摘要
This study presents the utility of remote sensing （RS）, GIS and field observation data to estimate above ground biomass （AGB） and stem volume over tropical forest environment. Application of those data for the modeling of forest properties is site specific and highly uncertain, thus further study is encouraged. In this study we used 1460 sampling plots collected in 16 transects measuring tree diameter （DBH） and other forest properties which were useful for the biomass assessment. The study was carded out in tropical forest region in East Kalimantan, Indo- nesia. The AGB density was estimated applying an existing DBH - biomass equation. The estimate was superimposed over the modified GIS map of the study area, and the biomass density of each land cover was calculated. The RS approach was performed using a subset of sample data to develop the AGB and stem volume linear equation models. Pearson correlation statistics test was conducted using ETM bands reflectance, vegetation indices, image transform layers, Principal Component Analysis （PCA） bands, Tasseled Cap （TC）, Grey Level Co-Occurrence Matrix （GLCM） texture features and DEM data as the predictors. Two linear models were generated from the significant RS data. To analyze total biomass and stem volume of each land cover, Landsat ETM images from 2000 and 2003 were preprocessed, classified using maximum likelihood method, and filtered with the majority analysis. We found 158±16 m^3.ha^-1 of stem volume and 168±15 t.ha^-1 of AGB estimated from RS approach, whereas the field measurement and GIS estimated 157±92 m^3.ha^-1 and 167±94 t.ha^-1 of stem volume and AGB, respectively. The dynamics of biomass abundance from 2000 to 2003 were assessed from multi temporal ETM data and we found a slightly declining trend of total biomass over these periods. Remote sensing approach estimated lower biomass abundance than did the GIS and field measurement data. The earlier approach predicted 10.5 Gt and 10.3 Gt of total biomasses in 2000 and 2003, while the later estimated 11.9 Gt and 11.6 Gt of total biomasses, respectively. We found that GLCM mean texture features showed markedly strong correlations with stem volume and biomass.
应用遥感技术、地理信息系统和野外观测数据,评估了热带森林环境下地上生物量和木材蓄积量。用于模拟森林属性的这些数据具有地理特异性和高度的不确定性,因此,这方面需要开展更多的研究工作。选取了16个试样地带1460个样地,测定树木胸径及其他用于评估生物量的其他森林属性。本实验在印尼加里曼丹东部的热带雨林开展。应用现有的胸径-生物量公式来评估地上生物量密度。估测值在研究区修正的GIS地图上重叠显示,计算各种地被物的生物量密度。用样品数据子集表达遥感方法来形成地上生物量和材积线性方程模型。皮尔森相关统计检验采用ETM条带反射率、植被指数、图像变化图层、主成分分析条带、缨帽变换、灰度共生矩阵纹理特征和DEM数据作为预报值。在显著的遥感数据中形成了两个线性模型。为了分析每块地被物总的生物量和材积量,对2000年到2003年卫星ETM图进行了预处理、最大似然估计法分类和主体分析过滤。遥感方法获得的结果表明:材积量为(158±16)m3·hm-2,地上生物量为(168±15)t·hm-2;而野外测定和地理信息系统估计的结果分别是材积量为(157±92)m3·hm-2、地上生物量为(167±94)t·hm-2。用多个瞬间ETM数据评估了从2000年到2003年间的生物量丰富度动态,结果发现这一时期总生物量呈略微的下降趋势。遥感技术评估的生物量丰富度低于地理信息系统和野外测定的结果。前一种测定方法估计2000年和2003年总生物量分别是10.47Gt和10.3Gt,而后一种则估计11.9Gt和11.6Gt。还发现,灰度共生矩阵纹理特征与材积量和生物量之间存在较强的相关性。图7表9参43。
