Statistical modelling for exploring diel vertical movements and spatial correlations of marine fish species: a supplementary tool to assess species interactions

Abstract

Abstract Species interactions are important for explaining community dynamics and population trends, but very time- and effort-consuming to assess. Although diel vertical movements (DVM) and spatial correlations for various fish species have been explored to assess species interactions in aquatic communities, there is lack of cost-efficient approaches to studying both aspects concurrently. In this study, we proposed a single-species spatio-temporal model to quantify diel variations in the probabilities of a non-empty catch (p) and catch rates given that the catch is non-empty (λ). The model is applied to investigate patterns in DVM of nine dominant fish and shrimp species in the Yellow Sea (YS) using bottom trawl survey data. Subsequently, we utilized the spatio-temporal models built for the nine species to hindcast the spatial distributions of p and λ by year and species, which were further used to calculate between-species spatial correlation coefficients (corspatial). Results showed the nine species exhibited three different patterns of diel variation in p or λ, indicating they utilize the water column swept by the bottom trawls at different times of a day. Moreover, for the nine species in the YS those with limited interactions generally have negative corspatial or positive corspatial < 0.2 in λ, and others linked by strong predator–prey relationships or major food overlaps have corspatial ≥ 0.46 in λ. The strength of the positive spatial correlations in λ of different fishes shows some correspondence with the intensity of predation and competition between species and may be a supplementary tool to assess species interactions.