Comparison of Noah simulations with eddy covariance and soil water measurements at a winter wheat stand

Weather and climate simulations are critically dependent on an accurate representation of land surface exchange processes. The present study tests the performance of the Noah land surface model (LSM) for energy partitioning and soil water dynamics on a winter wheat stand in southwest Germany. The model was parameterized field-specifically, i.e., data on leaf area index, green vegetation fraction, albedo, and soil texture were derived from measurements, and tested against a set of eddy covariance (EC) and soil water data collected in the 2009 season. With respect to energy partitioning, the field-specific parameterization performed fairly well during the main vegetation period. During ripening, starting in mid-July until harvest on 6 August, however, the sensible heat flux was distinctly underestimated. In August, the bias increased to −105.5Wm⁻². As a consequence, evapotranspiration was overestimated during this period. After introducing a time-variable minimum stomatal resistance (Rc,ₘᵢₙ), the model performed much better. The model also delivered acceptable results during crop maturing. The Noah LSM was unable to simulate the observed soil water dynamics. While the measured soil water content profiles showed a distinct gradient with depth, the Noah LSM tended to deplete the soil profile uniformly. Our study shows that for cereal-dominated croplands phenology and crop development play a crucial role in energy partitioning, especially during the ripening stage. Disregarding the dynamics of the physiological properties of the crops in the Noah LSM leads to significant bias in energy partitioning. Finally, we critically discuss the current approach to correct measured EC flux data for the energy residual based on the Bowen ratio with regard to LSM calibration and performance.