若尔盖亚高山草甸地上生物量与植被指数关系研究

为探讨亚高山草甸地上生物量和植被指数的关系，更好服务于草地生态建设，论文利用2008年7月覆盖若尔盖地区的TM影像，分别建立了7种植被指数（NDVI、RVI、DVI、SAVI、MSAVI、PVI、GVI）与地上生物量的线性和4种非线性（二次多项式、三次多项式、对数、幂函数）回归模型。研究结果表明植被指数（NDVI、DVI、SAVI、MSAVI、PVI、GVI）与地上生物量模型表现出三次多项式回归模型最优，再次是二次多项式模型、线性模型，相对较差的是指数模型；而基于RVI的地上生物量模型表现为指数模型最优，其次为三次多项式模型、二次多项式模型、线性模型。分析表明，基于RVI的地上生物量幂函数模型的模拟效果最好，复相关系数R²=0.817 7，精度检验结果表明该模型的平均误差为6.80%，拟合精度达93.20%，根据此模型模拟出若尔盖县草地地上生物量分布图，表明该县草地生物量东南部较高而西北部较低。

Study on Pertinence between the Vegetation Indexes and the Aboveground Biomass of Sub-alpine Meadow of Zoige County

To discuss the pertinence between the vegetation indexes and the aboveground biomass of sub-alpine meadow of Zoige(Ruoergai) County and serve better for the ecological improvement, the statistical analysis was conducted between seven vegetation indexes(NDVI,RVI,DVI,SAVI,MSAVI,PVI and GVI) of Zoige area in July 2008 and the field survey data. Linear and non-linear (second-degree polynomial, cubic polynomial, logarithm and exponential) models were established respectively between the seven indexes and the aboveground biomass. The results showed that, for the six vegetation indexes (NDVI,DVI,SAVI,MSAVI,PVI and GVI) and the aboveground biomass models, the cubic polynomial model was the best, followed by the second-degree polynomial model, linear model and exponential model; and for the RVI-biomass model, the exponential model was the best, followed by the cubic polynomial model, second-degree polynomial model and linear model. Analysis showed that the polynomial model based on RVI-biomass was the best model, with the multiple correlation coefficient R~2 being 0.8177; the precision test of the model indicated that the average error was 6.80%; the rating precision was 93.20%. The simulated biomass based on this model showed that the Zoige grassland biomass was higher in the southeast than in the northwest.