Probability maps for different phytoplankton species in the North Sea

This data product is a series of probability of occurrence maps for 200 phytoplankton species from 1995 to 2015 in the Greater North Sea, based on a neural network analysis. The input data files have been produced by Deltares (Luuk van der Heijden, Willem Stolte). They used EMODnet-Biology data as a basis. The selection of relevant datasets was optimized in order to find all planktonic species, and exclude all species that are not planktonic. The occurences from EMODnet Biology were complemented with absence data assuming fixed species lists within each dataset and year.

In addition to the occurrence data, the environmental variables bathymetry, silicate, phosphate and nitrate concentrations were used. Additionally the position (latitude and longitude) and the year are provided to the neural network. DIVAnd (n-dimensional Data-Interpolating Variational Analysis) and the neural network library Knet were used in this analysis.