首页 > , Vol. , Issue () : -
大气污染物浓度全方位动态监测是进行区域大气污染精细化防控的重要前提。为开展长三角地区小时分辨率PM2.5浓度无缝制图,本研究通过耦合AOD缺失信息重建以及多源异构数据融合分析技术,建立了一套能够有效集成卫星遥感、地面观测、模式模拟等多源异构数据的近地面PM2.5浓度无缝制图方案,并据此生产了2015–2020年长三角地区高分辨率无缝PM2.5浓度格点资料。结果表明：本研究生产的PM2.5浓度无缝格点资料与国控站点观测数据的交叉验证相关系数达0.9,平均偏差不超过10 μg m-3。同时,基于生产的长时序PM2.5浓度无缝格点资料分析发现,较于空间分布稀疏的站点PM2.5浓度观测,面域无缝PM2.5浓度格点数据更能有效揭示长三角地区PM2.5污染的时空变化特征,平均下降速率超过3 μg m-3 yr-1。本研究发展的PM2.5浓度无缝制图方法和生产的相关数据产品有望为区域PM2.5污染防控和风险暴露评估研究提供重要的方法参考和数据支撑。
Monitoring concentrations of atmospheric particulate matters is essential to regional haze pollution prevention and control. Satellite-based aerosol optical depth (AOD) data have been frequently used to map regional PM2.5 concentrations. However, the resultant PM2.5 concentration maps are always spatially incomplete due to significant data gaps in satellite-based AOD retrievals given excessive cloud cover and bright surface. To bridge the gap, a spatially contiguous PM2.5 concentration mapping approach was hereby developed by seamlessly gearing up the missing AOD imputation and multisource data fusion methods, aiming to generate spatially contiguous PM2.5 concentration maps on an hourly basis in the Yangtze River Delta. Specifically, hourly AOD maps from Himawari-8 during the daytime within a certain time period were used synergistically to reconstruct missing AOD information in each AOD map on each specific date. The ultimate goal is to maximize AOD coverage by taking advantage all available Himawari-8 AOD observations. Subsequently, AODs were estimated at each state-controlled air quality monitoring station by inferring from PM measurements. This enables us to extend the sparsely distributed aerosol monitoring network to nationwide, significantly improving the spatial coverage of AOD. Subsequently, the reconstructed AOD and PM inferred AOD were fused with AOD analysis from MERRA-2 via optimal interpolation to generate spatially contiguous yet far more accurate AOD reanalysis. Spatially complete PM2.5 concentration maps were finally generated on hourly basis over the study region using random forest, with a correlation coefficient equaling to 0.9 and mean absolute error of 9.87 μg m-3 compared against in situ PM2.5 measurements. Compared to sparsely distributed in situ PM2.5 measurements, this spatially contiguous dataset performs better in revealing PM2.5 variations in space and time in the Yangtze River Delta. Statistically significant decreasing trend are observed over the whole study area in the past five years, indicative of the effectiveness of the implemented clean air actions in reducing PM2.5 loadings over China. Overall, the proposed method can be used as a practical approach for future PM2.5 mapping practices and the generated spatially contiguous PM2.5 concentration dataset helps better assess the human exposure risk to haze pollution.