四川盆地臭氧浓度空间分异及驱动因子研究

为揭示四川盆地臭氧（O₃）的空间分布格局及其驱动因子，通过对2015~2017年四川盆地18城市82个国控环境监测站的O₃浓度观测数据，使用空间自相关，空间热点探测及地理探测器等地统计方法分析研究.结果表明：2015~2017年四川盆地O₃浓度总体呈现上升趋势（由2015年的（79.95±18.82） μg/m³上升到2017年的（88.64±11.67）μg/m³），污染态势逐年加重.盆地中西部成都，资阳，雅安等城市O₃浓度最高，且高值区范围逐年扩大.O₃浓度的空间分布呈现出显著的聚集性规律（空间正相关，Moran's I指数大于0），且年聚集区（H-H聚集区或L-L聚集区）主要呈现为H-H聚集区分布在盆地中西部，L-L聚集区分布在盆地东南部.此外，年聚集区的年际变化与浓度变化态势基本一致，即O₃浓度上升（降低）的区域转变为H-H聚集区（L-L聚集区）.基于地理探测器定量分析了二十个社会经济及自然驱动因子对O₃浓度空间分异的影响，结果显示社会经济因子占主导作用，其中城建用地占区面积比重（驱动值，q=0.5734），人口密度（q=0.5479）的驱动作用最为突出，自然因子中年降水量（q=0.4592）具有最显著的驱动作用；地理探测器双因子交互对O₃浓度空间分异的影响有明显的交互增强作用（每个单因子q值平均增大了1.5~2.1倍），且交互均值（即各因子交互q值的均值）与最大交互值逐年增大.在使得交互值达到最大驱动值的双因子中，出现频率最高的是人口密度（7次）和工业烟粉尘排放量（7次）.

Spatial differentiation and driving factors of ozone concentration in Sichuan Basin

In order to reveal the spatial distribution of ozone (O₃) and its driving factors in Sichuan Basin, the geostatistical methods such as spatial auto-correlation, spatial hot spot detection and geographic detector were used to analyze O₃ concentration data collected form 82 national control environmental monitoring sites in 18 cities. The results showed that there was an overall upward trend (from (79.95±18.82)μg/m³ in 2015 to (88.64±11.67)μg/m³ in 2017) of O₃ concentration in Sichuan Basin. The O₃ concentration reached highest in the midwest of the basin, including Chengdu, Ziyang, Yaan. And the heavy polluted area expanded as the years progressed.A significant annual clustering pattern (spatial positive auto-correlation, Moran's I was greater than 0) of the distribution of O₃ concentration was showed in Sichuan Basin, in which the H-H clustering areas in the Midwest and the L-L clustering areas in the Southeast of the basin. Besides, the variation of the annual clustering areas were nearly accordant with the variation of O₃ concentration, that is, where the O₃ concentration raised (decreased) were transformed into H-H (L-L) clustering areas. The driving force of 20 socio-economic and natural factors to the spatial variation of O₃ concentration was quantitatively analyzed by using geographic detector. It was found that the socio-economic factors provided a stronger drive force, especially the proportion of urban construction land (driving value, q=0.5734) and the population density (q=0.5479). In addition, the annual precipitation (q=0.4592) also had a significant effect on the distribution of O₃ concentration. The detection of interaction based on geographic detector showed that there was a significant interactions on the spatial variation of O₃ concentration, which made the q value of each factor increased by 1.5~2.1times on average. In addition, the average and maximum values of interactive q of each factor both increased year by year. Among the interaction factors that maximized the q value, the population density (7times) and the industrial dust emission (7times) were highest frequency.