北京市区绿色空间滞尘分布反演研究

北京市作为中国政治中心和京津冀特大型城市，近40年来城市化进程迅速，大气颗粒物和尘埃粒子污染问题较为突出，发挥绿色空间滞尘功能具有重要现实意义。将高光谱技术与遥感技术相结合，反演了市域尺度绿色空间滞尘分布。以北京市区绿色空间常见植被大叶黄杨(Euonymus japonicus)为研究对象，通过室外采样与室内试验获取叶片样本的滞尘量、光谱反射率和叶面积等数据，比较叶片滞尘前后的原始光谱曲线和反射率一阶导数特征，分析不同滞尘量对光谱反射率的影响，探究对叶片滞尘量高度敏感的波段。利用光谱响应函数将地面采集的窄波段光谱反射率数据分别转化遥感卫星的宽波段光谱反射率数据，建立对应卫星波段植被指数比值与滞尘量的回归模型，选取拟合效果最好的回归模型作为滞尘反演模型。结合GF-2影像所提取北京市区绿色空间范围，采用滞尘反演模型获取北京市区绿色空间滞尘分布。进而插值得到北京市区尘埃污染分布，采用空间自相关模型检验其空间聚集特性。结果表明：740~1 870 nm波段, 滞尘后光谱反射率明显低于滞尘前反射率，滞尘对红边、黄边、蓝边位置没有明显影响，对“红边幅值”和“红边面积”影响较明显，利用Sentinel-2影像计算的EVI指数与滞尘量相关性最高，所构建的线性与二次滞尘回归模型决定系数(R2)分别为0.705和0.751；利用2021年4月7日和2021年6月3日的Sentinel-2影像反演获取了北京市区绿色空间滞尘分布，其滞尘分布趋势表现为：市中心高于郊区，北部高于南部，东部高于西部。北京市区中部、北部和东部易产生尘埃污染。污染分布具有明显聚集性，并非完全随机。

Retrieval of Dust Retention Distribution in Beijing Urban Green Space Based on Spectral Characteristics

As the political center of China and a super large city of Beijing, Tianjin and Hebei, the urbanization process of Beijing has been rapid in the past 40 years, and the pollution problems of atmospheric particles and dust particles are prominent. It is of great practical significance to play the role of green space dust retention. This paper combines hyperspectral technology and remote sensing technology to retrieve the urban scale green space dust distribution. This study selected Euonymus japonicus, a common green space vegetation in Beijing, as the research object. The dust retention capacity, spectral reflectance and leaf area of leaf samples were obtained through outdoor sampling and indoor experiments. The original spectral curve and the first derivative of reflectance before and after dust retention were compared, and the effects of different dust retention on spectral reflectance were analyzed, To explore the band which is highly sensitive to dust retention of leaves. Using the spectral response function, the narrow band spectral reflectance data collected on the ground are transformed into the wide band spectral reflectance data of remote sensing satellite. The regression model of vegetation index ratio and dust retention capacity of corresponding satellite band is established. The regression model with the best fitting effect is selected as the dust retention inversion model. Combined with the green space range extracted from the GF-2 image, the dust retention distribution of Beijing urban green space was obtained using the dust retention inversion model. The spatial autocorrelation model is used to test the spatial aggregation characteristics. The results show that: in the 740~1 870 nm band, the spectral reflectance after dust retention is significantly lower than before dust retention. Dust retention has no obvious effect on the position of the red edge, yellow edge and blue edge but has pronounced effect on the “red edge amplitude” and “red edge area”. EVI index calculated by Sentinel-2 image has the highest correlation with dust retention. The coefficients of determination (R2) of the linear and quadratic regression models are 0.705 and 0.751, respectively. Based on the Sentinel-2 images on April 7, 2021, and June 3, 2021, the distribution trend of green space dust retention in the Beijing urban area is as follows: the city center is higher than the suburbs, the north is higher than the south, and the East is higher than the West. The central, northern and eastern parts of Beijing are prone to dust pollution. The pollution distribution is aggregated and not completely random.