Active tectonic plate boundary and fault mapping from SAR and optical
SAR backscatter, interferometric coherence, and high-resolution stereo DEMs let geologists trace active fault lines and fold scarps from orbit, even where field access is impossible. Resolution limits are real but quantifiable.
Sensors
- TanDEM-X: Global DEM at 12 m posting (0.4 arcsec), absolute vertical accuracy better than 10 m, relative accuracy around 2 m over gentle terrain. The standard product for extracting fault-scarp morphology and computing geomorphic indices such as mountain-front sinuosity and valley-floor width-to-height ratios.
- SRTM: 30 m global DEM (1 arcsec public release), single-pass C-band acquisition from 2000. Useful as a baseline and for regional-scale structural mapping, but scarp heights below roughly 20-30 m are poorly resolved and vegetation bias is significant in forested terrain.
- Sentinel-1 SAR: C-band (5.6 cm wavelength), 6-12 day repeat at mid-latitudes in IW mode, 5 x 20 m ground resolution. Coherence imagery between pre- and post-event pairs reveals newly ruptured fault traces as coherence-loss corridors. Backscatter mosaics highlight linear tonal anomalies associated with fault zones in arid terrain.
- WorldView stereo optical: 0.3-0.5 m panchromatic resolution; stereo pairs yield photogrammetric DEMs at roughly 1-2 m posting, resolving scarps of 1-2 m height where slope contrast is sufficient. Cloud-free tasking required; coverage is on-demand rather than systematic.
- ALOS-2 PALSAR-2: L-band (23.6 cm wavelength), 14-day repeat, 3-10 m resolution in spotlight mode. L-band penetrates dry vegetation and sand cover better than C-band, making it useful for detecting buried or mantled fault traces that Sentinel-1 misses.
What a fault scarp looks like from 600 km up
A fault scarp is, in essence, a topographic step. From orbit, its detectability depends entirely on whether the DEM resolves that step above the noise floor. TanDEM-X at 12 m posting and roughly 2 m relative vertical accuracy can resolve scarps of 4-6 m or taller on smooth alluvial fans, which covers many Holocene ruptures in arid settings. Scarps below that threshold, or those mantled by colluvium, require airborne lidar or very-high-resolution commercial stereo.
In practice, published studies of the Iranian plateau and the Kopeh Dagh have used TanDEM-X to compute mountain-front sinuosity indices, valley-floor width-to-height ratios, and stream-length gradient indices across hundreds of drainage basins simultaneously. These geomorphic proxies correlate with long-term uplift rate and allow tectonic geomorphologists to rank fault segments by relative activity without visiting each one. The approach does not replace field dating, but it prioritises where dating is worth the effort.
SAR coherence as a rupture detector
When a fault ruptures, the ground surface is disrupted: cracks open, blocks shift, and the fine-scale texture that sustained radar coherence between two passes is destroyed. A Sentinel-1 coherence map computed from a pre-event and a post-event image pair shows this as a corridor of low coherence aligned with the rupture trace, often visible within hours of the event once the next overpass occurs.
The method works best in arid or semi-arid terrain where coherence is naturally high between passes. In vegetated or seasonally wet environments, coherence decays rapidly for reasons unrelated to faulting, and the signal is ambiguous. The 6-day Sentinel-1 revisit in Europe and parts of Asia is adequate for most post-earthquake assessments; in some orbital geometries over Central Asia, the effective revisit is 12 days, which introduces a window of uncertainty if aftershocks rework the surface.
SAR backscatter mosaics also contribute. Linear tonal contrasts in C-band or L-band imagery often trace lithological boundaries controlled by faults, and azimuth-dependent illumination effects can highlight fault-parallel lineaments invisible in a single-look image. Multi-look averaging across several acquisition geometries suppresses speckle and sharpens these structural trends.
Stereo DEMs where 12 metres is not enough
For palaeoseismic work on individual fault segments, 12 m TanDEM-X is frequently insufficient. A 2 m scarp on a 15-degree alluvial fan fan surface is below the noise floor of any freely available global DEM. WorldView or Pleiades stereo pairs, processed through dense image matching, can yield 1-2 m posting DEMs that resolve such features, and published studies have used them to measure offset stream channels and terrace risers at the metre scale.
The honest caveat is cost and cloud. Commercial stereo tasking is priced per square kilometre and requires cloud-free acquisition, which in tectonically active mountain ranges with persistent orographic cloud (the Zagros in winter, the Tien Shan in spring) can mean multi-month waits. SRTM remains the baseline for regional structural mapping precisely because it is global, free, and was acquired in a single consistent pass in February 2000, giving a clean reference surface against which change can be measured.
Slip rates from geomorphology: what the numbers actually mean
Geomorphic slip-rate estimates derived from satellite DEMs are not the same as geodetically measured rates. A mountain-front sinuosity index gives a relative ranking of tectonic activity over timescales of 10,000 to 100,000 years. To convert it to a millimetres-per-year slip rate, you need independent age control on the geomorphic surfaces, typically cosmogenic nuclide dating or optically stimulated luminescence, neither of which comes from a satellite.
Published work on the Central Asian thrust belts has combined TanDEM-X-derived geomorphic indices with published GPS velocities from the Central Asian GPS network to cross-validate relative activity rankings. The satellite DEM identifies which fault segments are geomorphically fresh; the geodesy constrains the present-day strain budget. Neither dataset alone tells the full story. Buyers who want slip rates from space data should be clear that they are getting geomorphic proxies, not direct measurements, and that the uncertainty on any individual estimate is typically a factor of two or more without field calibration.
Resolution limits and the cases where satellites are not the right tool
Spaceborne methods have a hard floor. Fault traces expressed only as subtle soil-moisture contrasts, microseismic alignments, or sub-metre scarps in dense vegetation are not detectable from any current civilian satellite. Lidar, either airborne or terrestrial, resolves features that no orbital sensor currently matches in vegetated terrain. The Wasatch Front studies that redrew the hazard map of Salt Lake City relied on airborne lidar at 0.5 m posting, not on TanDEM-X.
Similarly, change-detection SAR identifies rupture corridors at scales of hundreds of metres to kilometres. It will not locate the exact surface trace to within 10 m, and it cannot distinguish a primary rupture from a triggered landslide or a liquefaction field without corroborating optical imagery or field verification. Treating coherence loss as a precise fault map is a common misapplication.
Where satellites are genuinely strong: regional reconnaissance over politically or physically inaccessible terrain, systematic coverage of entire plate boundary zones at consistent geometry, and multi-decadal archive comparison. The Sentinel-1 archive from 2014 onward and the ERS/Envisat legacy archive from the 1990s together give three decades of coherence and backscatter data over many active margins, a resource that field geology alone cannot replicate.
From data to a usable fault inventory
A practical fault-mapping workflow combines three layers: a TanDEM-X hillshade rendered at multiple illumination azimuths to reveal lineaments, a Sentinel-1 coherence mosaic to flag zones of recent surface disturbance, and a WorldView stereo DEM over priority segments where scarp height matters. The outputs feed into a GIS fault database with attributed geometry, scarp height estimates, geomorphic activity class, and confidence flags.
Satellize structures this kind of analysis as a delivered GIS layer with accompanying methodology notes, not as a raw data dump. The workflow draws on open Sentinel-1 and SRTM data, adds commercial TanDEM-X and WorldView products where the client's area of interest demands it, and documents the resolution limits explicitly so that downstream hazard assessors know what they are working with. The Tonga crop-estimation programme is a different domain, but the same principle applies: a number without an honest uncertainty bound is not analysis, it is noise.
Typical figures
| Best available DEM posting (global) | 12 m (TanDEM-X global DEM, 0.4 arcsec) |
| Best available DEM posting (commercial stereo) | 1-2 m (WorldView or Pleiades stereo, on-demand) |
| TanDEM-X relative vertical accuracy | ~2 m over smooth terrain; degrades on steep slopes |
| Minimum resolvable scarp height (TanDEM-X) | ~4-6 m on smooth alluvial fans; higher in rough terrain |
| SAR revisit (Sentinel-1, mid-latitudes) | 6 days (dual satellite); 12 days in some Central Asian geometries |
| SAR ground resolution (Sentinel-1 IW mode) | 5 x 20 m (range x azimuth) |
| SAR frequency | C-band 5.405 GHz (Sentinel-1); L-band 1.2 GHz (ALOS-2 PALSAR-2) |
| Archive depth | Sentinel-1 from 2014; ERS/Envisat legacy from ~1992; SRTM single epoch Feb 2000 |
| Optical stereo cloud constraint | Cloud-free acquisition required; wait times variable by region and season |
| Deliverable formats | GeoTIFF DEM, GeoPackage fault-trace vectors, PDF methodology report |
Analytics Satellize can run
| Fault-trace lineament map | Multi-azimuth TanDEM-X hillshade rendering and Sentinel-1 backscatter mosaic analysis; manual and semi-automated lineament extraction | GeoPackage vector layer with confidence attributes and illumination-direction metadata |
| Scarp height profile transects | DEM differencing along fault-perpendicular swath profiles; published geomorphic index methods (mountain-front sinuosity, Vf ratio) | CSV profile data and GIS polygon layer with scarp height estimates and uncertainty ranges |
| Post-rupture coherence-loss corridor | Sentinel-1 interferometric coherence differencing between pre- and post-event image pairs | GeoTIFF coherence-change map and vector corridor polygon, delivered within 48 hours of post-event overpass |
| Geomorphic activity ranking | Basin-scale extraction of stream-length gradient index, hypsometric integral, and mountain-front sinuosity from TanDEM-X; ranked against published benchmarks from Iranian plateau and Kopeh Dagh studies | Tabular fault-segment ranking with geomorphic index values and PDF interpretive report |
| High-resolution segment DEM | WorldView stereo photogrammetry using dense image matching over priority fault segments | 1-2 m GeoTIFF DEM with orthophoto, supplied with accuracy assessment against ground control where available |
| Multi-decadal lineament change detection | Backscatter and coherence comparison across Sentinel-1 and ERS/Envisat archive epochs to identify newly expressed or reactivated traces | Change-detection GIS layer with epoch-stamped observations and confidence flags |
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.