Mapping nematode diversity in the Southern Bight of the North Sea

In order to protect the biodiversity of the seas from, e.g., overexploitation, the spatial distribution of biodiversity and the mapping of biodiversity hotspots are of great importance. In the present paper we discuss different methods to develop full coverage biodiversity maps of free-living marine nematodes in the Southern Bight of the North Sea. A database with sampling data, gathered over 3 decades (1972 to 2004), combined with exhaustive environmental data, was employed to predict species richness and the expected number of species by different methods: ordinary kriging (OK) and regression kriging (RK) with ordinary least squares (OLS) and generalised least squares (GLS). The predictive value of these methods was evaluated by an independent validation set. Replicate samples were used to make an accurate estimation of the nugget variance, since replicates reveal local variability. Accordingly, it was feasible to find a spatial pattern in the residuals of the regression models. Our analysis pointed out that GLS improved the OK models substantially, while RK only slightly improved the GLS model. The diversity of marine nematodes is substantially influenced by the silt-clay fraction and the amount of total suspended matter, which is also reflected in the resulting map with a species-poor area near the coast line, especially near the south of the mouth of the Scheldt estuary. Off coast diversity and evenness are generally higher.In order to protect the biodiversity of the seas from, e.g., overexploitation, the spatial distribution of biodiversity and the mapping of biodiversity hotspots are of great importance. In the present paper we discuss different methods to develop full coverage biodiversity maps of free-living marine nematodes in the Southern Bight of the North Sea. A database with sampling data, gathered over 3 decades (1972 to 2004), combined with exhaustive environmental data, was employed to predict species richness and the expected number of species by different methods: ordinary kriging (OK) and regression kriging (RK) with ordinary least squares (OLS) and generalised least squares (GLS). The predictive value of these methods was evaluated by an independent validation set. Replicate samples were used to make an accurate estimation of the nugget variance, since replicates reveal local variability. Accordingly, it was feasible to find a spatial pattern in the residuals of the regression models. Our analysis pointed out that GLS improved the OK models substantially, while RK only slightly improved the GLS model. The diversity of marine nematodes is substantially influenced by the silt-clay fraction and the amount of total suspended matter, which is also reflected in the resulting map with a species-poor area near the coast line, especially near the south of the mouth of the Scheldt estuary. Off coast diversity and evenness are generally higher.A