10.6.2 — Dam Safety — maturity: live

Dam Wall Deformation (InSAR)

Measuring millimetre-scale displacement of dam walls and abutments using repeat-pass synthetic aperture radar interferometry to detect structural deformation before it becomes failure.

Millimetre-precision radar interferometry turns a constellation of small SAR satellites into a continuous structural health monitor for every dam a nation owns.

A dam wall that moves 10 millimetres in the wrong direction can be the first sign of imminent catastrophic failure. Ground-based geodetic instruments — prisms, tiltmeters, settlement gauges — cover discrete points and depend on site access, power, and telemetry links that floods or landslides routinely sever. A sovereign InSAR constellation revisits every major dam in the country on a short, predictable cadence, building a dense deformation time-series across the entire wall face, the abutments, and the reservoir rim simultaneously, regardless of weather or road access.

SAR satellites illuminate the dam structure with coherent microwave pulses and compare the phase of successive returns to resolve line-of-sight displacement to sub-centimetre precision. Processed against a digital elevation model and decomposed into vertical and horizontal components, the displacement field reveals settlement bowls, horizontal sliding, cracking patterns, and piping precursors weeks to months before they are visible to the eye or to a point sensor. C-band (5.6 cm wavelength) delivers broad coverage; L-band (24 cm) penetrates vegetation on earthen embankments and retains coherence over longer revisit gaps — the two are complementary and both should be in the national stack.

The operational outcome is a live deformation dashboard updated every few days for every registered dam structure in the national asset register. Anomaly thresholds trigger automated alerts to the dam safety regulator and the asset owner. For the Brumadinho or Banqiao class of event — tailings or earthen dam with no credible early warning — this capability is not an enhancement; it is the primary detection layer. A nation that relies on a foreign commercial InSAR provider accepts that data access can be throttled, delayed, or priced out of reach at exactly the moment the risk is highest.

What matters

InSAR detects precursor deformation at the millimetre scale weeks to months before macroscopic cracking or seepage becomes visible on site.
Coherence loss itself is a signal: sudden decorrelation over an earthen embankment face can indicate surface disruption or saturation-driven slumping.
L-band retains phase coherence over vegetated earthen dams where C-band loses it; a national fleet needs both frequency bands to cover the full dam typology.
When a dam fails in a jurisdiction served only by foreign SAR providers, export-control clauses, cloud-gap latency, or commercial triage can delay the warning that triggers downstream evacuation.

Quick facts

Global large dams at risk of failure: ~58,700 dams older than 30 years (2023) — ICOLD World Register of Dams — Statistics
InSAR displacement detection threshold: 1–3 mm per measurement cycle (2024) — ESA Sentinel-1 Mission Performance — SAR Interferometry
Temporal revisit — dual-satellite SAR constellation: 6-day repeat (single), 3-day repeat (dual) (2024) — ESA Sentinel-1 Revisit and Coverage
Economic loss — Brumadinho dam failure (Brazil, 2019): USD 7.5 billion total liability (2021) — World Bank — Mining Dam Failures: Risks, Consequences and Policy Responses
SAR microsatellite unit cost (100 kg class): USD 15–30 million per spacecraft (2024) — ESA — New Space Economy Report 2024
Lives at risk downstream of large dams globally: ~300 million people (2022) — UNEP — Dams and Development: Relevant Practices for Improved Integration of Dams

Sovereignty score: 9/10 — A nation cannot outsource early warning of dam failure to a foreign commercial data feed without accepting that the warning may arrive late, degraded, or not at all when political or commercial conditions change.

Dam failure downstream mortality is immediate and irreversible: the 1975 Banqiao cascade killed an estimated 170,000 people; the 2019 Brumadinho collapse killed 270 in under ten minutes — latency in the detection layer directly translates to body count.
Commercial InSAR providers operate under national export-control regimes (US EAR, EU dual-use regulation) that can restrict data delivery to specific countries or during declared national emergencies, precisely when the data is most critical.
A sovereign fleet can be tasked on-demand at conflict or crisis cadence — shortening revisit from 6 days to 1–2 days by adjusting the incidence angle schedule — whereas commercial tasking is subject to queue priority and cost-per-scene pricing.
National dam registers and deformation baselines constitute sensitive critical-infrastructure intelligence; routing them through a foreign commercial cloud creates a persistent geopolitical exposure that sovereign on-ground processing eliminates.

Reference architecture

Payload: Dual-frequency SAR: C-band (5.405 GHz, 5 m × 20 m IW resolution, 250 km swath) for urban and concrete structures; L-band (1.27 GHz, 10 m resolution, 150 km swath) for earthen and tailings embankments; both polarimetric (VV+VH) for coherence and surface-change analysis
Bus class: ESPA-class microsat, 150–200 kg, 900 W end-of-life solar power; deployable 3 m × 4 m SAR panel; S-band TT&C, X-band high-rate downlink at 400 Mbps
Orbit: Sun-synchronous LEO at 520–560 km, 6-satellite constellation (3 C-band, 3 L-band) in two planes separated by 30° RAAN, 4–6 day exact repeat for interferometric baseline control, 1–2 day revisit at national priority sites via multi-look scheduling
Ground segment: 3-station national network (X-band downlink, S-band TT&C) at geographically separated sites; on-site raw SAR processing cluster (FPGA-accelerated); Copernicus Sentinel-1 data ingested as free reference baseline; SatNOGS nodes for housekeeping telemetry backup
Data pipeline: On-board range-Doppler L0 compression → ground L1 SLC generation → GAMMA or ISCE-based InSAR chain (coregistration, interferogram, phase unwrapping, APS correction) → PS-InSAR and SBAS time-series on sovereign GPU cluster → millimetre displacement rasters and velocity maps at 10 m posting, updated per pass
End-user delivery: Deformation dashboard (web GIS) for the national dam safety regulator showing velocity maps, cumulative displacement time-series per structure, and automated red/amber/green status per registered dam; push alerts via secure API to asset owners and emergency management agencies when threshold exceedance is detected; raw SLC archive available to national research institutions
Time to launch: First demonstrator satellite (C-band, 3U-equivalent SAR pathfinder or commercial piggyback) within 18 months to establish baseline time-series; full 6-satellite dual-frequency constellation operational within 42 months from contract signature
Caveats: SAR payload and processing IP is export-controlled from US vendors under EAR; procure SAR payload from European (Airbus, OHB, ICEYE Finland), Indian (SAC/ISRO), or South Korean (LIG Nex1) primes; decorrelation over water surfaces makes InSAR ineffective on the reservoir face itself — reservoir level (§10.6.1) and spillway monitoring (§10.6.3) require optical or altimetric payloads, not SAR

Frequently asked

How is InSAR different from placing sensors directly on the dam wall?

Traditional in-situ instruments — crackmeters, piezometers, geodetic prisms — measure deformation at discrete point locations and require physical access and power infrastructure. InSAR provides spatially continuous deformation maps across the entire dam wall and abutments from orbit, with no on-site hardware, at millimetre-level accuracy. The two approaches are complementary: InSAR gives the full-surface picture; in-situ sensors give depth and causation.

What revisit frequency is needed for operational dam safety?

ICOLD Bulletin 158 recommends that anomalous deformation trends be identified within days, not weeks, to allow emergency intervention. A dual-satellite SAR constellation delivering 3-day revisit at X-band meets this threshold for most slow-onset failure modes. Rapid-onset failures (e.g., piping, overtopping) develop in hours and require integration with in-situ alarms; satellite revisit alone is insufficient for those scenarios.

Why should a government own the satellites rather than subscribe to ICEYE, Capella or Planet?

Commercial providers can and do reprioritise tasking for higher-paying customers, restrict data under export-control regimes (US ITAR/EAR applies to Capella and BlackSky), and may discontinue services or alter pricing at contract renewal. A dam failure that kills thousands is a sovereign emergency: the state cannot accept that its monitoring feed is subject to a vendor's commercial calculus or a foreign government's export licence. Owning the constellation guarantees priority tasking, data custody, and the ability to share raw data with national emergency services without third-party approval.

Which radar frequency band is best for dam wall InSAR?

X-band (9.6 GHz, ~3 cm wavelength) is the default choice for hard-surface infrastructure: it delivers the highest spatial resolution (sub-metre) and best coherence on concrete and rock-fill faces. C-band (5.4 GHz) — used by ESA Sentinel-1 — offers wider swath and better penetration through moderate vegetation but coarser resolution. L-band (1.27 GHz) excels on vegetated earthen embankments. An ideal sovereign constellation deploys X-band microsatellites for primary monitoring and ingests open C-band Sentinel-1 data as a redundant backstop.

What is a typical end-to-end data latency from satellite pass to deformation alert?

With a direct downlink to a national ground station and an automated processing pipeline (SNAP or proprietary processor), latency from acquisition to a displacement map can be under 4 hours. Issuing a calibrated alert with quality flags adds another 1–2 hours if a human analyst reviews the output. Fully automated pipelines using machine-learning anomaly detectors can compress total latency to under 90 minutes, though at the cost of higher false-positive rates.

Can InSAR work in heavy cloud cover or at night?

Yes. SAR is an active microwave sensor: it generates its own radar pulses and is unaffected by cloud, rain (at X-band, heavy precipitation causes modest signal attenuation but does not prevent acquisition), or darkness. This all-weather, day-and-night capability is one of InSAR's principal advantages over optical satellites for dam safety, where storms — the very events most likely to stress a dam — are also when cloud cover would blind an optical sensor.

How many satellites does a sovereign InSAR constellation for dam safety require?

A minimum viable constellation for national dam monitoring is 4 SAR microsatellites in two complementary orbital planes, delivering 2–3 day revisit with ascending and descending pass coverage for full line-of-sight decomposition. For a nation with more than 500 large dams spread across varied topography, 6–8 satellites are recommended to maintain 1–2 day revisit at critical assets. Incremental deployment — launching 2 satellites, then expanding — is technically feasible and spreads capital cost.

What international frameworks govern dam safety data sharing?

There is no binding global treaty specifically on dam safety data. The UN Sendai Framework for Disaster Risk Reduction 2015–2030 encourages data sharing for early warning, and the WMO's Hydrological Status and Outlook System (HydroSOS) provides a voluntary coordination mechanism. Bilaterally, nations sharing transboundary rivers (governed by the UN Watercourses Convention 1997) have obligations to notify downstream states of imminent hazards, which makes sovereign InSAR data politically sensitive and reinforces the case for keeping it under national control.

Glossary

InSAR: Interferometric Synthetic Aperture Radar — a technique that compares the phase of radar echoes from two or more satellite passes to measure ground-surface displacement at millimetre precision.
PSInSAR (Persistent Scatterer InSAR): An advanced InSAR processing method that identifies stable radar reflectors (buildings, rocks, corner reflectors) across a long time series to extract deformation signals with reduced atmospheric noise.
SBAS (Small Baseline Subset): An InSAR processing algorithm that combines pairs of images with short spatial and temporal baselines to produce deformation time series while mitigating decorrelation effects.
Line-of-sight (LOS) displacement: The component of surface movement measured along the radar beam direction, which is a projection of true 3-D deformation and must be decomposed using multi-geometry acquisitions for full vector analysis.
Coherence: A measure (0 to 1) of how similar the radar backscatter is between two SAR acquisitions; high coherence is required for reliable InSAR phase unwrapping and displacement measurement.
Phase unwrapping: The computational process of converting the ambiguous, cyclically wrapped interferometric phase (−π to +π) into a continuous displacement field measured in absolute distance.
Corner reflector: A trihedral metallic device placed on a dam wall that produces a highly stable, known radar return, used to anchor InSAR displacement measurements and validate satellite-derived results.
Atmospheric Phase Screen (APS): The component of InSAR phase delay caused by spatial variations in atmospheric water vapour and ionospheric electron content, which can mimic or obscure real ground deformation signals.
SAR (Synthetic Aperture Radar): An active microwave imaging radar that synthesises a large virtual antenna by combining returns from many positions along the satellite's orbit, achieving high spatial resolution independently of actual antenna size.
Creep: Slow, continuous deformation of a dam structure over time, often measured in millimetres per year; accelerating creep rate is a key precursor indicator of potential structural failure.

References

Sentinel-1 InSAR for Infrastructure Deformation Monitoring — ESA documents the operational use of Sentinel-1 C-band InSAR for monitoring large civil infrastructure including dams, demonstrating sub-centimetre displacement sensitivity at 6-day revisit across Europe and globally.
ICOLD Bulletin 158 — Dam Surveillance: Interpreting Monitoring Data to Assess Structural Behaviour — The definitive ICOLD guidance on dam instrumentation and monitoring data interpretation, establishing thresholds for anomalous deformation and recommending integration of remote sensing data into national dam safety programmes.
Mining Dam Failures: Risks, Consequences and Policy Responses — World Bank analysis of 50 major dam failures since 1915 quantifies cumulative economic losses exceeding USD 35 billion and argues that systematic remote monitoring — including satellite InSAR — is the most scalable risk-reduction measure available to middle-income nations.
Sendai Framework for Disaster Risk Reduction 2015–2030 — Priority 4 of the Sendai Framework calls on governments to substantially increase availability and access to multi-hazard early warning systems by 2030; dam deformation monitoring via InSAR is explicitly cited in UN guidance as a qualifying technology.
PSInSAR and SBAS Techniques for Large Dam Monitoring: A Comparative Study — This peer-reviewed study compares Persistent Scatterer and Small Baseline Subset InSAR processing on a dataset of 12 large dams, finding PSInSAR superior for rigid concrete structures and SBAS preferable for earthen-fill embankments with lower coherence.
ITU-R RS.1166-4 — Performance and Interference Criteria for Spaceborne SAR — This ITU-R Recommendation establishes interference protection criteria and technical performance parameters for spaceborne SAR systems operating in Earth Exploration Satellite Service frequency bands, providing the regulatory baseline for national SAR constellation frequency planning.
FAO — Dams, Agriculture and Water Management in Developing Countries — FAO estimates that large dams supply irrigation water to 30–40% of irrigated land globally; the report identifies structural failure of ageing dam stock as a critical food-security risk and recommends space-based monitoring as a cost-effective supplement to in-situ inspection regimes.
OGC Sensor Observation Service 2.0 — Implementation Standard — The OGC SOS 2.0 standard defines interoperable web service interfaces for querying time-series sensor observations, enabling InSAR-derived deformation data to be published alongside in-situ dam instrumentation feeds within national dam safety information systems.

Related applications

10.6.4 — Catchment Inflow Forecasting (Dam Safety)
10.6.5 — Downstream Flood Risk Modelling (Dam Safety)
10.1.1 — Rail Track Geometry Monitoring (Railway Intelligence)
10.1.2 — Encroachment Detection (Railway Intelligence)
10.1.3 — Slope Stability Around Tracks (Railway Intelligence)
10.1.4 — Construction Progress Tracking (Railway Intelligence)
10.1.5 — Right-of-Way Vegetation Monitoring (Railway Intelligence)
10.2.1 — Highway Construction Progress (Road Corridor Monitoring)