Inclusion of explicit measures of geodiversity improve biodiversity models in a boreal landscape

Measures of geodiversity may provide a potentially useful surrogate for biodiversity patterns in insufficiently surveyed areas. However, their reliability in modelling the spatial variation in species richness is inadequately understood. We investigated whether the explanatory and predictive power of species richness models can be improved by considering explicit measures of geodiversity (variability of earth surface materials, forms and processes) in addition to climate and topography variables. Vascular plant species richness was modelled in two study areas in Northern Europe, Finland at the resolution of 500 or 1000?m, and as a function of three geodiversity (geological, geomorphological and hydrological diversity) variables, and six climate and topography variables. Variation partitioning was used to identify the independent and shared contributions of the geodiversity, climate and topography variable groups in explaining the spatial patterns of species richness. Generalized additive models were used to explore the ability of the different explanatory variables in predicting plant species richness within and between the study areas. In both the study areas, the inclusion of measures of geodiversity improved the explanatory power, predictive ability and robustness of the plant species richness models. In conclusion, the explicit measures of geodiversity appear to be promising surrogates of biodiversity, which reflect important abiotic resource factors, and may thus provide an equally, or even more reliable basis for transferring biodiversity models to new areas than models based on climate and topography variables.