Abstract The accurate retrieval of land surface vegetation properties under varying environmental conditions from time series of moderately high spatial resolution satellite observations is challenging. By coupling various Radiative Transfer (RT) models one can describe the soil, vegetation and atmosphere contributions in a “bottom-up” approach and, thereby, simulate top-of-atmosphere (TOA) spectral radiance data comparable to satellite-observed TOA radiances. This makes it possible to retrieve vegetation properties directly from TOA radiances rather than from atmospherically corrected top-of-canopy (TOC) reflectance data. The advantages of this approach are that a separate atmospheric correction of the satellite images is not necessary, and that the anisotropic surface reflection can also be taken into account effectively. In this study, we coupled various RT models, including the brightness – shape – moisture (BSM) reflectance model of the soil, the optical radiative transfer (RTMo) model of vegetation and the ‘MODerate resolution atmospheric TRANsmission’ (MODTRAN) model of the atmosphere, to simulate an annual time series of Landsat satellite TOA radiances observed during a drought episode in California Mediterranean grasslands in 2004. The inversion of this coupled system through an optimization technique against Landsat TOA radiances resulted in direct retrieval of vegetation properties. We accommodated the surface anisotropic reflection in our coupled modeling and also defined a novel anisotropy index to quantitatively express the importance of this phenomenon in satellite image analysis for the first time. Our study showed that the coupled use of RT models was able to accurately reproduce the time series of observed TOA radiances collected under varying soil moisture contents during an extended drought episode. The proposed coupling approach is useful for successful retrieval of vegetation properties from time series of satellite TOA radiance data to produce maps of land surface properties and to monitor vegetation properties variations in an operational straightforward way. The approach can also be easily adapted for conducting multi-sensor time series studies, creating a much denser temporal sampling than would be possible for separate single sensors.

中文翻译：

---

使用耦合辐射传输模型从干旱事件期间 Landsat TOA 辐射度的年度时间序列中检索地表特性

摘要 从中等高空间分辨率卫星观测的时间序列中准确反演不同环境条件下的地表植被特性具有挑战性。通过耦合各种辐射传输 (RT) 模型，人们可以以“自下而上”的方法描述土壤、植被和大气的贡献，从而模拟与卫星观测的 TOA 辐射率相当的大气顶 (TOA) 光谱辐射率数据. 这使得可以直接从 TOA 辐射而不是从大气校正的冠层顶部 (TOC) 反射数据中检索植被属性。这种方法的优点是不需要对卫星图像进行单独的大气校正，并且还可以有效地考虑各向异性表面反射。在这项研究中，我们耦合了各种 RT 模型，包括土壤的亮度 – 形状 – 水分 (BSM) 反射模型、植被的光辐射传输 (RTMo) 模型和大气的“中等分辨率大气传输”(MODTRAN) 模型，以模拟在 2004 年加利福尼亚地中海草原干旱期间观测到的 Landsat 卫星 TOA 辐射度的年度时间序列。通过针对 Landsat TOA 辐射度的优化技术对该耦合系统进行反演导致直接检索植被特性。我们在耦合建模中适应了表面各向异性反射，并定义了一个新的各向异性指数，以定量表达这种现象在卫星图像分析中的重要性。我们的研究表明，RT 模型的耦合使用能够准确地重现在长期干旱期间在不同土壤水分含量下收集的观测到的 TOA 辐射的时间序列。所提出的耦合方法可用于从卫星 TOA 辐射数据的时间序列中成功检索植被特性，以生成地表特性地图并以操作简单的方式监测植被特性变化。该方法还可以很容易地适用于进行多传感器时间序列研究，创建比单独的单个传感器更密集的时间采样。所提出的耦合方法可用于从卫星 TOA 辐射数据的时间序列中成功检索植被特性，以生成地表特性地图并以操作简单的方式监测植被特性变化。该方法还可以很容易地适用于进行多传感器时间序列研究，创建比单独的单个传感器更密集的时间采样。所提出的耦合方法可用于从卫星 TOA 辐射数据的时间序列中成功检索植被特性，以生成地表特性地图并以操作简单的方式监测植被特性变化。该方法还可以很容易地适用于进行多传感器时间序列研究，创建比单独的单个传感器更密集的时间采样。