Forest inventories are a valuable data source for habitat modelling of forest species: an alternative to remote-sensing data

Habitat and species-distribution models are important tools in ecology, conservation and wildlife management. For the modelling of the species habitat on small spatial scales, the availability of suitable habitat variables is crucial. Collecting data in the field often incurs considerable costs. Forest agencies, however, collect data on forest structure and composition on a large extent and with a fine grain (200 m 200 m) during their forest inventories. Despite this the potential use of these data for habitat modelling and therefore conservation and wildlife management is still broadly neglected. Here, we evaluated and compared the performance of models based on environmental data derived purely from forest inventories, inventory data plus additional ecological information gained during the inventory, and aerial photographs to model the habitat of a forest species, the capercaillie (Tetrao urogallus L.). The best models developed from each of the three data sources showed a similar performance over the area (28 450 ha) and the considered grain (50 ha grids); all models had an area under the receiver operating characteristic curve (AUC) value 0.85, and various other metrics used for evaluating the models showed little differences. In all calibration plots, the predicted probabilities of presence were very similar to that of the observed probabilities, which indicated that our maps of the habitat suitability are a good representation of the real occurrences. This modelling of the capercaillies habitat using existing forest inventory data was successful because the variables collected in standard inventories also have some meaning for the ecology of the species. To model the habitat of forest species that need special resources outside the scope of forestry, possible additional variables could be sampled at low extra costs during such large inventories. There is therefore the potential for forest inventories, currently undertaken to support commercial forestry, to be applied in conservation, thereby supporting a multifunctional forestry.