GISAT S.R.O. (CZ)
The dense spatial and temporal coverage of the Amazon basin with Sentinel-1 Synthetic Aperture Radar scenes has opened vast potential for capturing the complexity in tropical forest loss and regrowth. Sentinel-1 for Science Amazonas aims to:
The methods developed in Sentinel-1 for Science Amazonas integrate the current state-of-the-art research on SAR-based monitoring in the fields of not only forests, but also agriculture, wetlands and grasslands. The primary dataset used in the development of the Multi-temporal forest Change Detection (MCD) algorithm is the complex Sentinel-1 Interferometric Wide (IW) swath Single Look Complex (SLC) time series. These time-series are analysed by break-point detection, moving-window trend-fitting in the temporal and spatial domain.The algorithms are enriched with complementary data, such as L-band SAR (e.g. ALOS-2) or spaceborne LiDAR (e.g. GEDI). Forest carbon losses and gains are estimated through a rigorous statistical comparative analysis of existing datasets to the newest in-situ field data and LiDAR data, or local-scale calibration and fusion of such datasets.
The Sentinel-1 for Science Amazonas project is expected to publicly release forest loss and gain maps produced using the Multi-temporal forest Change Detection (MCD) algorithm. The maps will incorporate spatially and temporally explicit estimates of change, the type of change (e.g. deforestation, degradation, areas of natural or assisted regrowth), severity of change, forest carbon gains/losses and the associated uncertainty in estimates.The study areas of the project include pilot sites in Madre de Dios, Mato Grosso and Manaus, and will be up-scaled to the extent of the Amazon basin.