Assessments of Weather Research and Forecasting Land Surface Models in Precipitation Simulation Over the Tibetan Plateau.

Precipitation is a key hydrometeorological variable for understanding surface energy partitioning and water budget over the Tibetan Plateau (TP). A substantial proportion of summer precipitation falls as rain. The effects of different land surface models (LSMs) on the TP's precipitation and their inner mechanisms remain unclear. Therefore, the assessments of different LSMs coupled with the Weather Research and Forecasting model in precipitation simulation were investigated over the TP during June 28–29 of 2008. The simulated results were evaluated with the merged Climate Prediction Center (CPC) MORPHing technique (CMORPH) precipitation data set developed by the China Meteorological Administration. The assessment demonstrated that precipitation simulated by the Community Land Model Version 4 (CLM4), Noah‐Multiparameterization (Noah‐MP), and Pleim‐Xiu LSM schemes was wider and stronger compared with the merged CMORPH over the central and western TP but was underestimated over the eastern and southern regions. Generally, both CLM4 and Noah‐MP schemes exhibited higher forecasting quality and accuracy in simulating precipitation over the TP. The optimal precipitation simulation was achieved by applying the Noah‐MP scheme, with a lowest root mean square error of 9.53 mm/day, mainly attributed to its corrections of overforecasting for precipitation that did not occur. Further mechanism analysis indicated that soil moisture‐energy flux‐precipitation feedback play an important role in different LSM schemes. Key Points: The Weather Research and Forecasting model with different land surface parameterization schemes was used to simulate a precipitation event over the Tibetan PlateauBoth the Community Land Model Version 4 and Noah‐Multiparameterization (Noah‐MP) schemes showed higher forecasting quality and accuracy, whereas the best performance was achieved by Noah‐MPA soil moisture‐energy flux‐precipitation feedback mechanism was identified to explain the discrepancies among different model performances [ABSTRACT FROM AUTHOR]