Lithological Information in Remotely Sensed Images and Surface Weathering in Arid Regions

Pontual, A (1990). Lithological Information in Remotely Sensed Images and Surface Weathering in Arid Regions. PhD thesis The Open University.



The main objective of this thesis is to understand the factors controlling the amount of lithological information present in remotely sensed images acquired by satellite-borne instruments. This research concentrated on Landsat TM image data of the Oman Ophiolite Complex in the arid Oman Mountains and of two areas in the hyper-arid northern Chilean Andes; a volcanic debris avalanche and a plutonic complex. The synoptic view provided by satellite-derived remote sensing images, and the spectral information beyond the visible wavelengths, facilitate geological mapping of rock type, in poorly accessible regions such as these.
The TM data were enhanced by decorrelation stretching, which is a technique that increases colour saturations on an image with little hue distortion. Another technique employed is the pseudoratio technique, which is a ratioing technique ideal for highly correlated data.
In the Oman, it was possible to establish a consistent colour scheme for recognising the ophiolite lithologies on the enhanced images by comparison of these with published geological maps. Study of the enhanced images revealed many more complexities than were present on the maps. Most of these were verified by fieldwork, allowing confident extrapolation of lithologies from well-known areas into more doubtful regions.
In addition to differentiating rock types, the TM data also revealed details within lithological units suggesting that rather more specific and subtle mineralogical variations were being sensed, as opposed to merely coarse lithological changes. This was evident in the Oman in the subtle variations in the gabbros and in the subdivision of the mantle sequence into uppermost and lowermost units. In Chile also, the field relations within a plutonic complex (Cerro Colorado) are clearly depicted on the enhanced images. Subtle changes are observed within the Colorado anorthosite gabbro that are evidently controlled by large-scale cumulate variations in the labradorite proportions. These variations observed on the images are typically not clear in the field, in both the Oman and Chilean examples, and became explicable only through detailed laboratory investigation of thin sections and visible to NIR spectra of collected samples. The effect of sub-pixel mixing of lithological components was also investigated by study of images of a volcanic debris flow deposit (in the northern Chilean Andes) and of laboratory spectra of samples collected from this area convolved across the TM bands.
The first problem addressed in this study was that of the significance of weathering on the spectral responses of the rock surfaces and whether the observed variations on the images were related to this or to true mineralogical controls. Detailed laboratory analyses of the varnish coatings on the Oman rocks were also carried out in order to characterise their compositions and thicknesses and to determine the influence of these coatings on the spectral responses of the rock surfaces.
It was found that red-weathered surfaces do not mask the characteristic spectrum of the substrate rock and tend to influence only the visible wavelengths, with spectral information from the underlying rock being most apparent in the NIR. Furthermore, significant spectral information over a pixel-sized area is still present in spite of modulation by the vamish coatings and dark red-weathering surfaces, as the masking effect of these are reduced by mixing with other surface types. In addition, it is notable that the deep red-weathered cortices of the Colorado Gabbro in the Chilean Andes have tended to improve the resolution of significant spectral absorption features that influence the image data. This is shown to be the result of magnetite oxidation within the weathering cortices.
It was also investigated whether calibrated pixel data 'spectra' could be used to clarify spectral relationships and how they compare with laboratory spectra convolved across the TM bands. The objective of this was to determine how the lithological units discriminated on the images may be related to mineralogically controlled spectral absorption features. It was found that it is the broad spectral shape that influences the TM data and that this shape is controlled by mineralogical factors. These are delineated for the areas studied. In more general cases, for other TM scenes outside of this study, it would not be possible to identify the lithologies uniquely without supplementary information. It is shown, however, that there is significant mineralogical information present on the images that facilitate an almost complete remotely sensed interpretation of the geology of a poorly- known and region, constraining the lithological types to within a few possibilities.

Viewing alternatives

Download history


Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions