Role of spatial variability, data availability and operative application in the production of lake evaporation maps. A combined remote sensing and numerical modeling study

We present an Earth Observation based methodology (EO-LSEv) for the retrieval of evaporation rates from water surfaces. The proposed model uses satellite derived surface water temperatures and meteorological data (i.e. wind speed, air temperature, relative humidity) for estimating the heat flux associated to evaporation at satellite overpasses, based on the bulk transfer theory (Dalton's law). A simplified energy balance equation is also computed exploiting satellite reflectances in the VIS-SWIR spectral range. Through the integration of the energy balance equation over 1hr time intervals, EO-LSEv can also compute daily evaporation rates from a lake. In this work we applied the EO-LSEv model to Lake Garda (Italy) and compared its outputs with the results of the hydro-thermodynamic model (Delft3D) forced by an atmospheric model (WRF). In order to achieve the best benefit from remote sensing techniques (e.g. repeatability and independency from in-situ measurements), we investigated the sensitivity of EO-LSEv to different meteorological data sources (i.e. in situ weather stations and freely available climatic models). The evaporation rates estimated from EO-LSEv are consistent with those modelled with Delft3D when the same weather forcing is used. The role of the spatial distribution of the weather forcing (especially wind) is investigated under an operational perspective.