摘要
利用2007至2011年雷达影像时序特征监测卧龙和草坡自然保护区森林动态变化。即首先将16景PALSAR数据按年度分成4组;提取时序影像特征,结合J48数据挖掘方法确定决策树实施4期森林精细成图,总体分类精度和Kappa系数分别为94.31%和0.89;然后对4期森林分类图进行统计分析,结果发现震后研究区森林覆盖率从56.66%下降至48.11%,降幅为8.55%。森林覆盖震后第二年即开始恢复,截止2011年3月,森林覆盖率提高到48.52%,新增面积为9.19km2,占受损面积的13.24%。研究表明:2008年汶川地震对大熊猫栖息地的森林造成损害,然而通过自然恢复和人工干预,目前栖息地森林覆盖正逐渐恢复。
Time series features of SAR backscattering in the period of 2007 to 20011 were used to monitor the change of forest coverage in Wolong and Caopo Nature Reserve in Sichuan Prov. China. Firstly, PAL- SAR data were separated into 4 groups;then,a decision tree approach was derived using the J48 algorithm for the forest fine mapping,resulting in the overall classification aecuracy(OA)and kappa eoeffieient(KC)e- qualing to 94.31% and 0.89, respectably.After that, four time-series maps of forest coverage in study area were obtained in the observation period of 2007-2011.Change detection demonstrated that the proportion of the forest coverage was reduced from 56.66% to 48.11% (that is 8.55% declining)caused by the Wenchuan Earthquake. However, the ecological system was recovering since the second year of the event, such as the forest coverage increased up to 48.52% (equals to a recovered square of 9.19 km2 that is amounting to 13.24% of the damaged area).This study implied that Wenchuan Earthquake could be a negative impact on the deforestation of the study area; nonetheless the ecological system is recovering nowadays jointly con- tributed by the combination of natural and anthropogenic impactors.
出处
《遥感技术与应用》
CSCD
北大核心
2017年第6期1100-1106,共7页
Remote Sensing Technology and Application
基金
中国科学院"百人计划"项目(Y5YR0300QM)
国家重点研发计划项目(2016YFB0501502)资助