Spatiotemporal patterns in marine fish and cephalopods communities across scales: using an autoregressive spatiotemporal clustering model. A study of fish and cephalopods of the Eastern English Channel
In marine ecology, spatial and temporal fish community structure is important to understand ecosystem’s responses to anthropogenic and environmental factors. Although spatial distribution patterns and fish time series of abundance have been studied in the past, little research has addressed the impact of spatial scale on mapping spatiotemporal patterns. To achieve this goal, the first step is to develop a method describing spatiotemporal communities statistically. At three spatial resolutions, from 1043 km2 to 522 km2, we compared the output of a new framework, the Autoregressive Spatiotemporal Clustering analysis (ASTEC) and we finally retained the most parsimonious scale: 522 km2. ASTEC combines a zero-inflated Poisson model with latent predictors alongside a hierarchical clustering analysis using multiscale bootstrap resampling. It estimates the spatial and spatiotemporal communities formed by 16 species in the Eastern English Channel using 20 years of data from the Channel Ground Fishery Survey. Community structure was strongly dependent upon the spatial scale of analysis and the smallest scale provided the most parsimonious model. Moreover, integrating temporality in spatial studies defined a stable composition and richness community structure. Combining a spatial and spatiotemporal approaches highlighted cluster composition shifts over time. It shows that small-scale data should be used for studies on spatial connectivity and meta-community when integrating temporality. Moreover, the scale of the study did not matter anymore in a spatiotemporal case. A combined spatial and spatiotemporal approach defines community dynamics, and may help management and conservation efforts facing climate change and anthropogenic pressures.