Synthetic aperture radar remote sensing for landfill monitoring

Despite today’s intensive efforts directed at the recycling and recovery of solid wastes, the controlled disposal of refuse into land remains an important and necessary means of effective waste management. The work presented in this thesis investigates the use of Synthetic Aperture Radar (SAR) data to monitor solid waste landfills. The end-users’ interests vary from detecting the presence of a landfill to more specifically monitoring on-site operations and environmental conditions. Following a general literature review on the application of Earth Observation data for landfill monitoring, the identified research objectives are to: 1) assess whether SAR data can support the identification of landfill sites by distinguishing them from other disturbed areas which present similar optical spectral signatures, and 2) assess the possibility of correlating SAR data with onsite operational procedures. Data acquired for the research are: ground observations and measurements examining the spatial, temporal and biophysical characteristics of a landfill that can influence SAR data; historical and new programmed SAR scenes obtained from the ESA ERS-1 and -2 satellites and from Envisat ASAR instrument; ground based SAR (GB-SAR) acquisitions; simulations based on the RT2 backscatter model; additional space-based and airborne optical data to support the analysis and discussion. The examination of both the SAR amplitude spatial structure and the temporal decorrelation of these sites shows that there are three key characteristics that can distinguish them from other disturbed areas with similar optical spectral signatures: the presence of anisotropic features that strongly affect the SAR backscatter; the fact that the coherence magnitude images of these sites are characterised by large decorrelated areas with transient attributes; and their distinctive positive topography. The analysis highlights that one single-polarisation acquisition can hardly provide correct land-cover information, and consequently knowledge on land-use. The research demonstrates the key value of merging together complementary information derived from both the space and time dimensions, achieving fairly accurate land-use classification results. The research also provides an appreciation of the applicability of the developed techniques in an operational framework. These can suffer a number of limitations if a landfill site is located in a particular environment, and/or if meteorological conditions can significantly affect the radar signal, and/or unusual landfilling procedures are applied by the operators. Concluding remarks on the end-users needs point out that there are a number of aspects, ranging from practical and managerial matters to legal and technical issues, that often discourage the utilisation of EO data by new potential users.