边界层方案与垂直混合对京津冀地区O3模拟的影响研究

采用WRF-Chem模式中的3种边界层方案YSU、MYJ和ACM2对2019年6月京津冀及周边地区典型O₃污染月份开展模拟研究.详细对比了各方案对地面气象要素、NO₂和O₃浓度时空分布，以及温湿风要素和O₃浓度垂直分布的模拟效果.结果表明：3种方案对地面气象要素的时空分布和温湿风要素的垂直变化模拟较为合理.MYJ方案模拟地面气象要素整体效果最佳.各方案对边界层高度的日变化特征模拟较好，相关系数为0.58~0.69，但存在白天偏高、夜间偏低的现象，YSU方案相比效果最佳.3种边界层方案对NO₂浓度模拟普遍高估，而O₃模拟结果则出现低估.白天模拟偏差较小而夜间偏差较显著.模拟最佳的是ACM2，其次为YSU和MYJ.3种方案均较好地模拟出了O₃的垂直分布特征，但整体低估了O₃浓度.对上午O₃垂直分布的模拟差异较下午更为明显.此外，基于YSU方案设置了3个敏感实验，通过调整化学模块所用的湍流扩散系数阈值，对比分析了垂直混合过程改变对O₃浓度模拟的影响，模拟的变化只反映由于边界层的垂直混合过程改变造成的污染差异，而不是由于热动力场的调整造成的变化.模拟结果表明3个方案均可改善区域上地面NO₂和O₃的模拟性能，尤其是对原3种边界层方案模拟O₃均明显低估的华北平原地区提升效果最显著，平均偏差降低了23.7%.在垂直方向上，湍流扩散系数阈值的调整增加了早间近地面模拟的O₃浓度，改善了模拟偏低的现象，但同时增大了高层O₃浓度的负偏差.敏感性方案显著改善了夜间的模拟，白天则并不明显.这些结果显示出湍流扩散系数对O₃垂直混合的重要影响.因此，改进湍流扩散系数的参数化对O₃模拟是必要的.

The study on the impact of boundary layer schemes on O3 simulations in the Beijing-Tianjin-Hebei region

Three boundary layer schemes, YSU, MYJ, and ACM2, in the WRF-Chem model were used to simulate O₃ during June 2019, a typical pollution month, over Beijing-Tianjin-Hebei and surrounding areas. The spatial-temporal distribution of simulated surface meteorological variables, NO₂ and O₃ concentrations, and the vertical profiles of temperature, humidity, wind components, and O₃ concentrations were compared. The simulations showed favorable performance in the spatial-temporal distributions and the vertical profiles of these meteorological variables in all three schemes, with the best performance of MYJ. All schemes could well simulate the diurnal cycle of planetary boundary layer height (PBLH), with correlation coefficients ranging from 0.58 to 0.69, but with overestimation and underestimation in daytime and nighttime, respectively. The YSU scheme performed best in PBLH simulation. The simulated NO₂ concentrations were generally overestimated by three boundary layer schemes, while the O₃ simulations were underestimated. The deviations were smaller for daytime simulations and more significant for nighttime. The best simulation was from ACM2, followed by YSU and MYJ. All three schemes gave vertical distributions of O₃ but underestimated their concentration. The simulating differences of profiles were more significant in the morning than in the afternoon. Additionally, three sensitivity experiments based on YSU were set up in this paper to compare and analyze the effect of the change of vertical mixing process on the simulation of O₃ concentration by adjusting the turbulent diffusion coefficients threshold values (TDCTs) used in the chemistry module, and the simulated changes only reflected the difference of pollution due to the change of vertical mixing process in the boundary layer rather than the change due to the adjustment of the thermodynamic field. The simulation results showed that all three experiments could improve the simulating performance of surface NO₂ and O₃ at the upper surface of the region. In particular, the most significant improvement was found in the North China Plain, where O₃ is significantly underestimated by the original three boundary layer schemes, and the mean bias is reduced by 23.7%. Vertically, the adjustments of TDCT increased the O₃ concentration near-surface in the morning and improved the simulation bias, but at the same time increased the negative bias of the underestimated O₃ concentration in the upper levels. The sensitivity experiments significantly improved the simulation performance at nighttime but not significantly during the daytime. This study showed the importance of turbulent diffusion coefficients on the vertical mixing of O₃. Therefore, improving the parameterization of turbulent diffusion coefficients is necessary for O₃ simulations.