Active fault and seismic hazard mapping for tower foundation siting
InSAR time-series and high-resolution optical stereo can locate active fault traces and measure interseismic creep before a tower foundation is poured, steering civil works away from rupture risk. Neither method is infallible: vegetated terrain breaks InSAR coherence, and blind faults leave no surface signature at all.
Sensors
- Sentinel-1 A/B (ESA, C-band SAR): 5.6 cm wavelength, 5 x 20 m ground range resolution in Interferometric Wide Swath mode, 6-day revisit at mid-latitudes with both satellites. Persistent-scatterer and SBAS time-series processing resolves line-of-sight displacement to roughly 1-2 mm/year over stable, coherent ground. Coherence degrades sharply in dense vegetation and in areas of rapid agricultural change.
- ALOS-2 PALSAR-2 (JAXA, L-band SAR): 23.6 cm wavelength penetrates light vegetation canopy better than C-band, improving coherence in scrub and open woodland. Stripmap mode delivers 3 x 3 m resolution; ScanSAR mode 60 m at 350 km swath. Revisit is 14 days for the same geometry. Particularly suited to slow interseismic creep on locked faults where C-band loses coherence over months-long baselines.
- Pleiades Neo (Airbus, optical): 30 cm panchromatic resolution. Stereo and tri-stereo acquisition in a single pass produces digital surface models with 50-80 cm vertical accuracy over bare or sparsely vegetated terrain. Used to confirm fault scarps, measure scarp height and map geomorphic lineaments that corroborate InSAR displacement signals.
- WorldView Legion (Maxar, optical): 30 cm panchromatic resolution, up to 15 revisits per day over a given point depending on latitude. High revisit cadence is useful for detecting post-seismic surface changes and for cloud-gap-filling in regions with intermittent cloud cover. Stereo capability supports independent DSM generation for cross-validation.
Why a tower foundation engineer needs a seismologist's dataset
A tower foundation that crosses an active fault trace is not merely at risk during a major earthquake. Slow aseismic creep, which can accumulate at rates of a few millimetres to several centimetres per year on faults such as the Hayward in California or the North Anatolian in Turkey, imposes differential settlement that shears pile caps and tilts monopoles over years, not in a single dramatic event. The structural consequence is the same; the warning is quieter.
Standard geotechnical site investigation, even with trial pits and boreholes, rarely covers the spatial extent needed to confirm whether a fault trace passes beneath a proposed foundation footprint. Remote sensing changes the economics: a single Sentinel-1 persistent-scatterer analysis over a 250 km swath costs a fraction of a drilling campaign and can flag candidate fault segments before a site visit is scheduled.
What InSAR time-series actually measures, and where it fails
Persistent-scatterer InSAR stacks tens or hundreds of SAR acquisitions over the same area and isolates stable reflectors, typically buildings, rock outcrops or bare soil, whose phase history encodes ground displacement. Over two to three years of Sentinel-1 data, the method can resolve creep rates above roughly 1-2 mm/year in the satellite line-of-sight direction. Combining ascending and descending pass geometries allows decomposition into approximate vertical and east-west components, though north-south motion remains poorly constrained by C-band geometry.
The method has hard limits. Vegetated terrain, particularly tropical forest, breaks interferometric coherence within days at C-band; even L-band PALSAR-2 loses coherence over dense jungle. Blind faults, those that terminate below the surface without a surface rupture or scarp, produce no displacement signal until they break through, which is precisely when it is too late for foundation planning. Urban areas with high scatterer density work well; rural scrubland is marginal; rainforest is largely opaque to the technique.
What a fault scarp gives away in stereo
A fault scarp is a topographic step, sometimes only a metre or two high, that marks where repeated surface ruptures have displaced the ground. In arid and semi-arid terrain these features persist for thousands of years and are clearly visible in high-resolution DSMs derived from Pleiades Neo or WorldView stereo. Scarp height, aspect and continuity along strike are diagnostic: a linear, consistently-facing step that truncates alluvial fan surfaces is a reliable geomorphic indicator of a Holocene-active fault.
The practical workflow pairs InSAR displacement maps with stereo-derived DSMs. Where InSAR shows a sharp gradient in line-of-sight velocity and the DSM shows a coincident linear scarp, confidence in fault activity is high. Where only one signal is present, the interpretation is weaker and warrants ground-truthing. Optical stereo alone cannot confirm that a scarp is currently active; InSAR alone cannot confirm the geomorphic structure. The two datasets are complementary, not interchangeable.
Interseismic creep: the slow signal that PALSAR-2 catches best
On creeping fault segments, ground displacement accumulates continuously between earthquakes. ALOS-2 PALSAR-2, with its 23.6 cm wavelength, maintains coherence over longer temporal baselines than Sentinel-1 in moderately vegetated terrain, making it better suited to detecting creep rates below about 5 mm/year in areas where C-band decorrelates. JAXA operates a systematic global acquisition strategy, and archived PALSAR-2 data extends back to 2014, giving roughly a decade of displacement history for many seismically active regions.
Creep rates vary significantly along fault strike. The central section of the San Andreas, for example, creeps at 25-35 mm/year, while adjacent locked sections accumulate strain silently. A tower sited on a creeping segment faces progressive differential displacement; one sited on a locked section faces lower chronic risk but higher catastrophic rupture risk. InSAR time-series can distinguish these regimes, but the interpretation requires a geophysicist, not just a GIS analyst.
Translating displacement maps into foundation setback recommendations
The analytic output of this workflow is not a pass-or-fail verdict. It is a ranked hazard map that shows, for each candidate tower location, the proximity to mapped fault traces, the measured or inferred creep rate, the confidence level given terrain type and coherence quality, and the data gaps where blind-fault risk cannot be assessed from above.
Setback distances are a matter of national building codes and engineering judgement, not satellite data alone. Many jurisdictions specify minimum setbacks from active fault traces, ranging from 15 metres in some US Alquist-Priolo zones to 50 metres or more in higher-risk classifications. The satellite-derived fault map provides the spatial input; a licensed geotechnical engineer applies the code. Satellize, whose analytics work has included geospatial analysis for sovereign clients such as the Kingdom of Tonga, structures deliverables as GIS layers and accompanying interpretation reports that are explicitly designed to feed into that engineering process rather than replace it.
One practical note on archive depth: Sentinel-1 data is freely available from the Copernicus Data Space back to 2014 for most of the globe. That gives a decade of displacement history at no data-acquisition cost, which is a material advantage over commissioning a new ground-based geodetic survey.
Honest limits, summarised plainly
InSAR cannot see blind faults. It cannot maintain coherence in dense vegetation. It measures displacement in the satellite line-of-sight, not true three-dimensional ground motion, and decomposing that into fault-parallel slip requires geometric assumptions that introduce uncertainty. Optical stereo cannot confirm that a scarp is currently active, only that the topography is consistent with past rupture. Both methods work best in arid, tectonically active regions with sparse vegetation and good scatterer density.
In humid tropical settings, where many telecoms build-outs are occurring, the honest answer is that satellite geodesy provides useful but incomplete hazard context, and it should be combined with published national seismic hazard maps, historical earthquake catalogues and field reconnaissance rather than treated as a standalone site-clearance tool.
Typical figures
| Sentinel-1 spatial resolution (IW mode) | 5 x 20 m ground range (multi-looked to ~10 m for standard products) |
| Sentinel-1 revisit | 6 days at mid-latitudes (both satellites); 12 days with one satellite |
| ALOS-2 PALSAR-2 spatial resolution | 3 x 3 m (Stripmap); 60 m (ScanSAR) |
| ALOS-2 PALSAR-2 revisit | 14 days (same geometry) |
| Pleiades Neo / WorldView optical resolution | 30 cm panchromatic; stereo DSM vertical accuracy 50-80 cm over bare terrain |
| Minimum detectable creep rate (PS-InSAR, stable terrain) | Approximately 1-2 mm/year line-of-sight, with 2-3 years of data |
| Sentinel-1 archive depth | 2014 to present (Copernicus Data Space, open access) |
| ALOS-2 PALSAR-2 archive depth | 2014 to present (JAXA AUIG; commercial licensing for non-research use) |
| SAR frequency bands used | C-band 5.405 GHz (Sentinel-1); L-band 1.2578 GHz (PALSAR-2) |
| Typical deliverable format | GeoTIFF displacement maps, GeoPackage fault-trace vectors, PDF interpretation report |
Analytics Satellize can run
| Persistent-scatterer velocity map | PS-InSAR time-series (StaMPS or similar published algorithm) over multi-year Sentinel-1 stack | GeoTIFF mean line-of-sight velocity raster with per-scatterer uncertainty, coloured by displacement rate |
| SBAS displacement time-series | Small-Baseline Subset (SBAS) InSAR applied to Sentinel-1 or PALSAR-2 interferogram network | Per-pixel displacement time-series exported as NetCDF or GeoTIFF stack, with seasonal signal separated from secular trend |
| Fault trace vector layer | Geomorphic lineament extraction from Pleiades Neo or WorldView stereo DSM, cross-validated against InSAR velocity gradients | GeoPackage polyline layer with confidence classification (confirmed active, probable, possible) and scarp-height attribute |
| Co-seismic displacement field | Two-pass differential InSAR using pre- and post-earthquake SAR acquisitions (published method, widely applied to events including 2023 Kahramanmaras sequence) | Wrapped and unwrapped interferogram GeoTIFFs showing surface displacement, with fault rupture trace mapped from fringe discontinuities |
| Candidate site hazard ranking | Spatial overlay of PS-InSAR velocity map, fault trace layer and national seismic hazard zone polygons; scored by proximity and creep rate | Site-by-site ranked table with setback distance to nearest mapped active trace, confidence tier and data-gap flags; delivered as PDF report and CSV |
| Terrain-coherence suitability assessment | Mean coherence map derived from interferogram stack; identifies zones where InSAR is unreliable and blind-fault risk cannot be assessed remotely | GeoTIFF coherence mask overlaid on site portfolio, with written caveat noting areas requiring ground-based investigation |
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.