基于遥感机理模型的多源数据山地森林生物量反演研究

Accurate estimation of forest biomass holds the key in reducing the uncertainties of carbon cycling analysis and providing a reliable proof for ecological studies. Statistical modeling or non-parameter methods are extensively used in previous studies to retrievel forest biomass from massive remote sensing datasets, while studies exploring remote sensing mechanism models for forest biomass estimation are rarely rare. In this proposed study, we will establish a simulated database of coherent radar backscattering (calculated by an improved radar backscattering model) and canopy BRDF (calculated by an optical model based on radiative transfer equation), with three-dimensional (3D) forest scenes under various conditions as the model driver. Based on the simulated radar, optical dataset and the corresponding biomass dataset obtained from the simulated 3D forest scene, a combined look up table would then be constructed for forest biomass retrieval using multi-sensor data. Using look up table method, we will estimate forest biomass of a study site located in Daxinganling forest area at two different growth phases. Biomass estimated with airborne or satellite waveform Lidar data would be used to test biomass estimations generated using look up table method and analyze the biomass dynamics and the correlation with drivers. Through combined use of radar and optical models with multi-source for regional biomass revision at differ growth phases, this study may provide a more effective way for forest biomass estimation and change detection at regional or even larger scale, and thus laid a solid foundation for policy-making.

精确估算森林生物量，可以减少碳循环研究中的不确定性，也可以为其它相关生态研究提供可靠的论据。目前已形成海量的遥感数据集，利用遥感技术估算森林生物量的方法也种类繁多，但多是基于遥感数据的统计分析或非参数方法，而少有从遥感机理模型的角度进行估算森林生物量。本课题将基于相同的三维真实场景得到的植被结构参数作为驱动，在改进现有雷达后向散射模型的基础上，联合辐射传输光学模型建立雷达相干散射和冠层BRDF模拟数据库。以三维真实场景模拟得到的森林生物量作为基准，以模型查找表的方式,发展基于雷达和光学模拟数据库的森林生物量反演方法。利用此反演方法，估算大小兴安岭林区两个时期的森林生物量，结合激光雷达反演结果验证并分析两个时期的生物量差异。本研究通过基于多源遥感数据与机理模型模拟库进行区域多时相的森林生物量反演，为国家或更大尺度上森林生物量估算和变化检测与政府决策提供支持。