Assessment and improvement of Noah-MP for simulating water and heat exchange over alpine grassland in growing season.

Alpine grassland is the main ecosystem of the Tibetan Plateau (TP), thus accurate simulation of water and heat exchange in the grassland will significantly enhance the understanding of the land-atmosphere interaction process on the TP. In this study, we assessed and improved the ensemble numerical simulations of the community Noah land surface model with multiparameterization options (Noah-MP) by using observations collected from four alpine grassland observation sites. The four observation sites belong to the upper Heihe River Basin Integrated Observatory Network located in the northeastern part of the TP. First, an ensemble of 1008 numerical simulation experiments, based on multiparameterization options of seven physical processes/variables in the Noah-MP, was carried out for the vegetation growing season. The Taylor skill score was then used to assess the model performance and select the optimal combination of parameterization options for a more exact simulation of the water and heat exchange in alpine grassland. The accuracy of Noah-MP simulation was further improved by introducing new parameterizations of thermal roughness length, soil hydraulic properties, and vertical root distribution. It was found that: (1) Simulation of water and heat exchange over alpine grassland in the growing season was mainly affected by the parameterizations of dynamic vegetation, canopy stomatal resistance, runoff and groundwater dynamics, and surface exchange coefficient for heat transfer. Selection of different parameterization options for these four physical processes/variables led to large differences in the simulation of water and heat fluxes. (2) The optimal combination of parameterization options selected in the current Noah-MP framework suffered from significant overestimation of sensible heat flux (H) and underestimation of soil moisture (θ) at all observation sites. (3) The overestimation of H was significantly improved by introducing a new parameterization of thermal roughness length. Furthermore, the underestimation of θ was resolved by introducing a new parameterization of soil hydraulic properties that considered the organic matter effect and a new vertical distribution function for the vegetation root system. The results of this study provide an important reference for further improving the simulation of water and heat exchange by using the land surface model in alpine grassland. [ABSTRACT FROM AUTHOR]