Understanding the effect of scale and resolution on predictive habitat maps

No abstracts are to be cited without prior reference to the author.

High intensity, localised, and targeted surveys using multibeam and drop down video have been carried out around Shetland aiming to map nationally and internationally important habitats of maerl and horse mussel (Modiolus modiolus) beds. These surveys have resulted in targeted closures to mobile dredging gear in order to conserve these habitats. However, the surveys were based on historical information and there is a growing requirement to identify additional suitable habitats which could be considered for protection. In order to achieve this, predictive habitat maps, using the MaxEnt species distribution model (SDM), were run and analysed at different spatial scales and data resolutions. Tide, bathymetry, and bathymetric associations such as slope and aspect were used as variables in the model. Comparisons were made at a large- (regional level) and small-spatial scale (site-specific) and between different bathymetry resolutions (e.g. localised high resolution versus broad scale lower resolution) to assess the impact of modelling resolution. Depth and tide had the greatest influence in the SDM outputs for both habitats. Access to high resolution data, in this case bathymetry, greatly increased the confidence of the SDMs.