Estimation of Evapotranspiration with a Dynamic Vegetation Approach Using the Noah Multiphysics Model and the CERES-Rice Model

Actual evapotranspiration (AET) over a cropland is the outcome of a complex interaction between soil, crop, and atmosphere. Land surface models and crop models have been used to estimate AET over a crop canopy. Crop models often depends on a simplified approach to compute evapotranspiration using daily data whereas land surface models simulates the behavior of stomata using sub-hourly data. Those differences in algorithm for calculation of AET would result in considerable discrepancy between those models. The objective of this study was to examine differences in the estimates of AET between the Noah Multi Physics (MP) model and the CERES-Rice model, which have been used for simulation of land surface processes and crop growth, respectively. The observations of AET obtained from 2003 to 2012 in Korea were compared with those estimates from the Noah MP model and the CERES-Rice model. The Noah MP model explained considerably large amount of variation in ET (66 %) than the CERES-Rice model (35 %) after the leaf area index was adjusted to represent actual plant density. The CERES-Rice model tended to overestimate ET, which is most likely due to the fact that the CERES-Rice model was dependent on Priestley-Taylor equation. These results indicated that an improvement of the CERES-Rice model would be needed for reliable assessment of the irrigation requirements for paddy-field rice, which would help water management in a region.