大气扩散应急预报的风场不确定性影响研究

Study of the Influence of Wind Field Uncertainty in Atmospheric Diffusion Emergency Forecast

使用WRF模式和CALMET模式,获得一个40 km区域的细网格气象预报场,同时利用加入地面观测资料的方法获得当地诊断风场。用随机粒子扩散模式模拟两种风场驱动下的扩散结果,并比较和评估预报模拟的偏差或不确定性。对4个季节的代表性月份(1,4,7和10月)的逐时排放情况以及4个排放高度的情景进行模拟分析。结果表明:1)预报模拟的烟云扩散形态(方向和宽度等)在当地大多数时段内(约占全年的80%)与诊断分析的实际扩散结果一致,且季节变化不大,其余时段为扩散形态有中度偏差和有明显差异的情况,二者各占10%左右;2)地面轴线浓度的不确定性随下风距离及污染排放高度而变化,20~100 m各高度的排放结果大致在2~4 km的下风距离出现最大偏差,但100 m的高源排放在约2 km以内的范围也有很大的预报不确定性;3)两种情况造成当地扩散预报结果的明显偏差,一是气象场预报的局地风场发生重要转变的时间不一致,使预报风场与实际风场处于转变前后的不同步状态,从而使污染扩散预报结果出现重大偏差,二是WRF模式对地面风速的预报系统性地偏大(50%左右),造成预报模拟浓度结果系统性地偏小。

This paper studies deviations and uncertainties of atmospheric diffusion caused by wind field forecasting, in conditions of emergency release. WRF and CALMET were used to create a 40 km fine-mesh meteorological forecast field and a diagnostic field with local data. In the simulation, we traced the emissions in January, April, July and October which are representative of four seasons and the emissions in four typical situations. The analysis shows that the forecasts are consistent with the diagnosis in 80% of the year and the change of seasons does not affect significantly, while the rest 20% is shared by different plume shape and significant deviations, each accounting for about 10%. Downwind concentration varies with emission height and downwind distance. The maximum deviation occurs when the height is 20-100 m and the distance is 2-4 km, while the result is highly uncertain when the height is 100 m and the distance is shorter than 2 km. The significant deviations occur in two situations. In the first, the time of the important transition of the local wind field predicted by the meteorological field is inconsistent, so that the forecast wind field and the actual wind field are in an asynchronous state before and after the transition, which causes a major deviation in the pollution diffusion forecast results. In the second, WRF, which systematically overestimates the wind speed(at 50% approximately), leads to systematically lower forecasted concentration.