Oil and gas field surface deformation from reservoir pressure change
Production-induced pore-pressure changes compact or rebound reservoir rocks, producing surface signals detectable by InSAR at millimetre precision. Sentinel-1 and archival ERS/Envisat data let operators quantify compressibility, anticipate casing damage, and assess induced-seismicity exposure before regulators do.
Sensors
- Sentinel-1 A/B (C-band SAR, ESA): 5.6 cm wavelength; Interferometric Wide Swath mode gives 5 x 20 m resolution across a 250 km swath. Six-day repeat at mid-latitudes with both satellites operating, enabling time-series with sub-centimetre line-of-sight precision after atmospheric correction. Free and open archive from 2014.
- Envisat ASAR (C-band SAR, ESA archive): Same C-band wavelength as Sentinel-1, 30 m resolution in Image Mode. Archive spans 2002-2010, providing the backbone for published Groningen and Valhall studies. No longer operational, but the archive underpins multi-decade deformation histories.
- ERS-1/2 (C-band SAR, ESA archive): 35-day repeat, archive from 1992. Resolution approximately 25 m. Used in the earliest published InSAR subsidence work over Wilmington and North Sea fields. Tandem phase (1995-1996) allowed 1-day interferograms that largely eliminate atmospheric noise.
- COSMO-SkyMed (X-band SAR, ASI): 3 cm wavelength; Spotlight mode reaches approximately 1 m resolution with a 16-day constellation repeat, compressible to shorter intervals through multi-satellite tasking. X-band is more sensitive to small deformation increments but also more sensitive to vegetation and soil-moisture phase noise over agricultural land above reservoirs.
What a producing reservoir does to the ground above it
Withdrawing fluids from a porous rock formation reduces pore pressure. The grains compact under the overburden load they previously shared with the fluid. That compaction propagates upward through hundreds or thousands of metres of overburden and appears at the surface as a bowl-shaped subsidence signal, typically centred above the reservoir and attenuated by depth. The reverse happens during water or gas injection: pore pressure rises, the rock rebounds, and the surface lifts.
The horizontal extent of the surface bowl is always wider than the reservoir itself. A simple nucleus-of-strain model, treating the compacting interval as a point or distributed source of volumetric strain, predicts that the bowl radius at surface is roughly equal to the reservoir depth. Groningen, at approximately 3 km depth, produces a bowl tens of kilometres across; Wilmington, shallower at around 600-900 m, produced a much tighter and steeper signal that reached 9 metres of cumulative subsidence at its peak in the 1960s before pressure maintenance began.
The depth-volume ambiguity InSAR cannot resolve alone
InSAR measures surface displacement with high spatial density but cannot directly see the reservoir. The observed surface deformation pattern is consistent with a range of combinations of reservoir depth, lateral extent, and volumetric compaction. A deep, large compacting volume and a shallower, smaller one can produce nearly identical surface bowls. This is the classic source-parameter ambiguity in potential-field and elastic-deformation inversion.
In practice, operators resolve the ambiguity by coupling InSAR observations to reservoir models constrained by well logs, seismic reflection data, and production history. The InSAR time-series then acts as an independent check on the modelled compaction coefficient. At Valhall in the North Sea, published studies using Envisat ASAR data showed that the observed subsidence rate of roughly 30-50 mm per year in the early 2000s was consistent with reservoir compaction models, but the surface signal alone could not distinguish between different vertical distributions of compaction within the chalk reservoir. Pressure data from wells was necessary to break that degeneracy.
A practical implication: InSAR is most powerful when it is integrated with subsurface data, not used in isolation. An operator who treats the surface bowl as a direct readout of reservoir volume change without accounting for depth and elastic overburden properties will systematically mis-estimate compressibility.
What the published field studies actually show
Groningen gas field in the Netherlands has been monitored by InSAR since the Envisat era. Published analyses show a subsidence bowl exceeding 35 cm of cumulative settlement over the production lifetime, with rates that tracked production volumes before deliberate output reductions. The spatial pattern of subsidence has been used to infer lateral variations in reservoir compressibility and to assess whether subsidence gradients correlate with induced earthquake locations. The correlation is imperfect: seismicity clusters near faults, and the relationship between compaction and fault reactivation involves fault geometry and stress state, not just the magnitude of surface settlement.
Wilmington oil field in California accumulated nearly 9 metres of subsidence between the 1920s and 1960s, causing infrastructure damage across Long Beach. Water injection for pressure maintenance arrested and partially reversed the subsidence. ERS-1/2 and later sensors captured the rebound phase. This field is a canonical example of how pressure management, verified by geodetic monitoring, can reverse production-induced deformation.
Valhall in the Norwegian North Sea, a chalk reservoir at roughly 2.4 km depth, showed subsidence rates of the order of 30-50 mm per year in Envisat studies published in the mid-2000s. The high compressibility of chalk makes it particularly susceptible. Seafloor subsidence at Valhall has also been measured by repeated bathymetric surveys, providing an independent cross-check on InSAR-derived surface rates.
Casing damage and induced seismicity: why the signal matters operationally
Subsidence bowls impose horizontal tensile strain on the surface and near-surface layers at the bowl rim, and compressive strain at the centre. Wells passing through compacting intervals experience differential vertical displacement along their length. Where compaction is uneven, shear strain develops. Published studies from Groningen and from Venezuelan heavy-oil fields document casing deformation and failure correlated with zones of high compaction gradient. Identifying those gradients early, from the spatial pattern of InSAR displacement, gives operators the option to prioritise integrity inspections before failure occurs.
Induced seismicity is a related but distinct concern. Reservoir compaction changes the stress state on pre-existing faults in and around the reservoir. Whether a fault slips seismically depends on its orientation relative to the stress field, its frictional properties, and the magnitude of stress change. InSAR cannot directly predict seismicity, but the spatial gradient of surface subsidence is a proxy for the compaction gradient, which in turn informs fault-stress modelling. Regulators in the Netherlands have used this logic to link production levels at Groningen to seismic hazard assessments.
Limits of the method and what they cost you
C-band InSAR is sensitive to tropospheric water vapour. Over onshore fields, a single interferogram can carry atmospheric artefacts of 5-20 mm that mimic or mask real deformation. Time-series methods (SBAS, PS-InSAR) average down atmospheric noise over many acquisitions, but they require a coherent target network. Agricultural land above reservoirs loses coherence seasonally. Deserts and industrial infrastructure maintain coherence well. Offshore platforms are point targets that maintain coherence but provide only sparse spatial sampling of the deformation bowl.
Vertical resolution after time-series processing is typically 1-3 mm per year for well-maintained PS or SBAS networks over stable terrain. Detecting the onset of a new deformation signal, rather than a long-running trend, requires more acquisitions and is harder. In areas of rapid deformation exceeding roughly 10-15 cm between acquisitions, C-band interferograms decorrelate or phase-wrap ambiguously; X-band is worse in this respect because the shorter wavelength wraps at smaller displacements.
Satellize processes Sentinel-1 time-series for deformation monitoring and can integrate archival ERS and Envisat stacks for fields where a multi-decade baseline matters. Our Groningen-class workflow is the same approach we apply in the Tonga crop-estimation programme: open-constellation data, rigorous atmospheric correction, and outputs in formats an engineering team can actually use.
Typical figures
| Spatial resolution (Sentinel-1 IW mode) | 5 x 20 m (range x azimuth); multi-looked to ~20-40 m for interferometry |
| Revisit interval | 6 days at mid-latitudes with Sentinel-1 A and B; 12 days single-satellite; 35 days for ERS-1/2 archive |
| Minimum detectable deformation rate (time-series) | 1-3 mm/year over stable, coherent surfaces after atmospheric correction |
| Single-interferogram precision | ~5-20 mm (limited by tropospheric noise); time-series reduces this significantly |
| Radar frequency / wavelength | C-band: 5.405 GHz / 5.6 cm (Sentinel-1, Envisat, ERS); X-band: ~9.6 GHz / 3.1 cm (COSMO-SkyMed) |
| Swath width | 250 km (Sentinel-1 IW); 100 km (Envisat IS2 mode); 40 km (COSMO-SkyMed Stripmap) |
| Archive depth | ERS-1/2 from 1992; Envisat 2002-2010; Sentinel-1 from 2014 (open access) |
| Coherence requirement | Industrial infrastructure and bare soil: high; seasonal vegetation: low to moderate; open water: none |
| Delivery formats | GeoTIFF displacement maps, CSV time-series per point, GIS-ready shapefiles, PDF interpretation report |
Analytics Satellize can run
| Cumulative subsidence / uplift map | SBAS (Small Baseline Subset) time-series InSAR over Sentinel-1 IW stack | GeoTIFF raster of total line-of-sight displacement over user-defined epoch, with uncertainty layer |
| Deformation rate map | PS-InSAR (Persistent Scatterer) over coherent targets; linear and seasonal decomposition | GIS point layer of mm/year rates at each persistent scatterer, colour-coded by magnitude |
| Surface bowl geometry and volume-change estimate | Nucleus-of-strain forward modelling fitted to observed InSAR surface displacement | PDF report with best-fit source parameters (depth, radius, volume change) and uncertainty ranges |
| Compaction gradient map for casing-risk screening | Spatial gradient of vertical displacement field derived from decomposed ascending/descending InSAR | GeoTIFF of horizontal strain proxy; well-location overlay flagging high-gradient zones |
| Multi-decade deformation history | Cross-calibrated stack combining ERS-1/2, Envisat ASAR and Sentinel-1 time-series | CSV time-series per benchmark or PS point from 1992 to present; change-point detection annotations |
| Atmospheric-corrected interferogram series | ERA5 or GACOS tropospheric delay correction applied per acquisition | Corrected interferogram stack in GeoTIFF format, ready for operator's own inversion workflow |
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.