Improving the Noah Land Surface Model in Arid Regions with an Appropriate Parameterization of the Thermal Roughness Length

Daytime land surface temperatures in arid and semiarid regions are typically not well simulated in current land surface models (LSMs). This study first evaluates the importance of parameterizing the thermal roughness length (z(0h)) to model the surface temperature (T(sfc)) and turbulent sensible heat flux (H) in arid regions. Six schemes for z(0h) are implemented into the Noah LSM, revealing the high sensitivity of the simulations to its parameterization. Comparisons are then performed between the original Noah LSM and a revised version with a novel z(0h) scheme against observations at four arid or semiarid sites, including one in Arizona and three in western China. The land they cover is sparse grass or bare soil. The results indicate that the original Noah LSM significantly underestimates T(sfc) and overestimates H in the daytime, whereas the revised model can simulate well both T(sfc) and H simultaneously. The improved version benefits from the successful modeling of the diurnal variation of z(0h), which the original model cannot produce.