Validation and scale dependencies of the triangle method for the evaporative fraction estimation over heterogeneous areas

Remote sensing has proved to be a consistent tool for monitoring water fluxes at regional scales. The triangle method, in particular, estimates the evaporative fraction (EF), defined as the ratio of latent heat flux (LE) to available energy, based on the relationship between satellite observations of land surface temperature and a vegetation index. Among other methodologies, this approach has been commonly used as an approximation to estimate LE, mainly over large semi-arid areas with uniform landscape features. In this study, an interpretation of the triangular space has been applied over a heterogeneous area in central Spain, using Landsat5-TM, Envisat-AATSR/MERIS and MSG-SEVIRI images. Some aspects affecting the model performance such as spatial resolution, terrain conditions, vegetation index applied and method for deriving the triangle edges have been assessed. The derived EF estimations have been validated against ground measurements obtained with scintillometer on a winter crop field during 2010–2011. When working with large spatial windows, removing areas with different topographic characteristics (altitude and slope) improved the performance of the methods. In addition, replacing the typically used NDVI with Leaf Area Index enhances the performance of the triangle method allowing a better characterization of the wet edge. Finally, results showed a relatively good performance for the EF estimates, with an RMSE of 0.11, 0.15 and 0.23 and R2 of 0.77, 0.41, and 0.24 for Landsat, Envisat and MSG satellites respectively, showing a scale dependency on the accuracy.

This research was initialized during a visiting stay at the Department of Geosciences and Natural Resource Management of the University of Copenhagen, Denmark, thanks to a grant within the BACH project funded by IMDEA Water Institute. This study is likewise part of the INTEGRATOR Project (CTM2011-27657) funded by the Spanish Ministry of Science and Innovation.

Peer reviewed