Improving the Treatment of the Vertical Snow Burial Fraction over Short Vegetation in the NCAR CLM3

One deficiency of the NCAR Community Land Model （CLM3） is the disappearance of the simulated snow even in the middle of winter over a boreal grassland site due to unrealistically modeled high downward turbulent fluxes. This is caused by the inappropriate treatment of the vertical snow burial fraction for short vegetation. A new snow burial fraction formulation for short vegetation is then proposed and validated using in situ observations. This modification in the CLM3 largely removes the unrealistic surface turbulent fluxes, leading to a more reasonable snowmelt process, and improves the snow water equivalent （SWE） simulation. Moreover, global offline simulations show that the proposed formulation decreases sensible and latent heat fluxes as well as the ground temperature during the snowmelt season over short vegetation dominant regions. Correspondingly, the SWE is enhanced, leading to the increase in snowmelt-induced runoff during the same period. Furthermore, sensitivity tests indicate that these improvements are insensitive to the exact functional form or parameter values in the proposed formulation.

Development and test of a multifactorial parameterization scheme of land surface aerodynamic roughness length for flat land surfaces with short vegetation

Aerodynamic roughness length is an important physical parameter in atmospheric numerical models and microme- teorological calculations, the accuracy of which can affect numerical model performance and the level of micrometeorological computations. Many factors influence the aerodynamic roughness length, but formulas for its parameterization often only con- sider the action of a single factor. This limits their adaptive capacity and often introduces considerable errors in the estimation of land surface momentum flux （friction velocity）. In this study, based on research into the parameterization relations between aerodynamic roughness length and influencing factors such as windrow conditions, thermodynamic characteristics of the sur- face layer, natural rhythm of vegetation growth, ecological effects of interannual fluctuations of precipitation, and vegetation type, an aerodynamic roughness length parameterization scheme was established. This considers almost all the factors that af- fect aerodynamic roughness length on flat land surfaces with short vegetation. Furthermore, using many years＇ data recorded at the Semi-Arid Climate and Environment Observatory of Lanzhou University, a comparative analysis of the application of the proposed parameterization scheme and other experimental schemes was performed. It was found that the error in the friction velocity estimated by the proposed parameterization scheme was considerably less than that estimated using a constant aero- dynamic roughness length and by the other parameterization schemes. Compared with the friction velocity estimated using a constant aerodynamic roughness length, the correlation coefficient with the observed friction velocity increased from 0.752 to 0.937, and the standard deviation and deviation decreased by about 20% and 80%, respectively. Its mean value differed from the observed value by only 0.004 m s-l and the relative error was only about 1.6%, which indicates a significant decrease in the estimation error of surface-layer momentum flux. The test results show that the multifactorial universal parameterization scheme of aerodynamic roughness length for flat land surfaces with short vegetation can offer a more scientific parameteriza- tion scheme for numerical atmospheric models.