天津大气稳定度特征及基于湍流扩散系数优化的空气质量数值模拟研究

Analysis of atmospheric stability characteristics in Tianjin and air quality modeling study based on the modification of minimum eddy diffusivity

空气质量模式中湍流引起的垂直混合与湍流扩散系数K密切相关.为避免针对强稳定边界层计算中可能出现“无湍流大气”(即K=0)的异常结果,模式通过预设最小湍流扩散系数Kzmin,定义了K值的下限.检验表明天津空气质量模式整体模拟效果较好,但02:00—08:00存在系统性偏高的问题.针对这一问题,本文在天津大气稳定度特征分析和模式评估基础上,利用气象塔和系留获取真实湍流扩散系数,试验性修正重污染期间Kzmin取值,以期探索提升稳定层结条件下天津空气质量模式PM_(2.5)模拟能力.结果表明:天津不同大气层结稳定度占比分别为4.91%(强不稳定)、9.33%(不稳定)、18.86%(弱不稳定)、50.29%(中性)、13.00%(较稳定)和3.61%(稳定).大气稳定条件下PM_(2.5)浓度(62.6µg·m^(-3))相较不稳定和中性条件(38.8µg·m^(-3))升高61.3%,02:00—08:00稳定大气层结占比显著提升(28.2%),模式对该时段PM_(2.5)浓度模拟存在系统性高估现象(9.3%).以2017年12月一次重污染过程为例开展模式敏感试验,该过程天津出现长时间逆温过程,高度达600 m以上,逆温强度达2.9℃·100 m^(-1),逆温层的持续存在抑制了污染物的垂直扩散.系留气艇和气象塔湍流观测表明:此次过程垂直湍流扩散速率存在明显日变化,白天(35 m^(2)·s^(-1))普遍高于夜间(<14 m^(2)·s^(-1)),比值为2.5∶1,基于模式过程分析技术估算湍流混合使PM_(2.5)浓度下降白天为-30µg·m^(-3)·h^(-1),夜间显著减弱,低于-5µg·m^(-3)·h^(-1),比值为6∶1,结合探空实测数据,空气质量模式中预设的Kzmin很可能造成对垂直湍流扩散作用的低估,并因此导致对夜间稳定层结PM_(2.5)浓度的高估.根据观测对模式Kzmin取值优化,改进后模式对城市和郊区站PM_(2.5)浓度模拟绝对误差分别下降36.6和46.1µg·m^(-3),且夜间PM_(2.5)浓度极大值的改进效果更加明显,基于Kzmin的修正,可能有助于稳定层结条件下天津空气质量模式模拟能力提升.

The simulated vertical mixing flux caused by turbulence largely depends on the turbulent diffusion coefficient K in the air quality model.To avoid the unrealistic diagnosis of‘no-turbulence atmosphere’(i.e.,K=0)in the case when the planetary boundary layer(PBL)is strongly stable,air quality model presets Kzmin to limit the minimum of K.The simulation of PM_(2.5) using preset Kzmin limit in Tianjin air quality modeling platform generally shows satisfactory performance except that the concentration of PM_(2.5) is systematically overestimated during 02:00 to 08:00.To address this issue,this study modified the value of Kzmin during heavy pollution episodes based on the turbulent diffusion coefficient observed from meteorological tower and tethered-balloon sounding in addition to the analysis of atmospheric stability characteristics and model evaluation.This development of model improved the performance of Tianjin air quality modeling platform in simulating the PM_(2.5) concentrations under stable conditions.Results indicate that the occurrence frequency of different atmospheric stability is 4.91%(strongly unstable),9.33%(unstable),18.86%(weakly unstable),50.29%(neutral),13.00%(weakly stable)and 3.61%(stable).The average PM_(2.5) concentration under stable conditions(62.6μg·m^(-3))is~61.3%higher than that under unstable and neutral conditions(38.8μg·m^(-3)).The occurrence frequency of stably-stratified atmosphere raises by 28.2%during 02:00 to 08:00,leading to the PM_(2.5) concentrations in the air quality model systematically overestimated by 9.3%during this period.Numerical sensitivity experiment was conducted for a heavy pollution episode in December,2017.The temperature inversion(TI)structure in Tianjin during this episode last for a long time.The height of TI layer was elevated up to about 600 m and the intensity of TI reached 2.9℃·100 m^(-1).The persistence of the TI layer consistently restrained the vertical diffusion of air pollutants.Tethered-balloon sounding and tower-based turbulence observation indicated that the vertical turbulent diffusion exhibited clear diurnal characteristics which is 2.5 times higher in the day(35 m^(2)·s^(-1))than at night(<14 m^(2)·s^(-1)).Process analysis shows that the reduction of PM_(2.5) concentration induced by turbulent mixing was-30μg·m^(-3)·h^(-1) during the daytime and weakened at night(less than-5μg·m^(-3)·h^(-1))with the ratio of 6.According to the simulation and sounding observation,the value of Kzmin preset in the air quality model is likely to cause the underestimation of vertical turbulent mixing,hence lead to the overestimation of PM_(2.5) concentrations under stably stratified nighttime conditions.The modification of Kzmin could significantly improve the model performance in the simulation of surface PM_(2.5) concentrations during air pollution episodes,which could lead to the decrease of 36.6 and 46.1μg·m^(-3) in the MABs in urban and rural PM_(2.5) concentrations,respectively.This effect was found to be more pronounced for the cases with high PM_(2.5) concentrations during the nighttime.Hence the modification of Kzmin would help to improve the forecast performance of Tianjin air quality model under stable conditions.