Using predictive models as a spatially explicit support tool for managing cultural landscapes

Due to the high sensitivity of mountain landscapes to environmental changes, the study of land cover dynamics has become an essential tool for guiding management policies. Since the second half of the twentieth century, the Cantabrian Mountains (NW Spain) have been substantially altered by the loss of traditional management practices and, more recently, by the new environmental schemes developed by the Regional Government. This area is a biodiversity hotspot, representing the south-western-most distribution limit for a large number of species in Europe. Therefore, small changes in landscape patterns can result in biodiversity losses. In this study, we analyzed land cover changes in the Cantabrian Mountains from 1991 to 2004 by means of remote sensing techniques, identifying the main driving forces and classifying the territory according to its risk of land cover change. Forest expansion and loss of shrublands were the two major trajectories of change apparent during this period. When modeling the occurrence of these land cover changes, we found that performance of models was related to the nature of the change. The most accurate models were associated with processes of secondary succession, i.e. forest expansion (78.6%), while the least accurate models related to changes linked with management decisions, i.e. loss of shrubs (61.8%). The main drivers of change were variations in the number of goats (for the forest expansion model) and changes in the number of head of sheep and cattle (for the loss of shrubs model). Topographic conditions (altitude and slope) were relevant in both models. Our approach proposes an explicit decision support tool for landscape managers, allowing better identification of the areas where they should focus their attention.

Research highlights► LANDSAT-based classification and modeling identify land cover changes and drivers. ► Forest expansion and loss of shrublands are the major trajectories of change. ► Forest expansion is mostly linked to land abandonment and livestock changes. ► Loss of shrub cover is mostly linked to management decisions (shrub clearance). ► Predictive models can be used as a spatially explicit support tool for land managers.