Hymap images as a base for geologic interpretation and gypsum variation detection in Carboneras, southern Spain

Abstract

The research question for this study is: is it possible to create a map of an area that can provide information of the geology of that area using remote sensing? The aim of this study is to be able to make an interpretation of the geology around the area of Carboneras, southeastern Spain using remote sensing. The primary objective of this study is to create a surface mineral map of an area in the vicinity of Carboneras, southern Spain to be indicative for the underlying geology and to be used as a cheap precursor to a geologic map. For this purpose remote sensing was used. The inherent problems arising from using such methods were then subsequently analyzed. The secondary objective was to determine chemical variations within one of the layers on this surface mineral map (gypsum), along with its inherent problems. The Carboneras area is dry with little vegetation, and is dominated by Late Miocene marls, Messinian gypsums and volcanics, and the area is intersected by the Carboneras fault. Creating a geologic map using remote sensing requires interpreting a surface mineral map as being representative of the underlying geology. The best way of acquiring an image-based surface mineral map from this particular HyMap image is to first cut out the piece of the HyMap that contains the GPS locations taken from the field and then filter out the vegetation using an NDVI operation to create a base map. Then subsequently apply a spectral angle mapping (SAM) operation on this map using ROIs based on GPS locations taken in the field. The best way of acquiring a rock sample-based surface mineral map is to perform SAM on the base map using the rock sample spectra from the rock samples acquired in the field. The gypsum variation mapping proved to be inconclusive based on the limited resources at hand. Creating a surface mineral map of the Carboneras area is thus possible, but extensive ground truth knowledge is needed to produce maps that best resemble reality. Encountered problems for both the creation of the geologic map and the gypsum mapping are mainly of the technical sort. Some methods are better suited to take advantage of particular sets of data than others, and for this HyMap the spectral angle mapping (SAM) method works best for the creation of the geologic map, and PCA works best for the gypsum variation mapping.