| Popis: |
We analyzed the applicability of forest planning data in predicting the occurrence of the Siberian flying squirrel (Pteromys volans) in managed northern boreal forests, in northeast Finland. Forest planning data is a source of information about forest characteristics for forest managers that may be used in estimating the availability of certain habitats for species conservation. Flying squirrel populations have declined in Finland, most probably due to habitat change and loss and maintenance of suitable habitats can be seen as a fundamental task in species conservation. First, we surveyed 715 ha of older spruce-dominated forest consisting of 91 stands, of which 35 were found occupied by flying squirrel. Flying squirrels inhabited larger stands, which had a higher volume of spruce and birch. Occupied stands also had more good quality forests surrounding them than the unoccupied stands. We based the model building on already existing knowledge of the habitat preferences of the species and built four alternative predictive models with logistic regression. Forest planning data seemed useful in estimating the forest quality and the suitability of habitats for the flying squirrel, with a model fit of ca. 72% with the original data. Second, we built the predictive models similarly with data from another study area (n = 98) situating ca. 150 km south from the first area. Third, we evaluated the first models using data of the second study area and also using new independent data from three municipalities situating almost in between the two study areas. Moreover, we reciprocally evaluated one model of the second study area using data of the first study area and of the three municipalities. The prediction success of our models indicated some applicability to other areas. The results also showed that the structure of the surrounding landscape is more important in a coarse-grained landscape than in a fine-grained landscape. However, because of some inaccuracies, predictive occupancy models built for the flying squirrel cannot replace field surveys and their generalizations to other areas must be made with caution. |
Full Text from ScienceDirect