Behaviour Impact Analysis of Tuna Purse Seiners in the Western and Central Pacific Based on the BRT and GAM Models

Frontiers in Marine Science

Huimin Shi¹  Yingjie Fei²  Sanling Yuan²  Han Zhang²  Shenglong Yang⁴  Wei Fan⁴ 

[1] College of Information, Shanghai Ocean University, Shanghai, China;College of Science, University of Shanghai for Science and Technology, Shanghai, China;Key Laboratory of Fisheries Remote Sensing, Ministry of Agriculture and Rural Affairs, Shanghai, China;Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, China;

关键词: tuna purse seiners;    BRT model;    GAM model;    habitats;    environmental factors;    spatial distribution;   

DOI  :  10.3389/fmars.2022.881036

来源: DOAJ

【 摘 要 】

Understanding the spatial pattern of human fishing activity is very important for fisheries resource monitoring and spatial management. To understand the spatial distribution of tuna purse seiner operations in the western and central Pacific Ocean and its relationship with the marine environment, this paper uses the AIS data of the western and central Pacific Tuna purse seiners from 2015 to 2020 to excavate spatial fishing effort information, which is combined with 24 marine environmental factors in the same period, including sea surface and subsurface levels using the boosted regression trees (BRT) model and general additive model (GAM) to construct the nonlinear relationship between the spatial distribution of fishing effort and marine ecological environmental factors and to discuss and analyse the niche of tuna purse seiners in the high seas. The results show that the average score of cv-AUC (cross-validated area under the curve) obtained by the BRT model training reaches 0.93, the average accuracy rate is 0.84, and the explained deviance is 43%; the average score of AUC (area under the curve) obtained by the GAM model training reaches 0.81, the average accuracy rate is 0.77, and the explained deviance is 34%. The results of BRT prior to GAM model. Using the BRT model for prediction, the results show that the average cv-AUC score for forecasting fishing effort in 2020 reaches 0.83, and the average accuracy rate of overall classification reaches 0.77. The results of factor analysis show that the water temperature at 100 m depth and longitude are the most important factors affecting the fishing effort of tuna purse seiners, and their contribution rates to the fishing effort of vessels are 12.38% and 9.76%, respectively, followed by sea surface temperature, latitude and DSH. The contribution to the fishing effort of tuna purse seiners was also large, accounting for 9.57%, 8.75%, and 7.11%, respectively; the 100-meter-deep chlorophyll and temperature gradient value contributed the least, 1.44% and 1.16%, respectively; tuna purse seiners are more likely to operate in the 100-metre water temperature of 25-29°C and sea surface temperature of 29-31°C. In terms of space, tuna purse seiners are more likely to operate in the 5°S-5°N latitudinal region and near the western sea area of 180°E. It is predicted that the modelled fishing effort of fishing vessels in 2020 and the actual fishing effort of fishing vessels have a relatively good spatial distribution. Research helps to understand the environmental impact of changes in the spatial distribution of tuna purse seiners and provides support for the management of tuna purse seine vessels in the western and central Pacific.

【 授权许可】

Unknown