A Study On Comparison Of The Thermal And Dynamic Structural Simulation In Boundary Layer With Different Boundary Layer Parameterizations

High resolution mesoscale WRF model was firstly used to conduct some experiments by choosing different boundary layer parameterizations, which was to evaluate model performance of simulating structural characteristics in boundary layer at the forested areas of northern United States and compare the thermal and dynamic structural characteristics in boundary layer with five different boundary layer parameterizations. The results showed that for the simulation of boundary layer structure at the forested areas, WRF model with different boundary layer schemes could successfully simulate the characteristics of strong turbulent mixing in boundary layer at daytime and strong inversion, inverse wet and low level jet in stable boundary layer at nighttime except some individual schemes. Compared to local schemes, MYJ and UW, nonlocal schemes, YSU and ACM2, showed strong turbulent mixing and entrainment at daytime, producing higher temperature, lower humidity, higher mixing layer and larger sensible heat flux in convective boundary layer, which was closer to observations. This indicated that considering the nonlocal transport of large eddy was more reasonable in the instable stratification. But local schemes had some advantages in the prediction of wind speed and direction. The scheme, TEMF, predicted the weakest locally turbulent mixing in all schemes, which was a disadvantage to form the mixing layer, and could not describe the characteristics of vertically homogeneous distribution of meteorological elements in convective boundary layer. For nocturnal stable boundary layer simulation, the discrepancies from different schemes were small. But to a certain extent, the scheme, YSU, overestimated mechanical turbulence at nighttime, leading to strong locally turbulent mixing, and could not well simulate stable boundary layer. The differences of simulating structural characteristics in boundary layer among five schemes at forested areas in northeast China were the same as in northern United States.Many local closure boundary layer parameterizations could be used in WRF model, so it was necessary to compare the differences of simulating structural characteristics in boundary layer among these local schemes. We compared five local TKE closure schemes using single column model in high resolution mesoscale WRF by observed surface sensible heat, latent heat and momentum fluxes driving the boundary layer process. The results showed that five local schemes could successfully simulate the characteristics of strong turbulent mixing in boundary layer at daytime and the simulation of potential temperature, vapour mixing ratio, wind speed and direction were close to observation during DICE experiment. The BouLac and UW schemes showed stronger turbulent mixing and entrainment than other three schemes and mixed layers simulated by BouLac and UW schemes were deeper than the others at daytime. MYJ, MYNN2and MYNN3schemes simulated turbulent mixing and entrainment were weak compared to observation. For example, these three schemes simulated entrainment at top of boundary layer was close to zero, which was inconsistent with the negative entrainment at top of convective boundary layer. Five local schemes could successfully simulate strong inversion, but failed to simulate inverse wet and simulated low level jet was weaker than observation in stable boundary layer at nighttime. The differences of simulating structural characteristics in boundary layer among these local schemes at nighttime were smaller than daytime. There were large differences in predicted TKE by schemes at daytime and predicted TKE by MYJ and UW schemes was smaller than observation.