Forest Aboveground Biomass Estimation using ICESat/GLAS and Imagery Remote Sensing Data in the Greater Mekong Subregion: 1st result from Yunnan Province, China

摘要

This study aims to develop a forest aboveground biomass (AGB) mapping method in the Greater Mekong Subregion (GMS). Vertical structure of forest parameters of two forest farms in Yunnan province, China were derive using airborne LiDAR system (ALS). Regression models were built between field data of forest AGB and percentiles of canopy height, canopy density which derived from ALS point cloud data. The high accuracy ALS estimated forest aboveground biomass (AGB) were used as training data for building forest AGB estimation model with ICESat GLAS waveform indices. Then the forest ABG was estimated at ICESat GLAS footprint level in the whole province. The regression tree and MAXENT methods were investigated to extend the AGB estimation from GLAS footprint to continuous mapping using imagery remote sensing data of ENVISAT MERIS and EOS MODIS data. The preliminary results showed that: 1) The integrated method based on field measurements, airborne and spaceborne LiDAR data can be used to estimate forest aboveground biomass effectively. 2) The estimation agreed well with inventory based results, and the average difference was about 10%. 3) Both regression tree and MAXENT methods predicted AGB spatial distribution well. 4) These methods will be investigated further and used to the entire Greater Mekong Subregion with more reference training data.