Modelling semi-natural habitat area on lowland farms in western Ireland

Remote sensing, farmland statistics and species data have been used in some EU countries to create maps to estimate the extent of semi-natural habitat cover but these are acknowledged as being too broad scale. In this study, we examined a method of fine-scale prediction of the spatial coverage of semi-natural habitats in lowland farms. A generalized additive model (GAM) was used to investigate the relationships between landscape and farm management variables and the lowland farmland habitat biodiversity on 32 farms outside of conservation designations, in a region of western Ireland. Semi-natural habitat cover on lowland farms could be predicted with a model using stocking density, soil diversity and river and stream length. It is proposed that this model could be used to predict the coverage of semi-natural habitats on farms in other regions of Ireland with similar land-use and landscape. A similar modelling approach could be adapted for application in other regions of Ireland and across Europe with different landscapes to predict semi-natural habitat coverage.