Given the large number of land surface process models (LSMs), extensive and rigorous model intercomparison is helpful to enhance our understanding of model behaviors and provide references for the improvement of LSMs. Due to the simplicity in selecting alternative schemes within one modeling framework, the Noah-MP LSM has been attracting increasing attention in intercomparison work among multiple parameterizations at point and watershed scales. However, there is still a lack of comparative studies on soil hydrothermal processes in permafrost regions. 

(1) Noah-MP tends to overestimate snow cover, which is most influenced by the snow/soil temperature time (STC) and snow sublimation (SUB) processes. Such overestimation can be greatly resolved by considering the from wind, SUB(2) and the semi-implicit snow/soil temperature time scheme, STC(1). 

(2) Soil temperature is largely underestimated by the overestimated snow cover and, thus, dominated by the STC and SUB processes. The canopy gap for radiation transfer (RAD), surface layer drag coefficient (SFC) and runoff and groundwater (RUN) processes play an important but secondary role. Systematic cold bias and large uncertainties in soil temperature still exist after eliminating the effects of snow, particularly in the deep layers and during the cold season. The combination of the Y08-UCT schemes contributes to resolving the cold bias of soil temperature. 

(3) Noah-MP tends to underestimate the soil liquid water content. Most physical processes have a limited influence on the soil liquid water content, among which the RUN process plays a dominant role during the whole year. The STC and SUB process have a considerable influence on topsoil liquid water during the warm season. 

This study has been published on Geoscientific Model Development recently titled with Assessing the simulated soil hydrothermal regime of the active layer from the Noah-MP land surface model (v1.1) in the permafrost regions of the Qinghai–Tibet Plateau.