Abstract
Evapotranspiration (ET) driven by thermal infrared remote sensing provides spatiotemporally distributed information on water availability at large scale. However, cloudy-sky conditions result in gaps in satellite-based ET, especially in complex landscapes such as the Alps. In this regard, we investigate different reconstruction approaches to fill missing ET at six alpine flux sites in Italy and Switzerland. The gapfilling is performed by three main steps: establishing a relationship between clear-sky actual ET and different estimates of reference quantities, interpolation of these dependencies to overcast conditions, and predicting ET by applying interpolated factors to reference quantities. The performance of the reconstruction strongly depends on landcover, achieving the best results over grasslands (RMSE = 0.38 mm day-1) and less performant outcomes over forest (RMSE = 0.70 mm day-1) when reference quantities from Penman-Monteith and Priestley-Taylor formulations are applied. The findings of this work give next-future perspectives for all-sky ET retrieval using remote sensing.