| 近20 a黑龙江省PM2.5时空分布变化及驱动力分析 |
| 摘要点击 1550 全文点击 201 投稿时间：2023-11-17 修订日期：2024-03-02 |
| 查看HTML全文
查看全文 查看/发表评论 下载PDF阅读器 |
| 中文关键词 黑龙江省 PM2.5 时空变化 驱动因素 地理探测器 多尺度地理加权回归模型 |
| 英文关键词 Heilongjiang Province PM2.5 spatial-temporal variation driving factors geographical detector multi-scale geographically weighted regression model |
| 作者 | 单位 | E-mail | | 乔璐靖 | 东北林业大学林学院, 森林生态系统可持续经营教育部重点实验室, 哈尔滨 150040 | 1457455430@qq.com | | 栾宜通 | 东北林业大学林学院, 森林生态系统可持续经营教育部重点实验室, 哈尔滨 150040 | | | 曾艳丽 | 东北林业大学林学院, 森林生态系统可持续经营教育部重点实验室, 哈尔滨 150040 | | | 琚存勇 | 东北林业大学林学院, 森林生态系统可持续经营教育部重点实验室, 哈尔滨 150040 | qucy09@nefu.edu.cn | | 陶金涛 | 东北林业大学林学院, 森林生态系统可持续经营教育部重点实验室, 哈尔滨 150040
山东华宇工学院招生办公室, 德州 253034 | |
|
| 中文摘要 |
| PM2.5是衡量空气污染程度的重要指标,研究其时空变化特征和影响PM2.5浓度空间分异的关键驱动因素对于治理大气污染,提升区域空气质量具有重要意义.基于黑龙江省2000~2021年PM2.5遥感数据,采用Theil-Sen Median趋势分析、Mann-Kendall显著性检验和空间自相关,分析PM2.5浓度的时空变化特征,利用地理探测器结合多尺度地理加权回归模型,探究影响PM2.5空间分异的关键驱动因子及其影响程度和作用方向.结果表明,①2000~2021年黑龙江省ρ（PM2.5）均值在22.01~41 μg·m-3之间,2008~2015年PM2.5均值高于《环境空气质量标准》中可吸入颗粒物（粒径≤2.5 μm）二级浓度限值（35 μg·m-3）,2013年为PM2.5浓度变化转折点,总体呈先升后降的变化趋势.冬季是PM2.5污染的高发季. PM2.5浓度空间上呈南高北低的分布格局,高值区常年以哈尔滨市、大庆市及周边地区为主,低值区则分布在大兴安岭和黑河市等北部地区. ②因子探测结果表明,年均气温是影响PM2.5空间分异最主要的驱动因子,其余的关键驱动因子按解释力大小依次为：高程、人口密度、年均风速、土地利用、夜间灯光、年降水量、坡度、年均相对湿度和NDVI.交互探测表明各驱动因子在交互作用后对PM2.5分异性的解释力均大于单一因子,说明影响PM2.5空间分异是各驱动因子共同作用的结果,自然因子间的交互作用比社会经济因子间的作用更加明显. ③不同影响因子对PM2.5的作用呈现明显的空间差异,年均气温、年均相对湿度、人口密度和夜间灯光对PM2.5污染起促进作用,高程、坡度、年降水量、年均风速、NDVI和土地利用对PM2.5污染起抑制作用；PM2.5与各影响因子的作用尺度具有显著差异,年均气温、年均风速和NDVI的影响尺度最小,变量带宽为43；人口密度和土地利用的作用尺度最大,变量带宽为140. |
| 英文摘要 |
| PM2.5 is an important indicator for measuring the degree of air pollution. Studying the space-time variation and the driving factors of spatial heterogeneity is important for controlling air pollution and improving regional air quality. Based on PM2.5 remote sensing data from 2000 to 2021, the Theil-Sen Median trend analysis, Mann-Kendall significant inspection, and spatial auto correlation were used to analyze the characteristics of space-time changes in PM2.5 concentration, and geographical detectors were combined with a multi-scale geographical weighted regression model to explore the key driver factor and its influence and direction of the impact and role of PM2.5 spatial differences. The results showed that： ① The average PM2.5 value of Heilongjiang Province was between 22.01 and 41 μg·m-3 from 2000 to 2021. From 2008 to 2015, the average PM2.5 value was higher than the secondary concentration limit （35 μg·m-3） of the “Environmental Air Quality Standard.” The turning point of the PM2.5 concentration change that occurred in 2013 generally showed the trend of change and then a downward trend. Winter was the high incidence season for PM2.5 pollution. The PM2.5 concentration space was a distributed pattern in the south and north and the high-value zone was mainly based on Harbin, Daqing City, and the surrounding area. The low-value areas were distributed in the northern regions such as the Great Khingan Mountains Region and Heihe City. ② Factor detection results indicated that the average annual temperature was the most important driving factor that affected PM2.5 spatial differences. The remaining key driver factors were in turn： high-end, population density, average annual wind speed, land use, night lights, annual years precipitation, slope, annual relative humidity, and NDVI. Interactive detection showed that the interpretation of PM2.5 points after interaction was higher than a single factor after interaction, indicating that affecting PM2.5 spatial difference was the result of the common effect of each driver factor. The effect of natural factors was more obvious than that of social and economic factors. ③ The effect of different influence factors on PM2.5 had a significant spatial difference. The average annual temperature, average annual relative humidity, population density, and night lighting played a promotion effect on PM2.5 pollution and NDVI and land use played an inhibitory effect on PM2.5 pollution. PM2.5 was significantly different from the action role of various influencing factors and the average annual temperature, annual average wind speed, and NDVI impact scale were the smallest, with a variable bandwidth of 43； population density and land use impact scale were the largest, with a variable bandwidth of 140. |
|
|
|
