中国典型城市PM2.5浓度时空演绎规律及影响因素分析

为探讨空气中ρ（PM_2.5）的空间集聚特征和气候、大气成分变量对空气中ρ（PM_2.5）的影响，利用首批纳入PM_2.5监测的74个城市的ρ（PM_2.5）数据计算Moran's I指数，并选取其中38个典型城市进行计量分析.在基于引力模型的空间权重矩阵基础上，构建面板数据SDM（空间面板杜宾模型）.结果表明:ρ（PM₁₀）、ρ（SO₂）、ρ（CO）、ρ（O₃）、RH（relative humidity，相对湿度）与城市ρ（PM_2.5）呈正相关，而T（temperature，温度）和WS（wind speed，风速）与城市ρ（PM_2.5）呈负相关；ρ（PM₁₀）、ρ（CO）、RH是位于前3位影响城市ρ（PM_2.5）的关键性因素，其总效应分别为0.720 1、0.241 7、0.133 9.地理上邻近城市ρ（PM_2.5）具有明显的外部空间溢出效应，即邻近城市ρ（PM_2.5）每增加10百分点，将导致该地区ρ（PM_2.5）增长6.12百分点.300 km左右是保证PM_2.5区域"联防联控"最佳效果的最大门槛距离，超过该门槛距离，区域"联防联控"的力度和效果会随着距离的增加而逐渐减弱；当门槛距离大于500 km时，ρ（PM_2.5）的空间自相关性不显著.气候变量中，RH和ρ（PM_2.5）呈同方向变化，而T、WS与ρ（PM_2.5）呈反方向变化.研究显示，关注单一地区或单一因素（气候或大气成分）均不能有效控制PM_2.5污染，在保持经济稳定增长的前提下，各地治理PM_2.5应从调整产业结构、优化能源结构、完善防控机制等多个维度共同推进，促使经济增长方式早日从"粗放型"向"集约型"转变. 

Spatio-Temporal Characteristics of PM2.5 and Influence Factors in Typical Cities of China

In order to describe the influence of spatial characteristics, climate and atmospheric composition variables on ρ(PM_2.5) in the air, this paper uses the ρ(PM_2.5) in 74 Chinese cities to calculate the Moran' I index. An econometric analysis of the climate and atmospheric composition data of the 38 typical cities has been made. Based on the spatial weight matrix of gravity model, a SDM model of panel data has been constructed. The results indicate that ρ(PM₁₀), ρ(SO₂), ρ(CO), ρ(O₃) and relative humidity have positive correlations with the ρ(PM_2.5) in the typical cities. However, wind speed and temperature have negative correlations with the ρ(PM_2.5) in typical cities. ρ(PM₁₀), ρ(CO) and relative humidity are the key factors affecting the ρ(PM_2.5), and the total effect of the three factors are 0.7201, 0.2417, 0.1339, respectively; The ρ(PM_2.5) in the geographically neighboring cities has a significant external space spillover effect, which means that an increase of 10% in the ρ(PM_2.5) in neighboring cities will result in an increase of 6.12% in the ρ(PM_2.5) in this area. The maximum threshold distance for PM_2.5 is about 300 km 'joint prevention and control'. The strength and the 'joint prevention and control' effect will gradually become weaker and weaker as the growth in distance. When the threshold distance is longer than 500 km, the spatial autocorrelation of the ρ(PM_2.5) is not significant. The research shows that focusing on single area or single factor (climate or atmospheric composition) cannot effectively control PM_2.5 pollution. With the sustainable development of economics, each city should focus on different perspectives to explore haze treatment plans, such as adjusting industrial structures, optimizing the energy structures and so on. These efforts will transform the mode of economic growth from 'extensive form' to 'intensive mode' in the future.