Evapotranspiration shed of key agricultural crops: combining agro-hydrological estimates with atmospheric moisture dynamics.

The increasing global demand for food, feed and flexible crops is exerting unprecedented pressure on the global hydrological cycle through landscape conversion and increasing irrigation demand, which altogether contribute to the alteration of land-atmosphere feedbacks. These feedbacks influence evaporation and precipitation patters through atmospheric flows. Atmospheric moisture flows connect sources of evaporation to sinks of precipitation, from local to regional and continental scale, up to thousands of kilometres away. Terrestrial sources of evaporation are crucial for global food production, regulating precipitation and climate patterns by redistributing water and latent heat. At the same time, the alteration of evapotranspiration dynamics from these sources is mainly driven by land-use conversion for pasture (cattle meat production), and feed crops (such as soy, and maize) and agricultural practises, such as irrigation. Current water use assessment disregard these feedbacks and the role played by atmospheric moisture connection in redistributing evaporation from agricultural parcels to precipitation in downwind areas. This understanding is particularly key to better assess the water-related implication of pivotal crops such as soy, maize and wheat which account for 33% of global harvested land. Addressing this gap, this thesis aims to advance the understanding of how evapotranspiration from agricultural areas contributes to precipitation whether or not to other agricultural area. It emblematically presents the cases of soy, maize and wheat. The first part of this thesis updates actual evapotranspiration estimates for soy, maize and wheat production for the period 2008–2017 by means of the agro-hydrological model waterCROP, which solves the daily soil water balance on a global 5 arc-minute grid, with global coverage for both irrigated and rainfed conditions. In the present work, the model is updated to a newer version, made consistent with daily climatic data from ERA5 reanalysis. In the second part, the evapotranspiration estimates are combined with atmospheric connections by means of the RECON dataset, a 4D matrix of annual moisture flow connections between any cell in the world at the spatial resolution of 0.5°. In the present work, each cultivated cell of soy, maize and wheat is linked to its blue and green evapotranspiration shed (i.e. the downwind area receiving precipitation from irrigated or rainfed crop production). Evaporation sheds are finally classified according to their land use category to analyse potential synergies and trade-off between land and water use between the sites at the origin of evaporation and at the fate of precipitation. By characterizing these connections, the thesis sheds light on the hidden global links between cultivated land and downwind areas. Ultimately, this thesis contributes toward a more comprehensive evaluation of the interplay between water and land use at the site of production with atmospheric feedbacks with local and distant link in the global water cycle.