基于Landsat TM/OLI影像的兖州煤田水域面积动态监测

    Dynamic monitoring of water areas in Yanzhou coalfield based on Landsat TM/OLI images

    • 摘要: 为揭示东部高潜水位矿区地表水体的时空演变规律,该文以兖州煤田为研究区,在比较改进归一化差异水体指数法、单波段阈值法、谱间关系法、K-T变换4种水体提取方法的精度及优缺点基础上,采用基于阈值分割的改进的归一化差异水体指数法提取了研究区1990-2014年的水体信息并分析了其时空变化特征。研究结果表明:从1990-2009年,兖州煤田地表水域面积增加了1 696.14 hm2,且2001-2005年水域面积扩张速度较快,年均增长率达14.0%,2009-2014年水域面积减少,年均下降率为3%;煤炭开采是研究区内水域面积变化的主要驱动因子。该文实现了煤矿区地表水体变化的动态监测,并分析了其变化原因,有助于客观与定量地评估煤炭开采的生态累积效应。

       

      Abstract: Abstract: Coal mining leads to some surface subsidence, which changes from farmland, abandoned low-developed and other agriculture land into water area in eastern China with high level underground water, so it is necessary to monitor the surface water change in those areas, which is beneficial to evaluate ecological accumulative effect of coal mining quantitatively and is good for corresponding policy making for coal mining and environment protection. Taking Yanzhou coalfield located in Shandong Province as a study area, the objective of this paper was to find comparatively accurate method to extract open-water information, to tell how much and where the water area changed and to reveal the correlation between water area change and rainfall or coal yield. The research data included 5 Landsat TM/OLI images which were captured on September 11, 1990, September 9, 2001, September 4, 2005, August 30, 2009 and October 15, 2014, coal yield data, coal mine boundary and rainfall data by local surveying or from public authoritative website. Firstly, the author compared different methods of open-water identification, which included modified normalized difference water index (MNDWI) method based on the threshold segmentation, single-band threshold method, spectrum relationship method and K-T transform method by using Landsat TM image captured on August 30, 2009. Accuracy validation with 190 randomly generated points showed the MNDWI method was the best method with an overall accuracy of 94.7% and a Kappa coefficient of 0.899. Then the best method, MNDWI, was used to extract the open-water from 1990 to 2014 with those 5 Landsat TM/OLI images, which were preprocessed to ensure the high quality prior to water information extraction. It showed the water area in Yanzhou coalfield was 551.79, 820.17, 1385.46, 2247.93 and 1927.35 hm2 in 1990, 2001, 2005, 2009 and 2014 respectively. Finally, the temporal and spatial characteristics of surface water were analyzed and its driven force was revealed from the aspects of rainfall and coal mining. The results showed the open-water area comparatively kept an increase trend from 1990 to 2014. It increased rapidly with an annual growth rate of 14.0% from 2001 to 2005, and 12.9% from 2005 to 2009 respectively. From the year of 2009 to the year of 2014, the open-water area decreased with a rate of 3% annually. The increase of the open-water area was mainly concentrated in Dongtan mine, Beisu mine, Xinglongzhuang mine, Nantun mine and Baodian mine. The correlation coefficient between rainfall and the open-water area was -0.537, which showed inter-annual variation of rainfall was not the main reason for the increase of open-water area in the whole study area. The correlation coefficient between coal yield and the surface water area was 0.904, which suggested that coal mining activity was the main driving force of water area change in this study. The total open-water area increases by 5.3% annually in Yanzhou coalfield, which has high positive correlation and middle negative correlation with coal yield and rainfall respectively. The study result will do benefit for understanding how much and what change has happened in similar area and better understanding the drive force for this kind of change. High spatial resolution images and object-oriented classification method will be used, and more driving factors and models for water area changing will be considered in further study to describe and explain the change more objectively and accurately.

       

    /

    返回文章
    返回