Modelling of rare plant species richness by landscape variables in an agriculture area in Finland

A multivariate linear regression model is proposed for predicting and mapping rare vascular plant species richness in Finnish agricultural landscapes according to landscape variables. The data used in developing the model were derived from a floristic inventory from 105 0.5 km × 0.5 km grid squares. Using a stepwise multiple regression technique, four landscape variables were found to explain 71.8% of the variability in the number of rare plant species. The results suggest that the local `hotspots' of rare plants (squares with 5 rare taxa) are mainly found in heterogeneous river valleys, where extensive semi-natural grasslands and herb-rich forests occur on the steep slopes. According to other similar studies, intermediate human disturbance increases the number of rare species in agricultural landscapes. It appears that empirical models based on landscape variables derived from digital maps can provide relatively accurate surrogates for extensive field surveys and fine-scale observations on the distributions of rare taxa in agricultural landscapes. Potential reasons for the performance of the model and the ecology and habitats of the species concerned are discussed.