Forest Biomass Estimation Based On Microwave Scattering Model And Model Visualization

Forest is the main part of the earth’s biosphere and plays an important role in connecting, the earth system to the cycles of water, carbon and energy. As one of the primary components on terrestrial ecosystems, forest ecosystem covers more than 4 billion hm2 in area that accounts for about 31% the total terrestrial land. Therefore forest has great research significance ecological systems. Radar(active microwave) remote sensing technology has been developed and applied since the 1960 s in many countries. From the 1990 s to now, radar remote sensing has been widely used as an important means of satellite image acquisition and analysis. Microwave imagery, because of its longer electromagnetic wavelength and better capability of penetration in forest canopies, is very useful in quantitative forest studies such as biomass extraction, however, still have lots of problems. The commonly applied physical models are made by C, C++ and FORTRAN program language and are mostly console-programming. Without visual interface windows, users need to manually change model parameters, which makes it difficult and inconvenient for those without programming background. This research aims to establish a visual interface of a forest scattering model and to apply it for forest biomass estimation in a case study of the Greater Hing’an Mountain Forest, northeast China. By integrative use of C# 、 FORTRAN and MATLAB programming languages, forward modeling of the microwave scattering model and its inversion simulation of forest biomass are performed using the ALOS PALSAR radar image. The results of the study as follow:(1)The Graphic User Interface(GUI) of the microwave scattering model is developed using C# language and the programming editor platform of Microsoft Visual Studio 2010 editor. Model algorithm uses FORTRAN language and the programming editor platform uses Intel Visual Fortran 2011. The effective visual interface windows greatly improve the efficiency of the model operation, and protect the kernel algorithm of the model and encapsulation.(2)This study makes full consideration to the L-band microwave signal and its interaction with forest canopy components. Scattering characteristics of different forest parameters are simulated, and the improved model is greatly improved in the visual aspects.(3) Using genetic algorithm optimization toolbox of MATLAB( Genetic Algorithm,GA),model inversion is performed and forest structure parameters are extracted from the ALOS PALSAR image data. Forest biomass is finally estimated.Uncertainties in model inversion deserved further investigation.(4)The ALOS/PALSAR image products have a variety of spatial resolution(6.25-100m) and spatial coverage(35-350 km in width). With the implement of the expected ALOS- 2, its spatial resolution(1-3m) of PALSAR data and temporal resolution(14d) will be greatly improved. These will improve the application of the improved model in this research to carry out quantitative estimation of forest parameters such as tree height and density. These will be the focus of my next-step research for more accurate forest monitoring.