Multitask Learning with Statistical Parametrization for Ecohydrological Analysis

Abstract This paper proposes a novel multi-task statistical learning framework which aims to concurrently address all the environmental challenges in the Alps. The goal is to analyse the effects of lichen and fog on water balance. The objective is the analysis of water balance mechanisms by investigating the contribution of fog and the role of forest age in the water cycle. The methods include advanced multitask learning with statistical modelling techniques. The results shown that interception plays a dominant role in the precipitation and evapotranspiration partitioning, enhanced by lichens. Trees transpiration as lower in the young stand and the evapotranspiration of soil and understory contributed considerably to the water balance at both stands. Moreover, fog caused additional throughfall in mixed fog and rain precipitation.