C10 – Continuous mapping of land surface categories

Current wall-to-wall global land cover mapping in the CLMS is restricted to annual updates of discrete land cover and cover fraction layers at 100m resolution. To increase the thematic flexibility and timeliness of the service this candidate envisages to develop continuous mapping (daily to 10-daily) of land surface characteristics (LSC; e.g. snow/no-snow; water/no-water; vegetation/no-vegetation) at 10m resolution based on Sentinel-1/2 data. To do so we will ingest and demonstrate the new LSC methodology output from T2.6 together with the AI-tailored training data (T2.2) as well as the generic embeddings (T2.4). Additionally, we will investigate the feasibility of further improvements of the future CLC+ Backbone raster updates, already pointing to an upcoming yearly update via the LSC continuous monitoring.