Definition of crop development stage to estimate evapotranspiration using FAO-56 approach and remote sensing.

Abstract

This study analyzed the temporal patterns of biophysical and spectral variables - such as leaf area index (LAI), aerial cover (fv) and vegetation index cinematically modified and adjusted for soil (IV_CIMAS) - in crops with medium leaf density (cotton and corn) and high leaf density (sorghum). In the case of sorghum, the relationship was analyzed between these temporal patterns and the baseline crop coefficient (Kcb). The variables fv, LAI and IV_CIMAS were modelled using truncated (ELT) and asymmetric (ELA) expolinear models. While both models showed a good statistical fit for the three crops, the ELT model was more suitable because it did not require the maximum value of the variable. Of the three variables modelled, IV_CIMAS better represented the quantity and quality of the foliage for a pixel or parcel, since it is a function of the leaf area, spatial foliar distribution and optical properties of leaves and background soil. In order to estimate crop evapotranspiration according to FAO-56, three different methods were analyzed, which characterize the developmental stages of the sorghum crop. The methods defined the length of this stage based on ground cover, the flowering stage and uses spectral information (IV_CIMAS). The analysis of errors (RMSE and ERM) for the Kcb estimates of the vegetation development stage was conducted using the three methods, in comparison to Kcb estimates obtained with a weighing lysimeter. The best results were found for the IV_CIMAS method and the worst corresponded to FAO-56-Flowering.