灌木扩张背景下的干旱/半干旱草原灌草混合景观植被盖度遥感反演研究

The global warming and climate change has been resulting in significant alterations in arid and semi-arid grasslands. Shrub expansion (also called shrub encroachment or shrub invasion) into arid and semiarid grasslands or savannas has been occurring around the world. It presents a serious issue to the earth system because the consequence is not only for the natural environment but also for the socio-economic system. In order to reveal the process and mechanism of shrub extension, and to manage the soil erosion and grassland degradation, it is important to identify the shrub, estimate the coverage of the shrub, and monitor the change of the shrub in the grass-shrub mixture landscape in the grassland. In this research, we select the grassland at Xilin Gole, Inner Mongolia, where the extension of Caragana microphylla is very popular, as the study area to study the methodology of extracting shrub and estimating the vegetation coverage using moderate spatial resolution remotely sensed data. The methodology includes 1) field measurement of vegetation parameters to build the canopy 3D model and dynamic spatial model of grass-shrub mixture landscape, 2) based on a 3D radiative transfer model, developing the LAI-coverage synchronous inversion technology and grass-shrub separation inversion algorithm to extract shrub from grass-shrub mixture and estimate its coverage, and 3) comparison of the inversion results among the experienced modeling, spectral mixture analysis and radiative transfer model to find out the limitations and conditions of the methods. The results of this research can help to monitor and understand the shrub extension in the context of global warming, and can also provide the guidelines for grassland management to prevent shrub extension and grassland degradation.

在全球变化背景下，干旱/半干旱草原区域的灌木扩张对当地生态系统产生了重大影响，研究灌草混合植被景观的灌木识别、提取植被盖度及其变化信息对于揭示灌木扩张的发展态势和机理、治理土壤侵蚀及草原退化具有重要意义。本研究拟选取内蒙古锡林郭勒典型草原的灌木扩张区域，通过实测盖度数据和植被冠层多角度光谱测量，根据植被生长周期，建立动态灌草混合植被冠层场景模型及3D辐射传输模型；同步考虑LAI与植被盖度对光谱的综合影响，利用Landsat ETM+及MISR数据，发展LAI-植被盖度同步反演及灌草分离反演技术，实现对灌木的识别并估算其盖度；对辐射传输模型、经验模型及混合像元分解法估算植被盖度进行比较研究，揭示各种方法的适用范围和条件。该研究可为全球气候变暖背景下草原灌丛化的监测和模拟提供科学依据，为防治灌木扩张和治理草原退化提供指导。

在全球变化及人类活动综合影响下，干旱和半干旱草原的灌丛化过程已成为全世界广泛关注的生态问题，对草原生态系统及其经济发展造成重大影响，研究灌草混合植被景观的灌木识别、提取植被盖度及其变化信息对于揭示灌木扩张的发展态势和机理、治理土壤侵蚀及草原退化具有重要意义。本项目选取内蒙古锡林郭勒典型草原的灌木扩张区域，通过实测盖度、LAI、灌木结构等数据和光谱测量，根据植被生长周期，建立动态灌草混合植被冠层场景模型及3D辐射传输模型；在利用非线性混合像元分解反演灌木盖度时，针对灌木结构参数对多次散射项的影响尚不明确的问题，通过3D辐射传输模型FLIGHT模拟灌木（小叶锦鸡儿）的各种结构参数及覆盖度，研究了灌木参数对多次散射的影响及其效应，为非线性混合像元分解反演灌木盖度提出了理论依据；针对非线性混合像元分解中存在的灌木端元不纯的问题，提出端元提纯的新算法，利用Landsat影像对研究区的灌木盖度进行了反演；本研究还针对较大区域的灌木盖度反演，利用粗分辨率、多时相的MODIS数据，提出了基于生长状况和物候期指标的小叶锦鸡儿遥感识别方法，为区域尺度的小叶锦鸡儿识别和盖度反演提供了有效方法。