Abstract Remotely sensed hyperspectral datasets were integrated with petrographic data to map the distribution of sedimentary rocks in order to infer grain size variations within the siliciclastic, Paleoproterozoic Bresnahan Group in Western Australia. Finer sandstones have greater compositional variation compared to coarser sandstones, with higher modal proportions of mudstone intraclasts, K-feldspar, and muscovite laths, and lower modal quartz. Partial Least Squares modelling indicates a link between visible to shortwave infrared spectral data collected on samples and their average grain size. Specific electronic absorptions are important for defining this relationship, particularly in the blue-green visible spectra, wavelengths related to water or hydroxyl bonds, along with certain wavelengths in the shortwave infrared (SWIR) spectra. The results from field samples were compared to airborne hyperspectral datasets which showed broad scale regional variations in composition could be discerned from processed white mica abundance maps and images that use spectral bands 2350 nm and 2200 nm. The remotely sensed compositional variations are relatable to changes in sandstone grain size. Diagenetic alteration and surface weathering may influence the absorption spectra, although the effect of the latter can be reduced by using processed white mica abundance maps. Overall, the results show that airborne hyperspectral datasets, particularly the SWIR part of the spectra, can be useful to remotely map spatial compositional variations associated with sandstone grain size. This shows the method can be used to remotely map siliciclastic successions in other sedimentary basins by aiding recognition of grain size trends and significant stratal surfaces associated with changes in composition. However, the important wavelengths may differ due to other controls on sandstone composition such as uplift history of the source region, changes in sediment dispersal, and transport mechanisms.

中文翻译：

---

使用光谱数据集推断砂岩粒度：来自西澳大利亚 Bresnahan Group 的示例

摘要 遥感高光谱数据集与岩相数据相结合，绘制沉积岩的分布图，以推断西澳大利亚古元古代布列斯纳汉硅质碎屑岩群内的粒度变化。与较粗的砂岩相比，较细的砂岩具有更大的成分变化，泥岩内碎屑、钾长石和白云母板条的模态比例较高，而模态石英的模态比例较低。偏最小二乘模型表明在样品上收集的可见光到短波红外光谱数据与其平均晶粒尺寸之间存在联系。特定的电子吸收对于定义这种关系很重要，特别是在蓝绿色可见光谱、与水或羟基键相关的波长以及短波红外 (SWIR) 光谱中的某些波长中。将现场样本的结果与机载高光谱数据集进行比较，结果表明，可以从使用 2350 nm 和 2200 nm 光谱带的处理过的白云母丰度图和图像中辨别出成分的大范围区域变化。遥感组成的变化与砂岩粒度的变化有关。成岩作用和地表风化可能会影响吸收光谱，尽管后者的影响可以通过使用处理过的白云母丰度图来减少。总体而言，结果表明，机载高光谱数据集，尤其是光谱的 SWIR 部分，可用于远程绘制与砂岩粒度相关的空间成分变化图。这表明该方法可用于通过帮助识别粒度趋势和与成分变化相关的重要地层表面来远程绘制其他沉积盆地中的硅质碎屑层序。然而，重要的波长可能因砂岩成分的其他控制而不同，例如源区的抬升历史、沉积物扩散的变化和输送机制。