Modeling the Oceanographic Impacts on the Spatial Distribution of Common Cephalopods During Autumn in the Yellow Sea
The importance of the role of cephalopods in marine ecosystems and commercial fisheries has increased over recent years. There is now evidence that the distribution of cephalopods is expanding latitudinally. Nevertheless, information about the spatial distribution of cephalopods and its implications in the Yellow Sea (YS) is not well known. In an attempt to redress this deficiency, we firstly conducted a simple analysis of geospatial patterns in fishing effort in the YS during 2012–2016 to ascertain if changes in fishing intensity (across all species) might be responsible for creating an apparent latitudinal shift in cephalopod distribution. Although fishing intensity increased in the YS over the 5-year period, there are no significant differences among years within each latitude, implying that all latitudes respond in a similar way each year. We then used long-term scientific survey data (2000, 2009, 2014, and 2017) of cephalopods (Todarodes pacificus, Loliolus spp., Octopus variabilis, Octopus ocellatus, Sepiola birostrata, and Euprymna spp.) collected each October in the YS, combined with oceanographic variables including sea surface temperature (SST) and chlorophyll-a concentration (CHLA), to model relationships to establish habitat suitability indices (HSIs) using both an arithmetic mean method (AMM) and a geometric mean method (GMM). Cross-validation, standard deviation, and mean squared error of prediction (MSEP) were used to evaluate the performance of the HSI. Abundance index data from surveys of 2018 were overlaid on maps of predicted HSI for the same year to visualize the correspondence of the modeled HSI. Spatiotemporal mapping of oceanographic variables showed that SST and CHLA change dramatically around 34°N, which may relate to the spatial distribution of cephalopods. CHLA is the most important oceanographic variable for most squid and octopus species, while SST is the most important for bobtail squid. The MSEP showed that the AMM-based HSI performed better than the GMM-based HSI. Future studies should take weighting of oceanographic variables into account and ideally integrate them into a more holistic model to obtain increased precision in predictions when establishing HSIs.