations were within ±10'^f of the constant recruitment result. By definition, population weight-specific fecundity rep- resents the number of larvae produced by one gram of female biomass (including immature 1-yr-olds). We ex- panded female biomass to total biomass including males. Results presented in Table 3 show that age-specific male cohort biomass, like that of females, is calculated as the product of numbers-at-age and individual weights-at- age, which when summed over all ages yields the total equilibrium male biomass (347.64 g/male recruit). Total population biomass (i.e. females-t-males) is then 759.01 g, of which females comprise 54.2'7f by weight. Thus, the total biomass is estimated by applying a 1.845 expansion factor to female biomass. Larval production Sampling with different mesh-size bongo nets (333 and 505 pm) allowed an assessment of whether a portion of smaller larvae retained by the smaller mesh was lost with the standard 505-pm mesh. Although shortbelly rockfish are relatively large and stout at parturition (-5.0 mm, Moser et al., 1977), undersampling of small, young larvae could seriously bias larval production estimates. Results presented in Figure 9 show, however, that the standard- ized catches of larvae ( number/| 10 m'-^l ) in the two net sizes were quite similar. A paired t-test of catches (333 minus 140 Fishery Bulletin 101(1) — I — I — I — I — I — I — I — r — I — I — I — I — 1 — I — I — I 3.0 4.0 5.0 60 70 loQc (catch witti 505-(.im net) Figure 9 Paired comparison of log^-transformed larval shortbelly rockfish catches taken in bongo nets with different mesh sizes (all samples collected during February 1991). Also shown is the line of equality. 505) resulted mx= -0.1458, n = 23J = -1.24, and P = 0.23, indicating no difference in the catch of the larger mesh net in relation to the smaller net. We conclude that the 505-pm net was an effective sampler of rockfish larvae. Examination of the standardized catches from the MOC- NESS tows conducted in 1992 revealed that shortbelly rockfish larvae were not caught in depths below the 120- 160 m interval (Fig. 10). The mean catch rate in that depth range was 16.8 larvae/1000 ni'^ which amounted to only l.S'/f of the combined average catch at all depths, i.e. 98.7% of all larvae were captured at depths < 120 m. Be- cause the bongo net was deployed to a maximum depth of 140 m in 1991, and because the mixed layer depth was de- pressed during MOCNESS sampling in 1992 due to strong El Nino conditions (Lynn et al., 1995), we concluded that the 1991 bongo net survey sampled the entire depth range where shortbelly rockfish larvae occurred. For this assessment, a total of 2203 shortbelly rockfish larvae were subsampled from the 505-pm bongo net catch- es and were aged by using optical microscopy (see Ralston et al., 1996). Results from that work indicated that the ages were quite precise (849^ agreement among three readers to within ±1 d) and there were, moreover, no increment in- terpretation differences between optical observations and those made with a scanning electron microscope (SEM). The horizontal distribution of very young (0-2 d) short- belly rockfish larv-ac was strongly associated with the continental shelf break (Fig. 11) and, latitudinally, with the Pioneer Canyon area. Due to the coincidence of this distribution with the locus of adult sampling sites (Fig. 1), we concluded that the trawl survey sampled the adults that produced the lai-vae captured in the iclithyoplankton survey. This finding supports the coupling of our popula- tion weight-specific fecundity statistic (102.17 larvae/g of female) with larval production to estimate the total bio- Sample depth (m) Figure 10 Mean catch rate of shortbelly rockfish larvae in seven depth strata taken by a MOCNESS sampler during March 1992. Means are bracketed bv ±1.0 standard error mass of shortbelly rockfish in the Pioneer Canyon region. Although not shown, older age classes of larvae tended to be more dispersed and to occur increasingly to the north- west — a pattern consistent with hydrographic conditions at the time of the survey (Sakuma et al., 1995). When Sette-Ahlstrom weights were used to expand age- specific standardized bongo net catches to the entire study region, the composite age-frequency distribution of short- belly rockfish larvae appeared to support a log^-linear model with constant exponential mortality (Z |/d| ), i.e. iV^ = AT^e^^r iog^.(7Vj,) = log,,(A^o) - ZT. where Nj = the total abundance of larvae of age Ttd); Z = the instantaneous larval mortality rate (/d); and 7V„ = the lai^val renewal rate (i.e. daily production of larvae). The model was fitted over the first 25 days of life, which largely represents the preflexion stage. In addition, the N-j. were first coded by scaling all obsei-vations to 1x10" and 0.5 was added to all larval ages as a continuity correction (see Ralston et al., 1996). Results show (Fig. 12, Table 2) a satisfactory fit, although there is some suggestion of an aberrant, serially correlated pattern in the residuals from age 0-9 d. The ,v-mtercept term (log.liV,,! = -0.3775), when back-transformed iwith bias-correction 1 yields N,, = 7.562x10'" age-0 lai-vae. Given estimates of daily larval production (7.562x10'" larvae) and population weight-specific fecundity (4>= 102.17 lai-vae/g of female), we calculated from the ratio of these two quantities that 740 t of female shortbelly rockfish spawned each day during the bongo net cruise (DSJ-9102), which is equivalent to a daily biomass of 1366 t/d (sexes combined). Ralston et al : An approach to estimating rockfish biomass from larval production 141 38- OJ ■D 37- O 124 00000 d