Military fuel depot tank-volume and stock-change estimation
Floating-roof depression geometry in very-high-resolution optical imagery, combined with SAR double-bounce signatures, lets analysts estimate liquid fuel stock levels at military storage facilities without setting foot inside the perimeter.
Sensors
- Maxar WorldView-3: Panchromatic resolution of 0.31 m, multispectral at 1.24 m. At this resolution the shadow cast by a floating roof's pontoon ring is measurable to within a few centimetres of vertical displacement, enabling volume inference. Revisit roughly 1 day at mid-latitudes with off-nadir tasking.
- Airbus Pléiades Neo: 0.30 m panchromatic, 1.2 m multispectral, with a daily revisit capability across the constellation. Stereo and tri-stereo collection modes allow direct 3-D measurement of roof depression, reducing dependence on solar-angle geometry alone.
- Capella Space SAR (X-band): Spotlight modes achieve 0.35 m resolution. X-band double-bounce returns from cylindrical tank walls confirm tank presence, provide diameter estimates, and are unaffected by cloud or darkness. Cannot directly measure roof height, but wall-return intensity changes when a tank is empty versus full due to altered dielectric boundary conditions.
- Planet SkySat: 0.50 m panchromatic, up to 0.72 m multispectral. Lower resolution than WorldView-3 or Pléiades Neo limits shadow-angle precision, but the constellation's high revisit rate (multiple collections per day over tasked areas) is useful for detecting rapid stock drawdown events.
What a floating roof gives away
Most large liquid-fuel storage tanks above roughly 10,000 cubic metres use a floating roof: a pontoon deck that rides on the fuel surface and rises or falls with it. The design exists to suppress vapour loss, not to frustrate intelligence analysts, but the physical consequence is the same either way. When the tank is full, the roof sits near the rim. When it is nearly empty, the roof sits several metres lower. That vertical displacement is measurable from space.
The measurement relies on shadow geometry. A satellite image taken at a known solar elevation angle records the shadow cast by the roof rim onto the tank wall. Trigonometry converts shadow length to roof depression, and roof depression to approximate volume, given the tank diameter. Tank diameter is itself readable from the same imagery to within the resolution floor of the sensor. WorldView-3 at 0.31 m panchromatic makes this calculation tractable for tanks of 20 m diameter and above. Below that threshold, shadow-length uncertainty grows quickly relative to the quantity being estimated.
The civilian precedent and its defence application
This is not an exotic method. Commodity-market intelligence firms have applied floating-roof shadow analysis to commercial petroleum storage for years, estimating crude inventories at Cushing, Oklahoma and comparable hubs to inform energy trading. The physics is identical whether the tank holds West Texas Intermediate or military jet fuel. The defence application simply points the same workflow at facilities that do not publish their inventory figures.
The analytic value in a defence context is change detection over time, not a single absolute reading. A series of images spanning weeks or months can reveal drawdown rates consistent with fuelling a surge operation, resupply patterns that suggest convoy scheduling, or a sudden stock reduction that precedes offensive action. No single image is conclusive. A time series, cross-referenced against vehicle activity at the depot perimeter, is considerably more informative.
SAR as the all-weather complement
Optical shadow geometry fails completely under cloud. Military depots in humid or temperate climates can remain cloud-covered for days or weeks at a time. X-band SAR from systems such as Capella Space fills that gap, though it answers a different question.
A cylindrical metal tank wall produces a characteristic double-bounce return in SAR imagery: the radar pulse reflects off the ground and then off the vertical wall, or vice versa, creating a bright linear arc at the base of the tank. This signature confirms tank presence and allows diameter estimation from the arc's angular extent. The intensity of the return is also sensitive to whether the interior surface above the fuel line is metal or liquid, because the two have different radar reflectivities. This does not give a precise fill level, but a consistent reduction in double-bounce intensity across multiple passes is a credible indicator of drawdown. The method is published in the remote-sensing literature for industrial tank monitoring and carries across to military targets with the same geometry.
SAR and optical are therefore complementary rather than redundant. SAR maintains detection continuity through cloud; optical provides the higher-precision volume estimate when skies clear.
Hard limits the buyer must understand
Fixed-roof tanks produce no volume signal by either method. Many military facilities, particularly older Soviet-era or Chinese-pattern depots, use fixed-roof or underground cistern storage specifically because it is harder to monitor. The presence of fixed-roof tanks in a compound should be noted as a gap in coverage, not treated as evidence of empty tanks.
The method cannot distinguish fuel type. Jet fuel, diesel, petrol and lubricating oil all look identical from orbit. A depot that switches from aviation fuel to vehicle diesel shows no spectral signature of that change. Quality and contamination are similarly invisible.
Shadow-angle measurement introduces geometric uncertainty. At low solar elevations the shadow is long and small angular errors translate into large volume errors. At high solar elevations the shadow is short and harder to measure precisely. The practical working range is roughly 20 to 60 degrees solar elevation. Analysts should report volume estimates with an explicit uncertainty band, typically plus or minus 10 to 20 percent for well-characterised tanks, wider for smaller or partially obscured targets.
Archive depth matters. If a facility has not been systematically collected before, there is no baseline against which to measure change. Establishing a baseline requires commissioning a historical tasking review or, where commercial archive coverage exists, purchasing retrospective imagery.
Putting the workflow into practice
A practical monitoring programme begins with facility characterisation: identifying each tank, measuring its diameter from the sharpest available image, classifying it as floating-roof or fixed-roof, and recording its GPS coordinates. This one-time step defines the analytic universe for all subsequent collections.
Routine monitoring then tasks optical sensors at whatever revisit the client's budget and the facility's operational tempo justify, typically weekly to monthly for steady-state monitoring and daily during periods of elevated interest. Each new image is processed to extract shadow lengths, which feed into a volume model. Outputs are expressed as percentage fill and absolute volume in cubic metres, with uncertainty bounds. SAR passes, collected on a separate schedule, provide the cloud-gap fill and a qualitative drawdown indicator.
Satellize structures this kind of programme around open-constellation baselines supplemented by commercial tasking on client licence, the same architecture used in its crop-estimation work for the Kingdom of Tonga, adapted here for a facility-monitoring rather than agricultural context. Delivery formats are GIS layers with timestamped volume attributes, plus a periodic written assessment that contextualises stock changes against other observable indicators at the same facility.
What the numbers can and cannot tell a commander
A stock-change estimate from satellite imagery is an indicator, not a measurement. It belongs in an all-source picture alongside signals intelligence, human reporting and logistics modelling. Used in isolation it can mislead: a tank that appears to be drawing down may simply be undergoing maintenance with the roof temporarily lowered for cleaning. Corroborating evidence, such as tanker truck activity at the facility gate or pipeline flow changes visible in adjacent imagery, raises confidence materially.
The method is most reliable when applied consistently over time to a known facility, least reliable when applied once to an unfamiliar target. Buyers who want a single-image assessment of an unknown depot should expect wide uncertainty bounds and treat the result as a hypothesis to be tested rather than a finding to be acted upon.
Typical figures
| Best optical spatial resolution | 0.30 to 0.31 m panchromatic (Pléiades Neo, WorldView-3) |
| Minimum tank diameter for volume estimation | Approximately 20 m (shadow length becomes unreliable below this at typical solar angles) |
| Volume estimation uncertainty | ±10 to 20% for well-characterised floating-roof tanks; wider for partial occlusion or low solar elevation |
| Optical revisit (commercial tasking) | 1 day or better at mid-latitudes with WorldView-3 or Pléiades Neo off-nadir; Planet SkySat can provide multiple passes per day over tasked areas |
| SAR spatial resolution (spotlight mode) | 0.35 m (Capella Space X-band spotlight) |
| SAR revisit | Capella Space: sub-daily revisit possible with multiple satellites; standard tasking typically 12 to 24 hours |
| Cloud penetration | SAR: full; optical: none. Cloud cover is the primary operational constraint on volume estimation |
| Solar elevation working range | 20 to 60 degrees for reliable shadow-angle measurement |
| Commercial archive depth | WorldView series: back to 2009 in many regions; Pléiades: 2012 onwards; Capella SAR: 2020 onwards |
| Delivery formats | GeoTIFF orthoimage, GIS vector layer with timestamped volume attributes, PDF/structured assessment report |
Analytics Satellize can run
| Per-tank fill-level time series | Floating-roof shadow-angle photogrammetry using published trigonometric volume model; tank diameter from same imagery | GIS point layer with percentage fill and cubic-metre volume per tank per collection date, with uncertainty bounds |
| Stock-change alert | Automated comparison of successive volume estimates against a user-defined drawdown threshold | Email or API alert when any tank crosses threshold, with supporting image chip and computed change value |
| Facility characterisation report | Manual and semi-automated tank detection and classification (floating-roof vs fixed-roof) from best-available optical imagery | PDF baseline report with annotated imagery, tank inventory table, and coverage-gap assessment for fixed-roof or underground storage |
| SAR double-bounce tank presence and diameter map | X-band SAR double-bounce arc detection and angular-extent measurement for diameter estimation | GIS vector layer of confirmed tank locations with diameter estimates and per-pass intensity values; cloud-gap fill indicator |
| Multi-source drawdown assessment | Fusion of optical volume estimates, SAR intensity trends, and perimeter vehicle-activity counts from the same imagery | Periodic written intelligence assessment contextualising stock changes against observable logistics indicators |
| Historical baseline construction | Retrospective tasking review and shadow-angle processing of commercial archive imagery | Volume time series extending back to earliest available archive coverage for the facility, delivered as CSV and GIS layer |
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.