Contribution of optical and SAR imagery in the identification and distribution of land cover features in the lower delta of Parana's river, Argentina

Using remote sensing, natural occurring vegetation data can be gathered and processed for planning purposes and environmental impact assessments. In contrast to agricultural crops, natural vegetation communities are present in the form of heterogeneous mosaics with fuzzy limits. This is observed in the delta of Parana's river, a major fresh water wetland, located very close to Buenos Aires city, characterized by a hydrologic regime determined mainly by the lunar and wind tides of De La Plata River waters, and the Parana and Uruguay river regimes. Natural vegetation communities coexist with forest plantations, and recreational areas. This paper presents the procedures used for the study of this region using satellite remote sensing data (optical and radar). With respect to optical data, a conceptual classification scheme based on the flooding regime, human intervention and the structural and fluoristics characteristics of vegetation was developed. Next a sequential masking unsupervised classification procedure was applied to a multitemporal data set to be able to link "information classes" with "spectral classes" to obtain a land cover map. With respect to radar data, the conceptual classification scheme was slightly modified to underline the structural characteristics of vegetation. The overall objective was to establish the interaction mechanisms between the different landscape elements of the delta and the SAR system as a function of sensor parameters and scene characteristics. Backscatter profiles for different angles of incidence and for different times were obtained, and discussed.