基于融合MODIS AOD和LME模型的京津冀地区PM2.5季节浓度模拟

Seasonal concentration simulation of PM2.5 in Beijing-Tianjin-Hebei region based on fused MODIS AOD and LME model

为探究京津冀地区PM2.5浓度的时空差异,基于线性混合效应模型,在该地区分别建立不同季节融合MODIS AOD与PM2.5的关系,并对其季节浓度进行模拟.研究结果表明:模型拟合所得四季的可决系数R2分别为0.738、0.668、0.644和0.760,留一交叉验证后R2分别为0.733、0.658、0.636和0.756,模型仅存在轻微过度拟合,且PM2.5浓度较高的春冬季拟合效果优于夏秋季;在气候、地形、经济和人口等多种因素的综合影响下,京津冀PM2.5浓度存在季节差异和地区差异,即四季PM2.5浓度冬季最高,春秋次之,夏季最低,空间上基本呈现北部低、南部高的特点.

In order to explore the spatial and temporal differences of PM2.5 concentration in Beijing-Tianjin-Hebei region,the relationship between the fusion of MODIS AOD and PM2.5 in different seasons was established in the region based on linear mixed effect model,and the seasonal concentration was simulated.The results show that R2 of the four seasons which are fitted by the model are 0.738,0.668,0.644 and 0.760,and R2 change to 0.733,0.658,0.636 and 0.756 after leave-one-out cross-validation,and the model has only slight overfitting.The fitting effects of higher PM2.5 of spring and winter are better than those of summer and autumn.There are seasonal and regional differences in PM2.5 concentration in Beijing-Tianjin-Hebei under the combined influence of climate,terrain,economy,population and other factors.The PM2.5 concentration generally peaks in win-ter,closely followed by spring and autumn,and reaches the minimum in the summer.And the PM2.5 concentration is lower in the north and higher in the south in the spatial distribution.