The Impacts Of The Entrainment Process On GRAPES Model

With the rapid development of numerical weather prediction and high performancecomputing technology, the resolution of NWP models is enhanced increasingly, but the finitemodel grid cannot resolve all the scales in the atmosphere explicitly. Since turbulent mixing isoften active at scales smaller than the horizontal grid, parameterization schemes are thennecessary in order to properly describe the impact of these subgrid-scale mechanisms. In otherwords, the effect of the subgrid-scale processes has to be formulated in terms of the resolvedgrid scale variables. While planetary boundary layer represents the turbulent transfer of heat,momentum and moisture between the surface and the lowest model level, so PBL influencesthe modeling performance directly both for numerical weather prediction models and climatemodels.In this paper, in order to resolve the existed problems of MRF PBL scheme in GRAPES(Global and Regional Assimilation and Prediction System), a revised scheme called YSU PBLscheme is incorporated to GRAPES model, and the single-column model and real-timeforecast are conducted to validate the performance of the scheme.Compared with MRF PBL scheme, the main improvement of YSU PBL scheme isconsidering the explicit treatment of entrainment processes. The SCM (Single Column Model)experiment indicates that YSU PBL increases the turbulent mixing in the thermally inducedfree convection regime and decreases in the mechanically induced forced convection regime.The enhanced diurnal diffusivity produces a more neutralized wind profile and potentialtemperature profile when the surface flux is positive. The larger negative vertical flux near thePBL top reveals that the explicit treatment of entrainment processes in YSU PBL enhances theeffect of entrainment between mixing layer and free atmosphere, which can prevent theupward flux to across the PBL top. In addition, a more reasonable PBL structure has beenrealized when the critical bulk Richardson number is set to0.25. A case simulation of typhoonFANAPI indicates that the typhoon track forecast error and the precipitation variation of YSUPBL are smaller than that of MRF PBL scheme, and the temporal and spatial distribution ofphysical elements are closer to the analysis. In the early stage, YSU PBL scheme slows downthe collapse of typhoon FANAPI because of the warm and moist environment. In the late stage, the sensitivity of boundary layer parameterization is prominent. Moistening below the PBL topand drying above the top is the consequence of the enhanced entrainment, which reduces toomuch upwards transfer of water vapor, and this alleviates the dry bias in the low-level in MRFPBL scheme.Based on the above, further analysis also reveals that the YSU PBL scheme has notestablished the nocturnal stable BL’s parameterization. So some improvements have beenconducted:(1) by changing the calculation of the velocity scale in stable layer, the too muchmixing has been reduced, and the temporal and spatial distribution of meteorological elementsis also improved.(2) Decrease of Prandtl number enhances the PBL mixing in daytime.(3)The asymptotic length scale is no longer set as a constant, the abnormal increase of mixingabove the PBL top after collapse of boundary layer is dismissed.