A Study On Potential Habitat Distribution Models Of Pacific Saury In The Northwest Pacific Ocean

The Pacific saury is one of the main fish species caught in the Northwest Pacific Ocean in China,with important economic value,mainly caught by China,Taiwan,Japan,South Korea,Russia and other countries and regions.Spatial-temporal factors and habitat factors are of great significance for studying the formation of Pacific saury fisheries.In this study,the Biomod2 model was used to analyze the potential habitat distribution of Pacific saury in the Northwest Pacific Ocean,in order to provide a more scientific theoretical reference for the efficient fishing of Pacific saury fisheries.This study combines the information of commercial Pacific saury fishing statistics(longitude、latitude、catch per unit effort)obtained in the high seas of the Northwest Pacific Ocean from July 2013 to November 2019 with the habitat marine environmental factors sea surface temperature(SST),sea surface salinity(SSS),sea surface height(SSH),and near primary productivity(NPP)data.The BRT model and the Biomod2 model were used to analyze the influence weight relationship between environmental factors and catch per unit effort(CPUE)and potential habitat distribution of Pacific saury.The results are as follows:(1)The BRT model showed that the contribution rate of Marine environmental factors in different habitats to CPUE ranged from 16.27%to 33.30%,and the weight distribution of Marine environmental factors in different habitats in different months was different.The influence weights of habitat marine environmental factors in July were SSS(27.94%),SST(27.16%),SSH(24.68%)and NPP(19.22%),respectively.In August,the influence weights of habitat marine environmental factors were SSH(33.30%),SST(28.99%),NPP(21.44%)and SSS(16.27%).The influence weights of habitat marine environmental factors in September were SSS(27.74%),SSH(26.24%),SST(24.43%)and NPP(21.59%)in descending order.The influence weights of habitat marine environmental factors in October were SSS(28.67%),SSH(26.55%),SST(23.26%)and NPP(20.93%)in descending order.The influence weights of habitat marine environmental factors in November were SSS(27.99%),SSH(26.46%),NPP(22.85%)and SST(22.70%)in descending order.(2)According to the results of the potential habitat distribution of saury between months generated by 10 single-model in the Biomod2 model,it can be seen that the spatial distribution results of potential habitats from July to November show that the changes in Pacific saury habitats show the characteristics of changing from northeast to southwest.The suitable habitat area in July is distributed at 41°N～51°N,155°E～175°E;the distribution range of suitable habitat sea area in August is41°N～50°N,150°E～170°E,the suitable habitat area in September is mainly concentrated in 41°N～47°N,150°E～170°E,the suitable habitat sea area in October is distributed between 37°N～45°N,145°E～165°E,and the suitable habitat sea area in November is mainly 35°N～45°N,142°E and 165°E.The embedded evaluation indexes TSS and AUC values of the single model showed that the accuracy of random forest(RF)and generalized boosted models(GBM)ranked first,and the accuracy of surface range envelope(SRE)was the worst among the 10 single models involved in the modeling.The results of random forest(RF)suitable habitat area were the smallest and most convergent among the 10 species distribution models.The suitable habitat area was 75.85×10~4km~2in July,49.76×10~4km~2in August,59.45×10~4km~2in September,52×10~4km~2in October and 45.06×10~4km~2in November.The suitable habitat area was concentrated between 45×10~4km~2and 76×10~4km~2.However,the suitable habitat area of the generalized boosted models(GBM)is relatively large.The suitable habitat area was 110.7×10~4km~2in July,103.48×10~4km~2in August,92.07×10~4km~2in September,107.32×10~4km~2in October and 90.29×10~4km~2in November.The suitable habitat area is concentrated between 90×10~4km~2and 120×10~4km~2.(3)The Ensemble Model(EMwmeanby ROC)was constructed based on the single Model whose AUC value was greater than 0.925 as a result of single Model accuracy test in Biomod2 from 2013 to July to November 2018.The habitat distribution of Pacific saury is predicted from July to November,2019.The results showed that the TSS and AUC model of the combined model was better than that of the single model,and the potential suitable habitat distribution of Pacific saury was simulated with higher accuracy.The prediction accuracy of the combination model ranged from 0 to 94%,with an average of 70.3%.In each month,the prediction accuracy of the high suitable area was higher,while the prediction ability of the medium and low suitable area was lower.