| Summary: | Predictive habitat modelling is a tool which provides a means to assess the current and potential distribution of organisms by utilising environmental data, combined with the known presence of the species. It is therefore important that the resolution of environmental data used reflects the scale most relevant in determining the species distribution. This investigation assesses whether environmental data at a resolution of 750m, 200m or 30m is best for modelling the distribution of the cold-water coral Lophelia pertusa (L. 1758) in north east Atlantic waters. Each of the three models was created using MaxEnt, utilising bathymetrical derived environmental data along with data on the known presence of Lophelia pertusa reefs. This allowed for a comparison between the three resolutions of model performance, the most important environmental parameters in the model formation, and the spatial output of the three models. The 30m model used all three resolutions of environmental layer, allowing for a direct comparison of the contribution of the different resolution of layers to the model formation. The investigation found the lowest resolution model had a lower model performance, and a much larger area of predicted presence than the two higher resolution models. The 30m and 200m models displayed no significant difference in model performance or spatial output. This is due to the 30m model relying on the 200m layers in its calculation, with analysis of the contributing environmental layers showing that the 30m environmental layers contributed very little to the model gain. This shows that an increase in terrain detail does not always make for a better fitting model, but that a resolution which displays the scale of environmental data most important in determining an organism’s distribution is more important.
|
|---|