The aim of this internal project of the Institute for Earth Observation, involving people from three groups (EOEM, ACEO, CDR) is to consolidate an Evapotranspiration (ET) algorithm for mountainous regions. ET is a key component of the water and energy budget at the Earth’s surface. Measurements of ET are scarce and do not characterize the spatial variability of energy and water fluxes. Consequently, at spatial scales from the landscape to the global, ET is modelled by different methods. One approach is the Two Source Energy Balance model (TSEB) that represents the surface where there is exchange of heat and moisture as including ground and vegetation. Latent heat is partitioned into soil evaporation and plant transpiration, and sensible heat is divided into that from the soil and that from the foliage. Each process is represented by different equations that can be solved iteratively for the unknown vegetation and soil temperatures, that are constrained by remote sensing observations of directional radiometric temperature (LST), based on the vegetation cover fraction. An open source version of the TSEB algorithm was developed within the ESA project sen-et, sentinels for evapotranspiration. However, there are issues in the algorithm, and a careful debugging is required. In addition, the proposed technique for sharpening moderate resolution (1 km) LST is not satisfactory for the inhomogeneous landscape of the Alps. In this project we will debug the code, test the algorithm, and implement a new sharpening layer based on machine learning.