Disturbance patterns in Sacco (Latium, Italy) river basin: relationships among urban and industrial areas and ecological changes

The landscape patterns are the result of biotic and abiotic interactions. Particularly, human activity can modify landscape patterns and processes, changing structural integrity of landscapes and modifying ecological flows. In this study structural changes in Sacco river basin were identified by the elaboration of remote sensing data (Landsat TM 5; July 1984-July 2009), the obtained classifications of land use were implemented in fragstats 3.3 software to elaborate opportunely chosen landscape metrics. Metrics were derived for class level (characteristics of all patches in a given class) and landscape level (integrated over all patch types or classes over the extent of the data). To detect landscape metric changes at local scale was used the moving window technique. The image differencing technique applied to NDVI (1984-2009) and derived from pixel by pixel value substraction revealed the vegetation and land cover dynamics and the regime of disturbance. The rates of changes for the main land use classes were also determined. The integration of both remote sensing and GIS methods allowed to assess the spatial relationships (Euclidean distance) among disturbance, anthropized areas, altitude and main river bank. A regular grid of 5 km cells was overlaid to NDVI image difference and the density of changes were calculated in each cell (distinctly for losses, stable areas and increase of NDVI). Density values of each cell were used to evaluate the data autocorrelation using variography technique. Anthropized and industrialized areas seem to be strongly spatial related to NDVI decrease on distance basis as like altitude. Variography highlights high autocorrelation and crosscorrelation for NDVI gain and stable area while no autocorrelation was found for NDVI losses. Moran's correlogram for losses data seems to indicate a randomly distributed disturbance regime (losses follow the chaotic distribution of urbanized areas). Stable and increased areas density values resulted high auto and cross correlated, probably because of stability driving phenomena that enhanced restoration process.