Estimating characteristics of forest structure using vegetation indices from SPOT data in a secondary forest of Nanjenshan, Southern Taiwan. [Chinese]

摘要

Integrating remote sensing and field data can be used to effectively monitor forest structure on a large scale. Two SPOT images acquired in humid and dry seasons were used, and a normalized-difference vegetation index (NDVI) and normalized-difference water indexes (NDWI) were derived from those 2 images. In this paper, we evaluated relationships between forest structure and vegetation indices, and then use parameters of forest structure (height, diameter at breast height, basal area, and volume) as dependent variables and vegetation indices as independent variables to construct an estimative linear model in a secondary forest at Nanjenshan, southern Taiwan. The results showed correlations of the NDWI with those forest structure variables were higher than those of the NDVI, and the NDWI in the dry season was sensitive for detecting dense canopies in this tropical forest. Furthermore, the linear models between forest structure and vegetation indices found that the NDWI in the dry season showed the best ability to interpret diameter at breast height. As a result, these derived vegetation indices can be valuable auxiliary tools for forest resource inventories, and helped us effectively estimate characteristics of forest structure on a large scale.