Publication on data.gouv.fr
Species and habitats - Scyliorhinus canicula - All ages - Potential habitat in October modelled by Quantile Regression and its uncertainty with CGFS data
Modelised abundance of species or prediction uncertainty.
In short, models based on GLMs predict the mean response of the species to environmental factors whilst models based on RQ predict the maximal response. When GLM uses abundance data, the preferential habitat is predicted, whilst the probable habitat is predicted when GLM uses binary presence-absence data. Generalised Linear Modelling (GLM) describes and predicts the "preferential habitat", i.e. the portion of the potential habitat that is used on average over time, or, in the case of presence-absence species data, the "probable habitat", i.e. where the species may be present. RQ tends to describe potential spatial patterns or the "potential habitat" of species, i.e. all possible areas with conditions suitable for the presence or high abundance levels of a species.
Purpose: not specified