Estimation of latent heat flux of pasture and maize in arid region of Northwest China based on canopy resistance modeling

Abstract

Estimating the latent heat flux (λET) accurately is important for water-saving irrigation in arid regions of Northwest China. The Penman-Monteith model is a commonly used method for estimating λET, but the parameterization of canopy resistance in the model has been a difficulty in research. In this study, continuous observation of λET during the growing period of maize and grassland in Northwest China was conducted based on the Bowen ratio energy balance (BREB) method and the Eddy covariance system (ECS). Two methods, Katerji-Perrier (K-P) and Garcıá-Santos (G-A), were used to determine the canopy resistance in the Penman-Monteith model and the estimation errors and causes of the two sub-models were explored. The results indicated that both models underestimated the λET of grassland and maize. The K-P model performed relatively well (R2 > 0.94), with the root mean square errors (RMSE) equaled 37.3 and 28.1 W/m2 for grass and maize, respectively. The accuracy of the G-A model was slightly lower than that of the K-P model, with the determination coefficient (R2) equaled 0.90 and 0.92, and the RMSE equaled 46.2 W/m2 (grass) and 42.1 W/m2 (maize). The vapor pressure deficit (VPD) was the main factor affecting the accuracy of K-P and G-A sub-models. The error of two models increased with the increasing in VPD for both crops.