Energy-balance Based Model Accurately Reproduces the Spatial Distribution of Evapotranspiration Over the Andean Páramo
Monday, 6 June 2016
Galo Carrillo-Rojas1,2, Brenner Silva1, Mario Córdova2, Rolando Célleri2 and Jörg Bendix1, (1)Philipps-Universität Marburg, Laboratory for Climatology and Remote Sensing, Faculty of Geography, Deutschhausstr. 12, 35032, Marburg, Germany, (2)Universidad de Cuenca, Departamento de Recursos Hídricos y Ciencias Ambientales, Facultad de Ciencias Químicas & Facultad de Ciencias Agropecuarias, Av. 12 de abril s/n, EC010207, Cuenca, Ecuador
Abstract:
Understanding of actual evapotranspiration (ET) processes over mountain environments is crucial, particularly due to the importance of these regions in delivering water-related ecosystem services. In this way, ET can be considered as a functional indicator in terms of ecohydrologic processes, which have been poorly studied in the Andes highlands. Our study assessed the use of satellite-based imagery over an area of neotropical alpine grasslands of the Andes (Páramo) for ET mapping. The study site was a catchment in Southern Ecuador, with an altitudinal gradient ranging 2800 to 3900 masl. We implemented the energy-balance based model METRIC (Mapping EvapoTranspiration at high Resolution with Internalized Calibration) with cloud-free sets of Landsat 7 ETM+ and MODIS-Terra images (2013 ~ 2014). Besides the model implementation, we included suitable algorithms that account for the topography of the Andean valley. Additionally, a comparison with MOD16, which is a global coarse-resolution ET product in monthly scale, was performed. These spatial ET estimates were validated with ET calculated from water balance. Results of METRIC ET (Landsat-based) showed good agreement with the outputs from water balance ET, with monthly and annual differences of 30% and 10% respectively. However, the analysis of retrievals from METRIC ET (MODIS-based) and MOD16 revealed a poor performance due its coarse resolution with a lack of spatial fidelity. Findings support the plausibility of the METRIC model in the Páramo when Landsat product is used in combination with adequate topography corrections. Our finding is essential for the spatial estimation of similar functional indicators, like water use efficiency (WUE). The latter will be obtained in upcoming research on atmospheric CO2 fluxes, using the eddy covariance technique for the first time over a Páramo ecosystem.
