基于HMSC模型分析山东近海夏季底层鱼类的环境适应性与种间关系

Environmental adaptability and interspecific relationships of demersal fishes in the coastal waters of Shandong in summer explored by HMSC models

传统的物种分布模型很少将种间关系纳入建模框架中,妨碍了对物种栖息分布的准确预测。近年来联合物种分布模型(JSDMs)越来越受到关注,但在海洋领域实际应用仍较为缺乏。本研究根据2017年夏季山东近海底拖网调查数据,结合水深、底层水温和底层盐度等环境数据,采用物种群落层次模型(HMSC)方法研究了山东近海17种底层鱼类与环境因素之间的关系和种间相关性。本研究根据生物与环境之间的线性或非线性关系以及随机效应构建了5种HMSC,并利用广泛适用信息准则(WAIC)等指标以及交叉验证方法,评价了模型拟合程度和预测效果。结果表明,最优模型为包含随机效应的非线性模型(模型五),非线性模型优于线性模型,且在模型中考虑种间关系能明显地提高模型的拟合效果。温度是影响山东近海底层鱼类分布的主要因素,占平均可解释方差的51.4%,其次是水深和随机效应,分别占35.7%和12.8%。山东近海大部分底层鱼类与水深存在显著线性正相关关系,而与水温存在显著的非线性关系。底层鱼类种间具有显著相关性,按其相关性的正负可大致分为3组,表明种间关系在预测物种分布方面的作用不容忽视。本研究建议,在建模中应同时考虑非生物因素和生物之间的相互关系,研究结果为预测渔业资源栖息分布提供了重要参考。

Traditional species distribution models rarely incorporate interspecific relationships into the modeling framework,which hinders their predictions of habitat distributions.In recent years,joint species distribution mod-els(JSDMs)have drawn increasing attentions,but their practical applications remain rare in the marine realm.In this study,we used the HMSC(hierarchical modelling of species communities)method to study their relationships between 17 demersal fish species and environmental factors and the interspecific correlation.The model was built on the basis of bottom trawling data collected in the coastal waters of Shandong in summer,2017,including the en-vironmental data of water depth,bottom water temperature and bottom water salinity.Five variants of HMSC mod-els were developed with respect to the linear or nonlinear relationships between species and the environmental vari-ables and the exists of random effects,and WAIC and other indicators as well as cross-validation were used to eval-uate the performances of fitting and prediction of these models.The results showed that the optimal model was the one incorporating nonlinear relationships and random effects(Model 5).The nonlinear models were generally su-perior to the linear models,and including the interspecific relationships in the model could improve model fitting performances.Temperature was the main factor influencing the distribution of demersal fishes in the coastal waters of Shandong,accounting for 51.4%of the mean explained variance,followed by water depth and random effects,which accounted for 35.7%and 12.8%explained variance,respectively.There were significant linear positive cor-relations between most demersal fishes and water depth,and significant nonlinear relationships with water temperat-ure.There were significant interspecific correlations among the demersal fishes,which could be roughly divided in-to three groups according to the sign of the correlations,indicating that the interspecies relationships played an im-portant role in shaping species distributions.This study suggested that the abiotic factors and biotic factors should be integrated in species distribution modeling,and our results might provide a guideline for the prediction of habit-at distribution of fishery resources.