Desarrollo de un modelo de evapotranspiración global con datos de satélite y de re-análisis
Evapotranspiration quantification is fundamental to hydrologic process study, climatic and agronomic characterization, as well as, to apply in climate change models, and the management of water resources. It is important to point out the knowledge of this parameter in the control of floods and droughts, in the planning of infrastructure works, among others. The factors that intervene in the ET process are diverses and variables in time and space. Therefore, satellite data gain relevance since they can be used to obtain a spatio-temporal variation of the hydrological cycle components, at different scales and resolutions with high precision. In particular, the MOD16A2 product (based on Penman-Monteith), has shown to have problems for its direct use in different study areas, according to several authors; therefore, one of the objectives of this Thesis is to analyze and adjust the MOD16A2 product in different areas of the planet. The next objective is to develop a global product of potential ET (ETp) and actual ET (ETa) with satellite data, with 0.25 km of spatial resolution and 8 days of temporal resolution. Using Priestley and Taylor equation, and as input data: CERES (Clouds and the Earth's Radiant Energy System), MODIS (Moderate Resolution Imaging Spectroradiometer) and the soil moisture product GLEAM (Global Land Evaporation Amsterdam Model). Combining the different products, ETp and ETa values were obtained and validated with local data from eight stations distributed in different environments (from arid to humid): 1.Torgnon, Italy; 2.Mead, USA; 3. Demokeya, Sudan; 4.Santarem, Brazil; 5.Fogg Dam, Australia; 6.Skukuza, South Africa; 7.Changling, China (FLUXNET); and 8.Tandil, Argentina who belongs to the Oficina de Riesgo Agropecuario which determines ET with the soil water balance method. The results of the method validation show that the associated error for ETp varies between 0.4 and 1.8 mm d-1 for warm and humid areas. For arid areas, good results are not obtained due to the low sensitivity of the method in aerodynamic conditions prevailing in these environments. As for ETa, the method has larger errors in equatorial climates and dry winters than in warm and humid climates. It is concluded that this model has worse performance in areas with arid or dry climates, while better results are observed in humid and temperate climates. In relation to the analysis of the MOD16A2 product, a systematic error was found for ETp, and a correction was proposed to be able to use it in the Pampean region of Argentina. For ETa, the results indicate that its performance is better in arid areas of the region. Comparing both products in different areas of the planet, ETa error is between 1.5 to 1.6 mm d-1. While for ETp the error is bigger for MOD16A2 product (2 mm d-1) than for the model proposed in this Thesis (1.4 mm d-1).