Impacts of climate change on the distribution of blue marlin (Makaira. nigricans) ) as inferred from data for longline fisheries in the Pacific Ocean

Impacts of climate change on the distribution of blue marlin (Makaira nigricans) ) as inferred from data for longline fisheries in the Pacific Ocean Nan-Jay Su 1*, Chi-Lu Sun 1, Andre Punt 2, Su-Zan Yeh 1, and Gerard DiNardo 3 1 Institute of Oceanography, National Taiwan University, Taipei, Taiwan 2 School of Aquatic and Fishery Sciences, University of Washington, Seattle, WA, USA 3 NOAA Fisheries, Pacific Islands Fisheries Science Center, Honolulu, HI, U.S.A. 26-29 April 2010, Sendai, Japan 1

Contents 1. INTRODUCTION - Background information - Objectives of this study 2. MATERIALS AND METHODS - Fishery data - Environmental data - Statistical modeling - Predict distribution 3. RESULTS AND DISCUSSION - Model fitting - Annual variation - Regional variation -Conclusions 2

1. Introduction Blue marlin is a large, pelagic species Source: NOAA 3

1. Introduction Distribution 30 N 0 30 S Source: Nakamura (1985) 4

1. Introduction Various fisheries Longline Harpoon Gillnet Source: TTA, CTW, MAFF 5

1. Introduction Catch in the three oceans Source: FAO 6

1. Introduction Landing in the Pacific Ocean Source: FAO 7

1. Introduction Migration and movement Source: Hinton (2001) 8

1. Introduction Migration and movement a single stock in the Pacific 4500 miles 4 years Source: Hinton (2001) 9

1. Introduction Tagging experiment Source: Block et al. (1992) 10

1. Introduction Source: Dr. Hsu HH, NTUAS 11

1. Introduction Blue marlin inhabit certain preferred habitats in the open ocean. The preference of this species for particular habitats may impact its distribution and vulnerability to being caught. We could infer relative density through catch-rate information. The objectives of this study To examine 1) the relationship between spatial patterns of abundance and environmental factors, 2) the impact of climate change on the distribution of blue marlin, and 3) the potential influence on the catch-rates for the longline fisheries. 12

Contents 1. INTRODUCTION - Background information - Objectives of this study 2. MATERIALS AND METHODS - Fishery data - Environmental data - Statistical modeling - Predict distribution 3. RESULTS AND DISCUSSION - Model fitting - Annual variation - Regional variation -Conclusions 13

2. Materials and Methods Two fishery data sets of longline 5x5 grids a) SPC, aggregated over longline fisheries (PLL) for 1998-2004 b) OFDC, Taiwanese longline fisheries (TLL) for 1998-2007 Covariates: Year, Month, Lat, Lon, Fishing effort in hooks, and number of blue marlin caught 14

2. Materials and Methods Four data sets of environmental factors for 1998-2007 a) chlorophyll a concentration (CHL): SeaWiFS b) mixed layer depth (MLD): OSU c) sea surface height anomaly (SSH): AVISO d) sea surface temperature (SST): PODAAC - Average to 5x5 grids to match the fishery data 15

Statistical modeling 2. Materials and Methods - delta GAM (Generalized additive model) a) P/A (present/absent) model: 0 or1, assuming binominal P/A ~ s(year) + s(month) + s(chl) + s(mld) + s(ssh) + s(sst) + s(year,) + s(year,) + s(,) b) CPUE (abundance) model: catch-rates > 0, assuming lognormal CPUE ~ s(year) + s(month) + s(chl) + s(mld) + s(ssh) + s(sst) + s(year,) + s(year,) + s(,) Diagnostic analysis: residual distribution and quantile-quantile plot 16

2. Materials and Methods Predict relative density Consider both information from the P/A and CPUE models Prediction = probability of present relative density given presence Predict the distribution of relative density a) temporally: 1998-2007 for each month b) spatially: 5x5 latitude and longitude grids Define the hotspot areas, the relative densities in the top 5% 17

2. Materials and Methods Impacts of climate change Track temporal variations of the hotspot areas - by and Remove yearly variations - medians of environment data in 1998-2007 - predict year-invariant relative density distribution Examine climate impacts, by six regions - separate the Pacific into temperate and tropical areas 18

2. Materials and Methods The six regions of the Pacific Ocean 40 N 20 N NW Pacific NE Pacific 0 W Trop Pacific E Trop Pacific 20 S 40 S SW Pacific SE Pacific 120 E 140 E 160 E 180 160 W 140 W 120 W 100 W 80 W 19

Contents 1. INTRODUCTION - Background information - Objectives of this study 2. MATERIALS AND METHODS - Fishery data - Environmental data - Statistical modeling - Predict distribution 3. RESULTS AND DISCUSSION - Model fitting - Annual variation - Regional variation - Conclusions 20

3. Results and discussion Nominal CPUE of Pacific longline fisheries (PLL) < 0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 > 0.5 21

3. Results and discussion Nominal CPUE of Taiwanese longline fisheries (TLL) TLL for 1998-2007 < 0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 > 0.5 22

3. Results and discussion Diagnostic analysis: residual and Q-Q plots a) PLL for 1998-2004 Density 0.0 0.2 0.4 Sample Quantiles -6-2 2-4 -2 0 2 4 Residuals -4-2 0 2 4 Theoretical Quantiles b) TLL for 1998-2007 Density 0.0 0.2 0.4 Sample Quantiles -3-1 1 3-4 -2 0 2 4 Residuals -4-2 0 2 4 Theoretical Quantiles 23

3. Results and discussion ANOVA table of P/A models for PLL and TLL PLL for 1998-2004 TLL for 1998-2007 Resid dev % dev explain AIC %AIC P(Chi) Resid dev % dev explain AIC %AIC P(Chi) PA model NULL 24477 24479 13816 13818 +s(year) 24459 0.2% 24465 0.2% <0.01 13566 14.8% 13575 15.6% <0.01 +s(month) 24415 0.5% 24426 0.5% <0.01 13516 3.0% 13530 2.9% <0.01 +s(chl) 23953 5.6% 23971 5.6% <0.01 13288 13.5% 13309 14.2% <0.01 +s(mld) 22593 16.5% 22622 16.7% <0.01 13177 6.6% 13206 6.6% <0.01 +s(ssh) 22448 1.8% 22483 1.7% <0.01 13104 4.3% 13139 4.2% <0.01 +s(sst) 18487 48.0% 18526 48.8% <0.01 12641 27.4% 12682 29.3% <0.01 +s(year,) 17663 10.0% 17737 9.7% <0.01 12380 15.4% 12454 14.6% <0.01 +s(year, 16878 9.5% 16990 9.2% <0.01 12243 8.1% 12348 6.7% <0.01 +s(,longitu 16234 7.8% 16376 7.6% <0.01 12128 6.9% 12255 6.0% <0.01 Tot dev expl 33.7% 12.2% 24

3. Results and discussion ANOVA table of CPUE models for PLL and TLL PLL for 1998-2004 TLL for 1998-2007 Resid dev % dev explain AIC %AIC P(Chi) Resid dev % dev explain AIC %AIC P(Chi) CPUE model NULL 27312 53497 6633 16923 +s(year) 27098 1.9% 53380 1.4% <0.01 6529 7.2% 16840 6.5% <0.01 +s(month) 27036 0.5% 53349 0.4% <0.01 6441 6.0% 16771 5.4% <0.01 +s(chl) 26469 4.9% 53022 3.9% <0.01 6414 1.8% 16755 1.3% <0.01 +s(mld) 25190 11.2% 52248 9.1% <0.01 6378 2.5% 16729 2.1% <0.01 +s(ssh) 24966 2.0% 52114 1.6% <0.01 6320 3.9% 16684 3.6% <0.01 +s(sst) 18791 53.8% 47630 53.0% <0.01 5963 24.5% 16358 25.7% <0.01 +s(year,) 17689 9.6% 46713 10.8% <0.01 5512 31.0% 15953 31.8% <0.01 +s(year, 17046 5.6% 46166 6.5% <0.01 5356 10.7% 15822 10.3% <0.01 +s(,longitu 15844 10.5% 45028 13.4% <0.01 5177 12.3% 15651 13.5% <0.01 Tot dev expl 42.0% 22.0% 25

3. Results and discussion Partial effects of environmental factors for PLL s(chl) -1.0 0.0 1.0-1.2-0.6 0.0 0.0 0.2 0.4 0.6 0.8 1.0 CHL 50 100 150 200 250 MLD s(ssh) -1.0 0.0 s(sst) -2-1 0 1 s(mld) -0.2-0.1 0.0 0.1 0.2 0.3 SSH 15 20 25 30 SST 26

3. Results and discussion Partial effects of environmental factors for TLL s(chl) -2.0-1.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 CHL 0.0 0.5 1.0 50 100 150 MLD s(ssh) -0.8-0.4 0.0 s(sst) s(mld) -0.5 0.5-0.2-0.1 0.0 0.1 0.2 SSH 15 20 25 30 SST 27

3. Results and discussion Yearly variation of environmental factors (15 S-15 N) 2008 Chlorophyll a concentration -0.05 2008 Mixed layer depth -20 2006-0.10 2006-30 -40 2004-0.15 2004-50 2002-0.20 2002-60 2000-0.25 2000-70 1998-0.30 1998-80 110E 145E 180 145E 110W 75W 110E 145E 180 145E 110W 75W 28

3. Results and discussion Yearly variation of environmental factors (15 S-15 N) 2008 Sea surface height anomaly -0.15 2008 Sea surface temperature 23 2006-0.10 2006 24-0.05 25 2004 2004 26 0.00 2002 0.05 2002 27 28 2000 0.10 2000 29 1998 0.15 1998 30 110E 145E 180 145E 110W 75W 110E 145E 180 145E 110W 75W 29

3. Results and discussion Year-invariant environmental factor (median of 10 years) Dec Chlorophyll a concentration -0.05 Dec Mixed layer depth -20 Nov Nov Oct -0.10 Oct -30 Sep Sep Aug Jul Jun May Apr -0.15-0.20 Aug Jul Jun May Apr -40-50 -60 Mar -0.25 Mar -70 Feb Feb Jan -0.30 Jan -80 110E 145E 180 145E 110W 75W 110E 145E 180 145E 110W 75W 30

3. Results and discussion Year-invariant environmental factor (median of 10 years) Dec Sea surface height anomaly -0.15 Dec Sea surface temperature 23 Nov Oct -0.10 Nov Oct 24 Sep Aug -0.05 Sep Aug 25 Jul Jun 0.00 Jul Jun 26 27 May Apr 0.05 May Apr 28 Mar Feb 0.10 Mar Feb 29 Jan 0.15 Jan 30 110E 145E 180 145E 110W 75W 110E 145E 180 145E 110W 75W 31

3. Results and discussion PLL 1998 to 2007 Jan 1998 Jan 2001 Jul 1998 Jul 2001 Jan 2004 Jul 2004 Jan 2007 Jul 2007 32

3. Results and discussion TLL 1998 to 2007 Jan 1998 Jan 2001 Jul 1998 Jul 2001 Jan 2004 Jul 2004 Jan 2007 Jul 2007 33

3. Results and discussion Year-invariant relative density without climate change Jan Jan 2002.5 PLL Jul Jul 2002.5 PLL Jan Jan 2002.5 TLL Jul Jul 2002.5 TLL 34

3. Results and discussion Track the hotspot areas for PLL and TLL -5 0 5 15 25 PLL 120 160 200 240 PLL 1998 2000 2002 2004 2006 2008 1998 2000 2002 2004 2006 2008 Year Year -5 0 5 15 25 TLL 120 160 200 240 TLL 1998 2000 2002 2004 2006 2008 1998 2000 2002 2004 2006 2008 Year Year 35

3. Results and discussion Relative density by region for PLL NW Pacific NW Pacific CPUE 0.0 0.1 0.2 0.3 0.4 CPUE 0.00 0.10 0.20 0.30 Yr-variant Yr-invariant 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year CPUE 0.2 0.3 0.4 0.5 NW Pacific CPUE 0.2 0.4 0.6 0.8 1.0 NW Pacific 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year CPUE 0.00 0.10 0.20 0.30 NW Pacific CPUE 0.00 0.10 0.20 0.30 NW Pacific 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year 36

3. Results and discussion Relative density by region for TLL NW Pacific NW Pacific CPUE 0.0 0.4 0.8 CPUE 0.0 0.4 0.8 Yr-variant Yr-invariant 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year NW Pacific NW Pacific CPUE 0.2 0.6 1.0 CPUE 0.2 0.6 1.0 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year NW Pacific NW Pacific CPUE 0.0 0.4 0.8 CPUE 0.0 0.4 0.8 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year 37

Conclusions 3. Results and discussion The blue marlin population apparently moved eastwardly during the 1997-1998 El Niño in the tropics Yearly patterns in distribution appear to be associated with the El Niño event and related to shifts in SST Future directions Increase spatial and temporal coverage of data set Include more explanatory variables, e.g. fishing gear 38

Thank you!! 39

3. Results and discussion PLL Jan 1998 Jul 1998 Jan 2001 Jul 2001 Jan 2004 Jul 2004 Jan 2004 Jul 2004 40

3. Results and discussion TLL Jan 1998 Jul 1998 Jan 2001 Jul 2001 Jan 2004 Jul 2004 Jan 2007 Jul 2007 41

3. Results and discussion Predicted catch = predicted CPUE x observed fishing effort W Trop Pacific E Trop Pacific Catch (1000 fish) 0.0 1.0 2.0 Observed Yr-variant Yr-invariant Catch (1000 fish) 0.0 1.0 2.0 1998 2000 2002 2004 2006 2008 Year W Trop Pacific E Trop Pacific Catch (1000 fish) 0.0 1.0 2.0 Catch (1000 fish) 0.0 1.0 2.0 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year 1998 2000 2002 2004 2006 2008 Year 42