作者: Maria A. Terres , Emma Lawrence , Geoffrey R. Hosack , Michael D. E. Haywood , Russell C. Babcock
DOI: 10.1371/JOURNAL.PONE.0136799
关键词:
摘要: Baited Underwater Video (BUV) systems have become increasingly popular for assessing marine biodiversity. These provide video footage from which biologists can identify the individual fish species present. Here we explore relevance of spatial dependence and park boundaries while estimating distribution habitat associations commercially recreationally important snapper Chrysophrys auratus in Moreton Bay Marine Park during a period when new National Parks zoned as no-take or “green” areas (i.e., with no legal fishing) were introduced. BUV studies typically enforce minimum distance among sites, then assume that observations different sites are independent conditional on measured covariates. In this study, additionally incorporated into modelling framework. This approach allowed us to test whether not incorporation highly correlated environmental covariates geographic placement produced dependence, if unaccounted could lead model bias. We fitted Bayesian logistic models without random effects determine available had an effect presence preference. Adding component little resulting parameter estimates emphasized positive association particular coastal types by snapper. Strong relationships between rock habitat, particularly rocky substrate composed indurated freshwater sediments known coffee rock, kelp reinforce consideration availability reserve design any associated monitoring programs.