大气物理过程方案对PM2.5预报的敏感性试验

利用WRF-Chem大气化学模式,选择2015年12月中旬发生在我国的大范围空气污染过程,在采用同样化学方案条件下,针对模式中不同物理过程及其参数化方案开展了地面PM_(2.5)预报的敏感性试验。结果表明:该模式能较好展示此次PM_(2.5)污染的演变过程,与实况也较接近,但对青海经宁夏至内蒙的PM_(2.5)高值区出现了漏报现象,这可能是模式外边界未对污染物做更新所致。对地面PM_(2.5)的预报,各物理过程的敏感度不同,边界层(含近地面层)过程的影响要明显大于积云对流及微物理过程的影响,不同的参数化方案会造成不同的预报误差。边界层过程QNSE和与其配套的近地面层方案的组合是预报的较佳组合;而TEMF和与其配套的组合以及ACM2和Pleim-X的组合则不佳。合理的物理过程参数化方案有助于提高PM_(2.5)预报质量。模式预报对排放源也有适应过程,其Spin-up时间较气象要素长。

Sensitivity of PM2.5 forecast for atmospheric physical processes

The sensitivity of the ground PM_2.5 forecast to physical process parameterization schemes was investigated by using the WRF-Chem atmospheric chemistry model for an air pollution case that occurred in China in December 2015. Results show that PM_2.5 pollution can be well simulated in the model, which is close to observations. The pollution in Qinghai and Ningxia is underestimated mainly because of the lack of contaminant update in the boundary of the model. The physical process schemes impact the ground PM_2.5 forecast differently. The influence of the selected boundary layer (including near-ground layer) schemes is dramatically greater than Cumulus convection parameterization and micro-physical process schemes. In comparison, the combination of the boundary layer process QNSE and its relative near-ground layer schemes yields the best forecast of PM_2.5 pollution forecast. While the TEMF and its relative schemes make the worst forecast. What''s more, the obvious adjustment has been found in the model containing emission, in which the spin-up time is longer than the other meteorological elements.