The Method Of Research On Estimation Of Biomass By Synergy Optical And SAR Remote Sensing Data

Forest, covering about 30% of the land area, is the key component of terrestrial ecosystem and also the green treasury that humans depend on. The forest above ground biomass(AGB) is one of important assessment indicators of ecosystem productivity and forest eco-environment. Meanwhile, it’s also the key role in the global carbon, water and energy cycles, which has remarkable significance on the study of global climate change. However, with the rapid development of RS technology, multi-source RS data including multi-optical images, radar and Li DAR images has been used for observing forest, which provides a cheap, fast and convenient way to estimate the large scale AGB and other characters. To make full use of different data for varied information and improve estimating precision, the fusion of multi-source RS data has been tried to obtain AGB, which is the focus of current research and has great development potential.Against the Daxing’anling mountain landforms environment, this paper discussed on the feasibility of estimating the larch forest AGB and other characters in Gen He by the fusion of TM and L-band PALSAR, which has achieved the estimating task of biomass by the 3D Coherent Radar Model(Pol SARpro SIM). This paper shows that combination of optical and microwave RS data based on the 3D Coherent Radar Model can make full use of the advantages of multi-source RS data, which has some applicability for the estimation of biomass and other forest characters. The main findings and conclusions are as follows:1. The leaf area index(LAI) in this paper was estimated from TM multispectral data. The least squares method that combining with Regression Model from different vegetation index extracted from TM data was used to build the joint inversion model for LAI, from which has been improved the accuracy of estimates in a certain extent compared with other methods such as empirical methods and physical model.2. Extend Fourier Amplitude Sensitivity Test(EFAST) for global sensitivity analysis was used in the model, which has simplified the complex input parameters. For the PALSAR(1.27GHz), the sensitivity parameters of Pol SARpro SIM are mostly as incidence angle, tree density, canopy height and the surface roughness for HH polarization with the global sensitivity index over than 0.1.3. The localization of PolSARproSIM model has been done based on the measured dataand information of forest canopy structure from multispectral data, where the parameters used in model consistent with the forest scene on the ground. By using simulated backscattering coefficient, the different scattering mechanisms and the simulated biomass from model directly validates the feasibility of model.4. The method of the biomass and canopy height about the larch forests in northeast China was found, based on the scattering coefficient from the simulation of Pol SARpro SIM. Analysis showed that there’s a better relationship between canopy height and biomass, of which determination coefficient of the regression model reached 0.827. Meanwhile, to get the biomass, the method has been put forward. It found that the estimated canopy height from Pol SARpro SIM(RMSE=2.34 m, MBE=1.82m) could be used in estimating the biomass, of which the accuracy of biomass has been improved to some degree(RMSE=24.67 t/ha, MBE=20.05 t/ha).