Abstract
To mitigate the impacts of global warming, afforested plantations have a considerable amount of wood stock volume with environmental benefits for carbon sequestration. Applying remote sensing technology linked to ground plot data has been an important international topic to precisely measure wood volume in large areas of afforestation. This study focused on Fraxinus griffithii (Fg) plantations, owned by Danong and Dafu Farms of the Taiwan Sugar Company, in Guangfu Township,Hualien County, Taiwan. The soil-adjusted vegetation index (SAVI) from a FORMOSAT-2 (FS2) satellite image acquired on February 13, 2009 was classified into 5 separate grades to perform a stratified purposive sampling scheme for the setting up of 60 spatially independent ground plots. In each plot, the Fg wood volume (VOL) was determined by a field survey of each tree, and 2 leaf area indices (LAI4 and LAI5, i.e., LAIs) and a canopy closure index (CCI) were measured by utilizing gap light analyzer (GLA) method. SAVI and LAIs or CCI were adopted as independent variables in curve regression analyses to establish optimal regressions on the VOL estimation. In addition, the SAVI was used in curve regression analyses to estimate LAI and CCI. The advantage of these 2 optimal equation sets was saving time and labor in the field for ground plot surveys and VOL mapping. Results showed that the optimal regressions could explain more than 94% of the variation of the estimations. Linking SAVI derived fromthe FS2 image or ground measurements of LAI5 or the CCI to the VOL from ground plots is well suited to infer the VOL, LAIs, or CCI of Fg plantations. Using these optimal regressions has the benefit for mapping the VOL, CCI, and LAIs by providing their distributions of spatial heterogeneities. In the future, the time and labor required to investigate, estimate, and map of the Fg stand volume should be greatly reduced by using the LAI calculated from the GLA and linking it to multiphase sampling derived from the SAVI.