Perspective of Sentinel-1 for Land Applications
Defourny, Pierre1; Bériaux, Emilie1; Gamba, Paolo2; Leonard, Aline1; Pathe, Carsten3; Santoro, Maurizio4; Schmullius, Chris3; Wegmüller, Urs4
1Earth and Life Institute - Université catholique de Louvain, BELGIUM; 2Department of Electronics - University of Pavia, ITALY; 3Earth Observation - Friedrich-Schiller-University Jena, GERMANY; 4Gamma Remote Sensing, SWITZERLAND
Since 20 years the all-weather observation capabilities of SAR instruments have been successfully illustrated for many potential land applications. In particular, the intensity of the backscattering signal in C-band was found related to various terrestrial variables, such as the water surface, fresh biomass, the soil moisture, the soil roughness, the forest type, etc. However the development of potentially operational monitoring system has been often hampered by scares time series and lack of continuity in SAR acquisition plan. Indeed, while their all-weather capabilities is a critical asset for monitoring activities, many SAR systems did not succeeded to deliver systematically the requested homogeneous time series on a regular basis due to conflicting configurations between users and competition between instruments. The Sentinel-1 mission is expected to change significantly the SAR time series availability at global scale. Furthermore, open and free access of such data will significantly boost the development of operational applications. This paper reviews various recent land applications of interest which will take much advantage of such data continuity and availability
The agriculture monitoring systems have among the highest requirements for continuous observations over large regions. In countries with high cloud occurrence, the previous controls for the Common Agriculture Policy had to rely on SAR acquisitions to insure timely crop discrimination over large areas. Similarly, very early acreage estimate of winter versus summer crops can only rely on C-band time series for a robust information delivery. For crop growth model adjustment, the fast growing period of the crop cycle requires a weekly effective observation which is hard to insure with optical data. This is particular true in food insecure countries where the cropping cycle corresponds to the rainy season. Biophysical variables, such as Leaf Area Index and biomass, are successfully estimated by inversion of radiative transfer model whenever dense homogeneous time series is available.
Tropical deforestation estimate derived by satellite remote sensing have been an important achievement in the recent years. However, large parts of countries like Cameroon, Gabon, Equatorial Guyana and Congo in Central Africa, were hardly covered by several wall-to-wall coverage over last decades due to the poor quality of optical data. In the context of the REDD+ initiative, the capabilities of regular deforestation mapping will be a key element for the Monitoring-Reporting-Verification process.
More recently, global LC_CCI inland water bodies' product was derived from the 5-year ENVISAT ASAR archive pre-processed on the ESA cloud computing facilities. The impressive accuracy of such a global product opens the way for an operational monitoring system based on Sentinel-1. However, it was found that only very dense time series allow discriminating automatically the inland water body’s extent. Along this line, delineation of urban areas from SAR time series was found sometimes more efficient than from optical data, in particular in some semi-arid contexts.
Some other applications are also illustrated and their prospects are discussed based on the contribution of Sentinel-1 mission potentially providing consistent and dense time series over large areas.