Comparison of four models on forest above ground biomass estimation based on remote sensing

Abstract

The estimating of forest biomass can help us to find out the carbon storage of our country, gradually raise the level of carbon emissions constraints, reduce the concentration of the greenhouse gases in the atmosphere, slow down the global warming process and will be conducive to the formation of resource saving, environment friendly modes of production, lifestyles and consumption patterns, and realize sustainable development. Here, several common estimation methods of forest above ground biomass with remote sensing were used respectively, and were compared. 150 sample data which received by CASA model and 45 variables achieved by remote sensing and meteorological data in Mount Tai scenic area are applied in these models. The results showed that the fittings, estimating precisions and root mean square errors of multi-stepwise regression model, traditional BP neutral network model, RBF neutral network model and K-NN method were 72.1%, 79.2%, 22.805 t·m-2; 79.3%, 84.7%, 20.854 t·m-2; 81.6%, 87.1%, 19.195 t·m-2 and 73.2%, 79.6%, 24.092 t·m-2 respectively. Above all, this paper provides the reference for forest above ground biomass estimation with high precision of the further research. © Springer-Verlag Berlin Heidelberg 2013.