IMPROVING PAIRWISE DSM WITH 3SGM: A SEMANTIC SEGMENTATION FOR SGM USING AN AUTOMATICALLY REFINED NEURAL NETWORK

The EVapotranspiration Assessment from SPAce (EVASPA) algorithm that will generate TRISHNA evapotranspiration (TRISHNA ET) products shows good agreement with observations. Analyses conducted on 4 variables used in contextual ET modelling showed that land surface temperature and evaporative fraction methods introduce the largest uncertainty in ensemble estimates, followed by radiation. Soil heat flux methods contribute the least. Robust ET products will require optimizing the sensitive variables.