Aircraft dispersal revetment and hardened shelter construction monitoring
New revetments and hardened shelters signal a shift from peacetime to wartime basing posture months before aircraft arrive. Shadow geometry, SAR backscatter change and construction-sequence analysis make that shift legible from open and commercial imagery.
Sensors
- Maxar WorldView-3: Panchromatic resolution of 0.31 m allows individual shelter walls, blast-deflector angles and revetment berm profiles to be measured directly. Shadow lengths at known solar elevation angles yield berm height estimates to within roughly 0.5 m. Revisit over a fixed target is typically 1 to 4.5 days depending on latitude and tasking priority.
- Airbus Pléiades Neo: 0.30 m panchromatic, 0.75 m multispectral. Stereo and tri-stereo collection modes allow photogrammetric digital surface models to be derived, giving direct volumetric estimates of earthwork spoil and berm height without relying solely on shadow geometry. Revisit is 1 to 2 days at mid-latitudes with the four-satellite constellation.
- Planet SkySat: 0.50 m panchromatic, video collection available. Useful for confirming construction activity between higher-resolution revisits and for detecting equipment presence. Constellation of 21 satellites provides flexible tasking, though the multispectral bands (4 bands, 1 m) are less useful for structural mensuration than for vegetation-clearance mapping around new dispersal areas.
- Sentinel-1 SAR (C-band, 5.405 GHz): Interferometric Wide Swath mode delivers 10 m ground-range resolution with 6-day repeat (12-day for a single satellite) over most landmasses, free of charge. New concrete pads and compacted gravel aprons produce a measurable increase in C-band backscatter, often detectable before optical imagery confirms the surface change. Coherence change between sequential passes isolates disturbed ground from stable surroundings.
What a revetment programme reveals about intent
A hardened aircraft shelter or earth revetment is not a defensive afterthought. It is a capital investment that takes months to complete and requires heavy plant, concrete, drainage and electrical infrastructure. Governments do not build them speculatively. The construction sequence, its pace and its geographic distribution across a country's airfield network therefore carry genuine intelligence weight, distinct from the order-of-battle questions covered elsewhere in this library.
The public record is useful here. CSIS and IISS analyses of Chinese airfield upgrades at bases including Kashgar, Hotan and Ream have shown that revetment and shelter construction programmes precede significant increases in aircraft basing capacity by twelve to thirty-six months. Russian airfield hardening in Kaliningrad and along the western military district has been tracked similarly through open-source imagery. These published analyses establish the interpretive baseline: construction geometry, sequencing and rate-of-effort are the signals, not simply the presence of finished structures.
Shadow geometry: reading height from a flat image
High-resolution optical imagery does not directly measure height, but shadows do the work if you know when the image was collected. At a solar elevation angle of 30 degrees, a shadow 10 metres long indicates a feature approximately 5.8 metres tall. WorldView-3 and Pléiades Neo imagery carries precise collection timestamps and sensor geometry metadata, so shadow-derived height estimates for revetment berms and shelter walls can be computed to within roughly 0.5 to 1 metre under good conditions.
The honest limits matter here. Shadow mensuration degrades when solar elevation is above roughly 60 degrees (shadows become very short and measurement error dominates), when shadows fall on uneven terrain, or when adjacent structures create overlapping shadow fields. Stereo-derived digital surface models from Pléiades Neo tri-stereo collection are more reliable for volumetric work, but they require dedicated tasking and add cost and latency. For routine monitoring, shadow analysis on single-pass imagery is the practical workhorse.
Revetment spacing is a secondary signal. Standard NATO dispersal doctrine places aircraft revetments far enough apart that a single weapon cannot destroy more than one aircraft. Spacing tighter than roughly 60 metres between centrelines suggests either a space-constrained site or a different doctrinal standard. Measuring spacing from imagery is straightforward at sub-metre resolution and requires no special processing.
SAR backscatter change: seeing concrete before the optical window opens
Cloud cover and tasking latency mean optical imagery does not always catch early construction phases. Sentinel-1 fills part of that gap. When a gravel apron or concrete pad is laid, the surface dielectric constant and roughness both change relative to bare soil or grass. C-band backscatter from Sentinel-1 typically increases by 2 to 5 dB over newly compacted gravel and by a larger margin over fresh concrete, depending on moisture content and surface texture. That change is detectable in a simple log-ratio of sequential GRD (Ground Range Detected) images even at 10 m resolution.
The method has honest limits. C-band does not penetrate more than a few centimetres into dry soil, so it cannot see subsurface drainage or foundation work. Seasonal moisture changes, agricultural activity and urban construction nearby can all produce similar backscatter increases, so SAR change alone is not sufficient for confident attribution. It is most useful as a cueing mechanism: a backscatter anomaly triggers a commercial optical task, rather than serving as the primary evidence product.
Coherence change detection adds a complementary layer. Between two Sentinel-1 passes separated by 6 or 12 days, areas of active construction lose interferometric coherence because the scatterer geometry changes. A coherence map therefore highlights the active construction footprint even when the backscatter change is ambiguous. Published studies in Remote Sensing (MDPI) have demonstrated this approach for urban construction monitoring; the physics transfers directly to airfield earthworks.
Construction sequencing as a rate-of-effort indicator
A single image tells you what exists. A time series tells you how fast it is being built, and rate-of-effort is often more informative than presence alone. Measuring the number of completed shelter bays per month, the expansion of the compacted apron area per quarter, or the progression of dispersal taxiway paving gives an estimate of the resources being committed and a projected completion date.
Planet SkySat's frequent revisit is useful for this, even though its resolution is lower than WorldView-3. Detecting the presence or absence of construction equipment, spoil heaps and concrete mixers on a near-daily basis allows activity days to be counted and work stoppages to be identified. A programme that pauses for several weeks and then accelerates sharply is a different signal from one that proceeds at a steady pace.
Archive depth matters for establishing the baseline. Sentinel-1 data is freely available from 2014 onwards via the Copernicus Data Space Ecosystem. Commercial optical archives from Maxar extend back to the late 1990s for some sites. Comparing current construction against a multi-year baseline distinguishes a genuine programme from routine maintenance.
What this method cannot do
Sub-metre optical imagery cannot see inside a completed hardened aircraft shelter. Once the doors are closed, the shelter's contents are opaque to any passive optical sensor. SAR can detect large metallic objects through some building materials at lower frequencies (L-band, P-band), but Sentinel-1's C-band does not penetrate reinforced concrete meaningfully. Inferring what aircraft type a shelter is sized for requires measuring the shelter's external footprint and comparing it against known type dimensions, which is a reasonable but indirect method.
Revisit gaps remain a genuine vulnerability in any monitoring regime. A six-day Sentinel-1 repeat means that a rapid construction sprint, common in Chinese military infrastructure projects documented by CSIS, can advance significantly between passes. Commercial tasking on WorldView-3 or Pléiades Neo is constrained by cloud, competing demand and cost. No current open or commercial constellation provides daily sub-metre coverage of a specific airfield as a routine product.
Satellize's analytics work is built on exactly these trade-offs: knowing which sensor to task when, how to combine SAR cueing with optical confirmation, and how to present uncertainty honestly to decision-makers. Our Tonga crop-estimation programme is a different domain, but the underlying discipline of multi-sensor time-series analysis and honest uncertainty quantification is the same.
Delivery and integration
The practical output of a revetment monitoring programme is not a raw image stack. It is a structured change log: which sites changed, by how much, over what period, with what confidence. That log needs to integrate with existing geospatial workflows, which in practice means GeoJSON or GeoPackage layers, georeferenced GeoTIFF change maps and periodic written assessments that contextualise the numbers.
Alert latency is a function of Sentinel-1 revisit (6 days minimum) plus processing time, which can be reduced to hours with automated pipelines. For commercial optical tasking, latency depends on cloud conditions and satellite availability, typically 1 to 5 days from tasking request to delivered image. Clients should plan monitoring cadence around the 6-day SAR floor rather than assuming daily optical availability.
Typical figures
| Best optical resolution (panchromatic) | 0.30 m (Pléiades Neo), 0.31 m (WorldView-3), 0.50 m (SkySat) |
| SAR resolution (Sentinel-1 IW mode) | 5 x 20 m (azimuth x range) in SLC; 10 m in GRD product |
| Optical revisit (commercial tasking) | 1 to 4.5 days (WorldView-3); 1 to 2 days (Pléiades Neo four-satellite constellation) |
| SAR revisit (Sentinel-1) | 6 days (two-satellite constellation); 12 days (single satellite) |
| Shadow-derived height accuracy | Approximately ±0.5 to 1 m at solar elevations of 20 to 50 degrees; degrades above 60 degrees |
| Minimum detectable surface change (SAR backscatter) | Typically 2 to 5 dB increase over new concrete or compacted gravel at C-band; site-dependent |
| Archive depth | Sentinel-1: 2014 to present (free). Maxar optical: late 1990s to present (licensed). Planet: 2016 to present (licensed). |
| Spectral bands (optical) | Panchromatic plus 4-band multispectral (blue, green, red, NIR) standard; WorldView-3 adds 8 VNIR and 8 SWIR bands |
| Delivery formats | GeoTIFF change maps, GeoJSON or GeoPackage site-change layers, PDF/structured assessment reports |
| Alert latency (SAR-cued pipeline) | 6 days minimum (Sentinel-1 repeat) plus processing; typically 6 to 24 hours post-acquisition with automated pipeline |
Analytics Satellize can run
| Revetment and shelter construction change log | Multi-date optical image differencing and manual mensuration of shadow geometry at sub-metre resolution | Quarterly GeoPackage layer with per-site change records, estimated berm heights and confidence ratings |
| SAR backscatter change alert | Log-ratio of sequential Sentinel-1 GRD images; threshold-based anomaly flagging over monitored airfield polygons | Automated alert (email or API push) within 24 hours of Sentinel-1 pass processing, with flagged GeoTIFF |
| Coherence-loss construction footprint map | Interferometric coherence change detection between 6-day or 12-day Sentinel-1 SLC pairs | GeoTIFF coherence-change map per acquisition pair, highlighting active earthwork and paving zones |
| Rate-of-effort time series | Automated apron-area and shelter-bay counting from multi-date commercial optical imagery using object detection on known shelter geometries | Monthly chart of completed bays, paved area (hectares) and projected completion date with uncertainty range |
| Stereo-derived digital surface model of earthworks | Photogrammetric processing of Pléiades Neo tri-stereo collections to produce a 0.5 m DSM; differenced against baseline DSM | GeoTIFF DSM difference layer with volumetric spoil and berm height estimates per revetment cell |
| Network-level basing posture assessment | Aggregation of per-site change logs across a defined airfield network; comparison against published doctrinal dispersal standards (NATO, PLA) from open IISS and CSIS analyses | Written assessment report with annotated imagery, site-by-site status table and analyst confidence ratings |
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.