Radar imagery for environmental geology study of coastal zones in southern Italy

Summary form only given, as follows. The high spatial resolution of fine beam Radarsat imagery makes it suitable for detailed analysis of coastal zones. For this study three SAR images were acquired in different seasons and a stereoscopic pair was acquired with different incidence angles. The images were visually and automatically interpreted to derive the geological, geomorphologic and hydrogeologic setting of the Ionic coastal zone of Salento Peninsula (southern Italy). Mesozoic age limestone underlying calcarenite, sand and clay of Cenozoic-Quaternary age characterize the geological setting of the study area. The structural-geological setting largely determines the morphology of coastal areas, the erosion-accumulation processes and the groundwater flow patterns. Several coastal environment types can be recognized in this area with different geomorphologic setting and dynamic evolution trend, such as cliffs, sandy beaches, rocky low coasts, gravel bays, coastal dunes, coastal lakes and wetlands. The sea level changes in recent times have produced a general geomorphologic instability in the coastal zones, confirmed by active erosion and accumulation processes. A texture classification of the radar images was carried out in order to produce a land use map of the study area. This map provides useful information about underlying geology and geomorphologic setting, besides giving a comprehensive view of the anthropic activities that can interfere with natural processes. The radar stereoscopic imagery acquired in the F2 and F4 far beam mode was visually interpreted using a mirror stereoscope. The 3D view considerably improves the reconnaissance of geological structures and landforms characterizing the study area. The stereo pair was also interpreted to analyze the geomorphology of the coastal line, and local erosion and accumulation processes. The results highlighted the fact that the coast is mainly affected by erosion, with local high regression rate. These destructive processes affect both high cliff coast and sand beaches with coastal dunes. Besides their stereoscopic vision, these images were used for radargrammetric measurement of topographic surfaces. The DEM was calculated by means of a fully digital radargrammetric process based on an automatic pixel matching on both images for parallax measurement. Several DEMs were produced using the original and filtered images in order to evaluate the influence of speckle on the results. The accuracy of the radargrammetric DEMs was assessed by comparing them to a DEM interpolated from the official contour map at scale 1:50000. The difference between the elevation measured by radargrammetry and the topomap data was calculated on each pixel to produce error maps. The mean values of the error (generally below 2 m) are satisfactory, while their standard deviation (about 15-20 m) is slightly high considering the flat morphology of the study area. In this study, fine resolution radar images were interpreted using different approaches. The results obtained testify that these imagery- single, stereo pairs and multitemporal-provide a large amount of information on several environmental variables that are essential for producing a global model of coastal zones.