Summary:
Day-1 covered the use of Land Surface Temperature, the importance of Surface types and surface water cover in the morning
The afternoon has covered different LDAS systems, a common challenge seems to have enough spread near the surface.
Recommendations:
On LST and IST:
On snow products/DA
-Boreal snow
-Orographic snow
-Forest snow (snow cover products in forested areas do not have much fractional resolution?)
-Understory snow (need innovative observations, e.g. cameras/drones)
-Satellite snow products (IMS, Hydrology- SAF, Copernicus Global Land, CCI-Snow, GlobSnow)
-SnowPEX compared different products
On LDAS methodology:
-"Flow dependence" or "Errors of the day" are a recipe for improving snow, soil moisture and other land/near surface variables
-Moving away from OI towards Ensembles and EnVAR
-On T2m/RH2m assimilation aliasing atmospheric errors
-Direct Radiance Data Assimilation complex but may bring benefits (better representation)
-TSMM Terrestrial Snow Mass Mission
On LSM shortcoming:
-Diurnal cycles
-Verify Day and Night separately
-Regional bias (Boreal Forest is an issue)
-Snow Fraction (need to represent hysteresis)