Airbase order-of-battle change detection
Sub-metre optical imagery and Sentinel-1 SAR coherence change let analysts track aircraft presence, shelter occupancy, and apron activity at foreign airbases without setting foot near them. Readiness states can be inferred from open and commercial sources alone.
Sensors
- Planet SkySat: 0.5 m panchromatic, 0.8 m multispectral. Can be tasked for same-day or next-day revisit over a specific airbase. Aircraft as short as 12 m are resolvable; shadow geometry is legible enough to distinguish fighter-class from transport-class airframes.
- Maxar WorldView-3: 0.31 m panchromatic, 1.24 m multispectral, 16 SWIR bands. The sharpest routinely available commercial optical sensor. Parked aircraft, open or closed shelter doors, and ground-support equipment are all individually countable. Revisit is roughly 1 day at mid-latitudes with off-nadir tasking.
- Airbus Pléiades Neo: 0.30 m panchromatic, 0.75 m multispectral, stereo and tri-stereo capable. Stereo pairs allow height estimation of parked aircraft and shelters, which helps distinguish open-top revetments from closed hardened shelters. Revisit is 1 to 2 days with constellation scheduling.
- Sentinel-1 IW SAR: C-band, 5 × 20 m ground range resolution in Interferometric Wide Swath mode, 6-day repeat at the equator, 3-day at higher latitudes with both satellites. Penetrates cloud and darkness. Coherence change between repeat passes flags new or removed metal objects on aprons, including aircraft and support vehicles, even when optical collection is blocked.
What a parked aircraft gives away
A fighter aircraft on an apron is a surprisingly legible object from 500 km. At 0.31 m resolution, WorldView-3 resolves the planform well enough to separate a fourth-generation twin-engine type from a single-engine trainer by wingspan and fuselage length alone. Even at 0.5 m, the shadow is often more diagnostic than the aircraft itself: a shadow cast at a known solar elevation angle has a calculable length, which gives you aircraft height, and height narrows the type considerably.
Counting is the easy part. The harder analytical question is what absence means. An empty hardstand could mean the aircraft is airborne, relocated to a dispersal strip, in a maintenance bay, or destroyed. Temporal stacking, comparing the same apron across a sequence of images over days or weeks, turns a single ambiguous frame into a pattern. Persistent absence combined with increased ground-support vehicle activity near hangars suggests maintenance. Sudden absence across all hardstands simultaneously is a different signal entirely.
Shelter occupancy and the closed-door problem
Hardened aircraft shelters (HAS) are designed to frustrate exactly this kind of observation. A closed blast door tells you nothing directly about what is inside. Analysts have developed several indirect indicators. Blast deflectors or jet-blast erosion marks on the apron in front of a shelter suggest recent engine runs. Fuel hoses and ground-power cables snaking to a closed door indicate an occupied shelter. Thermal infrared, available on WorldView-3's SWIR bands and on some Pléiades Neo tasking configurations, can detect residual heat signatures from recently operated engines, though the detection window is narrow and dependent on ambient temperature.
SAR coherence adds a different dimension. A metal aircraft inside a shelter changes the radar return of the structure measurably, because the aircraft scatters C-band energy differently from an empty concrete floor. This effect is subtle and requires careful baseline selection, but published open-source intelligence methodology has documented it as a supporting indicator rather than a primary one. It is most useful when optical is cloud-blocked and you need any signal at all.
SAR coherence change for all-weather continuity
Sentinel-1's six-day repeat cycle is a genuine operational constraint. A lot can change at an airbase in six days. But the sensor's value is its indifference to weather. During sustained cloud cover over a contested region, optical collection may be impossible for days or weeks. SAR keeps collecting.
Coherence change detection works by comparing the phase correlation between two SAR passes over the same area. Stable surfaces, concrete, tarmac, buildings, maintain high coherence. New objects, or objects that have moved, decohere. An apron that was empty in the first pass and now holds a dozen aircraft will show a coherence drop in the aircraft footprints. The method is sensitive to any change in surface scattering, which means wind-blown debris or standing water can produce false positives. Experienced analysts filter these by cross-checking with meteorological records and by requiring the decoherence to persist across multiple passes. At 5 × 20 m resolution, individual aircraft are not resolved, but clusters and dispersal patterns are.
Commercial SAR constellations, including ICEYE and Capella Space, offer sub-metre resolution and much shorter revisit, though they are not discussed in detail here as their open-access documentation is less extensive than Sentinel-1's.
Building a readiness inference from multiple collections
No single image answers the readiness question. The methodology that has emerged from open-source intelligence practice, documented in academic remote sensing literature and publicly by organisations such as the Federation of American Scientists, treats airbase monitoring as a time-series problem. Analysts build a baseline: normal aircraft count, typical dispersal pattern, expected shelter occupancy for a given unit. Deviations from baseline, not absolute counts, are the signal.
Indicators that collectively suggest elevated readiness include: increased aircraft count on aprons relative to baseline, fuel bowsers positioned at multiple hardstands simultaneously, removal of protective covers from parked aircraft (visible as a colour change in multispectral imagery), and the appearance of additional ground-support equipment. Indicators of stand-down or degraded readiness include persistent shelter occupancy with no apron activity and the accumulation of maintenance equipment around specific airframes.
The honest limit here is type identification. At 0.5 m, you can count and roughly classify. Positive identification of a specific tail number or variant requires 0.3 m imagery with favourable sun angle and aircraft orientation. Even then, it is an inference, not a certainty.
Archive depth and the retrospective baseline
One underappreciated asset is the commercial archive. WorldView-3 has been collecting since 2014. Planet's archive extends to 2016 for medium resolution and to 2017 for SkySat. Sentinel-1 data is publicly available from 2014. This means that for most significant airbases in the world, a multi-year baseline of imagery already exists and can be queried retrospectively.
Retrospective analysis is often more valuable than near-real-time collection for establishing what normal looks like. It also allows analysts to date the introduction of new aircraft types or infrastructure changes, which has treaty-verification implications. Satellize runs this kind of archive-first baseline construction as a standard step before any operational monitoring programme begins.
The practical limits of archive analysis are access latency and tasking gaps. Commercial operators do not task every airbase every day. An archive query may reveal that a specific facility was collected only four times in a given year, which is enough for trend analysis but not for detecting a 48-hour surge deployment.
What this method cannot do
Cloud is the most obvious limit and SAR only partially solves it. Sentinel-1's 5 × 20 m resolution does not resolve individual aircraft; it detects change in clusters. Sub-metre SAR from commercial providers is available but expensive and not yet on daily revisit over arbitrary targets.
Camouflage and concealment degrade optical detection significantly. Netting over hardstands reduces shadow contrast. Aircraft painted to match apron colour reduce spectral contrast. These are not hypothetical concerns; they are standard dispersal and concealment practices documented in open military manuals. A well-disciplined adversary who understands satellite collection windows, which are increasingly predictable from published orbital parameters, can time movements to avoid collection.
Finally, counting aircraft on an apron tells you about disposition, not intent. An airbase with all aircraft sheltered and no visible activity could be at high readiness with engines warm, or it could be stood down for a national holiday. Satellite imagery is one input to an all-source picture, not a substitute for it.
Typical figures
| Best optical spatial resolution | 0.30 m panchromatic (Pléiades Neo, WorldView-3) |
| SAR resolution (Sentinel-1 IW) | 5 m range × 20 m azimuth (ground range detected) |
| Optical revisit (commercial tasking) | 1 day or less at mid-latitudes with off-nadir scheduling (WorldView-3, SkySat, Pléiades Neo) |
| SAR revisit (Sentinel-1 A+B) | 6 days at equator, approximately 3 days at latitudes above 45° |
| Minimum countable aircraft (optical) | Approximately 10 m wingspan at 0.5 m resolution; smaller targets at 0.3 m |
| Spectral bands (WorldView-3) | Panchromatic, 8-band VNIR, 8-band SWIR (1195–2365 nm), CAVIS atmospheric bands |
| SAR frequency (Sentinel-1) | C-band, 5.405 GHz |
| Archive depth | WorldView-3 from 2014; Sentinel-1 from 2014; Planet SkySat from 2017 |
| Typical collection latency (commercial) | 2 to 24 hours from tasking to analyst delivery depending on operator and licence |
| Delivery formats | GeoTIFF orthorectified imagery, GeoJSON change polygons, PDF intelligence summary, GIS-compatible vector overlays |
Analytics Satellize can run
| Aircraft count and type-class assessment | Object detection on sub-metre optical imagery using shadow-geometry analysis and planform measurement against published aircraft dimension references | Structured count table per apron zone, with type-class confidence rating, delivered as PDF report and GeoJSON point layer |
| Shelter occupancy map | Multispectral contrast analysis for open/closed door state combined with apron-activity indicators (fuel hoses, ground-power cables, blast deflector marks) visible at 0.3–0.5 m resolution | Per-shelter occupancy status layer (occupied / unoccupied / indeterminate) updated on each new optical collect |
| SAR coherence change alert | Interferometric coherence differencing between consecutive Sentinel-1 IW passes; decoherence polygons on apron surfaces flagged as new-object or removed-object events | Automated GeoJSON alert feed with coherence-difference magnitude score, triggered when change exceeds calibrated threshold |
| Baseline deviation score | Time-series comparison of aircraft count and dispersal pattern against a multi-month archive baseline; z-score deviation calculated per collection epoch | Weekly trend chart with deviation score and annotated imagery, delivered as PDF and CSV time series |
| Retrospective baseline construction | Archive query across Sentinel-1, WorldView-3, and Planet holdings; manual and semi-automated aircraft counting across all available epochs to establish normal operating patterns | Historical baseline report covering available archive depth, with annotated imagery for key change events |
| Surge or dispersal event report | Multi-source fusion of optical count change, SAR coherence shift, and ground-support equipment movement; cross-referenced against meteorological records to filter false positives | Single-event intelligence summary with supporting imagery, delivered within agreed latency window after trigger collection |
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.