Models-3/CMAQ和CAMx对珠江三角洲臭氧污染模拟的比较分析

Evaluation and intercomparison of ozone simulations by Models-3/CMAQ and CAMx over Pearl River Delta

运用Models-3/CMAQ和CAMx两个区域空气质量模式同时对珠江三角洲2004年10月的臭氧污染进行模拟,利用监测网12个站点的观测数据对两个模式的臭氧模拟效果进行统计评估,并应用过程分析技术,量化各种大气物理、化学过程对臭氧浓度变化的影响,探讨两个模式结果偏差产生的主要原因.CMAQ与CAMx均能较好地模拟出珠江三角洲大多数站点的臭氧浓度水平和变化趋势,与监测值的相关系数分别为0.73,0.74,标准化平均偏差分别为-8.5%、8.8%、标准化平均误差分别为36.7%和37.9%;两个模式的模拟结果具有很高的相关性(相关系数R为0.92)和较为一致的空间分布,但CMAQ的臭氧模拟浓度总体上较CAMx的结果偏低约17%.两个模式对珠江三角洲不同地区、不同类型站点的模拟结果一致性不尽相同.在干沉降、化学反应参数、垂直输送等方面处理方法上的差异共同造成了两个模式对珠江口沿岸站点臭氧模拟结果上的差别;细网格边界浓度差异是造成上风向区域站点CAMx模拟浓度比CMAQ结果偏高的主要原因.CAMx有必要提供更多干沉降的算法供选择,改善光解速率等计算方法可能改善CMAQ臭氧模拟结果.

Ozone pollution over the Pearl River Delta （PRD） in October 2004 was simulated by the regional air quality models Models-3/CMAQ and CAMx. The results from both models were evaluated and compared with the observed concentrations from 12 monitoring stations. By integrated process rate analysis, the influences of different physical and chemical processes were quantified, and the causes of the deviations between the two models were investigated. Both CMAQ and CAMx reproduced the magnitudes and variations of ozone at most stations over the PRD. The correlation coefficients （R） between the simulated results and monitoring data were 0.73 for CMAQ and 0.74 for CAMx. The NMBs for CMAQ and CAMx over the 12 sites were -8.5% and 8.8% on average, respectively. The NMEs for CMAQ and CAMx were 36.7% and 37.9%, respectively. The correlation between the results of two models was very high （R = 0.92）, and their simulated ozone spatial distributions exhibited common features. But the simulated result of CMAQ was about 17% lower than that of CAMx on average. The simulated results from the two models were not identical in certain regions and for different types of monitoring stations. The differences in dry deposition, reaction parameters and vertical transport accounted for the diversity of the results from the two models near the Pearl River Estuary. The discrepancy in the boundary concentration of the finest nest was the main cause for the higher CAMx results than those of CMAQ in the upwind areas. There is a need for CAMx to provide more choices of dry deposition algorithms. Improvement of the calculation methods for photolysis rates would improve the ozone simulation of CMAQ.