The short answer

In geology, GIS is the system that turns field observations and expert interpretation into structured, queryable spatial data — not just a way to draw a coloured map. Geologists use it to digitise and attribute geologic maps, store structural and sample data, integrate terrain and remote-sensing evidence, run spatial analysis (buffering, overlay, proximity), and build predictive models such as mineral prospectivity. Standards like the USGS GeMS schema exist so that these maps are consistent, archivable, and machine-readable across organisations.

The thread running through every use is the separation of observation from interpretation, with provenance and confidence preserved. A digital geologic map that records who measured what, at what scale, and how certain they were is far more valuable than one that records only colours.

Digital geologic mapping

A geologic map in GIS is a set of related feature classes, not a single coloured layer. At minimum:

  • Map unit polygons (MapUnitPolys) — the formations and lithologies, each linked to a unit code.
  • Contacts and faults (ContactsAndFaults) — the lines, attributed by type (contact, normal fault, thrust), existence and identity confidence, and location confidence (a metres value tied to map scale).
  • Description of Map Units (DescriptionOfMapUnits) — the legend as a table: unit name, age, lithology, description.
  • Data sources — every feature points to where it came from.

This is exactly the structure the GeMS (Geologic Map Schema) formalises. GeMS, maintained by the USGS, defines these feature classes and tables so that a map produced by one survey can be read, archived, and combined with another. The point of the schema is that the attributes carry the geology — symbology is just a rendering of the data model, never the data itself.

Maps are typically stored in a GeoPackage or a file/enterprise geodatabase, with PostGIS used when several geologists edit concurrently or when the data drives web services.

Storing field and structural data

Structural measurements — strike and dip, foliation, lineation, fold axes — are stored as point features with numeric attributes: an azimuth (strike or dip-direction) and a dip angle, plus a type code. The map symbol (the classic T-shaped strike/dip mark) is then rotated by the strike attribute using data-defined symbology in QGIS or rotation expressions in ArcGIS Pro. The discipline here is to store the angles as queryable numbers, so you can later select "all bedding dipping more than 60 degrees" or feed orientations into stereonet analysis, rather than baking the measurement into a static rotated icon.

Field collection increasingly feeds straight into GIS via mobile apps (QField for QGIS, Field Maps for ArcGIS), so observations arrive already geolocated and attributed against the project schema.

Integrating terrain and remote sensing

Geological interpretation rarely uses geology alone. GIS lets you stack:

  • Terrain derivatives from a DEM — hillshade and slope for landform and structural expression, aspect, and curvature. Faults and lithological contacts often show up as lineaments in a multidirectional hillshade.
  • Remote-sensing layers — Sentinel-2 or Landsat band ratios and indices for lithological and alteration discrimination, and SAR for structure. Much of this is computed in Google Earth Engine and brought into the desktop as a GeoTIFF.
  • Geophysics and geochemistry — gridded magnetics, radiometrics, and sampled element concentrations interpolated to rasters.

The integration requirement is a shared, metric CRS so layers actually overlay. A typical project fixes a projected CRS up front (a UTM zone such as EPSG:32633, or a national grid) and reprojects everything into it on ingest with gdalwarp for rasters and ogr2ogr -t_srs for vectors.

Spatial analysis on geological data

Once data is structured, standard spatial operations answer geological questions:

  • Buffer and proximity — distance to a fault or intrusive contact, often a key exploration vector.
  • Overlay/intersection — where a permissive lithology coincides with a structural corridor (ST_Intersects in PostGIS, or the Intersection tool).
  • Reclassification and map algebra — turning continuous layers (slope, geochemical grids) into ranked criteria.

A simple PostGIS query finding sample sites within 500 m of a mapped fault:

SELECT s.sample_id
FROM samples s
JOIN faults f
  ON ST_DWithin(s.geom, f.geom, 500)
WHERE f.fault_type = 'thrust';

A worked example: mineral prospectivity

Prospectivity mapping is one of the clearest demonstrations of GIS in geology.

  1. Common grid. Choose a project CRS and a target cell size (say 50 m) and resample all evidence to it.
  2. Build evidence layers. Distance-to-fault from the structural map; a lithology favourability raster reclassified from map units; a geochemical anomaly raster interpolated from samples; an alteration index from Sentinel-2 band ratios.
  3. Standardise. Rescale each layer to a common 0-1 favourability range.
  4. Combine. Use a weighted overlay (knowledge-driven) where an expert assigns weights, or a data-driven method (e.g. weights-of-evidence) trained on known occurrences.
  5. Rank and validate. Produce a prospectivity raster, then check that known deposits fall in high-scoring areas before using it to prioritise fieldwork.

Every input layer keeps its provenance and the model parameters are recorded, so the result can be defended and rerun.

Common pitfalls and why they happen

  • Treating colour as the data model. Digitising a scanned map into coloured polygons with no unit codes, confidence, or sources loses the geology; the map can be displayed but not queried. Attribute against a schema like GeMS from the start.
  • Dropping source scale and uncertainty. A 1:250,000 contact and a 1:10,000 contact are not equally precise; combining them without recording scale implies false confidence. Store location confidence and source scale per feature.
  • Mixing CRSs. Geology, terrain, and imagery in different CRSs will not overlay correctly. Reproject everything to one metric project CRS on ingest.
  • Baking measurements into symbols. A rotated strike/dip icon with no numeric attributes cannot be analysed. Always store the azimuth and dip as numbers.
  • Combining scales silently in analysis. Buffering or overlaying layers digitised at very different scales mixes precisions; note the limits in the output.

Validating geological GIS work

Before a geological GIS product is trusted, check topology (no slivers or gaps between unit polygons, contacts that close), confirm every feature has a valid unit/source attribution, verify the CRS and units are explicit and consistent across layers, and compare the result against an independent reference — a published map sheet, surveyed control, or known occurrences. Keep observed data separate from interpreted boundaries so a reviewer can see what was measured versus inferred.

Bathyl perspective

We treat a geologic map as a structured evidence base, not a picture. The value is in the attributes — units, contacts, confidence, sources — and in keeping observation distinct from interpretation so the work stays inspectable. Schemas like GeMS and a disciplined CRS make geology that others can query, combine, and trust years later.

Related reading

Sources