地理研究

地理研究 ›› 2005, Vol. 24 ›› Issue (3): 473-479.doi: 10.11821/yj2005030016

• 论文 • 上一篇    下一篇

热带森林植被生物量与遥感地学数据之间的相关性分析

杨存建1,2,3, 刘纪远2, 黄河1, 许辉熙1, 党承林3   

  1. 1. 四川师范大学遥感与GIS应用研究中心, 成都 610068;
    2. 中国科学院地理科学与资源研究所, 北京 100101;
    3. 云南大学地植物学与生态研究所, 昆明 650018
  • 收稿日期:2004-06-28 修回日期:2004-10-22 出版日期:2005-06-15 发布日期:2005-06-15
  • 作者简介:杨存建(1967-),男,四川成都人,研究员。现主要从事遥感和地理信息系统的应用研究,发表相关 论文多篇。
  • 基金资助:

    国家自然科学基金项目(40161007)、科技部863项目(2002a135230);中国科学院知识创新项目(CX10G-E01-02-03);四川省青年基金项目(03ZQ026-032)资助

Correlation analysis of the biomass of the tropical forest vegetation, meteorological data and topographical data

YANG Cun-jian1,2,3, LIU Ji-yuan2, HUANG He1, XU Hui-xi1, DANG Cheng-lin3   

  1. 1. Research Center of Remote Sensing and GIS Applications, Sichuan Normal University, Chengdu 610066, China;
    2. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China;
    3. Institute of Geobotany and Ecology, Yunnan University, Kunming 650018, Chain
  • Received:2004-06-28 Revised:2004-10-22 Online:2005-06-15 Published:2005-06-15

PDF (PC)

赞

被引次数

130 | 21

摘要：

以我国云南省西双版纳的热带森林为例,对热带森林植被生物量与遥感地学数据之间的相关性进行了分析。首先,利用森林资源连续清查的林业固定样地数据计算出各样地的森林植被生物量,并建立其GIS数据库。然后,对遥感图像进行几何校正,并对遥感图像进行主成分变换、缨帽变换以及植被指数的计算来产生其派生数据。其次,将样地数据、遥感数据及其派生数据,地形和气象数据转换到统一的坐标系和投影下,并将其内插为30米分辨率的格网数据。最后,进行样地森林植被生物量与其遥感地学数据之间的相关性分析。该分析表明,森林植被的生物量与年降雨量和第二主成分在0·01的水平上相关显著,而与中红外植被指数、LANDSATTM5、缨帽变换的亮度、湿度以及第一主成分在0·05的水平上相关显著。其中,与年降雨量的相关性最高,达到0·308;其次是与第二主成分,达到-0·231;再次是与中红外植被指数和LANDSATTM5,其相关系数分别为0·203和-0·201。

关键词: 热带森林植被, 生物量, 遥感地学数据, 相关性分析

Abstract:

This paper analyses the correlations between the biomass of the tropical forest vegetation and LANDSAT TM data, meteorological data and topographical data taking Xishuangbanna of China's Yunnan province as an example. It includes four steps. Firstly, the biomass for each forest sample is calculated by using the field inventory data of each sample. GIS Database is established according to the coordinate of each forest sample. Secondly, the LANDSAT TM images are geometrically corrected by using topographic maps. The derivative data are derived from the LANDSAT TM images by using principal component analysis, tasseled cap transform and vegetation index analysis. Thirdly, the data including Landsat TM data and its derivative data, the topographical data such as DEM and aspect, and the climatic data such as annual mean temperature, annual average accumulative temperature above zero degree, annual average precipitation and humidity are referenced to the same projection and coordination, and interpolated as the grid data with a resolution of 30 m. The Landsat TM data and its derivative data, the topographical data and the climatic data for the samples are achieved by overlay analysis. Finally, the correlation between the LANDSAT TM and its derived data, meteorological data, topographical data and the biomass are analyzed. It is shown that the biomass of the tropical forest vegetation is most obviously correlated with annual average precipitation and the second principal component of the principal component analysis of LANDSAT TM at 0.01 confident level. The correlation coefficients are respectively 0.308 and -0.231. The biomass is obviously correlated with spectral index VI3, Landsat TM5, spectral index such as brightness and humidity of the Tasscap transform, the first principal component at 0.05 confident level. The correlation coefficients are respectively 0.308, -0.231, 0.203 and -0.201.

Key words: tropical forest vegetation, biomass, remotely sensed geographic data, correlation analysis