基于时间序列遥感数据的森林火烧迹地提取

Forest burned scars area extraction using time series remote sensing data

森林火灾是当今世界性的重大灾害之一,也是影响北方森林演替的主要因素,其成灾后往往会造成很大的经济损失,并对该地区的森林生态环境产生很大影响。基于2000-2013年8 d合成的MODIS EVI和land surface temperature(LST)时间序列数据,综合应用表征植被生长状态的植被物候期特征指标及全球扰动指数(MGDI),提出了一种森林火烧迹地检测方法,并在大兴安岭及其附近林区进行了应用验证。结果表明,所提研究方法可以有效地提取火烧迹地信息;与基于Landsat的森林扰动监测结果相比,该方法对植被细微的异常变化更加敏感,所提着火点像元也更加集中,在时间尺度上也更加精确和可靠。

Forest fire is one of the major disasters over the world, which is also an important factor influencing seccession of forest in the north. Fire disasters may cause huge economic losses and significantly impact the future forest ecological environment in the region. In this paper, considering the character of the vegetation growth state and the MODIS Global Disturbance Index （MGDI）, we proposed a method for forest burned scars detection based on the 8 -day MODIS Enhanced Vegetation Index （EVI） and Land Surface Temperature （LST）. The proposed method was applied to the Daxing＇ an Mountains and its adjacent forest region. The result indicates that, this approach may effectiyely extract information about burned scar area. Compared with the results extracted from the Landsat disturbance products, this method is more sensitive to subtle abnormal changes in vegetation. The extracted ignition points are more concentrated and more accurate and reliable in the time scale.