Generation of a Species-Specific Look-Up Table for Fuel Moisture Content Assessment
Generation of a Species-Specific Look-Up Table for Fuel Moisture Content Assessment
复制标题
DOI：
10.1109/jstars.2009.2014008
复制
发表时间：
2009-02
影响因子：
5.5
通讯作者：
M. Yebra;E. Chuvieco
中科院分区：
文献类型：
--
作者：
M. Yebra;E. Chuvieco
本研究涉及的生成一个物种特定的查找表（LUT）的燃料水分含量（FMC）的检索占主导地位的冬青栎（霍尔姆橡树）的自然区域。干燥Q中观察到的参数组合。ilex样品被用作链接的前景和SAILH辐射传输模型（RTM）的输入，以避免LUT中不切实际的模拟光谱。Terra/MODIS反射率数据，提取了五个样地，主要是Q。ilex进行LUT反演。该反演是基于对LUT中发现的观测和模拟反射率之间的最小相对均方根误差（RMSErho*）的搜索。不同的反演选项进行了测试，以寻找准确估计FMC所需的最佳光谱采样。还研究了估计FMC的计算所需的最小数量的解决方案。检索性能进行了评价，在五个研究地块的FMC值测量。当使用归一化差分红外指数（NDII 6）并选择10个最佳情况作为解决方案时，获得了最准确的FMC估计（RMSE=26.28%）。最后，一个非橡树特定的LUT（通用LUT）以同样的方式使用，以评估是否物种特定的LUT检索FMC更准确。结果表明，种特异性LUT比通用LUT提供更准确的FMC估计。只有当解的个数大于35时，两种查找表的精度才相近。今后的工作将侧重于根据从实地测量和文献中提取的数据，编制适用于更广泛物种的对照表的可能性。
This study involved the generation of a species-specific Look-Up Table (LUT) for the retrieval of Fuel Moisture Content (FMC) in natural areas dominated by Quercus ilex (Holm oak). Parameter combinations observed in drying Q. ilex samples were used as inputs into the linked PROSPECT and SAILH Radiative Transfer Models (RTM) to avoid unrealistic simulated spectra in the LUT. Terra/MODIS reflectance data, extracted over five plots dominated by Q. ilex, were used to carry out the LUT inversion. This inversion was based on the search for the minimum relative root mean square error (RMSErho*) between observed and simulated reflectance found in the LUT. Different inversion options were tested in order to search for the optimal spectral sampling necessary for accurately estimating FMC. The minimum number of solutions required for the calculation of the estimated FMC was also investigated. The retrieval performance was evaluated with FMC values measured at the five study plots. The most accurate FMC estimation was obtained when using the normalized difference infrared index (NDII6 ) and selecting the ten best cases as the solution (RMSE=26.28%). Finally, a non-oak-specific LUT (generic LUT) was used in the same way to evaluate whether or not the species-specific LUT retrieved FMC more accurately. The results showed that the species-specific LUT provided more accurate FMC estimations than the generic LUT. Only when the number of solutions was higher than 35 was the accuracy of the two LUT similar. Future work will focus on the possibility of generating a LUT adapted to a wider range of species based on data extracted from field measurements and literature.