Comprehensive evaluation of typical planetary boundary layer (PBL) parameterization schemes in China. Part I: Understanding expressiveness of schemes for different regions from the mechanism perspective

Abstract. The optimal choice of the planetary boundary layer (PBL) parameterization scheme is of particular interest and urgency to a wide range of scholars, especially for many works involving models. At present, there have been many works to evaluate the PBL schemes. However, little research has been conducted into a more comprehensive and systematic assessment of the performance capability of schemes in key regions of China, especially when it comes to the differences in the mechanisms of the schemes themselves, primarily because there’s scarcely sufficient observational data, computer resources, and storage support to complete the work. In addition, there are many factors that influence the selection of schemes, such as underlying surface, initial and boundary conditions, near-surface layer scheme, horizontal/vertical resolution, etc. In this Part (i.e., Part I), four typical schemes (i.e., YSU, ACM2, BL and MYJ) are selected to systematically analyze and evaluate near-surface meteorological parameters, PBL vertical structure, PBL height (PBLH), and turbulent diffusion in five key regions (i.e., North China Plain, NCP; Yangtze River Delta, YRD; Sichuan Basin, SB; Pearl River Delta, PRD and Northwest Semi-arid, NS) of China in different seasons (i.e., January, April, July and October). The differences in the simulated 2-m temperatures between the nonlocal closure schemes are mainly affected by the downward shortwave radiation, but to compare the nonlocal closure schemes with the local closure schemes, the effect of sensible heat flux needs to be further considered. In terms of temporal variation, the simulated results for July are better than the other three months, and the simulated results for nighttime are better than daytime. In terms of regional distribution, the temperature at stations with higher elevation is easily underestimated, while overestimated with lower elevation. The variation of relative humidity corresponds to temperature. The 10-m wind speed is under the influence of factors like the momentum transfer coefficient and the integrated similarity functions at night. The wind speeds are more significantly overestimated in the plains and basin, while less overestimated or even underestimated in the mountains, as a result of the effect on topographic smoothing in the model. Moreover, the overestimation of small wind speeds at night is attributable to the inapplicability of the Monin-Obukhov similarity theory (MOST) at night. The model captures the vertical structure of temperature well, while the wind speed is outstandingly overestimated below 1000 m, largely because of the turbulent diffusion coefficient (TDC). The difference between the MOST and the mixing length theory, PBLH and Prandtl number is cited as the reason for the difference between the TDC of the YSU and ACM2 schemes. The TDCs of the BL and MYJ schemes are affected by the mixing length scale, which of BL is calculated on the basis of the effect of buoyancy, while MYJ calculates it with the consideration of the effect of the total turbulent kinetic energy. The PBLH of the BL scheme is better than the other schemes because of the better simulation results of temperature. The difference in the PBLH by the YSU and ACM2 scheme mainly comes from the Richardson number and the jagged PBLH of the MYJ scheme is due to the coarse vertical resolution and the threshold value.

In general, to select the optimal scheme, it is necessary to offer different options for different regions with different focuses (heat or momentum). (1) Temperature field. The BL scheme is recommended for January in the NCP region, especially for Beijing, and the MYJ scheme is better for the other three months. The ACM2 scheme would be a good match for the YRD region, where the simulation differences between the four schemes are small. The topography of the SB region is more complex, but for most of the areas in the basin, the MYJ scheme is proposed, but if more stations outside the basin are involved, the BL scheme is recommended. The MYJ scheme is applied to the PRD region in January and April, and the BL scheme in July and October. The MYJ scheme is counselled for the NS region. (2) Wind field. The YSU scheme is recommended if the main concern is the near-surface layer, and the BL scheme is suggested if focusing on the variation in the vertical direction. The Part II will analyse and evaluate the factors that may influence the choice of the schemes and the results of model. The final evaluation of the parameterization scheme and uncertainties will lay the foundation for the improvement of the modules and forecasting of the GRAPES_CUACE regional model developed independently in China.