基于SPOT5和PCA的柞树林地上碳储量的估测模型

Estimation Model of Aboveground Forest Carbon Storage in Xylosma racemosum Forests Based on SPOT5 and Principal Component Analysis

以北京市延庆县柞树林为研究对象,利用森林资源二类调查数据和2004年SPOT5遥感影像,选取SPOT5数据的4个单波段,提取差值植被指数(DVI)、比值植被指数(RVI)、归一化植被指数(NDVI)等3种植被指数以及海拔、坡度、坡向和郁闭度共11个遥感及样地因子,提取这11个因子的主成分,建立基于主成分分析的多元线性回归模型估测碳储量。结果表明:模型经方差分析以及相关性检验,达到显著相关水平,相关系数R=0.829,可用于柞树林地上部分碳储量估测。对30个独立样地进行配对样本t检验,结果达到显著相关水平,相关系数R=0.850,地上部分碳储量估算值为27.19 t·hm-2,模型估测精度可达到92.73%。

A study of remote sensing estimation of aboveground carbon storage was performed in Xylosma racemosum forests in Yanqing County of Beijing using forest resource inventory data and SPOT5 images in 2004. A total of 11 factors, including 4 multi-spectral bands （B1, B2, 133 and B4）, 3 types of vegetation indexes （DVI, RVI, NDVI） as well as altitude, slope, aspect and canopy closure, are analyzed by principal component analysis. Then a multiple linear regression model of forest carbon storage was set up based on principal component analysis and SPOT5 images. Result shows that the correlation coefficient （R） is 0.829, with a significant level of p〈0.01. The model is suitable for the estimation of aboveground carbon storage in X. racemosum forests. A t-test for 30 independent sample plots shows that the correlation coefficient （R） is 0.850. The average aboveground carbon storage in X. racemosum forests is 27.19 t./hm^2, and the accuracy of the regression model of carbon storage is 92.73%.