煤炭开发区植被扰动时空效应及影响范围界定:以宁东矿区为例

煤炭开采是一种具有大规模高强度性质的生态扰动行为,会产生区域地表压占与损毁、植被破坏等生态扰动和环境污染问题。西北干旱荒漠区气候干旱,植被稀疏,生态环境本身极其脆弱,宁东矿区是典型的干旱荒漠煤炭开发区,有重要的研究价值。本文选择宁东矿区以及与矿区自然条件相近的自然植被参照区作为研究对象,收集2008—2019年共276期MODIS归一化植被指数(normalized vegetation index,NDVI)长时间序列遥感数据,通过STL时间序列分解、函数主成分分析(function principal component analysis,FPCA)、像元二分模型植被覆盖度计算等方法,从时间维度上分析了矿区周边区域和参照区NDVI均值在研究期内的总体变化趋势,发现随着开采过程的持续,相比自然参照区植被而言,矿区周边植被生长受到了明显胁迫;从空间维度上分析了不同距离缓冲区NDVI数据的相关性特征,在此基础上通过函数主成分分析得到第一主成分函数,确定矿区生态影响范围边界在7 km左右;本研究还在像元尺度分析了植被覆盖度的变化情况,按照变化的程度进行了分级统计分析,发现矿区周边区域虽然NDVI均值总体长期趋势相对稳定,但植被中度和重度退化区域面积明显增加。

Spatio-temporal impact and the scope of vegetation disturbance from coal mining: A case of the Ningdong mining district

Coal mining is a large-scale, high-intensity ecologically disruptive activity causing considerable ecological and environmental problems, such as land occupation, land damage, and vegetation destruction. The northwest region of China has an arid climate, sparse vegetation, and an extremely fragile ecological environment. The Ningdong mining district is a typical arid coal mining area and has important research value. Therefore this area, along with a reference area with similar natural conditions, were selected as the study area. A total of 276 MODIS NDVI remote sensing images from 2008 to 2019 were used for this study. In the temporal dimension, seasonal and trend decomposition using Loess (STL), functional principal component analysis (FPCA), pixel binary model calculation of vegetation coverage, and other methods, were applied to analyze the temporal variation characteristics of NDVI. The results show that the prolonged mining operation caused noticeable stress to vegetation growth around the mining district compared with the reference area. In the spatial dimension, FPCA was used to analyze the correlation of NDVI data across different buffer zones. We found that the boundary of the ecological impact area lies about 7 km from the mining district. In addition, we analyzed the variation in vegetation coverage at the pixel level and performed a hierarchical statistical analysis of vegetation coverage for different variation ranges. It was found that although the average value of NDVI remained stable over time, the moderate to severely degraded vegetation areas have increased significantly.