Mulit-scale Analysis Of Spatial Patterns For Ommastrephes Bartramii And Their Relationships With Marine Environments In The Northwest Pacific

Marine commercial fishes swim horizontally and vertically and are variable to the environments.Their spatial distribution is complex and changes across time(year,month and day).Investigation of the spatial pattern of fishery resources and its environmental impacts is not only scientificlt valueable in exploring its spatial ecology,but also has practically significant in forecasting fishing grounds.Convertional studies of fishery distribution is commonly based on an individual spatial scale(fishing grid),hence the results are only valid under this scale.Studies have shown that spatial scale has a crucial influence on the spatial analysis of fisheries and their distribution patterns,that is,spatial models are sensitive to the choice of scale.However,several questions such as how the spatial scale affects the analysis of the spatial fishery models,how it affects the CPUE-environmental relationships,and how it influences the modeling and prediction of fishery habitats has not been adequately addressed in the literature.Therfore,there is an urgent need for systematic analysis of these issues to provide effective scale recommendations for spatial analysis of fishery resources and fishery surveys.To address the above research issues,this paper thesis investigated the spatial pattern of Ommastrephes bartramii and its relationships with the marine environments in the Northwest Pacific,under multiple spatial scales.The O.bartrami is one of the important marine commercial fishery in the Northwest Pacific Ocean and is an important fishing target for Mainland China,Taiwan,South Korea and Japan.The original fishery data presented in this thesis are production records of O.bartrami in the Northwest Pacific from July to November in 2004-2013.The marine environmental factors include sea surface salinity(SSS),sea surface temperature(SST),chlorophyll-a concentration(Chl-a)and sea surface height(SSH).To perform multi-scale analysis,the original commercial fishery data and oceanographic factors were tessellated to 12 spatial scales from 5' to 60',with an interval of 5',resulting in a total of 13 spatial scales including original scales.Under multiple spatial scales,the local spatial pattern of O.bartrami was analyzed using a spatial autocorrelation method,the CPUE-environmental relationships were detected using the generalized additive model(GAM),and the habitat suitability of the squid was evaluated using the habitat suitability index(HSI).(1)Under multiple spatial scales,this thesis examined the distribution patterns of cold spots and hot spots(central fishing ground)for the O.bartramii resources in different months(July-November),and analyzed their variations affected by spatial scale and the environmental factors(SSS,SST,Chl-a,and SSH).This thesis identified the most significant spatial scales according to the z-scores of the cold and hot spots.The results show that the most significant scales are quite different in each month,indicating the great effects of month on the spatial scale effects.The summary statistics in the cold spots and hot spots including average,standard deviation,skewness,and kurtosis of the four environmental factors strongly change with the changing spatial scales.However,no significant mathematical rules were observed for such changes because the poor goodness-of-fitting(R2).This indicates the spatial complexity of the hotspots and the environmental factors,while the complexity of the scaling reponse may come from the spatial data itself and resampling processing.(2)Under multiple spatial scales,GAM was used to establish the relationships between CPUE and the space-time and environmental factors.The fitting statistics(factor sort-order and deviance explained)of GAM are important indicators of the effects of factors on CPUE.This research further studied the spatial scale relationships of factor sort-order and deviance explained in GAM,and analyzed the spatial scale effects of the relationships between O.bartramii CPUE and factors.The factor sort-order shows that the time factors(Year and Month)have the greatest impacts on the CPUE,followed by the spatial factors(Lon and Lat)and the environmental factors(SST,SSS,SSH,and Chl-a).The deviance explained of Year,Month,SSS,and Chl-a followed the scale relationship of quadratic polynomials,Lat and SST followed the scale relationship of power laws,Lon followed linear scale relations,and SSH followed exponential scale relations.At different scales,the cumulative deviance explained of all factors followed the quadratic polynomial relationship.Based on the scale sensitivity of the sort-order and the cumulative deviance explained of GAM,this study identified that 30'-45' is the optimal spatial scale range for studying the CPUE-environments relationships.(3)Under multiple spatial scales,we selected the 2005-2013 fishery and environmental data establish HSI models to study the spatial scale relationships that affect the environmental suitability of O.bartramii using different environmental factors.This study then identified the optimal spatial scale for HSI.The 2004 fishery data was applied to validate the HSI model using a inverse distance weighted interpolation.The CPUE-based individual-factor suitability index established by SST and Chl-a followed the logarithm and quadratic polynomial relations,while those of SSS and SSH did not show significant mathematical relations.The SI established by SSS and Chl-a based the Effort followed quadratic polynomial relations,that of SSH followed the linear relations,and that of SST did not show obvious mathematical relations.HSI based on both CPUE and Effort were calculated using the arithmetic average method,which followed quadratic polynomial relations.Validation in 2004 showed that CPUE-based HSI modeling achieved the best results between 30'-55'(except 35'),and the optimal spatial scale of Effort-based HSI modeling was 20'.In general,the scale relationship of CPUE-based HSI has a higher degree of goodness-of-fit(R2=0.90).Meanwhile,the CPUE-based HSI values are higher than those of the Effort-based HSI,and CPUE is suggested as the preferrence when studying the O.bartramii through HSI.This study explored the scale effects and scale relations of the spatial distribution,CPUE-environments relations,and habitat suitability index of O.bartramii under multi-scale,and identified their possible optimal spatial scale ranges.The study therefore provides a basis to select the spatial scale in analyzing fishery resources,and theoretical supports for fishery resource surveys.