Global Mapping and quantification oflong-term annual forest loss is mainly assessed through 30-m resolution opticalimagery (Vancutsem et al., 2021; Hansen et al., 2013), providing large-scaleconsistent information, but does not allow neither short-period monitoring incloudy condition nor small-scale deforestation assessment from householdlogging and farming.  Recent earlywarning S1-based models (Reiche et al., 2020, Hoekman et al., 2021) tend toreduce this temporal and spatial window but create alarm system rather thanmapping specific soil use information. This seminar presents the early steps ofa sub-annual full C-band SAR model using Sentinel-1 data, combining both assetsof short-term notice and soil use delimitation and quantification.