基于GAM的西北太平洋日本鲭资源丰度预测模型建立

Establishment of forecasting model of the abundance index for chub mackerel(Scomber japonicus) in the northwest Pacific Ocean based on GAM

日本鲭(Scomber japonicus)是西北太平洋重要的鱼类资源之一,科学预测日本鲭的资源丰度有利于其资源的合理开发和利用。本研究依据日本渔业机构提供的1987–2012年日本鲭太平洋群体的资源量数据,结合产卵场和渔场的海洋环境数据以及气候因子,使用广义加性模型对影响日本鲭太平洋群体的海洋环境和气候因子进行分析,筛选出有显著影响的因子并建立该群体的资源量预测模型。结果表明,与该群体资源量有显著关系的影响因子有:北极涛动指数、太平洋年代际振荡指数、渔场海表面高度、渔场海表面盐度和渔场海表面温度。基于赤池信息准则筛选出的4个资源量预测模型分析表明,包含北极涛动指数、渔场海表面高度和渔场海表面温度的模型有较好的预测效果,该模型的验证结果也通过了t检验(P<0.05),可用于日本鲭太平洋群体资源量的预测。

Chub mackerel (Scomber japonicus) is one of the important fishery resources in the northwest Pacific Ocean. Building a scientific forecast model of abundance index to this species is beneficial for its exploitation and utilization. In this study, based on the biomass data of the Pacific-cohort of Scomber japonicus during 1987–2012 obtained from Japan Fisheries Institution, as well as the marine environmental data and climatic data of spawning ground and fishing ground, we analyzed the relationship between the environmental and climatic factors and the biomass of this cohort. The significant factors were selected and the forecast models were established by using the generalized addictive models (GAM). The result shows the significant factors affecting the biomass of this cohort conclude the Arctic Oscillation index (AOI), Pacific Decadal Oscillation index (PDOI) and sea surface height (SSH2), sea surface salinity (SSS2) and sea surface temperature (SST2) both in the fishing ground. Result based on Akaike’s Information Criterion (AIC) suggests that the model 1 which included AOI, SSH2 and SST2 has the optimal model impacts. The model 1 passes the significant test (P<0.05) and the t test (P<0.05) is also passed based the validation result of model 1. Therefore, we suggest that this model can be used to forecast the abundance of the Pacificcohort of Scomber japonicus.