| Forest biomass is an important source of material and energy for forest ecosystem, which is a basic quantity character of the forest ecosystem. Forest aboveground biomass is important for sustainable forest management and fuelwood supply monitoring, and the natural environment and human activity could influence its development. Alos, to some extent, it can reflect the change of structure and function of forest ecosystem, the influence human disturbance on forest, global climate, all kinds of pollution issues in regional extent. is the basic of the study of NPP and vegetation productivity. Research on forest biomass especially studying the forest biomass in large scale is significant important.In this study, Landsat 8 and ALOS remote sensing data, the measured biomass data and forest survey data were used to build remote sensing models of forest biomass by linear regression analysis and nonlinear regression analysis(neural network). The main research contents include the pretreatment of remote sensing image for reduction the influence from topographic factors; extracted varieties of remote sensing characteristic factors baesd Landsat8 and ALOS based data for biomass estimation; building biomass inversion model. This paper include following results:The mid-subtropics forest has complex structure. The correlation between forest biomass and a single remote sensing factor is lower than the correlation forest biomass and between principal component analysis,which could including more image information, and analysing the correlation between ABS and the variables show that the precision of foreat biomass model would improved by using Tasseled Cap and Texture variables.The model builded by different time and types of remote sensing image had the different precision, but the model getting by combine the two had high precision than either of them. nonlinear regression(neural network) model own high correlation in comparison with linear regression model.With the help of neural network model, the biomass of Jinggangshan can be estimated, the verage biomass was 132.571 t/ha. The overall distribution of forest biomass is highly related to human activities. Due to the impact of humanactivities, the biomass value of traffic convenient area is low than the lower elevation valleys and complex terrain regions. For eight types of forest, the verage biomass from high to low is broad-leaved forest, evergreen, deciduous broad-leaved forest, deciduous broad-leaved forest, coniferous and broad-leaved mixed forest, coniferous forest, bamboo forest, shrub forest and montane elfin forest.The analysis of topographic distribution of forest biomass showed: with the increase of the altitude and slope, forest biomass first increased and then decreased. 550 ~ 1200 meters showed large verage biomass values, with the value 152.005 t/ha. A high correlation exists in biomass and the slope area, the larger the slope area, the higher the total biomass. the average biomass of slightly side is higher than that of shady side as a result of human social and economic activities. |
