摘要
以江西省兴国县为研究区域,基于不同时相的LandsatETM+地面反射率图像,计算了RS、NDVI和RSR3种植被指数,并与野外观测的叶面积指数(LAI)数据建立相关关系,从而进行了LAI的反演研究.研究发现,对于针叶林地区,一月份图像也可用来反演LAI,只是预测值较五月偏低.而同一时相的原始图像和反射率图像的反演结果表明,去除传感器自身和大气辐射影响的地面反射率图像,更能真实地反演地表植被覆盖度.此外,在研究区森林覆盖度较高,林种较单一的情况下,RSR同LAI的关系比其他植被指数的相关性好,反演的精度也较高.
<Abstrcat>Leaf area index (LAI) is a parameter of the vegetation structure, and is important for a quantitative analysis of many physical and biological processes related to the dynamic change of vegetation and its effects on carbon cycle, hence the change of the global environment and climate. In this paper, LAI is estimated in Xingguo County based on the correlation between the field-measured LAI and the vegetation indexes (VI). After making the geometric and atmospheric corrections of the asynchronous high-resolution Landsat ETM+ images, three VIs (SR, NDVI, RSR) are derived, and their separate correlations with LAI are investigated. According to the analysis of non-liner relationships between the VIs and LAI, it was found that the precision of the retrieved LAI was high, especially when using the atmospheric-corrected reflectance image. So, it is feasible to retrieve LAI in great region by using the remote sensing data. The correlation between RSR and LAI was higher than that between LAI and other two VIs. It is because the forest coverage of Xingguo county is high, and the forest type is not complicated. Moreover, for coniferous forest, though there is defoliation in winter, the LAI could still be estimated by using the remote sensing data due to the lack of grass and shrub. In general, the predicted value in January was lower than that in May, which was consistent with the reality.
出处
《南京大学学报（自然科学版）》
CAS
CSCD
北大核心
2005年第3期253-258,共6页
Journal of Nanjing University（Natural Science）
基金
加拿大CIDA项目"中国碳循环研究"
关键词
遥感
森林
叶面积指数
植被指数
大气订正
remote sensing,coniferous forest,leaf area index,vegetation index,atmospheric correction
