Structural geological mapping for mineral and hydrocarbon exploration

Multispectral, hyperspectral, and SAR imagery can reveal lithological contacts, alteration halos, and structural lineaments across terrain that would take years to map on foot. The method is powerful and genuinely cost-reducing, but ambiguity between tectonic and non-tectonic features demands disciplined ground-truth before exploration capital is committed.

Sensors

ASTER (Terra): 14 bands from VNIR to TIR; SWIR bands 4-9 at 30 m resolution discriminate clay minerals (kaolinite, illite, alunite), carbonates, and silica by their diagnostic absorption features. TIR bands at 90 m resolve silicate versus carbonate lithologies. SWIR detector failed in 2008; archive data to that date remains the primary resource for mineral mapping.
Sentinel-2 MSI: 13 bands, 10-20 m resolution, 5-day revisit at the equator. Bands 11 and 12 (SWIR, 1610 nm and 2190 nm) give a coarser proxy for hydroxyl-bearing minerals and ferric iron. Adequate for broad lithological discrimination and vegetation-free contact mapping; cannot resolve individual mineral species the way ASTER SWIR can.
Sentinel-1 SAR (C-band): 6-12 day revisit, IW mode at 5x20 m ground range resolution. Backscatter texture and intensity differences reveal structural lineaments, fold limb orientation, and differential surface roughness between lithological units. Cloud-independent and illumination-independent, which matters in persistently overcast or equatorial terrain.
PRISMA (ASI): Hyperspectral imager, 30 m spatial resolution, 400-2500 nm range across 239 bands. Enables full spectral unmixing for mineral abundance mapping. Tasked on request; archive is growing but coverage of specific exploration targets is not guaranteed. Published studies have used PRISMA for alteration mapping in the Andes and North Africa.
EO-1 Hyperion (archive): 220 bands, 30 m, 7.5 km swath. Decommissioned in 2017 but the USGS archive provides hyperspectral coverage of many historical exploration targets. Useful for method development and comparison against ASTER classifications.

What spectral physics actually gives you

Clay minerals, carbonates, and sulphates each carry diagnostic absorption features in the shortwave infrared between roughly 2000 nm and 2500 nm. ASTER's six SWIR bands were designed to straddle these features, which is why ASTER band-ratio composites (particularly the alunite index, the kaolinite index, and the carbonate index derived from ratios of bands 4, 6, 7, and 8) became standard tools in epithermal gold and porphyry copper exploration through the 2000s. The Arabian Shield studies published by the Saudi Geological Survey and by academic groups using ASTER data demonstrated that hydrothermal alteration zones associated with known gold deposits showed statistically separable spectral signatures from background greenstones and granites at 30 m resolution.

Sentinel-2 cannot replicate that mineral-species discrimination. Its two SWIR bands are too broad and too widely spaced to distinguish kaolinite from illite. What it does well is broad lithological contact mapping across large areas quickly, at no data cost, with a five-day revisit that allows cloud-free compositing even in seasonally humid regions. The practical workflow in most exploration programmes is to use Sentinel-2 for regional screening across hundreds of thousands of square kilometres, then task ASTER archive pulls or PRISMA acquisitions over the shortlisted anomalies.

Lineaments: what SAR sees and what it cannot prove

Structural lineaments extracted from SAR backscatter are one of the most widely cited products in remote-sensing-assisted exploration, and one of the most frequently misused. Sentinel-1 C-band backscatter responds to surface roughness, dielectric constant, and slope orientation. Linear features in backscatter imagery can represent fault scarps, lithological contacts, dolerite dykes, wadi systems, agricultural field boundaries, pipeline routes, or processing artefacts from the SAR geometry itself. The physics does not distinguish between them.

Published work on the West African cratons (Birimian greenstone belts of Ghana, Burkina Faso, and Mali) has shown that lineament density maps derived from multi-azimuth SAR composites correlate spatially with known shear-zone-hosted gold deposits. The correlation is real but probabilistic. A lineament anomaly in a SAR image is a hypothesis about structure, not a confirmed fault. Multi-azimuth illumination (combining ascending and descending Sentinel-1 passes, or adding optical hillshade from a DEM) reduces the proportion of illumination artefacts, but it does not eliminate structural ambiguity. Ground-truth of at least a sample of lineaments before designing a drill programme is not optional; it is the minimum due diligence.

Fold axes, contacts, and the value of terrain context

In well-exposed arid terrain, the combination of a high-resolution DEM (SRTM at 30 m, ALOS World 3D at 30 m, or TanDEM-X at 12 m where licensed) with Sentinel-2 band composites can trace lithological contacts and fold axial traces across areas that would require months of field traverses. The key is spectral contrast between competent and incompetent units: quartzites versus phyllites, carbonates versus clastics, mafic intrusions versus felsic country rock. Where weathering has produced a regolith that masks bedrock, the spectral signal degrades sharply and structural interpretation becomes unreliable.

Vegetation is the other hard limit. In the Birimian terranes of West Africa, the dry season reduces canopy cover enough to allow partial bedrock exposure in laterite-stripped areas, but meaningful SWIR mineral mapping requires near-zero vegetation cover. The Arabian Shield, the Saharan cratons, and the arid cordillera of South America are the environments where these methods perform best. Tropical and boreal settings require a different approach, typically geophysical rather than spectral.

Hyperspectral mapping: genuine advance, genuine constraints

PRISMA and the forthcoming ESA CHIME mission represent a step change in spaceborne mineral mapping capability. Full spectral unmixing across 239 bands allows the generation of mineral abundance maps rather than simple classified outputs, which is a meaningfully different product for an exploration geologist. Published PRISMA studies have demonstrated separation of white mica polytypes (muscovite versus phengite) that carry information about hydrothermal temperature and depth, which is directly relevant to targeting within an alteration system.

The constraints are real. PRISMA's 30 m pixel is the same as ASTER's, so sub-pixel mixing remains a problem in heterogeneous terrain. Cloud cover invalidates any given acquisition. Tasking is not on-demand in the way Sentinel-2 is; acquisition requests must be submitted and approved, and repeat coverage of the same target takes weeks to months. For exploration programmes in equatorial regions, this can mean a six-to-twelve month wait for a cloud-free hyperspectral scene. Atmospheric correction quality also degrades at high aerosol loadings, which is common over active mining and smelting regions.

Building a defensible exploration target from imagery alone

The standard workflow that has emerged from published Arabian Shield and West African craton studies runs roughly as follows. First, Sentinel-2 band-ratio composites (false colour combinations emphasising iron oxide, hydroxyl, and carbonate responses) define the regional lithological framework and flag spectral anomalies at 10-20 m resolution. Second, ASTER SWIR band-ratio maps over the anomalies identify the mineral assemblage: argillic, phyllic, or propylitic alteration zones carry different mineral suites and point to different deposit types. Third, Sentinel-1 multi-azimuth lineament extraction defines the structural architecture, with the explicit caveat that every lineament is provisional. Fourth, the intersection of spectral alteration anomalies with structural lineament clusters defines a ranked list of targets for field reconnaissance.

Satellize applies this workflow using open-archive Sentinel and ASTER data, with PRISMA tasking added on client licence for high-priority targets. The Tonga crop-estimation programme demonstrated the analytics pipeline on a different problem class, but the underlying approach of combining open multispectral data with targeted hyperspectral acquisition transfers directly to exploration geology. A competent field geologist reviewing the output before any drill commitment is not a limitation of the method; it is the correct use of the method. Remote sensing narrows the search space. It does not replace the hammer.

Typical figures

Best spatial resolution (mineral mapping)	30 m (ASTER SWIR, PRISMA); 10-20 m (Sentinel-2 VNIR/SWIR)
Best spatial resolution (lineament extraction)	5x20 m ground range (Sentinel-1 IW); 12 m (TanDEM-X DEM hillshade)
Sentinel-2 revisit	5 days at equator (two-satellite constellation)
Sentinel-1 revisit	6 days (single satellite, mid-latitudes); 12 days at equator in standard IW mode
ASTER SWIR archive depth	2000-2008 (SWIR detector failure April 2008); TIR operational to 2023
PRISMA hyperspectral bands	239 bands, 400-2500 nm, 30 m resolution, ~30 km swath
Minimum detectable alteration zone	Practically 1-4 ha at 30 m pixel; sub-pixel mixing degrades confidence below ~5 ha
Cloud sensitivity	Optical and SWIR sensors fully blocked by cloud; SAR unaffected
Typical delivery format	GeoTIFF band-ratio composites, classified mineral maps (GeoPackage/shapefile), lineament vector layers, ranked target report (PDF)
Coverage	Global for Sentinel-1 and Sentinel-2; ASTER global archive; PRISMA tasked on request

Analytics Satellize can run

Regional lithological framework map	Sentinel-2 band-ratio compositing (iron oxide, clay/carbonate, ferrous iron indices); principal component analysis of SWIR bands	GeoTIFF and classified GeoPackage layer, covering client-defined AOI at 10-20 m
Hydrothermal alteration zone map	ASTER SWIR band-ratio indices (alunite, kaolinite, carbonate, propylitic assemblage) following published USGS and Saudi Geological Survey protocols	Classified mineral assemblage raster with confidence scores, delivered as GeoTIFF with accompanying PDF interpretation note
Hyperspectral mineral abundance map	PRISMA spectral unmixing using published reference spectra (USGS spectral library); white mica composition index from 2200 nm absorption position	Per-pixel mineral abundance GeoTIFF stack; summary report ranking alteration intensity across target polygons
Structural lineament map with confidence ranking	Sentinel-1 ascending and descending pass backscatter texture analysis; directional filtering; multi-azimuth composite lineament extraction; cross-validation against DEM hillshade	Vector shapefile of lineaments attributed by length, orientation, confidence tier (tectonic probable / ambiguous / likely non-tectonic), and intersection density
Ranked exploration target list	Spatial intersection of alteration anomaly polygons with high-confidence lineament clusters; scoring matrix based on alteration assemblage type, structural complexity, and proximity to known occurrences	PDF target report with ranked table, annotated maps per target, and recommended field-validation transects
Change detection over active exploration licence	Sentinel-2 time-series compositing to detect surface disturbance, new access tracks, or vegetation clearance indicating competitor or artisanal activity	Monthly GeoTIFF change layer with flagged polygons; optional alert feed via API

Sources

Who does the work

We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.