Integration of vegetation indices into a water balance model to estimate evapotranspiration of wheat and corn

Abstract

Vegetation indices (VIs) have been traditionally used for quantitative monitoring of vegetation. Remotely sensed radiometric measurements of visible and infrared solar energy, which is reflected or emitted by plant canopies, can be used to obtain rapid, non-destructive estimates of certain canopy attributes and parameters. One parameter of special interest for water management applications, is the crop coefficient employed by the FAO-56 model to derive actual crop evapotranspiration (ET). The aim of this study was to evaluate a methodology that combines the basal crop coefficient derived from VIs with a daily soil water balance in the root zone to estimate daily evapotranspiration rates for corn and wheat crops at field scale. The ability of the model to trace water stress in these crops was also assessed. Vegetation indices were first retrieved from field hand-held radiometer measurements and then from Landsat 5 and 7 satellite images. The results of the model were validated using two independent measurement systems for ET and regular soil moisture monitoring, in order to evaluate the behavior of the soil and atmosphere components of the model. ET estimates were compared with latent heat flux measured by an eddy covariance system and with weighing lysimeter measurements. Average overestimates of daily ET of 8 and 11% were obtained for corn and wheat, respectively, with good agreement between the estimated and measured root-zone water deficit for both crops when field radiometry was employed. When the satellite sensor data replaced the field radiometry data the overestimation figures slightly changed to 9 and 6% for the same two crops. The model was also used to monitor the water stress during the 2009 growing season, detecting several periods of water stress in both crops. Some of these stresses occurred during stages like grain filling, when the water stress is know to have a negative effect on yield. This fact could explain the lower yield reached compared to local yield statistics for wheat and corn. The results showed that the model can be used to calculate the water requirements of these crops in irrigated areas and that its ability to monitor water stress deserves further research.