The evapotranspiration (ET) was spatialized using SEBAL algorithm for a region of primary forest in the eastern Amazon (Caxiuanã, Pará). To this end, we used observational data of micrometeorológica tower (located in this forest) in combination with orbital source data (Modis/Acqua images). In qualitative terms, the first results indicated that, despite the overestimation, the SEBAL plays well the pattern of variability monthly evapotranspiration for the region, mainly for the months of dry season; as to quantitative terms, the results revealed there’s need for accuracy in the algorithm. Thus, we calibrated the SEBAL from net radiation (Rn), with adjustments in the albedo, atmospheric and surface emissivity. Estimates of ET generated from this modified SEBAL presented significant improvements in reproduction of daily variability of evapotranspiration for the region, especially in the days of the rainy season. That is, the settings made in the algorithm showed that the rates of ET estimated became more similar to those reported in the literature for the Amazon, agreeing better with evapotranspiration observed. Using the modified SEBAL was can also map the albedo, net radiation, NDVI and ET for two distinct vegetation in Caxiuanã. Spatial estimation of biophysical parameters was consistently played for the two vegetation types, demonstrating that if the SEBAL modified is applied to temporal and spatial data of high resolution, this technique can be routinely used, becoming a fundamental tool in the monitoring of atmospheric and water needs.
|
