Uncertainty with the scaling-up of remotely sensed evapotranspiration estimation

In this paper, the uncertainty issues in evapotranspiration estimation using remotely sensed data are discussed. The processes governing the mass, energy, and momentum exchange across the land-atmosphere interface are nonlinear, because of the interdependence of the dominant variables and parameters. Spatial resolution of remote sensing data also has an impact on the spatial pattern of evapotranspiration captured, due to the spatial variation of the surface. In order to study the nonlinear and spatial variation problems, evapotranspiration was first estimated at 30 meter and then aggregated to coarser scales using two approaches. One is to serve as true value of evapotranspiration, while the other one is assumed the linear resulted value. Results show that factors do have a nonlinear impact on evapotranspiration, however, this impact becomes very close to linear at 120m. The impact also varies with scale, as well as on different land cover types.