欢迎访问《生态学杂志》官方网站,今天是 分享到:

生态学杂志

• 研究报告 • 上一篇    下一篇

基于最大熵模型的神农架林区华山松大小蠹灾害遥感监测

马望1,2,房磊1,方国飞3,于跃4,张旭3,杨健1*
  

  1. (1中国科学院沈阳应用生态研究所, 沈阳 110016;  2中国科学院大学, 北京 100049;  3国家森林局森林病虫害防治总站, 沈阳 110034;  4沈阳师范大学, 沈阳 110031)
  • 出版日期:2016-08-10 发布日期:2016-08-10

Mapping the infestation of Dendroctonus armandi in Shennongjia forested region using Landsat and MaxEnt model.

MA Wang1,2, FANG Lei1, FANG Guo-fei3, YU Yue4, ZHANG Xu3, YANG Jian1*#br#   

  1. (1Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang 110016, China; 2University of Chinese Academy of Sciences, Beijing 100049, China; 3 General Station of Forest Pest Control, State Administration of Forestry, Shenyang 110034, China; 4Shenyang Normal University, Shenyang 110031, China).
  • Online:2016-08-10 Published:2016-08-10

摘要: 掌握森林病虫害的发生范围和危害程度,对于森林管理部门制定及时、有效的防治决策至关重要。本研究以2014年湖北省神农架林区华山松大小蠹(Dendroctonus armandi)灾害为背景,以野外调查数据、多光谱陆地资源卫星影像(Landsat)和数字高程模型(DEM)为基础数据源,结合最大熵(MaxEnt)模型和迭代阈值分割算法,提出了适用于复杂林区的森林病虫害遥感监测方法(MaxEntSegmentation),实现了神农架林区华山松大小蠹灾害空间分布范围和灾害程度的专题制图与精度评价。同时,为衡量所提出方法对于灾害程度评估的可靠性与准确度,本文还与传统光谱指数分析法进行了对比研究。结果表明:结合遥感光谱指数、海拔、坡度及有效太阳辐射等环境因子构建的MaxEnt模型能够较为准确地监测华山松大小蠹灾害发生范围,受试者工作特征曲线下面积(AUC)值为0.938;当分类类型包括健康、轻度和重度时,MaxEntSegmentation法分类精度最高达73.68%,明显高于传统光谱指数分析法(64.47%),表明该算法能够提高森林虫灾监测精度,适合用于植被类型多样、地形复杂林区的病虫害遥感监测。

关键词: GIS, 人口密度, 人口空间分布, 阴影长度法

Abstract: Accurate spatial information on the location and extent of the forest pest infestation is important for the manager to take prompt and effective preventative measures. In 2014, Pinus armandii in the Shennongjia Forestry District was largely attacked by Dendroctonus armandi, a typical tree trunkboring pest. In this study, we mapped the pest infestation based on the forest inventory data, Landsat images and DEM products. We proposed a novel method that employed a MaxEnt model and iteration threshold segmentation algorithm (MaxEntSegmentation) for this purpose. In order to evaluate reliability and accuracy of the proposed method, the traditional spectrum index analysis algorithm was also carried out and its performance was compared. The results showed that the MaxEnt model was capable of accurately mapping the infested area using spectral indices, elevation, slope, potential solar radiation, with the AUC as high as 0.938. MaxEntSegmentation algorithm had higher overall classification accuracy (73.68%) compared with the traditional spectral index algorithm (64.47%) when three classification classes (health, lowseverity infestation, and highseverity infestation) were included. The results suggest that this proposed algorithm can improve the accuracy of pest detection and is suitable for mapping forest pest infestation in areas with mixed forest stands and variable terrains.

Key words: population density, shadow length method, GIS., spatial distribution of population