SEBS validation in a Spanish rotating crop

• The SEBS algorithm was applied over non-irrigated agricultural cropland for 4 years.
• Components of the energy balance and EF were calculated using SEBS.
• Observed data and SEBS results were compared.
• The SEBS algorithm was improved by applying NDVI and Tsurf in EF calculations.

This paper focuses on calculating Evaporative Fraction (EF) and energy balance components, applying the Surface Energy Balance System (SEBS) model combined with remote sensing products and meteorological data over an agricultural rotating cropland from 2008 to 2011. The model is validated by comparing SEBS results with observed EF and surface fluxes obtained using an Eddy Covariance (EC) technique together with meteorological instrumentation. Three different approaches of the model are evaluated: SEBS-0 (original algorithm), SEBS–SM (algorithm modified with soil moisture), and SEBS–NDVI (SEBS-0 modified with the Normalized Difference Vegetation Index – NDVI – and the surface temperature – Tsurf). Based on current knowledge of the close relationship between EF and NDVI, a modified SEBS-0 algorithm, SEBS–NDVI, is proposed in this study. This new approach is developed so as to improve results for EF and latent heat flux (LE), and this paper presents the results of all three SEBS approaches used in this study. Modelled Rn is found to be in good agreement with observed data (R2 = 0.75), although SEBS-calculated G gave less satisfactory results (R2 = 0.38) and its seasonal dynamics shows discrepancies with observed data. An evaluation of SEBS-0 shows a clear underestimation of H (R2 = 0.54) and a marked overestimation of EF and LE. Comparison with ground-based data yielded the best correlation applying SEBS–NDVI, avoiding overestimation of EF and LE obtained with SEBS-0 and SEBS–SM. Results show that the proposed SEBS–NDVI, using a scale factor related to NDVI and Tsurf, is able to reproduce satisfactorily the EF (R2 = 0.65) and LE (R2 = 0.70) seasonal pattern better than the two previous approaches for our study plot.