ASSESSMENT OF ECOLOGICAL CONDITIONS OF PERMANENT MEADOWS IN THE ITALIAN ALPS: LOSS, BIODIVERSITY AND REMOTE SENSING CHANGE DETECTION

The monitoring of ecological condition of grasslands ecosystems in the European Alps is a main issue for mountain regions, since the abandonment of traditional and sustainable management practices has exposed grassland habitat to significant impacts in a context of global environmental change. The present research project was focused in assessment of the state of permanent meadows in the lowlands of Valtellina Valley (80 km2), Italian Alps, during the timeframe 1980-2000. In specific, it quantified the land use/land cover changes and identified main drivers behind permanent meadows loss; characterized the relationship between biodiversity in the meadows and the spatial-environmental conditions in the landscape and by last evaluated the use of satellite remote sensing data for fast change detection in landscape. To achieve such aims, the research project was organized in three different approaches presented in the four chapters of this thesis. Concerning the quantification of the land use/land cover and identification of main drivers behind permanent meadows loss, the results show a strong decrease in meadows (-18.5%) in a context of agricultural land decrease and human settlements increase. This was the land cover type with highest loss and conversion rate during the study period. Meadows were converted to human settlements (urban, industrial and roads), other agriculture uses (cultivation, orchard, vineyard), bushland and uncultivated land. Meadows loss occurred mainly in soils with good land capability, low slope, exposed to south and in proximity of roads, urban settlements and bushland. Densities of urban, industrial and bushland and land capability were the only significant drivers for meadows loss, while distance to meadow edge, meadows density, distance to roads and soil degradation were the only significant drivers for meadows preservation. Concerning the characterization of the relationship between biodiversity in the meadows and the spatial-environmental conditions in the landscape, the results evidenced that species richness and Shannon indices were best explained by regressive models including changes occurred in spatial environmental heterogeneity from 1980 to 2000. Species richness was negatively related to strong decrease in meadows habitat area and recent urban area, while Shannon index was positively related to the increase in landscape diversity. In contrast, species evenness was better explained by regressive model including recent spatial environmental heterogeneity and positively related to increase ineastness in the study area, and negatively affected both by the area of woody and soil pH (KCl). Concerning the evaluation of the use of satellite remotely sensing data for land cover mapping and change detection in landscape, the results show that the hybrid approach for land cover classification based of Landsat imagery was highly accurate. Image differencing is the technique which best detect changes in landscape as well as in urban, meadow and bush land. The accuracy of change detection was moderate. This thesis concludes that the conflict by land in locations densely occupied by other land cover types with good land capability is the major threat to meadows and avoidance of fragmentation may be a good strategy for its preservation. The meadows habitat needs a well-designed landscape and farming planning, which should account the economic value of the ecosystem services provided by this habitat. In addition, to conserve plant diversity in meadows it is necessary to avoid loss of meadows habitat, maintain landscape diversity and execute a sustainable meadow management. Remotely sensed imagery can be a reliable source of information for alps, although particular attention should be made to the image pre-processing and classification, as well as, to minimize topography effects in spectral information.