关键词: 雷竹林, 地上生物量, Landsat TM遥感数据, 偏最小二乘回归

Abstract:

Based on data collected with Landsat Thematic Mapper (TM), a remote sensing technique, and a field survey, a model established with the partial least squares (PLS) regression was used to estimate aboveground biomass (AGB) of Phyllostachys praecox forest in Lin'an City, Zhejiang Province. Results showed that AGB of individual Phyllostachys praecox was significantly correlated with diameter at breast height and AGB of Phyllostachys praecox forest was significantly correlated with culms density. The predicted accuracy of the PLS model has be improved through PLS-Bootstrap variable selection method, with a root mean square error (RMSE) of 3.45 t·hm^-2. The PLS model is an effective way of estimating AGB of Phyllostachys praecox forest in a large area. Most of Phyllostachys praecox in the forested area had AGB values between 13 and 25 t·hm^-2, and the average AGB density was 19.52 t·hm^-2.

Key words: Phyllostachys praecox forest, aboveground biomass, Landsat TM data, partial least squares regression