Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A global analysis of apparent trends in abundance and recruitment of large tunas and billfishes inferred.. Ahrens, Robert Norman Matthew 2010-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2010_spring_ahrens_robert.pdf [ 17.77MB ]
[if-you-see-this-DO-NOT-CLICK]
Metadata
JSON: 1.0069329.json
JSON-LD: 1.0069329+ld.json
RDF/XML (Pretty): 1.0069329.xml
RDF/JSON: 1.0069329+rdf.json
Turtle: 1.0069329+rdf-turtle.txt
N-Triples: 1.0069329+rdf-ntriples.txt
Original Record: 1.0069329 +original-record.json
Full Text
1.0069329.txt
Citation
1.0069329.ris

Full Text

A GLOBAL ANALYSIS OF APPARENT TRENDS IN ABUNDANCE AND RECRUITMENT OF LARGE TUNAS AND BILLFISHES INFERRED FROM JAPANESE LONGLINE CATCH AND EFFORT DATA by ROBERT NORMAN MATTHEW AHRENS  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) March 2010 © Robert Norman Matthew Ahrens, 2010  Abstract There has been substantial debate in recent years about the extent to which industrialized fishing has affected tunas and other large pelagic predator populations and altered the pelagic community. Variations in the type of data incorporated into assessments, statistical treatment of catch rate information, and different assessment methodologies have lead to diverging interpretations of stock levels and the sustainability of current large-scale industrialized fisheries. Simple nominal catch rates derived from Japanese longline catch and effort data paint a biased picture of the impact of industrialized fishing on the large pelagic tuna and billfish community, suggesting that abundance as of 2002 was only 10% of pre-1950 levels. Methods that correct for the spatial expansion, shift in distribution, and change in species targeting of the Japanese fleet, by averaging catch rates over spatial areas while imputing missing catch rate values, indicate a less severe decline with tuna and billfish stock reduced to an average of 50% of pre 1950 levels. For the majority of stocks, simple assessment methods indicate these relative abundance trends may still be biased and additional information sources are necessary to constrain assessments and evaluate the status of the stocks. With the incorporation of prior information on current fishing mortality rates and in some instances stock productivity, assessments indicate that a number of stocks are over-fished and experiencing over-fishing. Optimization models based on the same catch and effort data, aimed a redistributing fishing effort to maximize profits subject to fishing mortality constraints, suggest economic efficiencies can be gained in the long term if effort reductions are coupled with closed areas. Areas open to fishing should be placed where potential value and recruitments into the fishery are high. Fisheries are complex adaptive systems and it is not necessarily apparent how data resulting from fishing activities relate to the states of the assemblage of species captured. Careful consideration must be given to the nature of the sampling processes that give rise to these data. Without such consideration or alternative sources of information, inferences about impacts of fisheries on natural systems can be severely distorted.  ii  Contents Abstract.......................................................................................................................................... ii Contents ........................................................................................................................................ iii List of tables................................................................................................................................. vii List of figures................................................................................................................................ xi Acknowledgements ................................................................................................................. xxxii Acknowledgements ................................................................................................................. xxxii Chapter 1 General introduction .................................................................................................. 1 Context................................................................................................................................ 1 Thesis structure ................................................................................................................... 4 Chapter 2 Background information.......................................................................................... 10 Development of the Japanese longline fishery ................................................................. 10 Species biology and fisheries............................................................................................ 15 Albacore tuna (Thunnus alalunga) ....................................................................... 16 Biology...................................................................................................... 16 Fisheries .................................................................................................... 19 Bigeye tuna (Thunnus obesus) .............................................................................. 21 Biology...................................................................................................... 21 Fisheries .................................................................................................... 24 Yellowfin tuna (Thunnus albacares) .................................................................... 26 Biology...................................................................................................... 26 Fisheries .................................................................................................... 28 Southern bluefin tuna (Thunnus maccoyii) ........................................................... 30 Biology...................................................................................................... 30 Fisheries .................................................................................................... 32 Pacific bluefin tuna (Thunnus orientalis) ............................................................. 32 Biology...................................................................................................... 32 Fisheries .................................................................................................... 34 Atlantic bluefin tuna (Thunnus thynnus)............................................................... 34 iii  Biology...................................................................................................... 34 Fisheries .................................................................................................... 36 Blue marlin (Makaira nigricans) .......................................................................... 37 Biology...................................................................................................... 37 Fisheries .................................................................................................... 39 Striped marlin (Tetrapturus audax) ...................................................................... 40 Biology...................................................................................................... 40 Fisheries .................................................................................................... 43 Atlantic white marlin (Tetrapturus albidus) ......................................................... 44 Biology...................................................................................................... 44 Fisheries .................................................................................................... 45 Black marlin (Makaira indica) ............................................................................. 45 Biology...................................................................................................... 45 Fisheries .................................................................................................... 47 Swordfish (Xiphias gladius) ................................................................................. 48 Biology...................................................................................................... 48 Fisheries .................................................................................................... 50 Catch and effort data......................................................................................................... 51 Japanese 5°x5° monthly longline catch and effort data............................ 52 5°x5° monthly longline catch for all other countries................................ 52 5°x5° monthly catch by other gears all countries combined .................... 55 Annual nominal catch all gear all countries combined............................. 57 Chapter 3 Deriving relative abundance trends from spatial catch and effort data.............. 79 Introduction....................................................................................................................... 79 Methods............................................................................................................................. 82 Calculating ratio estimators ...................................................................... 83 Correcting for changes in catchability due to hook depth ........................ 84 Estimators derived assuming a stationary population distribution ........... 87 Results............................................................................................................................... 88 Comparison between methods .................................................................. 88 Comparison between methods with imputation........................................ 90 iv  Comparison with indices used in stock assessment.................................. 90 Discussion ......................................................................................................................... 91 Chapter 4 Assessment of recruitment and productivity using catch and relative abundance information ................................................................................................................................ 121 Introduction..................................................................................................................... 121 Methods........................................................................................................................... 123 Recruitment reconstruction................................................................................. 123 Stochastic stock reduction analysis..................................................................... 127 Relative fishing mortality forced numbers dynamic model................................ 129 Results............................................................................................................................. 130 Recruitment reconstruction................................................................................. 130 SRA and relative effort forced models ............................................................... 133 Discussion ....................................................................................................................... 134 Chapter 5 A spatially explicit population dynamics model to assess apparent recruitment, productivity, distribution, and movement patterns............................................................... 164 Introduction..................................................................................................................... 164 Methods........................................................................................................................... 165 Results............................................................................................................................. 171 Overall fits .......................................................................................................... 171 Estimated abundance distributions ..................................................................... 172 Estimated recruitment distributions .................................................................... 174 Estimated management reference points ............................................................ 175 Discussion ....................................................................................................................... 175 Chapter 6 A simulation approach to assessing a mosaic of closures to meet multi-species fishing rate constraints while maximizing profits.................................................................. 198 Introduction..................................................................................................................... 198 Methods........................................................................................................................... 200 Results............................................................................................................................. 204 Discussion ....................................................................................................................... 207 Chapter 7 Summary ................................................................................................................. 233 References.................................................................................................................................. 240 v  Appendix for Chapter 2............................................................................................................ 273 Appendix for Chapter 3............................................................................................................ 299 Appendix for Chapter 4............................................................................................................ 302 Appendix for Chapter 5............................................................................................................ 361 Appendix for Chapter 6............................................................................................................ 386  vi  List of tables Table 1.1 Stock breakdown and short identifier coded associated with each species. ................... 7 Table 1.2 A description of acronyms used in this document.......................................................... 8 Table 2.1 von Bertalanffy growth equation parameters, instantaneous annual natural mortality rates, age, or size at 50% maturity, and length-to-weight conversions parameters for tuna and billfish stocks by ocean. von Bertalanffy growth equation parameter K is an annual rate and t0 is in years. Where multiple values are presented, bold values indicate parameters used for conversions in this thesis. Unless otherwise stated, length for tunas is fork length (cm) and weight is round weight (kg). For billfish, length is lower jaw fork length (cm) and weight is round weight (kg). EFL indicates eye fork length. Male and female relationships are designated with m and f. Mortality rates designated with † are approximated from mortality at age schedules for longline vulnerable individuals. Repeated references are indicated with ("). ...................................................................... 59 Table 2.2 Values used to convert gear specific catches to number of longline vulnerable individuals in the catch. W is the mean weight in kilograms (kg) estimated by gear type. L90% is the average ocean specific length for which 90% of longline caught individual are larger. Proportion (p) > L90% is the estimated proportion of the gear specific catch greater than the minimum longline size, and treated as part of the longline-vulnerable catch . Weight for all species is round weight (kg). Length for tuna is fork length (cm) and length for billfish is lower jaw fork length (cm). For yellowfin and bigeye tuna, purse seine A, U and D indicate associated, unassociated and dolphin sets respectively. Pacific northern bluefin tuna conversions for purse seine are presented as decadal averages. Repeated references are indicated with ("). ...................................................................... 65 Table 3.1 Coefficients for determining changes in catchability due to hook depth for species considered in the analyses presented in this thesis; coefficient estimates from Ward and Myers(2005a).................................................................................................................. 102 Table 3.2 Apparent change in stock abundance using four different relative abundance indices calculated as the percentage difference between the averaged over the first decade of fishing and the last decade. First year used for the Indian Ocean is 1952, for the Pacific is vii  1950, and for the Atlantic is 1956. Columns represent the different methods shown in Figure 3.2 to Figure 3.4. Nominal is nominal cpue, mean fished is the average cpue from fished areas, Poisson is derived assuming a Poisson model and spatial is derived averaging over all areas with missing space/time strata imputed (see description of SF31 in methods)...................................................................................................................... 103 Table 3.3 Areas subdivisions and references for relative abundance trends used in stock assessments by RFMOs. Repeat references are marked with ("). .................................. 104 Table 4.1 Fixed parameter values and leading or derived parameter prior probability distribution values. k is the assumed age at first vulnerability to longline gear. For the recruitment reconstruction (figure legend letters A-F) M is fixed and three hypothesis of Fcur are explored. For the stochastic SRA and Effort Forced analyses, prior distributions, where applicable, are assumed lognormal with means given below and standard deviations shown in parentheses. References are provided for Fcur mean values. Stock identified with a * indicate substantial catches prior to 1950. ........................................................ 143 Table 4.2 Compensation ratio estimates for various stocks from sSRAs where process error was assumed low versus high and from F/q forced assessments. Stocks without values were assessed with priors on Fmsy. ........................................................................................... 144 Table 4.3 Point estimates of reference points from assessments produced by RFMOs. The year column indicates the assessment year. The B ratio column presents the ratio of current biomass to the biomass that produces MSY and the F ratio column is the ratio of current fishing mortality rate to the fishing mortality rate that would produce MSY. ............... 145 Table 6.1 Mean weights and relative ex-vessel prices for species considered in Chapter 6. ..... 214 Table 6.2 Optimization model results expressed as percentages relative to average values for 1998-2002 for relative change in average fishing mortality, value, areas status, as well as the number of stocks for which target fishing mortality was exceeded. Results are presented for various cost and movement rate (v) assumptions. In the final uniform cost scenario, billfish catchability (q) is halved. .................................................................... 215 Table 6.3 Linear model coefficients for predicting optimized cell effort, along with R-squared values and significance codes for optimization scenarios for various cost assumptions and initial movement rates. In the final uniform cost scenario, billfish catchability (q) is halved. Codes "***", "**", "*", "." indicate p-values of < 0.001, <0.01, <0.05, <0.1. .. 216 viii  Table 8.1 Proportion (p) of nominal catch reported spatially by longline fleets in the Atlantic Ocean. Fleet ID. was assigned to facilitate plotting and corresponds to the ICCAT fleet ID in the ICCAT task II database. .................................................................................. 274 Table 8.2 Proportion (p) of nominal catch reported spatially by longline fleets in the Indian Ocean. Fleet ID. was assigned to facilitate plotting and corresponds to the IOTC fleet ID in the IOTC task II database. .......................................................................................... 275 Table 8.3 Estimated total landings (number in thousands) of longline vulnerable individuals caught by all gears combined, for the stocks as defined in this study. ........................... 276 Table 9.1 Yearly average cpue values calculated using the SF31 spatial filling method........... 300 Table 10.1 Fmsy and MSY estimates from recruitment reconstructions assuming Beverton-holt (BH) and Ricker (R) recruitment relationships for 3 current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. Values are presented as mode (95% credible interval) and (-) indicates a point estimate.* indicates where estimates have exceeded maximum or minimum values. ....................................................................... 303 Table 10.2 Biological reference points Fratio, the ratio of current fishing mortality to the fishing mortality that produces MSY, and Nratio, the ratio of current stock size to the stock size that produced MSY when fished at Fmsy. Estimates are from recruitment reconstructions assuming Beverton-holt (BH) and Ricker (R) recruitment relationships for 3 current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. Values are presented as, mode (95% credible interval) and (-) indicates a point estimate............... 306 Table 10.3 Parameter estimates from stock reduction analysis (SRA) with process error assumed low or high and from the E* forced method. Parameters are presented as prior mode (95% interval) posterior 50% quantile as an approximation to posterior mode (95% interval). A single set of numbers indicates no prior was used. ..................................... 333 Table 10.4 Biological reference point estimates from stock reduction analysis (SRA) with process error assumed low or high and from F/q forced method. Parameters are presented as posterior 50% quantile (95% interval). Fratio is the ratio of current fishing mortality rate to Fmsy. Nratio is the ratio of current stock size to the stock size that would produce MSY if fished at Fmsy.................................................................................................... 336  ix  Table 11.1 Leading parameter MLE values and 95% credible interval for various initial movement rates, from fitting the spatial model to catch data. See Table 4.1 for prior distributions..................................................................................................................... 362 Table 11.2 Biological reference points estimated by fitting the spatial model to catch data, given various base mixing rates v. ............................................................................................ 364  x  List of figures Figure 1.1 Global catch of principal market tuna and billfish from 1950-2002. Bluefin sp. includes Atlantic, Pacific, and southern bluefin and marlin sp. includes blue, black, white, and striped marlin. .............................................................................................................. 9 Figure 2.1 Japanese catch (thousands of metric tonnes) of tuna and billfish from 1905-2002. Tuna catches excluded skipjack. Orange bars indicate the years of World War II. ......... 68 Figure 2.2 Map of the Northwest Pacific and MacArthur Lines. ................................................. 69 Figure 2.3 Contour plot of the approximate year when areas were first fished by the Japanese longline fleet. .................................................................................................................... 70 Figure 2.4 Proportion of each stock’s distribution area fished by Japanese longline gear from 1950-2002 measured as the proportion of 5°x5° areas fished. ......................................... 71 Figure 2.5 Temporal changes in the seasonal distribution of fishing effort for each species....... 72 Figure 2.6 Average number of months that areas were fished by Japanese longline after first receiving effort.................................................................................................................. 73 Figure 2.7 Decadal average 5°x5° effort distribution of the Japanese longline fleet from 19502002................................................................................................................................... 74 Figure 2.8 Proportion of nominal longline catch accounted for in spatial records (black bars) for countries reporting catch from the Indian and Atlantic oceans. ....................................... 75 Figure 2.9 Total estimated catch, in millions, of individuals of size vulnerable to longlines captured by Japanese longline (blue), the longlines of all other nations combined (orange), other gear types (light grey), and all other gear types not reported spatially (white) for the Indian Ocean (1950-2002). ....................................................................... 76 Figure 2.10 Total estimated catch, in millions, of individuals of size vulnerable to longlines captured by Japanese longline (blue), the longlines of all other nations combined (orange), other gear types (light grey), and all other gear types not reported spatially (white) for the Pacific Ocean (1950-2002). ...................................................................... 77 Figure 2.11 Total estimated catch, in millions, of individuals of size vulnerable to longlines captured by Japanese longline (blue), the longlines of all other nations combined  xi  (orange), other gear types (light grey), and all other gear types not reported spatially (white) for the Atlantic Ocean (1950-2002). .................................................................... 78 Figure 3.1 Relative changes in species catchability estimated as a function of hooks between floats................................................................................................................................ 105 Figure 3.2 Relative abundance indices developed using four different methods, standardized to their mean, for stocks in the Indian Ocean from 1952-2002. ......................................... 106 Figure 3.3 Relative abundance indices developed using four different methods, standardized to their mean, for stocks in the Pacific Ocean from 1950-2002. ........................................ 107 Figure 3.4 Relative abundance indices developed using four different methods, standardized to their mean, for stocks in the Atlantic Ocean from 1956-2002........................................ 108 Figure 3.5 Relative abundance indices developed using four variations of the spatial filling method, standardized to their mean, for stocks in the Indian Ocean from 1952-2002. .. 109 Figure 3.6 Relative abundance indices developed using four variations of the spatial filling method, standardized to their mean, for stocks in the Pacific Ocean from 1950-2002. . 110 Figure 3.7 Relative abundance indices developed using four variations of the spatial filling method, standardized to their mean, for stocks in the Atlantic Ocean from 1956-2002. 111 Figure 3.8 Comparison between relative abundance indices derived for stock assessment in the Indian Ocean using Japanese longline data and relative abundance trends developed using the SF31 variation of the spatial filling method.............................................................. 112 Figure 3.9 Comparison between relative abundance indices derived for stock assessment in the Pacific Ocean using Japanese longline data and relative abundance trends developed using the SF31 variation of the spatial filling method.................................................... 113 Figure 3.10 Comparison between relative abundance indices derived for stock assessment in the Atlantic Ocean using Japanese longline data and relative abundance trends developed using the SF31 variation of the spatial filling method.................................................... 114 Figure 3.11 Mean cpue trend for species in the Indian Ocean grouped by the years areas were first fished. Trends for groupings in the early years are shaded in blue progressing throught green, yellow, orange, and red for later years. ................................................. 116 Figure 3.12 Mean cpue trend for species in the Pacific Ocean grouped by the years areas were first fished. Trends for groupings in the early years are shaded in blue progressing throught green, yellow, orange, and red for later years. ................................................. 117 xii  Figure 3.13 Mean cpue trend for species in the Atlantic Ocean grouped by the years areas were first fished... .................................................................................................................... 118 Figure 3.14 Examples of various forward and backward filling conditions. Black vertical lines are observed cpue averaged across quarters and grey vertical lines are imputed values.119 Figure 3.15 Relative level of depletion as a function of fishing mortality rate relative to natural mortality (M=0.4) for three diffusive movement scenarios............................................ 120 Figure 3.16 Relative abundance trends developed for various stocks in the Pacific Ocean. The grey polygon demarks the range of indices developed for each quarter from the SF31 method prior to averaging............................................................................................... 121 Figure 4.1 Input data for assessments and reconstructions for Indian Ocean stocks.................. 146 Figure 4.2 Input data for assessments and reconstructions for Pacific Ocean stocks................. 147 Figure 4.3 Input data for assessments and reconstructions for Atlantic Ocean stocks. .............. 148 Figure 4.4 Stock recruitment reconstructions and best fits for stocks in the Indian Ocean........ 149 Figure 4.5 Stock recruitment reconstructions and best fits for stock in the Pacific Ocean. ....... 150 Figure 4.6 Stock recruitment reconstructions and best fits for stock in the Atlantic Ocean....... 151 Figure 4.7 Atlantic Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M.. .......................... 152 Figure 4.8 Southern bluefin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming BevertonHolt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M............................................ 153 Figure 4.9 Atlantic Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 154 xiii  Figure 4.10 Indian Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 155 Figure 4.11 Fit and confidence bounds for SRA and F/q forced models for stock in the Indian Ocean. SRA σv low column are results from SRA where variance in recruitment anomalies was assumed low. .......................................................................................... 156 Figure 4.12 Fit and confidence bounds for SRA and F/q forced models for stocks in the Pacific Ocean. SRA σv low column are results from SRA where variance in recruitment anomalies was assumed low. SRA σv high are from SRA where variance in recruitment anomalies was assumed high. ......................................................................................... 157 Figure 4.13 Fit and confidence bounds for SRA and F/q forced models for stocks in the Atlantic Ocean. SRA σv low column are results from SRA where variance in recruitment anomalies was assumed low. SRA σv high are from SRA where variance in recruitment anomalies was assumed high. ......................................................................................... 158 Figure 4.14 Atlantic Ocean northern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. .......................................................................... 159 Figure 4.15 Joint distributions for biological reference points for recruitment reconstructions for all hypothesized current fishing mortality rates as well as Ricker and Beverton-Holt recruitment functions. ..................................................................................................... 160 Figure 4.16 Joint distributions for biological reference points for SRA estimations and E* forced simulations combined. Stock names demark the centre of the 1% quantile................... 161 Figure 4.17 Observation residuals and 95% credible intervals from stochastic SRA assessment of each stock assuming a high proportion of total variability in relative abundance trends was due to observation error. .......................................................................................... 162  xiv  Figure 4.18 Estimated abundance trends, compensation ratio (Ω), and reference points for Pacific Ocean yellowfin tuna from stochastic SRA with alternate hypothesis of natural mortality (M) and current fishing mortality(Fcur). .......................................................................... 163 Figure 4.19 Point estimates of biological reference points from stock assessments produced by RFMOs............................................................................................................................ 164 Figure 5.1 Fit to catch and relative abundance trends for stocks in the Indian Ocean. Blue vertical bars indicate observed catch (Obs. Ct)............................................................................ 181 Figure 5.2 Fit to catch and relative abundance trends for stocks in the Pacific Ocean. Blue vertical bars indicate observed catch (Obs. Ct)............................................................... 182 Figure 5.3 Fit to catch and relative abundance trends for stocks in the Atlantic Ocean. Blue vertical bars indicate observed catch (Obs. Ct). Solid circles are relative abundance trends (Obs. yt). Black lines (solid, dashed, dotted, and dash-dotted) represent predicted catch (Ct) for various initial movement rates. .......................................................................... 183 Figure 5.4 Estimated population numbers by 5°x5° area in 1950 for stocks in the Indian Ocean. Colour scale key for each stock is located above the individual plots............................ 184 Figure 5.5 Estimated population numbers by 5°x5° area in 1950 for stocks in the Pacific Ocean. Colour scale key for each stock is located above the individual plots............................ 185 Figure 5.6 Estimated population numbers by 5°x5° area in 1950 for stocks in the Atlantic Ocean. Colour scale key for each stock is located above the individual plots............................ 186 Figure 5.7 Estimated population numbers by 5°x5° area in 1950 for southern bluefin tuna and Indo-Pacific black marlin................................................................................................ 187 Figure 5.8 Ratio of estimated population in 2002 to that in 1950 by 5°x5° area, evaluated at MLE estimates for leading parameters for all stocks. .............................................................. 188 Figure 5.9 Estimated recruitment in numbers by 5°x5° cell in 1950 for Indian Ocean yellowfin tuna.................................................................................................................................. 189 Figure 5.10 Estimated net flow (movement of individuals) into or out of each 5°x5° area given population size in 1950 for Indian Ocean yellowfin tuna............................................... 190 Figure 5.11 Estimated numbers by 5°x5° area in 1950 for Indian Ocean yellowfin tuna, estimated with different assumed mixing rates v as indicated. ....................................................... 191  xv  Figure 5.12 Joint distribution of biological reference points produced by combining the randomly generated points from all movement scenarios............................................................... 192 Figure 5.13 Simulated population trends of spatial cells aggregated by the year they were first fished for stocks in the Indian Ocean. Trends estimated using the spatial model with MLE parameter estimates and an assumed overall mixing rate v=0.7..................................... 193 Figure 5.14 Simulated population trends of spatial cells aggregated by the year they were first fished for stocks in the Indian Ocean. Trends estimated using the spatial model with MLE parameter estimates and an assumed overall mixing rate v=10...................................... 194 Figure 5.15 Simulated population trends of spatial cells aggregated by the year they were first fished for stocks in the Pacific Ocean. Trends estimated using the spatial model with MLE parameter estimates and an assumed overall mixing rate v=0.7. .......................... 195 Figure 5.16 Simulated population trends of spatial cells aggregated by the year they were first fished for stocks in the Pacific Ocean. Trends estimated using the spatial model with MLE parameter estimates and an assumed overall mixing rate v=10. ........................... 196 Figure 5.17 Simulated population trends of spatial cells aggregated by the year they were first fished for stocks in the Atlantic Ocean. Trends estimated using the spatial model with MLE parameter estimates and an assumed overall mixing rate v=0.7. .......................... 197 Figure 5.18 Simulated population trends of spatial cells aggregated by the year they were first fished for stocks in the Atlantic Ocean. Trends estimated using the spatial model with MLE parameter estimates and an assumed overall mixing rate v=10. ........................... 198 Figure 6.1 Average relative fishing mortality rate distribution from 1998-2002 (top panel). Estimated relative fishing cost by cell assuming cost is proportional to distance from port and total cost is 40% of total revenues (middle panel). Estimated relative fishing cost by cell assuming cost equals 40% of cell specific average revenue from 1998-2002 (bottom panel)............................................................................................................................... 217 Figure 6.2 Distribution of mean relative fishing mortality for profit optimization scenarios, for various initial movement rates (v) assuming area cost is 40% of area revenue and no penalty is applied if stock specific target fishing mortality rates are exceeded.............. 218 Figure 6.3 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is 40% of area revenue and a xvi  moderate penalty is applied if stock specific target fishing mortality rates are exceeded. ......................................................................................................................................... 219 Figure 6.4 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is 40% of areas revenue and a severe penalty is applied if stock specific target fishing mortality rates are exceeded... 220 Figure 6.5 Fishing mortality summaries for each stock, for profit optimization scenarios for various initial movement rates and penalties for exceeding target fishing mortalities assuming area cost is 40% of area revenue..................................................................... 221 Figure 6.6 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is a function of distance to port, total cost is 40% of total revenue and no penalty is applied if stock specific target fishing mortality rates are exceeded............................................................................................ 222 Figure 6.7 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is a function of distance to port, total cost is 40% of total revenue and a moderate penalty is applied if stock specific target fishing mortality rates are exceeded. .............................................................................. 223 Figure 6.8 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is a function of distance to port, total cost is 40% of total revenue and a severe penalty is applied if stock specific target fishing mortality rates are exceeded. .............................................................................. 224 Figure 6.9 Fishing mortality summaries for each stock for profit optimization scenarios for various initial movement rates and penalties for exceeding target fishing mortalities assuming area cost is a function of distance to port, total cost is 40% of total revenue. Blue and orange are estimated fishing mortality rates.................................................... 225 Figure 6.10 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is uniform, total cost is 40% of total revenue and no penalty is applied if stock specific target fishing mortality rates are exceeded.......................................................................................................................... 226 Figure 6.11 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is uniform, total cost is 40% of total xvii  revenue and a moderate penalty is applied if stock specific target fishing mortality rates are exceeded.................................................................................................................... 227 Figure 6.12 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is uniform, total cost is 40% of total revenue and a severe penalty is applied if stock specific target fishing mortality rates are exceeded.......................................................................................................................... 228 Figure 6.13 Fishing mortality summaries for each stock for profit optimization scenarios for various initial movement rates and penalties for exceeding target fishing mortalities assuming area cost is uniform, total cost is 40% of total revenue. ................................. 229 Figure 6.14 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is uniform, total cost is 40% of total revenue, billfish q is halved and no penalty is applied if stock specific target fishing mortality rates are exceeded............................................................................................ 230 Figure 6.15 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is uniform, total cost is 40% of total revenue, billfish q is halved and a moderate penalty is applied if stock specific target fishing mortality rates are exceeded. .............................................................................. 231 Figure 6.16 Distribution of mean relative fishing mortality for profit optimization scenarios for various initial movement rates (v) assuming area cost is uniform, total cost is 40% of total revenue, billfish q is halved and a severe penalty is applied if stock specific target fishing mortality rates are exceeded............................................................................................ 232 Figure 6.17 Fishing mortality summaries for each stock for profit optimization scenarios for various initial movement rates and penalties for exceeding target fishing mortalities assuming area cost is uniform, total cost is 40% of total revenue and billfish q is halved. ......................................................................................................................................... 233 Figure 9.1Estimated mean weight (kg) of species caught by longline in each ocean. Indian and Pacific oceans weights were calculated from monthly spatial 5°x5° Japanese biomass and numbers caught. .............................................................................................................. 279 Figure 9.2 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for albacore tuna of longline vulnerable size from 1950-1979 by decade. ..................................................... 280 xviii  Figure 9.3 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for albacore tuna of longline vulnerable size from 1980-2002 by decade. ..................................................... 281 Figure 9.4 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for bigeye tuna of longline vulnerable size from 1950-1979 by decade. ..................................................... 282 Figure 9.5 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for bigeye tuna of longline vulnerable size from 1980-2002 by decade. ..................................................... 283 Figure 9.6 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for yellowfin tuna of longline vulnerable size from 1950-1979 by decade. ..................................................... 284 Figure 9.7 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for yellowfin tuna of longline vulnerable size from 1980-2002 by decade. ..................................................... 285 Figure 9.8 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for southern bluefin tuna of longline vulnerable size from 1950-1979 by decade.......................................... 286 Figure 9.9 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for southern bluefin tuna of longline vulnerable size from 1980-2002 by decade.......................................... 287 Figure 9.10 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for Atlantic bluefin tuna of longline vulnerable size from 1950-1979 by decade.......................................... 288 Figure 9.11 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for Atlantic bluefin tuna of longline vulnerable size from 1980-2002 by decade.......................................... 289 Figure 9.12 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for North Pacific bluefin tuna of longline vulnerable size from 1950-1979 by decade.............................. 290 Figure 9.13 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for North Pacific bluefin tuna of longline vulnerable size from 1980-2002 by decade.............................. 291 Figure 9.14 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for blue marlin of longline vulnerable size from 1950-1979 by decade. ..................................................... 292 Figure 9.15 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for blue marlin of longline vulnerable size from 1980-2002 by decade. ..................................................... 293 Figure 9.16 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for Indo-Pacific striped marlin and Atlantic white marlin of longline vulnerable size from 1950-1979 by decade. ............................................................................................................................ 294  xix  Figure 9.17 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for Indo-Pacific striped marlin and Atlantic white marlin of longline vulnerable size from 1980-2002 by decade. ............................................................................................................................ 295 Figure 9.18 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for Indo-Pacific black marlin of longline vulnerable size from 1950-1979 by decade............................. 296 Figure 9.19 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for Indo-Pacific black marlin of longline vulnerable size from 1980-2002 by decade............................. 297 Figure 9.20 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for swordfish of longline vulnerable size from 1950-1979 by decade. ..................................................... 298 Figure 9.21 Estimated mean spatial (5°x5°) catch (in numbers of individuals) for swordfish of longline vulnerable size from 1980-2002 by decade. ..................................................... 299 Figure 11.1 Indian Ocean albacore tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 309 Figure 11.2 Indian Ocean bigeye tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 310 Figure 11.3 Indian Ocean yellowfin tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M.. .......................... 311 Figure 11.4 Southern bluefin tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 312 xx  Figure 11.5 Indian Ocean blue marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 313 Figure 11.6 Indian Ocean striped marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 314 Figure 11.7 Indo-Pacific black marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 315 Figure 11.8 Indian Ocean swordfish leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 316 Figure 11.9 Pacific Ocean northern albacore tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 317 Figure 11.10 Pacific Ocean southern albacore tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 318 xxi  Figure 11.11 Pacific Ocean bigeye tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 319 Figure 11.12 Pacific Ocean yellowfin tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 320 Figure 11.13 Pacific Ocean bluefin tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 321 Figure 11.14 Pacific Ocean blue marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 322 Figure 11.15 Pacific Ocean striped marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 323 Figure 11.16 Pacific Ocean swordfish leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 324 xxii  Figure 11.17 Atlantic Ocean northern albacore tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 325 Figure 11.18 Atlantic Ocean southern albacore tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 326 Figure 11.19 Atlantic Ocean bigeye tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 327 Figure 11.20 Atlantic Ocean yellowfin tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 328 Figure 11.21 Atlantic Ocean bluefin tuna leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 329 Figure 11.22 Atlantic Ocean blue marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 330 xxiii  Figure 11.23 Atlantic Ocean white marlin leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ............... 331 Figure 11.24 Atlantic Ocean swordfish leading parameter joint distributions (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for Fmsy and MSY (Cmsy) estimated from recruitment reconstructions assuming Beverton-Holt (BH) and Ricker (R) recruitment relationships for three current fishing mortality estimates (Fcur or F) and a known natural mortality rate M. ........................... 332 Figure 11.25 Indian Ocean albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 338 Figure 11.26 Indian Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 339 Figure 11.27 Indian Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 340 Figure 11.28 Southern bluefin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 341 xxiv  Figure 11.29 Indian Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 342 Figure 11.30 Indian Ocean striped marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 343 Figure 11.31 Indo-Pacific black marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 344 Figure 11.32 Indian Ocean swordfish leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 345 Figure 11.33 Pacific Ocean northern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. .......................................................................... 346 Figure 11.34 Pacific Ocean southern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. .......................................................................... 347 xxv  Figure 11.35 Pacific Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 348 Figure 11.36 Pacific Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 349 Figure 11.37 Pacific Ocean bluefin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 350 Figure 11.38 Pacific Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 351 Figure 11.39 Pacific Ocean striped marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing.. ............................................................................................... 352 Figure 11.40 Pacific Ocean swordfish leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 353 xxvi  Figure 11.41 Atlantic Ocean northern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. .......................................................................... 354 Figure 11.42 Atlantic Ocean southern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. .......................................................................... 355 Figure 11.43 Atlantic Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 356 Figure 11.44 Atlantic Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. .......................................................................... 357 Figure 11.45 Atlantic Ocean bluefin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 358 Figure 11.46 Atlantic Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 359 xxvii  Figure 11.47 Atlantic Ocean white marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 360 Figure 11.48 Atlantic Ocean swordfish leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing. ................................................................................................ 361 Figure 12.1 Indian Ocean albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 366 Figure 12.2 Indian Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 367 Figure 12.3 Indian Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 368 Figure 12.4 Southern bluefin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 369 xxviii  Figure 12.5 Indian Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 370 Figure 12.6 Indian Ocean striped marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 371 Figure 12.7 Indo-Pacific black marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 372 Figure 12.8 Indian Ocean swordfish leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 373 Figure 12.9 Pacific Ocean southern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ..... 374 Figure 12.10 Pacific Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 375 xxix  Figure 12.11 Pacific Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 376 Figure 12.12 Pacific Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 377 Figure 12.13 Pacific Ocean striped marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 378 Figure 12.14 Atlantic Ocean northern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ..... 379 Figure 12.15 Atlantic Ocean southern albacore tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ..... 380 Figure 12.16 Atlantic Ocean bigeye tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 381 xxx  Figure 12.17 Atlantic Ocean yellowfin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ..... 382 Figure 12.18 Atlantic Ocean bluefin tuna leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 383 Figure 12.19 Atlantic Ocean blue marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 384 Figure 12.20 Atlantic Ocean white marlin leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 385 Figure 12.21 Atlantic Ocean swordfish leading parameter joint distribution (lower triangular), marginal posterior distributions (diagonal) and biological reference points (upper triangular) for natural mortality (M), current fishing mortality (Fcur), fishing mortality that produced MSY (Fmsy), MSY (Cmsy) and the ratio of population in 1950 to that expected in the absence of fishing from fitting the spatial model to catch data. ........................... 386 Figure 13.1 Estimated total initial value (initial population size x weight x price) for various initial movement rate scenarios summed over all stocks................................................ 387 Figure 13.2 Estimated total initial recruitment for all species combined, under various initial movement rate scenarios................................................................................................. 388 Figure 13.3 Estimated tuna and marlin species richness.. .......................................................... 389 xxxi  Acknowledgements This work would not have been possible without the assistance and support of many individuals. I owe a great debt to my mentor Dr. Carl Walters for guiding me through this endeavor and for providing endless opportunities and challenges. He has been a great inspiration throughout my education. Dr. Villy Christensen has been exceptionally generous with his input and time on both this thesis and other endeavors. I thank Dr. Ussif Rashid Sumaila for his contributions. Dr. Steven Martell was a guiding light through much of the modeling in AD Model Builder and I am truly appreciative of the time he gave and the rigor of his intellectual contribution. I am grateful to Dr. John Sibert who provided me with funding and many opportunities early in this venture that allowed me to meet many of the individuals involved in tuna assessment and acquire the necessary data. The funding he provided was through Cooperative Agreement NA17RJ1230 between the Joint Institute for Marine and Atmospheric Research (JIMAR) and the National Oceanic and Atmospheric Administration (NOAA). Funding through the Apex Predator Workgroup from the University of Wisconsin, Madison was provided by Dr. Jim Kitchell and afforded me many opportunities to meet with colleagues in Hawaii. I am truly thankful for his support. Japanese longline data from the Pacific Ocean used in this thesis were provided by the Fisheries Agency of Japan. I am grateful to Dr. Ziro Suzuki and Naozumi Miyabe from the National Research Institute of Far Seas Fisheries, Shimizu, Japan, for their assistance in acquiring data. I am also indebted to Dr. Hiroyuki Kurota for many discussions and facilitating my communication with individuals at the Far Seas Laboratory. I have had the great pleasure of meeting many of the individuals involved in the assessment of tuna and billfish and thank them for their kindness and intellectual contributions: John Hampton, Shelton Harley, David Kirby, Adam Langley, Peter Williams, Simon Hoyle, Valérie Allain, Michael Hinton, Robin Allen, Mark Maunder, Pierre Kleiber, Keith Bigelow, and Chris Boggs. Public domain data was made available through the dedication of many other individuals at the SPC-OFP, IATTC, IOTC, ICCAT, and CCSBT. xxxii  Robyn Forrest and Nathan Taylor inspired me through this process and provided much advice. I also thank Bob Lessard for discussions and support in the early days. Many other friends and colleagues at the Fisheries Centre and Zoology department have made the process enjoyable and I am grateful. This thesis has been a long time coming and I thank my friends for enduring and my parent for their endless support and having faith that the end would come. Finally, I thank my beloved, Andrea Rambeau for her unwavering love and support.  xxxiii  "Our knowledge, our attitudes, and our actions are based to a very large extent on samples." William G. Cochran  "How should scientists operate when they must try to explain the results of history, those inordinately complex events that can occur but once in detailed glory?" Stephen Jay Gould  xxxiv  Chapter 1 General introduction Context Reported landings of principle market tunas and billfishes captured by industrialized fisheries steadily increased from 0.4 million tonnes in 1950, reaching ~4 million tonnes by 2002 (Figure 1.1). Approximately 40% of the cumulative ~100 million tonnes of landings is skipjack tuna, followed by yellowfin tuna (~30%), albacore and bigeye tuna (~10% each), bluefin tuna sp. (~4%), and billfish sp. (~6%). Removals from the Pacific Ocean account for ~65% of the reported catch, followed by the Indian (~20%) and Atlantic (~15%) oceans. The majority of catches were taken by purse seine, longline, pole-and-line, and troll vessels. During the 1940s and 1950s, Japan was the first to develop large-scale industrial fisheries within the western Pacific Ocean. Japanese fleets, particularly the Japanese distant-water longline fleet, advanced into and throughout the world’s oceans by the 1970s. Spain, France, the Republic of Korea, Taiwan Province of China, and the United States followed during the mid 1960s and developed large-scale operations. Driven by technological advances, market forces, and political constraints, tuna fisheries have undergone substantial changes over time. Excluding skipjack tuna fisheries, early-industrialized fishing was dominated by longlines, targeting older and larger individuals. During the 1980s, a rapidly increasing purse seine fishery, targeting smaller individuals, emerged as the dominant gear. Though most fleets target tuna, some specialized fleets targeting billfish, particularly swordfish, emerged. Attempts to assess potential yields of tuna stocks began early in the history of industrialized tuna fishing (Schaefer 1957, Shimada and Schaefer 1956). Assessments of the major tuna and billfish stocks are currently conducted in all major oceans by Regional Fisheries Management Organizations (RFMOs). Japanese longline catch and effort data, having broad temporal, spatial, and species coverage, has been used often as the primary information source to derive relative abundance indices for assessments. Recently, spatial catch and effort data have become available in the public domain, and variations in the type of data incorporated into assessments, statistical treatment of spatial catch rate information, and assessment methodology utilized have lead to widely divergent interpretations of stock levels and the sustainability of current large-scale 1  industrialized fisheries. At one extreme, Japanese pelagic longline catch (in numbers) and effort (in hooks) data were used as the sole indicator of large pelagic predator community status (Myers and Worm 2003). In that analysis, simple exponential model fits to trends in longline ‘nominal’ catch-per-unit-effort (cpue, calculated as total catch summed over species and areas divided by total effort), suggested large pelagic predators communities were rapidly depleted (by 80%) within 15 years of industrial fishing commencing and are currently at 10% of abundances before 1950. Such an analysis was contrary to current single species stock assessments, which generally utilize a broader suite of data. For the more abundant tuna stocks, recent assessments indicated fishing mortality rates are at or near levels that produce the maximum sustainable yield (Fmsy), and stock biomasses are at or near levels that produce the maximum sustainable yield (Bmsy) (IATTC 2008, ICCAT 2009, IOTC 2008, ISC 2007, WCPFC 2008). There are some exceptions; assessments suggest bluefin stocks are likely below Bmsy (‘over-fished’) and fishing mortality rates are above Fmsy (‘over-fishing’) (CCSBT 2006b, ICCAT 2009, ISC 2008c). Furthermore, when assessments are available, similar results are found for some marlin stocks (ICCAT 2006c, Uozumi 2003). Much literature warns of biases that may be introduced into stock assessment due to errors in understanding the mechanisms that give rise to relative abundance observations or the misspecification of processes that govern dynamic behaviour (Haddon 2001, Hilborn and Mangel 1997, Hilborn and Walters 1992, Quinn II and Deriso 1999, Walters and Martell 2004). The analysis presented by Myers and Worm (2003) has been criticized for presenting a misleading picture of the status of large pelagic communities and the severity of management actions implied (Hampton et al. 2005b, Polacheck 2006, Sibert et al. 2006). There is no disagreement concerning the trends presented. Nominal cpue aggregated over the major species captured in Japanese longline operations did decline rapidly after 1950 in the Pacific Ocean, after 1952 in the Indian Ocean, and after 1956 in the Atlantic Ocean. Rapid change in apparent abundance appears incongruous in relation to relatively small catch removals in the early 1950s. As a result, various unsubstantiated explanations for the rapid change in nominal cpue have been proposed (Polacheck 2006). Although the mechanisms that gave rise to changes in nominal catch rates are not well understood, a number of critical assumptions must be met if changes in nominal cpue are to be interpreted as indicative of changes in large pelagic predator abundance. 2  A number of papers suggest that nominal cpue is unlikely to reflect changes in tuna and billfish, given historical changes in the Japanese longline fishery (Hampton et al. 2005b, Maunder et al. 2006b). Developing abundance trends using methods that explicitly account for changes in the spatial distribution of the fishery has been recommended (e.g., Walters 2003). Kleiber and Maunder (2008) demonstrated the possibility of hyperdepletion of cpue aggregated over species, due to differences in catchability and productivity among species in the overall catch. Hampton et al. (2005b) point out that the spatial scale used by Myers and Worm (2003) is inappropriate. The definitions of eco-regions used by Myers and Worm (e.g., tropical, sub-tropical) do not necessarily span the core range of some species. Finally, stock assessments generally rely on data as catch removals or composition in addition to cpue to make inference about population scale and productivity relationships. Using cpue alone, particularly to infer impacts at a community level is generally not accepted; even the simplest of assessments should account for fishery impacts (Hampton et al. 2005b). The primary motivation for this thesis was to investigate the differing intepretations of Japanese catch and effort data that gave rise to such divergent views of population status and rates of decline of large pelagic predators. Concerns highlighted above regarding methodology and interpretation makes a reexamination of the Japanese longline catch and effort data a worthwhile exercise. Heeding the litany of advice regarding development of abundance indices and stock assessment, apparent impacts of industrialized fishing on tuna and billfish stocks globally were reassessed using Japanese longline data. More specifically, status of the component of those stocks that are of a size vulnerable to longline gear was reassessed. This thesis does not attempt to reproduce assessments performed by the various RFMOs. Instead, similar methods and assessments were applied to all stocks in all oceans to: 1) highlight how and why different methodologies produced widely different relative abundance trends; 2) demonstrate when such trends in conjunction with known removals are sufficient to assess stock status or when additional information is required; 3) highlight which stock-specific relative abundance trends derived from Japanese catch and effort data may not be proportional to abundance even when more appropriate methods are applied; and 4) assess apparent status of tuna and billfish stocks globally. Finally, given concerns regarding over-exploitation of a number 3  of stocks, a global spatially explicit multi-species model was developed to explore potential equilibrium effort distributions that maximize profit subject to the constraint of not over-fishing any stock within the multi-species fishery.  Thesis structure This thesis contains five main sections, presented in the same basic order as a fisheries assessment. Chapter 2 presents background information required for most assessments. The first section in Chapter 2 presents a brief history of the Japanese longline fishery, highlighting major developments over time that may have altered species-specific catchability or fishing power. The second section of Chapter 2 introduces biological, distributional and fisheries information for each species considered in this thesis: albacore tuna (Thunnus alalunga), bigeye tuna (Thunnus obesus), yellowfin tuna (Thunnus albacares), Southern bluefin tuna (Thunnus maccoyii), Pacific bluefin tuna (Thunnus orientalis), Atlantic bluefin tuna (Thunnus thynnus), blue marlin (Makaira nigricans), striped marlin (Tetrapturus audax), Atlantic white marlin (Tetrapturus albidus), black marlin (Makaira indica), swordfish (Xiphias gladius). Within each ocean, species were broken down in to stocks. Table 1.1 presents the stock designation used along with a short code to identify each stock. Stock definitions do not necessarily conform to those used by RFMOs. Sailfish and spearfish are captured by Japanese longline but have been excluded for analyses due to major changes in data reporting requirements during the first few decades of the fishery. The final section in Chapter 2 details the development of various data sets spanning 1950-2002 for the species listed above: global 5°x5° Japanese longline catch and effort, global 5°x5° longline catch of all other nations combined, global 5°x5° catch of longline vulnerable-sized individuals in other gear and nominal catches by species. In Chapter 3, various methods for constructing relative abundance trends for each stock are presented and compared. The emphasis in this chapter is to develop so-called ‘folly and fantasy’ spatial estimators (Walters 2003) for trends in relative abundance, so as to correct for gross violation of statistical assumptions about representative sampling of spatial density patterns by the fishing fleets. The basic approach in the development of such estimators is to recognize that the estimation procedure must assign a catch per effort index value to every spatial cell 4  frequented by each stock, for every historical time period used in assessment, whether or not the spatial cell was actually fished in every period. The ‘folly and fantasy’ methods involve trying to make better assumptions about unsampled cells than the hidden, common assumption in catch rate data analysis that such unsampled cells had the same average catch rate as the cells that were fished. In Chapter 4, relative abundance trends developed in Chapter 3 in conjunction with nominal catches are used to explore stock productivity and abundance using assessment models. These models range from simple stock-recruitment reconstructions based on recruitment trends estimated from relative abundance and catch data, to age-structured models fit to the relative abundance data using a stochastic approach to stock reduction analysis. In Chapter 5, a global, spatially explicit multi-species model with movement is developed and parameterized. For this model, spatial movement rates among 5x5 degree cells are estimated from trends in spatial relative abundance and catch, where the essential idea is to estimate the movement between cells that must have occurred to explain observed catches. Finally, in Chapter 6 the global spatial model is used in a gaming and optimization framework to identify optimized spatial distributions of fishing effort given constraints on allowable fishing mortality rates on all species considered in the analysis. A key aim in this chapter is to identify a spatial mosaic of large-scale closed areas (marine protected areas) that would meet target fishing mortality constraints at minimum cost in terms of lost profits from fishing. Throughout this thesis, a number of terms and acronyms that can have multiple interpretations are used. To avoid confusion these terms are defined here and a list of acronyms is presented in Table 1.2. The term ‘stock’ is used to define a management unit. This may be part of a population or multiple populations. Within this thesis, with the exception of albacore tuna, a stock is defined as the individuals of the same species within an ocean basin. Maximum sustained yield (MSY) is the largest yield that can be taken on a continuing basis from a stock. Associated with MSY is the stock abundance in biomass (Bmsy) or numbers (Nmsy) that produces, on average, the largest surplus. The fishing mortality rate that captures this surplus is termed the 5  fishing mortality rate that produces MSY (Fmsy). In this thesis, ‘overfishing’ indicates the fishing mortality rate (F) exerted on a stock is greater than the fishing mortality rate that would reduce the stock abundance to a level that produces the greatest yield and capture the surplus production (F>Fmsy). When a stock is referred to as ‘overfished’, stock abundance is below the level that would produce MSY (B<Bmsy or N<Nmsy).  6  Table 1.1 Stock breakdown and short identifier coded associated with each species. Species  Stock (short code identifier)  Albacore tuna  Southern bluefin tuna  Indian (IALB) North Pacific (PNAB) South Pacific (PSAB) North Atlantic (ANAB) South Atlantic (ASAB) Indian (IBET) Pacific (PBET) Atlantic (ABET) Indian (IYFT) Pacific (PYFT) Atlantic (AYFT) Global (GSBT)  Pacific bluefin tuna  Pacific (PBFT)  Atlantic bluefin tuna  Atlantic (ABFT)  Blue marlin  Atlantic white marlin  Indian (IBUM) Pacific (PBUM) Atlantic (ABUM) Indian (ISTM) Pacific (PSTM) Atlantic (AWHM)  Black marlin  Global (GBLM)  Swordfish  Indian (ISWO) Pacific (PSWO) Atlantic (ASWO)  Bigeye tuna Yellowfin tuna  Striped marlin  7  Table 1.2 A description of acronyms used in this document.  Acronym cpue RFMO IOTC CCSBT WCPFC SPC-OFP IATTC ICCAT ISC MSY Fmsy Fcur Nmsy Bmsy ALB BET YFT PBFT ABFT GSBT BUM BLM STM WHM SWO EEZ HPB MLE GLM GAM GIS MPA  Description Catch-per-unit-effort Regional Fisheries Management Organization Indian Ocean Tuna Commission Commission for the Conservation of Southern Bluefin Tuna Western and Central Pacific Fisheries Commission Secretariat of the Pacific Community Inter-American Tropical Tuna Commission International Commission for the Conservation of Atlantic Tunas Interim Scientific Committee for Tuna and Tuna-like Species in the North Pacific Ocean Maximum Sustainable Yield Fishing mortality rate that on average produces MSY over the long term Current fishing mortality rate Population size in numbers that on average produces MSY Population biomass that on average produces MSY Albacore tuna Bigeye tuna Yellowfin tuna Pacific bluefin tuna Atlantic bluefin tuna Southern bluefin tuna Blue Marlin Black Marlin Striped Marlin White Marlin Swordfish Exclusive Economic Zone Hooks-per-basket – the number of hooks between floats in a longline set Maximum Likelihood Estimate Generalized Linear Model Generalized Additive Model Geographic Information System Marine Protected Area  8  Catch (millions of tonnes) 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 1950  Skipjack Tuna Yellowfin tuna Albacore tuna Bigeye tuna  1960  Bluefin sp. Marlin sp. Swordfish  1970  1980  1990  2000  Year  Figure 1.1 Global catch of principal market tuna and billfish from 1950-2002. Bluefin sp. includes Atlantic, Pacific, and southern bluefin and marlin sp. includes blue, black, white, and striped marlin.  9  Chapter 2 Background information Basic information pertaining to species biology and fisheries as well as methods used to generate the data sets required for the construction of relative abundance trends and assessments of stocks are presented within this chapter. In the first section, a brief history of the Japanese longline fishery is presented highlighting major fishery developments related to fleet distribution, fleet composition, targeting practices, and technological advances. Changes in the fishery altered fishing power, species-specific catchability, and the spatial relationship between fishing effort and species distributions. These factors must be considered when developing relative abundance indices. The second section emphasizes aspects of species biology and fisheries. Various life history trait parameters required for the assessment models used in later chapters are presented. Attention is also given to aspects of species distribution, movement, and habitat use that should be considered when interpreting fisheries data. A brief description of the fisheries is also presented to emphasize changes in fishing intensity over time and the diversity of gears used. The final section of this chapter details the development of various global catch and effort data sets spanning 1950-2002.  Development of the Japanese longline fishery It is critical to recognize potential sources of change in species-specific catchability and fishing power when interpreting fisheries dependent catch and effort information. Failing to account for shifts in these components, which relate catch rate to stock abundance, may lead to erroneous interpretations of fishery impacts, population change, or population distribution. Catch and effort data from Japanese distant-water longliners are no exception. During the expansion of industrialized fishing by Japan after World War II, the distant-water fleet underwent changes in available technology, targeted species, and was subject to economic and political constraints. A number of authors have documented operational and procedural changes in the Japanese longline fleet (Beverly et al. 2003, Matsuda and Ouchi 1984, Miyake 2004, Miyake et al. 2004, Okamoto 2004, Ward and Hindmarsh 2007). Some of the major developments are reviewed in this chapter, as they influence the interpretation of catch and effort data. In particular, it is important to  10  recognize how such developments may have affected species-specific catchability, fishing power, and the spatial relationship between fishing effort and species captured over time. Governmental development and promotion of Japanese costal and distant-water tuna fisheries is traced to Meiji Era (1868-1912) policies responding to foreign fishery activity in waters around Japan. Catches for tuna fisheries around Japan were recorded as early as 1894 (Matsuda and Ouchi 1984). By 1906, the advent of power-driven vessels aided in the formation of distantwater fisheries and by 1914 offshore tuna fisheries emerged. The Japanese territory expanded in 1915 facilitating the expansion of tuna fisheries into the western central Pacific through development of overseas fishing bases. Shoulder season fishing by bait-boat operators targeting skipjack tuna (Katsuwonus pelamis) soon proved the viability of yellowfin and bigeye fisheries. In the 1930’s, longline fisheries that were initially restricted to more temperate species (albacore, northern Pacific bluefin, swordfish, and striped marlin) expanded into the western central Pacific. These ventures were mothership operations that primarily targeted yellowfin for canning. Distant-water fisheries continued to increase production until the commencement of World War II (Figure 2.1). In 1942, Japanese vessels came under military control and suffered significant losses during wartime activities. Post-war fleets were reduced by 60%, and all overseas fishing bases were lost (Matsuda and Ouchi 1984). Following the war, Japanese fishing activities were restricted to the use of wooden vessels within 12 nautical miles of the coast (Matsuda and Ouchi 1984). By September 22, 1945, all vessels were allowed access to waters within 12 nautical miles. Successive relaxations (Figure 2.2) of restrictions on Japanese fishing activities were demarked by ‘MacArthur Lines’ (Morita 1998). By September 22, 1945, vessels <100 gmt were allowed to fish outside the 12 nautical mile zone. On October 13, 1945, vessels >100 gmt were also given permission to expand beyond 12 nautical miles. These two expansions were combined on November 3, 1945, to form the first MacArthur Line. Fishing activities were further expanded on June 22, 1946, and September 19, 1949. The establishment of the third MacArthur Line in 1949 allowed Japanese vessels to access fully their historical grounds. On May 11, 1950, ‘mothership’ operations were allowed access to fishing ground to the south and on April 25, 1952, the MacArthur Lines were abolished. Concurrent with the removal of the MacArthur lines was the development of a vessel licensing program (1949) and refinement of catch reporting to the species level (1951, Okamoto 2004). Thus, available catch and effort 11  records from 1950 and 1951 are from mothership operations only and after 1952 from an increasing number of large (>100 gmt) longliners. Species-specific catchability and fishing power between and within vessel classes is likely to have changed during these first few years though it is unknown to what extent. Changes in license restrictions, government promotion, and programs aimed at reducing economic inefficiencies altered the distribution and composition of the Japanese fleet following 1952. Up to the mid 1960s, a shifting mosaic of vessel classes, many equipped with -25°C air blast freezers, expanding into the Indian and Atlantic oceans increasing landings (Figure 2.1) and the geographic range of the fishery (Figure 2.3). Pacific expansion was eastward as well as north and south from western equatorial waters. In the Indian Ocean, the fleet extended westward and to the south. In the Atlantic, the pattern was generally north and south from equatorial waters. As a result, in all oceans a greater portion of each stock’s range was sampled (Figure 2.4). During the first few years of fishing within each ocean, effort was concentrated within a few months (Figure 2.5). As fisheries expanded, so did the seasonal distribution of fishing effort. This effect is very prominent in the Pacific where the seasonal pattern of mothership effort in 1950 and 1951 was different from the seasonal effort distribution after 1952. Seasonal expansion of fishing effort also occurred in each new 5°x5° area; initially fished for only a few months (Figure 2.6), new areas were later fished for a greater period. These patterns hint at a methodical exploration: new areas were initially probed for fishing opportunity for a few months. If fishing was successful, effort expanded into the areas and the fishing season was extended. Spatial heterogeneity in stock distribution and seasonal migratory patterns relative to the effort distribution complicates the interpretation of catch rates from aggregated catch and effort data, particularly from the early expansion period. In the late 1960s and early 1970s, the expanding Japanese fleet underwent significant changes. Further modifications were made to the 1949 New Fisheries Law, to include area specific vessel size restrictions aimed at promoting fishery development (Matsuda and Ouchi 1984). Advances in freezing technology allowed longer trips but the majority of products were still destined for canning. Line-casting devices and other labor saving measures were introduced to improve efficiency and increase profit (Miyake 2004). By the late 1960s, increasing operational costs and increasing numbers of larger longliners with freezing technology resulted in a substantial 12  reduction in mothership operations. A single mothership operation was operating in the western Pacific after 1965 and the last mothership operation occurred in 1975 (Miyake 2004). Improvements in blast freezing technology (<-40°C), prevented product from browning during thawing and allowed frozen product to be sold as higher valued sashimi. By 1969, -55°C freezers were common on longliners and a large proportion of the fleet changed fishing strategy (Miyake 2004). When fisheries were targeting tuna for canning during the 1950s and 1960s, areas of high catch rate were a primary objective, and albacore and yellowfin tuna were target species. On the sashimi market, bigeye and bluefin tuna are preferred species and fetch higher prices. This economic incentive resulted in a targeting switch to these colder water species. During this relatively rapid transition period (late 1960s to late 1970s), significant changes occurred in the fleet. Many vessels began setting deeper longlines to target bigeye tuna, which altered the relationship between the depth of gear and captured species. Although new areas were still being explored during this period and later, particularly in the North Atlantic, the Japanese fleet had reached its greatest level of occupancy in each ocean (Figure 2.7). Seasonal effort distribution over the range of most species also changed during this time (Figure 2.5) though less apparent in the Pacific where southern albacore were targeted year round. General patterns in the Indian Ocean suggest a shift from a northern winter dominated distribution to dominance by summer fishing. Patterns in the Atlantic appear more haphazard, and effort tended to be spread throughout the year. The impact of seasonal effort shifts must be considered when interpreting aggregated data. The energy crisis in the mid 1970s resulted in several cost saving measures. Trip routes were optimized to reduce fuel consumption. Ship hulls were redesigned to improve efficiency through reduced drag. Freezing technology was improved to reduce energy loss and propeller design was changed to decrease fuel consumption. In addition, the Japanese government provided support in the form of fuel subsidies, product price support, and low interest loans (Matsuda and Ouchi 1984). The Fisheries Special Reconstruction and Adjustment Act (1976) provided loans and subsidies for the withdrawal of vessels from the longline fleets, and in the Pacific, ~20% of the longline fleet was removed. These cost-saving measures also helped to stimulate expansion of the more economical purse seine fleet. In the late 1970s and early 1980s, a significant number of countries declared 200 nautical mile exclusive economic zones (EEZ). With the exception of the 13  Banda Sea agreement with Indonesia, Japan possessed no bilateral agreements with other countries requiring Japan to develop agreements usually requiring new joint ventures or access fee agreements (Matsuda and Ouchi 1984). In the Indian and Atlantic oceans this transitional period resulted in a decline in area occupied (Figure 2.4), followed by increase again after 1980. Declines in area occupied in the Pacific were more gradual but continued after 1980. The general result of these changes was a gradual contraction of Japanese effort from northern and southern extremes, except in the North Atlantic. Seasonal effort distribution during this period remained stable until the late 1990s. Communication, navigation, and information systems improved dramatically over time (Ward and Hindmarsh 2007). Radio communication and echo sounders, prevalent by the late 1950s were replaced. Satellite navigation and communication, and weather facsimiles dramatically changed navigation, positioning, and communication by the 1980s. The introduction of global positioning systems, Doppler current profilers, multi-directional sonar, remote sensing images, and information integration further enhanced the information available to longliners into the 1990s, allowing for more accurate targeting of specific oceanographic features. Amidst advances aimed at improving targeting, navigation, and communication was the modernization of gear materials such as monofilament for mainline and leaders, and stainless steel for hooks. Refinements also occurred in setting techniques aimed at reducing setting time, increasing hook fishing time and hook exposure during crepuscular periods (Hayasi 1974, Ward and Hindmarsh 2007). Distributional changes in fishing effort both spatially and temporally, coupled with technological advances and shifts in target species, resulted in an appreciably different Japanese fleet by the mid 1980s and 1990s than that of the 1950s. The overall effect of changes in the Japanese fleet for interpreting catch rates and their relationship to species abundance is significant, and comparisons between methods that do or do not attempt to account for some of these changes are presented in Chapter 3.  14  Species biology and fisheries The biology and fisheries for tuna and billfish are diverse. Variability in terms of life-history strategy, habitat use, and the nature of the fishing gears utilized must be considered when interpreting fisheries data and developing assessment models given the multi-species nature of the longline fisheries and the historical changes in fishing practices that have occurred. Growth, mortality, maturity, and allometric relationships are essential information for any stock assessment, and though it is possible to estimate some of these quantities using composition information, in many instances values must be obtained from basic scientific studies or metaanalyses. Species horizontal and vertical distributions as well as seasonal and daily shifts in distribution must also be considered as changes in the spatial distribution of Japanese longline fisheries and change in species targeting altered the spatial relationship between species and gear. Consideration must also be given to the diversity of gears capturing tuna and billfish and the size of individuals captured. Longline gear tends to capture large individuals. However, there are overlaps in the size distributions between gears. Although large individuals may comprise only a small fraction of the catch in some gears, the biomass harvested by some alternate gears was substantial and the resulting number of large individuals removed must be considered. The removal of individuals by other gears of size vulnerable to longline gear is an important consideration for the analyses in later chapters. In the open ocean, tuna and billfish are commonly categorized by habitat use according to photic zone and surface water temperature. Yellowfin and marlin are categorized as epipelagic (shallow) and tropical, bigeye is mesopelagic and tropical, swordfish are mesopelagic and temperate, and albacore and bluefin tunas are epipelagic and temperate. Furthermore, marlins are generally considered a by-catch species while tuna and swordfish are targeted. Marlin and swordfish are mainly solitary, observed sometimes in spawning or feeding aggregations, while tuna aggregate in schools according to size. Much of this categorization stems from fisherybased observations of larger individuals, mainly captured by daytime longline sets. As knowledge of tuna and marlin biology has increased, particularly through acoustic, archival, and satellite tag development as well as a greater understanding of physiological constraints, a richer picture of life history strategies and spatial ontogeny has emerged. Billfish and tuna, at least 15  those examined here, use tropical waters to spawn and remain in epipelagic waters during the night and for brief periods during the day. Juvenile tuna are found in mixed-species schools in tropical waters, so niche differentiation appears to emerge as individuals grow. The daytime foraging behaviour of yellowfin and bigeye appear distinctively different, as does the behaviour of marlins and swordfish. Adult Atlantic bluefin are observed to forage where the Gulf Stream meets cooler sub-arctic waters, then return to the Mediterranean Sea and Gulf of Mexico to spawn. These differences in habitat use, migration, and life history are worthwhile mentioning as they complicate the interpretation of fisheries data particularly, as is the case with the Japanese longline fleet, when data are collected from a fleet that altered its distribution and target species over time. Furthermore, a wide variety of fleets and gears had developed to exploit theses species throughout their ontogeny. The subsections that follow provide some details of the biology and fisheries for individual species in each ocean. Data for tuna are more readily available than that for billfishes and information for species in the Indian Ocean is sparse. Life history parameters such a mortality, growth rates, length-to-weight conversions, and estimates of age at 50% maturity used in the creation of data sets as well as stock assessments and gaming simulations are presented at the end of each biology subsection, and are also summarized for all species in Table 2.1. In a number of instances, information was not available, and reasonable guesses had to be made as noted in the table. Brief descriptions of fisheries for each species are provided to highlight the variety of gears used and the magnitude of removals by each gear. A more detailed breakdown of catch by county and gear can be found in Miyake (2004) or on the FAO website. Albacore tuna (Thunnus alalunga) Biology Distributed in temperate to tropical waters of the Pacific, Atlantic, and Indian oceans as well as the Mediterranean Sea, albacore tuna are generally found in epi- to mesopelagic waters where surface temperature ranges between >10-25° C. Though commonly taken with surface gears particularly in more temperate waters, individuals are observed to depths of ~600 m (Collette and Nauen 1983) with an apparent preference for cooler temperatures (~15° C). Lower lethal 16  oxygen levels estimated for albacore vary depending on body size and model assumption (1.67 mg l-1 at 50 cm, 1.39 mg l-1 at 75 cm (Sharp 1978), 5 mg l-1 (Graham et al. 1989)). As with most Scombroids, albacore show a preference for surface waters during the night. Unlike yellowfin, which make repeated forays to depth during the day, albacore appear to spend a substantial proportion of the day at intermediate depths (Bard 2001, Chen et al. 2005, Dagorn et al. 2000, Saito 1973). In the Pacific and Atlantic oceans, warm equatorial waters are assumed to be a barrier to migration which has resulted in distinct northern and southern populations (Nakamura 1969). In the Indian Ocean, no such north-south differentiation exists. Generalized migratory patterns have been proposed for most populations. Mature adults migrate from temperate to subtropical waters to spawn. Young of the year rear for a time in tropical waters then migrate toward temperate waters. Sub-adults and adults outside the spawning season tend to migrate east west, concentrating along convergence zones (Jones 1991, Nakamura 1969, Wang 1988). In the Indian Ocean, albacore are distributed between 25° N-45° S with a main spawning area off eastern Madagascar (Koto 1969). Life history stages appear segregated by major current systems (Chen et al. 2005). Mature individuals are mainly found north of 10° S, where the monsoondriven current prevails, in temperatures >15° C. Spawning occurs in the austral summer between 10-30°S within the subtropical gyre zone at depth, where surface temperatures are >24° C and the mixed layer is deep (Ueyanagi 1969). Immature albacore are predominantly found >30° S within circumpolar currents in the austral autumn, migrating northward to 25° S in the austral winter, then to between 15-25° S in the austral spring and returning south of 30° S in the austral summer. Northern Pacific albacore appear to range between 0-50° N with immature individuals making large trans-Pacific migrations (Nakamura 1969). Kimura (1997) expanded upon previous migration models proposing an anti-clockwise migration for both immature and mature albacore. Immature albacore concentrate between 25-35° N between October to March, dispersing widely and up to 45° N in the North Pacific during the summer. Winter movement depends upon the Kuroshio Current meander. In years with little extension of the Kuroshio into eastern Pacific waters, immature individuals appear to migrate in an anti-clockwise direction during the winter from 120-180° E. This pattern extends eastward when the Kuroshio’s influence extends into the 17  eastern Pacific. Mature individuals appear to follow an anti-clockwise migration centered on 20° N and 170° E. During the summer, mature individuals are found spawning south of 20° N between 150° E-160° W moving north and east in the fall and westward between 25-35° N in the winter. Distribution and migration patterns of southern Pacific albacore are not as well understood (Nakamura 1969). Distributed between ~0-<50° S, southern Pacific albacore appear to spawn within the western Pacific during the southern summer with juveniles migrating south to the Subtropical Convergence Zone (Roberts 1980) before returning to spawn in more tropical waters when mature. Catch rates of mature individuals are high in the southern summer between 15-25° S and between 30-40° S in the southern winter (Wang 1988). Both immature and mature southern albacore are thought to migrate west to east within the subtropical convergence (Jones 1991). In the Atlantic Ocean, albacore are distributed between 45° S-50° N (Collette and Nauen 1983). North Atlantic albacore are assumed to be distributed between 5-<50° N (Bard 1981) with spawning grounds located in waters offshore of Venezuela, the Sargassum Sea, and the Gulf of Mexico. Spawning occurs from April-November, peaking in the summer (Nishikawa et al. 1985). Mature individuals are thought to migrate eastward in the winter months (Grant 1969). Juvenile individuals appear to migrate toward the eastern Atlantic. Immature albacore undergo a seasonal migration associated with warming surface waters (Bard 2001). In the spring, young albacore migrate northward and eastward from the Azores to the Bay of Biscay and Celtic Sea, potentially moving westward to concentrate on the front of the Gulf Stream. During the summer, individuals are found in two feeding areas, on the continental slope of Bay of Biscay and the Celtic Shelf as well as Gulf Stream fronts and south of the Grand Banks. Distribution and movement of albacore in the Mediterranean is poorly understood. Mature individuals are distributed discontinuously with concentrations in the Tyrrhenian, Ionian, Adriatic, and Aegean seas (Megalofonou 2000). Tagging studies suggest exchange between areas within the Mediterranean and limited exchange with the north Atlantic (Arrizabalaga et al. 2002) 18  South Atlantic albacore are assumed to be distributed between 50° S-5° N. Spawning is observed to occur off the eastern Brazilian coast during the austral summer (Koto 1969). Mature individuals appear to migrate eastward in the winter concentrating off Angola and Southwest Africa (Grant 1969). Coimbra (1995) proposed that all life stages of southern albacore follow the south Atlantic subtropical gyre. The Brazil current transports larvae southward to the intertropical convergence zone. Immature individuals move eastward and spread northward up the west coast of Africa toward the Gulf of Guinea in the Benguela current. Mature individuals migrate westward in the south equatorial current and Brazil current to the northeast Brazilian coast. Maximum life span is assumed 13 years, though Mediterranean albacore are estimated to reach a maximum age of 9 years (Megalofonou 2000). Age at 50% maturity is estimated to be around 80-90 cm (~ 5 years) in the Pacific, Atlantic and Indian oceans, and >62 cm in the Mediterranean Sea. Length-at-age varies with location and sex, with recent estimates for the von Bertalanffy curvature coefficients between 0.14-0.54 year-1 and mean asymptotic lengths from 78-142 cm. Natural mortality rate is commonly assumed to be between 0.2-0.4 year-1 for older ages, though estimates range as high as 0.56 year-1 (Chen and Watanabe 1988, Labelle and Hampton 2003). Sex ratio appears to diverge from 1:1 after the age of maturity, with males having apparent lower mortality, slower growth, and larger body size (Collette and Nauen 1983). Fisheries Albacore are captured using both surface (troll, pole-and-line/baitboat, purse seine, and drift gillnet) and subsurface (longline) gears with relative contribution of each gear to total catch depending on ocean and stock. Surface gears tend to target immature albacore, 40-80 cm, and longlines target fish >65 cm. Albacore fisheries in the Indian Ocean are dominated by longline. Japanese vessels began largescale longline operations in the Indian Ocean in 1952, primarily targeting yellowfin and albacore. Albacore catch increased until the early 1960s from less than 5000 tonnes to ~15,000 tonnes (Miyake et al. 2004). Through the late 1960s, catches declined rapidly as vessels outfitted 19  with super-cold freezers began targeting southern bluefin and bigeye. Coinciding with the decline in Japanese catch was a rapid increase in Taiwan’s catch. A number of other nations contributed to total longline removal, though Taiwan's catch contribution accounts for the majority of the ~40,000 tonnes caught in the late 1990s (Miyake et al. 2004). The only other fishery removals of note are a small amount of purse-seine by-catch reported by the European Community and a sizeable Taiwanese drift gillnet fishery that operated during the mid 1980s until 1992 with catches peaking around 20,000 tonnes. Unlike Indian Ocean fishery development, north Pacific albacore has had a long history of targeted exploitation. Japanese total catch of tuna increased steadily after the beginning of the 20th century to around 85,000 tonnes by 1940 (Matsuda and Ouchi 1984), and it is likely that a sizable proportion of this catch was comprised of albacore. By 1950, the Japanese albacore catch was ~75,000 tonnes (Miyake et al. 2004). The development of the U.S. commercial albacore fishery off the coast of California began decades prior to 1950, and was comprised of 3,000 vessels by 1950 (Laurs and Dotson 1992) catching ~20,000 tonnes. Miyake et al. (2004) note that prior to the late 1980s the largest component of north Pacific albacore catch was taken by the Japanese baitboat fleet. Peak catches of >100,000 tonnes occurred in the 1970s. In the mid 1970s to 1991 Japanese, Taiwanese and Chinese drift gillnet fishery catch increased, surpassing all other gears by the late 1980s. Longline fishery catch surpassed baitboat catch in the early 1990s though total catch had declined substantially. By 1999, catches had increased again to levels seen in the 1970s. Fisheries for albacore in the South Pacific Ocean are predominantly longline operations. Japanese longline catch increased from low levels in 1952 to ~30,000 tonnes by the early 1960s. As in the Indian Ocean, the Japanese catch declined as vessels shifted to targeting bigeye, and were replaced by Taiwanese and Korean longliners. Drift gillnet fisheries began in the late 1980s but ended in the high seas by 1992. Troll fisheries developed in the late 1970s between New Zealand and 140° W along the southern convergence zone. Total removals from the south Pacific stock has not exceeded ~50,000 tonnes (Miyake et al. 2004).  20  North Atlantic albacore were caught prior to 1950 mainly by French and Spanish troll fleets in the Bay of Biscay, producing a stable catch of 30-40,000 tonnes (Miyake et al. 2004). Troll fishery catch declined in the late 1950s as the Spanish fleet converted to baitboats and in the 1980s as French trollers converted to trawls and drift gillnets. After 1956, Japanese longliners began targeting albacore and yellowfin but, as in the Indian and Pacific oceans, switched to bigeye and bluefin in the late 1960s with fleets from Korea and Taiwan taking their place. Total longline catch from the 1960s through the 1980s varied between 10-20,000 tonnes declining in the 1990s with Taiwan remaining the dominant longline fleet. Total catch of northern albacore from the mid 1960s to the late 1980s varied between 50-60,000 tonnes. Miyake et al. (2004) remark that the southern Atlantic albacore fishery is predominantly a longline fishery with Japanese catches increasing in the late 1950s followed by Korea and Taiwan. By the early 1970s, longline catch peaked around 30,000 tonnes, varying between 10,000-30,000 tonnes since. Taiwanese catch accounts for a high proportion of the biomass captured in recent years. In the early 1980s, baitboat fisheries developed off Namibia and South Africa with catches quickly reaching 10,000 tonnes. Albacore fisheries in the Mediterranean Sea are also predominantly longline fisheries with a reported catch of ~5,000 tonnes. Bigeye tuna (Thunnus obesus) Biology Bigeye tuna occur in tropical and subtropical oceanic waters of the Pacific, Atlantic, and Indian oceans. Bigeye, an epi- and mesopelagic species, are found where surface waters range between >10-29° C (Collette and Nauen 1983) with an apparent adult temperature preference of 11-15° C (Brill 1994). Spawning is thought to occur year round in surface waters >24° C, with a seasonal peak during the summer months at depths >50 m (Miyabe 1994). Bigeye are found to have the lowest dissolved oxygen tolerance limit of the major tuna species though estimates vary depending on size and method (0.52 mg l-1 at 50 cm, 0.65 mg l-1 at 75 cm (Sharp 1978), 1.3 mg l1  (Hanamoto 1987)). Low temperature and oxygen tolerance afford bigeye tuna a distinct depth  distribution and vertical movement pattern. Remaining in surface waters during the night, individuals mirror the vertical movement of the deep scattering layer (Brill et al. 2005) reaching depth >500 m during the day but making short forays into warm shallow water possibly to 21  recover thermal or oxygen debts (Holland et al. 1990b, Matsumoto et al. 2005, Schaefer and Fuller 2002). Bigeye tunas appear to spend daytime hours below the thermocline. Arrizabalaga (2005) noted an exception to this behaviour in Azorean waters where bigeye made repeated dives to depth but remain in surface waters. A crude migratory pattern is apparent from purse seine and longline catch rate for bigeye tuna (Fonteneau et al. 2005). As spawning season approaches, adults move toward warm equatorial waters where the sea surface temperatures is >24° C. Juveniles rear within warm equatorial surface waters in mixed schools with yellowfin and skipjack until sufficient thermoregulatory ability is achieved allowing them to move to deeper colder waters. Adults tend to migrate from spawning areas to feeding zones in more temperate waters. In the Indian Ocean bigeye are distributed north of 40° S. Lee et al. (2005) note that movement patterns of bigeye within the Indian Ocean appear tied to the 6 month cycle of north-east and south-west prevailing monsoons as the inter-tropical convergence zone shifts north to south across the equator. The migration pattern is poorly defined though adults appear to concentrate between 15° S-10° N during the northeast monsoon season (austral summer) for spawning. Juveniles appear in the purse seine fishery in the western ocean in mixed species schools. Adults appear to move into temperate waters outside the spawning season. Information that is more detailed is available on the distribution of bigeye in the Pacific Ocean (Miyabe 1994). Pacific bigeye tunas are distributed between 40° S-40° N in the west and 30° S-40° N in the east, except near coastal waters off Central America. Gonadosomatic index measurements suggest year round spawning with peak spawning in June and July west of 140° E within the equatorial counter current. Size of individuals and proportion spawning tends to increase from west to east. However, bigeye tuna associate with the thermocline and such a pattern is explainable if catchability of larger individuals increases as the thermocline shallows west to east. Peak spawning is observed between 140-180° E from April-July shifting to between February-July at 140-100° W. Mature bigeye are often caught from April-September between 10° S-10° N in the west (Kikawa 1962) and from January-September at 0-10° N and January-June at 0-10° S in the east (Kume and Joseph 1966). Miyabe (1994) notes that along 30° N and off Chile, 22  immature individuals are caught in winter fisheries. Japanese baitboat fisheries in the northwestern Pacific targeting immature individuals move northeast from spring to summer and return south to southwest in the fall. This north-south movement is also apparent in the eastern Pacific as well as the Southern Ocean. Such observations support the notion of spawning (tropical) to feeding (temperate) ground migration with immature individuals tending to reside temperate areas moving north and south in response to seasonal changes in sea surface temperature. The main spawning area for Atlantic bigeye is between 10° S-10° N within the Gulf of Guinea (Alvarado Bremer et al. 1999). Fonteneau (2005) suggests mature bigeye return to this area to spawn during northern winter and summer months depending on feeding ground location (north or south). Juveniles appear to rear within the Gulf of Guinea as indicated by capture in the purse seine fisheries. Based on tagging studies and fisheries information, Hallier (2005) suggest immature individuals north of the Equator move northward of along the African coast during spring and summer returning south during winter months. Tagging studies suggest some individuals move northwest to the Azores and westward. Individuals tagged further south (Cape Lopez) have displayed southward movement along the coast toward Angola and westward movement in the south equatorial current. Proposed feeding grounds for mature and larger immature bigeye in the northern hemisphere are located between 10-20° N in the western Atlantic as well as spanning the Atlantic between 20-40° N in the east and 35-45° N in the west (Fonteneau et al. 2005). Southern feeding areas are hypothesized between 30-40° S in the west and 10-35° S in the east. Longevity estimates based on natural mortality rates suggest a maximum age of >12 years (Hampton and Williams 2005) to 16 years (Farley et al. 2006). Maturity is observed to occur at sizes >100 cm or >2 years though estimates depend on geographic location (Calkins 1980, Nikaido et al. 1991, Schaefer et al. 2005) with age at 50% maturity assumed to be around 3.5 years. Estimates of growth parameters are also highly variable depending on method and location. Estimates for the von Bertalanffy curvature coefficient range from 0.14-0.45 year-1 and estimates of asymptotic mean length from 166-250 cm. Natural mortality is estimated to decrease with size until maturity (~130 cm). As with albacore and yellowfin tuna, increasing male sex 23  ratios after maturity suggest an increase in female mortality or differential growth associated with spawning. As a result, estimates of mortality rate in some assessment models increase around the age of maturity and decline as the proportion of females decreases (Fonteneau et al. 2005, Hampton et al. 2005a, Maunder and Hoyle 2006a). Estimates of natural mortality for mature bigeye vary between 0.25-0.7 year-1 depending on size, though a reasonable estimate is around 0.4 year-1 (Fonteneau et al. 2005, Hampton et al. 1998). Fisheries Bigeye tunas are targeted in all three oceans using longline, baitboat, and purse seines though longline is the dominant gear type. Longlines target individuals >100 cm while baitboat and purse seine fisheries capture immature individuals <100 cm. In the Indian, Pacific and Atlantic oceans, fisheries for bigeye have undergone three major changes. In the late 1970s, longliners equipped with super-cold freezers (<-40° C) began targeting bigeye and bluefin for the sashimi market. This shift in target species resulted in gear configurations aimed at targeting bigeye at greater depth (Suzuki et al. 1977). Purse seine development in the 1980s resulted in increased catches of immature bigeye associated with natural floating object. The use of man-made fish aggregating devices (FADs) increased in mid to late 1990s resulted in a substantial increase in the catch of small juvenile bigeye in mixed schools with yellowfin and skipjack (Miyake et al. 2004). In the Indian Ocean, bigeye tuna were first captured by Japanese longline fleets targeting yellowfin and albacore for canning in 1952. Catch increased as fleets from Korea and Taiwan joined the fishery in the 1960s. Catches peaked at 40,000 tonnes in the late 1960s. When vessels switched to targeting bigeye for the sashimi market in the late 1970s, catches stabilized between 40,000-60,000 tonnes. With other nations entering the fishery and the use of ‘flags of convenience’, catches increased to over 100,000 tonnes by the late 1990s (Miyake et al. 2004). French, Spanish, and to a lesser extent Japanese purse seine fisheries for yellowfin and skipjack developed in the early 1980s resulting in an increase in juvenile bigeye catch because of mixed schooling. By the early 1990s, catch of juvenile bigeye increased to 20,000 tonnes. The development of FAD fishing in the early 1990s resulted in further increases, again of primarily 24  small bigeye, to approximately 40,000 tonnes by the late 1990s (Miyake et al. 2004). Baitboat and gillnet catches are small in comparison, accounting for <2,000 tonnes. As in the Indian Ocean, catches of bigeye in the Pacific are predominantly from longline operations though catch by purse seine and baitboat operations does occur. Bigeye in the western Pacific likely caught prior to 1950 in baitboat and longline operations off Japan (Matsuda and Ouchi 1984). With the expansion of the Japanese longline fishery, catch of bigeye increased rapidly to 60,000 tonnes by the mid 1960s. Targeting of bigeye in the late 1970s resulted in a further increase in the Japanese catch, peaking around 100,000 tonnes in the mid 1980s to mid 1990s then declining to 60,000 tonnes in the late 1990s. Taiwan, Korea and China also developed fleets in the mid 1960s and total longline catch peaked at ~160,000 tonnes during the mid 1980s to late 1990s declining to <100,000 tonnes by the late 1990s. Miyake et al. (2004) note that expansion of purse seining fleets started in the 1970s. Japanese, U. S., Korean, and Taiwanese fleets catch increased in the western Pacific to ~20,000 by the 1990s primarily targeting free schools or schools associated with natural floating objects. Eastern Pacific catches peaked in the mid 1970s at ~16,000 tonnes primarily using dolphin associated sets. In 1993, within the eastern Pacific, and 1998 in the western Pacific bigeye catch increased substantially with the development of FAD fishing. In the western Pacific, catch of bigeye also occurs in the Japanese baitboat fishery as well as artisanal fisheries in Indonesia and the Philippines. Prior to the expansion of the Japanese longline operations in the late 1950s, Atlantic bigeye, primarily large fish, were captured using baitboats within the Azores, Madeira and Canary islands (Miyake et al. 2004). Japanese catch increased as in other oceans as longliners switched to targeting bigeye in the late 1970s. Korean and Taiwanese vessels commenced longline operations in the 1960s. Total catch of bigeye varied between 40,000-60,000 tonnes through the 1970s and 1980s (Miyake et al. 2004). Attempts to account for IUU fishing in the 1990s resulted in a further increase in reported catch from longlines to ~70,000 tonnes. Purse seine fisheries developed in the Gulf of Guinea in the late 1960s with catches varying between 10,000-15,000 tonnes in the mid 1970s. Catches declined through the 1980s but increased during the 1990s with the introduction of FAD fisheries, to ~20,000 tonnes. In 1962, Japanese baitboats established a small fishery off Ghana (Miyake et al. 2004). Korea, Panama, and Ghana joined the fishery but currently only Ghana produces notable catches. 25  Yellowfin tuna (Thunnus albacares) Biology Yellowfin tuna occur in tropical and subtropical oceanic waters within the epipelagic zone. Distribution is generally restricted to where surface water range between 18-31° C (Collette and Nauen 1983) though yellowfin are most commonly caught in temperatures >20° C (Fonteneau 2005). Spawning is observed to occur year round in surface waters >24° C (Kikawa 1966) though seasonal peaks vary with location. Field observations and laboratory investigation show that yellowfin require higher dissolved oxygen concentration than bigeye, though potentially lower than albacore (1.49 mg l-1 at 50 cm, 2.32 mg l-1 at 75 cm (Sharp 1978), <2.5 mg l-1 (Dizon 1977), 2.1 mg l-1 (Bushnell and Brill 1992)). Archival tagging studies indicate that yellowfin are predominantly within surface waters during the night, and spend 90% of their daytime activity above 100 m making forays to depths near the thermocline provided temperature is no less than 8° C colder than surface temperature (Block et al. 1997, Brill et al. 1999, Holland et al. 1990b). However, Dagorn (2006) and Schaefer (2007) observed deep diving behaviour in yellowfin tuna, observing forays up to >1000 m in depth. During these dives water temperatures encountered were more that 8° C cooler that surface temperature though dives were terminated as body temperature approached 15° C. Korsmeyer (1996) and Brill (1998) point out that at 15° C cardiac function in yellowfin is impaired likely limiting vertical distribution. Indian Ocean yellowfin are mainly distributed north of 30° S. Spawning occurs year round concentrated between 10° S-10° N (Mimura 1963). Areas of high larval density can be found within this geographic band in the western ocean south to Madagascar and in the eastern ocean within the Malay Archipelago during the Austral summer (Conand and Richards 1982, Mimura 1963). North-South migration of adult individuals is likely tied to seasonal changes in water temperature and currents, with higher densities in the north during the austral winter. Juveniles also appear to migrate north to south in response to seasonal changes, though their distribution is closer to coastlines. Mimura (1963) notes an apparent geographic distribution in adult sex ratio with a higher proportion of males north of 10° S, equal ratio between 10-15° S and a higher female ratio south of 15° S. 26  In the Pacific Ocean yellowfin are distributed in tropical and sub-tropical waters between 35° S40° N in the western ocean and 33° S-35° N in the east (Cole 1980). Spawning occurs in the tropical Pacific year round and at higher latitudes when sea surface temperature is greater than 24° C (Suzuki 1994). Peak spawning in the western and central tropical Pacific is observed December-January between 120°-180° E, and April-May between 180° E-140° W, May-June in the Kuroshiro Current, and November-December in the East Australia Current (Cole 1980, Kikawa 1966). Suzuki (1994) notes that within Philippine waters, peak spawning occurs MarchMay and to a lesser extent November-December. Wild (1994), in a summary of eastern yellowfin spawning, indicates year round spawning off the coasts of Mexico, Central America as well as between 0-10° N and 130-190° W. Between 0-10° S and 130-190° W, spawning peaks from January-June and is suppressed in the second half of the year due to the intrusion of water colder than 26° C. As in the Indian and Atlantic oceans, north-south movements of both adults and juveniles are observed. Seasonal changes in thermal and current structure limit latitudinal extent of distributions with the highest latitudes reached in the summer (Suzuki 1994). Nakamura (1969) notes an apparent west to east cline in yellowfin tuna size distribution within the western and central Pacific. It is uncertain if such a cline is the result of size related ontogeny or differential exploitation rates. Several authors suggest western, central, and eastern individuals comprise separate stocks (Schaefer 1955, Suzuki et al. 1978) with potential further differentiation between eastern individuals north and south (Diaz-Jaimes and Uribe-Alcocer 2006). Yellowfin appear to show limited dispersal (Bayliff 1979, Fink and Bayliff 1970, Schaefer et al. 2007, Sibert and Hampton 2003) while recent tagging information indicates limited movement across the Pacific though large scale migratory behaviour is not apparent (Hampton et al. 2006b). Distribution in the Atlantic extends from 50° S-50° N, though predominantly in tropical and subtropical waters. Spawning occurs along the West African coast in the Gulf of Guinea (JanuaryMarch). Juveniles move north and south along the African coast within coastal currents and in response to seasonal temperature changes, ranging north to the Canary Islands and south off Angola. In the western Atlantic, spawning areas are found in the southeast Caribbean Sea (July27  September) and in the Gulf of Mexico (May-August) (Arocha et al. 2001, ICCAT 1991). In the western Atlantic, juveniles found off Brazil are likely spawned in the southeast Caribbean Sea. As in the eastern Atlantic, north-south movement is also observed. Adult yellowfin make transatlantic migrations within equatorial currents (ICCAT 1991) as well as north-south migrations along both the western and eastern Atlantic coasts. Yellowfin tuna likely live 7-10 years. Size and age at 50% maturity is estimated to occur at sizes of 85-100 cm or 2-3 years age, though estimates depend on geographic location (Fonteneau 2005, Maunder and Hoyle 2006b). Growth parameters are also highly variable depending on method and location. Estimates for the von Bertalanffy curvature coefficient range from 0.260.66 year-1 and estimates for asymptotic mean length from 166-230 cm. Departures from the standard von Bertalanffy growth model have been noted by a number of authors with growth apparently slowing around 40-70 cm (Lehodey and Leroy 1999, Wild 1986). Natural mortality is generally thought to decrease with size until maturity (~100 cm). As with bigeye and albacore, there is an almost linear increase in male sex ratio after maturity and potential differences in sex specific growth result in estimates of age specific natural mortality increasing with the onset of maturity. Thus, estimates of adult mortality vary considerably 0.55-1.2 year-1. Hampton (2006b) estimate a minimum mortality of 0.6-0.8 year-1 for sub-adults (<1.25 years). Similar estimates are made by Hoyle (2007) with sub-adult (<1.5 years) instantaneous mortality estimated at 0.8 year-1 increasing to 1.2 year-1 by age 3 and slowly declining back toward 0.8 year-1. Fisheries Yellowfin tuna are fished in the Indian, Pacific, and Atlantic oceans using primarily longline, baitboat, and purse seine gears though catch in the Pacific and Indian oceans is also taken using gillnet and artisanal gears. In general, longlines targets individuals >90 cm, while baitboat and purse seine fisheries capture immature individuals <90 cm. Free and dolphin-associated purse seine sets in the eastern Pacific Ocean capture larger individuals. Since the 1980s, purse seining has been the dominant method of capture, with the exception of the eastern Pacific where purse seining dominated since the 1960s (Miyake et al. 2004). Japanese longline fleets began targeting yellowfin and bigeye in the Indian Ocean in 1952. Taiwan and Korea followed by the 1960s. Catch rose to 20,000-60,000 tonnes from the late 1950s to the mid 1980s though Japanese catch 28  had declined to 10,000 in the mid 1970s due to vessels targeting bigeye and southern bluefin. Increased catch by Taiwan and Indonesia resulted in peaks of 160,000 tonnes in the early 1990s, fluctuating between 80,000-100,000 tonnes by the late 1990s. French and Spanish purse seine fisheries began in the early 1980s with catches rising quickly and fluctuating from 80,000100,000 tonnes through the 1990s with total purse seine catch fluctuating from 100,000-160,000 tonnes in the late 1990s. Purse seine catch in the Indian Ocean is primarily on FADs (Miyake et al. 2004). Baitboats have been operating out of the Maldives sine the early 1950s with catches around 10,000 tonnes in the 1990s. Sri Lankan and Iranian gillnet fisheries developed in the 1980s with catches >60,000 tonnes in the late 1990s. Differences in fishery development in the western and eastern Pacific are striking. Commercial ventures for yellowfin were in place well before 1950 in both the western and eastern Pacific (Laurs and Dotson 1992, Matsuda and Ouchi 1984). Japanese baitboat and longline vessels operated in the western Pacific and U.S. baitboats operated in the eastern Pacific. With the expansion of the Japanese longline fleet in 1952, yellowfin catch in the western and central Pacific increased quickly. Taiwan and Korea commenced longline operations in the late 1950s and early 1960s. By the early 1980s longline catch had peaked at >100,000 tonnes but with fleets switching to target bigeye and the expansion of purse seining, catch declined and fluctuated between 50,000-100,000 tonnes (Miyake et al. 2004). Japanese purse seine fleets targeting free schools were in operation in the western Pacific before 1950. Until the development of fleets by the U.S., Korea, and Taiwan, in the 1980s, catches were small. Expansion of purse seine fishing, mainly on free schools, resulted in catch increasing to >200,000 tonnes in the 1990s. FAD fishing which commenced in 1998 has been used extensively since (Miyake et al. 2004). Eastern Pacific longline fisheries contribute little to total catches. The development of the U.S. purse seining fleet in the late 1950s resulted in catch increasing to >200,000 tonnes by the mid 1970s primarily using dolphin associated sets. Catch restrictions and El Niño conditions resulted in catch declines as the U.S. fleet shifted to the western Pacific. Catches increased in the mid 1980s with the development of Latin American fleets and FAD fishing in the early 1990s. Catch rose to between 250,000-300,000 tonnes (Miyake et al. 2004). Artisanal fisheries of Indonesia and the Philippines in the western Pacific are poorly documented but catches in recent years are estimated to have reached 140,000 tonnes (Miyake et al. 2004). 29  Japanese longlines began targeting yellowfin in the Atlantic Ocean after 1956. Catch increased rapidly to 50,000 tonnes by the early 1960s and declined through the 1970s. Taiwanese and Korean vessels entered the fishery in the 1960s and catches have been stable ranging from 20,000-30,000 tonnes since the 1980s (Miyake et al. 2004). Baitboat fisheries by France and Spain developed in the late 1950s in the eastern tropical Atlantic with catches increasing to 20,000 tonnes. In the 1960s, Japan developed a baitboat fishery based in Ghana followed by Korea, Panama, and Ghana. With French and Spanish fleets switching to purse seining, catch by baitboat has been stable at around 20,000 tonnes (Miyake et al. 2004). In the mid 1960s, French and Spanish purse seining fleets developed in the Gulf of Guinea with catches peaking at 130,000 tonnes in the early 1970s. Catches declined through the 1970s and vessels moved to the Indian Ocean. Catches increased through the 1980s and peaked with the introduction of FADs at around 130,000 tonnes in the early 1990s (Miyake et al. 2004). Since 1995, catches have declined. Southern bluefin tuna (Thunnus maccoyii) Biology Southern bluefin tuna are a single stock with a near circumpolar distribution in the southern ocean (Proctor et al. 1995). Distribution is delimited by a broad temperature spectrum in epipelagic and mesopelagic waters (5-30° C) though outside of the spawning season, 5-20° C appears to be the preferred temperature range (Olson 1980). Cardiac performance of bluefin tunas allows them to exploit waters below 2° C (Blank et al. 2004). Southern bluefin remain in surface waters at night and, like bigeye, appear to follow the scattering layer during the day while returning to surface waters to warm (Gunn and Block 2001). Observation from tracking studies on southern bluefin (Davis and Stanley 2002, Gunn and Block 2001) and other bluefin (Block et al. 2001, Block et al. 2005, Brill et al. 2005, Lutcavage et al. 2000, Teo et al. 2007) suggest temperature tolerance is related to body size, allowing larger individuals to exploit deeper depths.  30  Southern bluefin are distributed between 10-50° S though individuals north of 30° S are either juveniles or spawning adults. The main distribution is thought to occur between 30-50° S (Nakamura 1969). Southern bluefin spawn in water temperatures between 25-30° C in a restricted area in the eastern Indian Ocean south of Java, with the major spawning area between 10-20° S and 110-125° E and a minor areas 20-30° S and 110-125° E (Nakamura 1969). Spawning occurs between September-March (Mimura and Warashina 1962, Serventy 1956) with peaks in abundance during October and February (Farley and Davis 1998). Distribution and migration of southern bluefin is reasonably well established (Farley et al. 2006, Olson 1980). Young of the year appear to take one year to migrate southward along the western Australian coast, within the Leeuwin Current, to the southwestern Australian coast. Juveniles move eastward into the Great Australian Bight (GAB) off southern Australia. Juveniles age 2-4 years appear to winter in the GAB making eastward migrations south of Tasmania and up the southeastern Australian coast, or westward migrations within the southern Indian Ocean gyre to the African coast. After the age of four, juveniles move into the West Wind Drift and migrate east and west centered on 35° S. Upon reaching maturity, adults undertake seasonal migrations back to the spawning ground in the eastern Indian Ocean. Distribution of southern bluefin is not continuous within the West Wind Drift. Areas of high abundance are found off Southern Africa, the southeast Indian Ocean, Tasmania, and New Zealand. Southern bluefin tuna are estimated to live >20 years (Caton 1994), with the oldest age observed being estimated at 41 years (Farley et al. 2007). Age at 50% maturity is estimated to occur at sizes 150-160 cm or 11-12 years (Davis et al. 2001). Growth parameters estimates for southern bluefin are reasonably well defined though asymptotic length is variable depending on the size range of individuals sampled. Estimates for the von Bertalanffy curvature coefficient range from 0.1-0.146 year-1 and estimates for mean maximum asymptotic length between 180-260 cm. Polacheck et al. (2004) suggest, given an apparent reduction in growth rate during the transition between juvenile to sub-adult, that a two stage growth model is more appropriate for southern bluefin tuna. Increases in growth rate and decreased asymptotic size since 1960 have also been observed (Polacheck et al. 2004). Farley (2006) notes an increase in male sex ratio with size after maturity suggesting potential differences in male and female growth with males reaching a larger 31  asymptotic size. Natural mortality is estimated to decrease with size (Polacheck et al. 2006) with adult natural mortality between >0.05-0.2 year-1 (CCSBT 2006a, Polacheck et al. 2006). Fisheries Southern bluefin are targeted using longline and purse seine gear. Reported catches from Japanese vessels commenced in 1952 with the expansion of the fleet into the Indian Ocean. Catches, primarily for canning, peaked at 20,000 tonnes in the late 1950s. With the development of freezing technology, southern bluefin was targeted for the sashimi market and catch peaked at 79,000 tonnes in 1961. By the early 1980s, catches had declined to 30,000 tonnes and decreased continuously to 6,000 tonnes by the 1990s. Taiwan began longline fishing in the late 1970s with catches reaching 1,000 to 1,500 tonnes in the 1990s. Reported catches from New Zealand longline vessels began in 1980 and fluctuated from 150-500 tonnes. Taiwan and Indonesia began longline operations in 1991 and 1986, with a combined catch of ~3,500 tonnes in the late 1990s. Australian catch (baitboat, troll, and purse seine) increased to 20,000 tonnes from 1950 through the early 1980s but declined rapidly to around 6,000 tonnes in the 1990s (Miyake et al. 2004). Australian catch prior to the 1980s was primarily for canning. However, because of the high price for southern bluefin on the Japanese sashimi market, farming operations have developed since the 1980s, where juveniles are purse seined and fattened in sea pens for a few months (Miyake et al. 2004). Total catch of southern bluefin in the 1990s was <20,000 tonnes. Perhaps the most alarming aspect of the southern bluefin fishery is the recent documentation that market assessments indicate the total reported catches may be only 50% of the actual catch removals (CCSBT 2006b).  Pacific bluefin tuna (Thunnus orientalis) Biology Pacific bluefin are a highly migratory species found in tropical and subtropical epipelagic waters of the Pacific Ocean in surface water temperatures ranging between 14-30° C, with larger individuals preferring 14-23° C outside the spawning season (Bell 1963, Uda 1957). In the western Pacific, bluefin are distributed 40° S-50° N from Sakhalin Island to southwestern 32  Australia and New Zealand (Bayliff 1994). In the eastern Pacific, the distribution is narrower and predominantly off the coasts of California and Mexico (roughly 20-35° N). Juveniles have been recorded along the Chilean, Oregon, and British Columbian coasts suggesting a potential range of 37° S-47° N, (Bayliff 1994). Archival tagging studies on immature pacific bluefin indicate a strong preference for surface waters with dives extending down to or below the thermocline (Itoh et al. 2003a, b, Kitagawa et al. 2006, 2007, Kitagawa et al. 2000, Kitagawa et al. 2002, Marcinek et al. 2001). However, Blank et al. (2004) demonstrated that Pacific bluefin are capable of tolerating very low temperatures so that larger individuals may exploit much deeper depths as they grow (Kitagawa et al. 2006). Such behaviour is also observed in Atlantic bluefin tuna (Block et al. 2001). Migration of Pacific bluefin is described in Bayliff (1994). Spawning is restricted to the western central Pacific between the Philippines and Japan (April-June), southern Honshu (July) and the Sea of Japan (August) (Nishikawa et al. 1985, Yamanaka 1963). Larvae, post larvae and juveniles are transported northward within Kuroshio Current waters, then migrate southward in winter months. After the first year of life, juveniles either remain in the western Pacific while seasonally migrating north-south, or make transpacific migrations to eastern Pacific waters within the Kuroshio meander and the north Pacific current where they remain for 1-6 years. Immature individuals that remain in the western Pacific show a north-south movement. After reaching maturity, Pacific bluefin appear to remain within the western Pacific while migrating north to south. Longevity is estimated to be >13 years (Hsu and Chen 2006). Age at 50% maturity is assumed around 150 cm or 5 years (Harada 1980). Growth parameters for the von Bertalanffy growth model are reasonably well defined though growth studies are limited. Asymptotic mean length is estimated between 320-325 cm and the curvature coefficient between 0.11-0.135 year-1. Natural mortality is estimated to be 0.276 year-1 for older individuals (Bayliff et al. 1991) and 1.6 year-1 for age 0 individuals (Takeuchi and Takahashi 2006).  33  Fisheries Pacific bluefin are mainly captured by Japanese fleets, using a variety of gears: purse seine, longline, baitboat, trolls, gillnet, and trap net. Purse seine is the dominant gear in both the western and eastern Pacific (Miyake et al. 2004). Pacific bluefin were likely caught in moderate quantities prior to 1950 particularly in the waters surrounding Japan (Matsuda and Ouchi 1984, Miyake 2004). Longline fisheries specifically targeting pacific bluefin only occur near the spawning grounds within the western Pacific, though non-targeted catch occurs throughout the Pacific in regions that are more temperate. Taiwanese and Korean catch of Pacific bluefin is relatively minor compared to Japan with a combined catch in the late 1990s of <5,000 tonnes. Japanese longline catch through the 1960s was significant with catches >5,000 tonnes. When compared to purse seine catches, longline catches in recent years are small. In general, purse seine vessels capture intermediate sized individuals, though substantial variation in length is observed depending on location, cohort strength, and time of year (Itoh 2006, Nakamura 1969). Juveniles are captured in the winter and adults in the summer. Japanese vessels account for a significant proportion of the total purse seine catch varying between 5,000-20,000 tonnes. Korean purse seines have operated since 1980. Within the Eastern Pacific, U.S. and Mexican purse seines catch <5,000 tonnes. Japanese baitboat, troll, trap, and gillnet fisheries account for around 10,000 tonnes though catches have been low through the 1990s (Miyake et al. 2004). Atlantic bluefin tuna (Thunnus thynnus) Biology Atlantic bluefin are a highly migratory species found in tropical to sub-arctic epipelagic waters of the Atlantic Ocean. Eastern and western Atlantic stocks are assumed to exist with separate spawning grounds in the Mediterranean Sea and the Gulf of Mexico (Fromentin and Powers 2005, Mather et al. 1995) though mixing occurs in the North Atlantic (Block et al. 2005). Atlantic bluefin have a wide distribution though predominantly in the northern hemisphere. Information of migration and distribution are summarized in Fromentin (2005) and Mather (1995). In the eastern Atlantic, bluefin are distributed as far north as 70° N down to the equator and from 25° S-50°N in the west with a west-east distribution from the Gulf of Mexico to the 34  Black Sea. Spawning occurs in the Gulf of Mexico and Mediterranean Sea where surface waters are >24° C (Baglin 1982, Richards 1976). Spawning individuals appear to show site fidelity (Block et al. 2005). As with other bluefin species, Atlantic bluefin spend nighttime hours as well as a large proportion of daylight hours in surface waters (Block et al. 2001, Brill et al. 2002, Lutcavage et al. 2000, Wilson et al. 2005). However, dives into deep (500-1000 m) cold (<4° C) water for prolonged periods have been observed; depth and temperature limits appear to depend on body size (Block et al. 2001, Brill et al. 2002, Lutcavage et al. 2000). Diving behaviour is thought to be associated with foraging in the deep scattering layer. Western Atlantic bluefin are assumed to originate from spawning in the Gulf of Mexico where spawning occurs between April-June within the Gulf or the Straight of Florida (Baglin 1982). In the Gulf of Mexico, larval densities are highest below 200m between 23°-30° N and 84°-94° W. Spawning in the Straight of Florida appears to occur west of Bimini Island (Mather et al. 1995). Young of the year individuals appear to spread widely through the Strait of Florida into waters surrounding the Bahamas and Greater Antilles, moving northward along the eastern coast of the U.S. to nursery grounds between Cape Hatteras and Cape Cod. After the first year of life, individuals seasonally migrate during the spring-summer feeding period along the coast and within frontal structures of the Gulf Stream. The extent of this southwest to northeast migration depends on size. During winter months individuals move into offshore waters. Transatlantic migrations have been observed for individuals older than age 1, with tag recaptures occurring predominantly in the Bay of Biscay. After leaving, the Gulf of Mexico, post-spawning adults move either northward along the coast to feeding areas in the North Atlantic reaching as far as the coast of Norway, or move southward to areas off Brazil. Tagging data suggest substantial mixing of the eastern and western stocks in the north Atlantic (Block et al. 2001, Block et al. 2005). Eastern Atlantic bluefin spawn within the Mediterranean Sea from June to August (Richards 1976). Spawning is thought to occur in both the west (Balearic Islands, Malta Island and the Tyrrhenian Sea) (Corriero et al. 2003, Medina et al. 2002, Susca et al. 2001) and east (Ibero-Moroccan embayment and Black Sea) (Karakulak et al. 2004). Young of the year move towards the Atlantic coast of Morocco by age 1. As individuals emerge from the Mediterranean, migration is to the north within the Bay of Biscay, toward the north coast of Spain, or south along the Atlantic coast of Africa. During winter months, individuals appear to move offshore. Older individuals move northward as far as the northern coast of Norway or southward crossing the Atlantic to the 35  coast of Brazil. Substantial movement into the middle of the north Atlantic is also seen (Block et al. 2005). Such migratory patterns are generalizations as individuals of all sizes are found within the Mediterranean throughout the year. Longevity is estimated to be > 25 years (Mather et al. 1995). Length at 50 % maturity for eastern Atlantic females is estimated at 103 cm or 3-5 years age (Cort 1991). Western Atlantic individuals appear to mature at older ages, 190-240 cm, or 8-12 years (ICCAT 1997). Growth parameters for the von Bertalanffy growth model are reasonably well defined though estimation of age for larger-sized individuals is problematic due to small sample size (Mather et al. 1995). Currently accepted values for asymptotic mean length are 318-325 cm with the curvature coefficient between 0.093-0.11 year-1. Differences in sex specific growth have been noted (Caddy et al. 1976) as well as strong seasonal growth pattern (Mather et al. 1995). Natural mortality is estimated at 0.165 year-1 for mature individuals in the eastern Atlantic and 0.14 year1  for the western stock (ICCAT 1997).  Fisheries Atlantic bluefin, particularly eastern Atlantic bluefin, were caught at significant levels prior to 1950 (Mather et al. 1995, Miyake et al. 2004). Records of trap catches of bluefin within the Mediterranean Sea date back for centuries (Ravier and Fromentin 2001) with estimates of total catch varying from 7,000-30,000 tonnes with 15,000 tonnes caught on average (Ravier and Fromentin 2002). Longline, purse seine, trap nets, baitboat, gillnet, harpoon, hand line and recreational gear are used to catch Atlantic bluefin. Since the 1970s, removals using purse seines have dominated the catch. Japanese longliners began targeting bluefin off the coast of Brazil in the late 1950s with catches increasing rapidly to 12,000 tonnes by the mid 1960s. Catches declined in the late 1960s as the fleet expanded to the Gulf of Mexico, New England Coast, and the Bay of Biscay. Stringent quotas on western Atlantic catch pushed the fleet to the Mediterranean in the early to mid 1980s. Much of the fleet left the Mediterranean in the 1990s for new fishing grounds in the North Atlantic (Miyake et al. 2004). Since the mid 1970s, Japanese catch has fluctuated between 2,000-6,000 tonnes. Other longline operations in the Mediterranean are operated by France, Spain, or classified as IUU. Purse seining in the western Atlantic by U.S. vessels increased during the mid 1960s peaking over 5,000 tonnes. Catch 36  restrictions since the 1980s resulted in a lower biomass taken. The only other significant fisheries for bluefin in the western Atlantic are recreational and trap fisheries, and catches under the current quota system are small. Purse seining in the eastern Atlantic from 1930 to the mid 1960s was by Norwegian vessels with catches in the 1950s between 5,000-15,000 tonnes. Declining catches in the 1960s resulted in the fishery shifting to other resources in the early 1970s. France started purse seining in the Mediterranean in the late 1960s followed by Spain and Italy. Greece, Croatia, Algeria, and Turkey also joined the fishery. Catches in the Mediterranean peaked in the mid 1990s around 25,000 tonnes but declined into the late 1990s. An increasing proportion of the purse seine catch since the 1990s has been used for farming operations (Miyake et al. 2004). French and Spanish baitboats have operated in the Bay of Biscay since the 1930s with catches varying, but usually <5,000 tonnes. Trap fisheries in the Mediterranean had peak catches of 20,000 tonnes in the late 1950s declining to <5,000 tonnes by the early 1970s. Blue marlin (Makaira nigricans) Biology Blue marlins occur in epipelagic tropical and sub-tropical oceanic waters of the Indian, Pacific, and Atlantic Oceans between 45° S-45° N. The 24° C isotherm is thought to delineate northern and southern range extents (Nakamura 1985). Blue marlins are solitary individuals with preference for open ocean waters. Tagging data from the Atlantic suggest blue marlin follow cyclical migration paths with spawning site fidelity (Ortiz et al. 2003). Migration patterns and spawning sites are not well determined for stocks in the Pacific or Indian oceans. Trans-oceanic migration and inter-oceanic migrations have been observed. Pepperell (2000b) notes that females appear to migrate further into subtropical waters than males in the Indian Ocean. The ability of females to tolerate colder temperatures is likely related to their larger size and potential thermal inertia. Tagging studies exploring habitat use indicate blue marlin spend a substantial proportion of time in surface waters >25 m making excursion to the top of the mixed layer with dive depth limited to 200-300 m depending on location (Block et al. 1992, Holland et al. 1990a, Saito and Yokawa 2006) though Goodyear (2006) noted dives up to 800 m. As with other billfish, blue marlin have evolved a ‘brain heater’ (Block 1986) which maintains brain and eye temperature above ambient temperature. 37  In the Indian Ocean, blue marlins are distributed north of 45° S in the west and north of 35° S in the east. Seasonal concentrations are found off Sri Lanka, the Maldives, and Laccadive Islands (December to August), around Mauritius (December to February), between 0-13° S off the east coast of Africa (April to October) and year round between Java and northwestern Australia with peak abundance from November to April (Nakamura 1985, Pepperell 2000b). Larval blue marlins have been observed around the Maldives, Mascalene Islands, Java, and Sumatra (Nakamura 1985). In the Pacific Ocean, blue marlin distribution extends from 35° S-45° N in the west and 25° S-35° N in the east though abundance declines along a west to east cline (Nakamura 1985). Within the western and central Pacific, concentrations have been observed between 8-26° S from December to March, between 2°-24° N from May to October and 10° S-10° N from April to November (Nakamura 1985). Larvae have been collected in both tropical and subtropical waters of the western and central Pacific suggesting year-round spawning in equatorial waters and seasonal spawning in more subtropical waters (Nakamura 1985). Tagging data from the Pacific indicate trans-oceanic migrations though no apparent cyclical migration patterns (Ortiz et al. 2003). Atlantic blue marlins are distributed from 40° S-45° N in the western ocean and 30° S- 45° N in the east. Nakamura (1985) notes two seasonal concentrations of blue marlin in the western Atlantic, from January to April between 5-30° S and from June to October between 10-35° N. Spawning has been observed from January to February at 17-23° S and 37-42° W (Amorim et al. 1998). Martins (2007) observed spawning between 20° S-7° N west of 15° W from June to August. Spawning around the Bahamas occurs in Exuma Sound around July (Serafy et al. 2003). In the eastern Atlantic where blue marlin abundance is lower, areas of higher abundance are off the African coast between 25° S-25° N (Nakamura 1985). In the western North Atlantic, significant movement occurs between the US mid-Atlantic coast and the Gulf of Mexico to Venezuelan waters, hinting at a cyclical migration between these two areas for feeding and spawning (Ortiz et al. 2003). Trans-Atlantic movements have been observed in a small fraction of recaptures as well as movement from the Atlantic Ocean into the Indian Ocean (Ortiz et al. 2003). 38  Life history information on mortality, growth, and maturity of blue marlin is sparse. Longevity for Pacific blue marlin in waters around Hawaii has been estimated at 27 years for females and 18 years for males (Hill et al. 1989). Blue marlin growth over the first year of life can be considerable, with individuals reaching >100 cm (Prince et al. 1991). Blue marlins are sexually dimorphic with females reaching larger body size than males. Pepperell (2000b) notes that males rarely exceed 180kg (~300cm) while females have been caught >800 kg (>400 cm). Skillman (1976) estimated asymptotic mean length in males between 368-370 cm and 626-660 cm for females with a curvature coefficients between 0.285-0.315 year-1 and 0.116-0.123 year-1 respectively. Age at 50 % maturity is poorly determined. Nakamura (1985) suggested males first mature at 130-140 cm (lower jaw fork length LJFL) and females >200 cm. Pepperell (2000b) indicates weight at first maturity for males around Hawaii was 31 kg (~170 cm) for males and 80 kg (~230 cm) for females, whereas Arocha (2006) observed female Atlantic blue marlin size at 50% mature around ~250 cm. These observations suggest males mature at age 2-3 years while females mature at age 4-5 years. Estimates of instantaneous natural mortality based on growth rates are higher for males. Boggs (1989) estimated male natural mortality at 0.53 year-1, female natural mortality at 0.21 year-1. Hinton (2001) estimated male mortality between 0.38-0.41 year-1 and female between 0.18-0.19 year-1 using meta-analysis in Pauly (1980) and growth equations derived by Skillman (1976). Fisheries Blue marlins are primarily caught on longlines as by-catch in tuna fisheries though catch is taken in gillnet, harpoon, purse seine, and recreational rod-and-reel fisheries. Longliners targeting blue marlins generally use modified tuna gear to fish shallow (Nakamura 1985). Prior to 1970, Japanese longline removals dominated blue marlin catch in the Indian Ocean with peak catches in the mid 1950s around ~5,000 tonnes declining after the early 1970s and varying around 2001,000 tonnes. As Japanese catch declined in the 1970s, Taiwanese catch increased and has varied between 1,000-4,000 tonnes. Catches in the Indian gillnet fishery increased in the early 1970s and accounts for a significant proportion of the gillnet catch from the Indian Ocean. By the late 1990s, reported catches varied between 1,000-2,500 tonnes. Total reported catch for the late 1990s fluctuated between 6,000-10,000 tonnes. Japanese catch of Pacific blue marlin peaked in 39  the early 1960s at >25,000 tonnes but declined thro