Short answer
Diesel expands as it warms, so the same mass of fuel occupies more litres when hot. Published correction factors put the change at roughly 0.00085 per degree Celsius, about 6.8 litres on a 400-litre tank across a 20-degree swing, which looks exactly like a slow theft on a chart.
Key takeaways
- Diesel volume is a function of temperature, so litres in a tank are not a fixed quantity of fuel.
- Published volume correction factors run 1.0000 at 15 °C, 0.9958 at 20 °C and 0.9873 at 30 °C.
- That is about 0.00085 per °C, roughly 6.8 litres on a 400-litre tank across a 20-degree swing.
- The industry standard is ASTM D1250 / API MPMS Chapter 11.1, with base temperatures of 15 °C and 20 °C.
- The expansion coefficient is not one constant: it is a density-dependent correlation, α60 = K0/ρ² + K1/ρ + K2.
01Litres are not a quantity of fuel
A litre is a volume, and volume is one of the properties of a liquid that changes with temperature. Warm diesel occupies more space than the same diesel cold, with no molecule added or removed. This is why a fuel tank that read 400 litres at dawn can read fewer litres at dawn the next day after a hot afternoon and a cool night, without anyone touching it. What has actually stayed constant through that cycle is the mass of fuel, and mass is what burns, what an engine consumes, and what a supplier has genuinely delivered. Volume is the convenient thing to measure and the wrong thing to treat as conserved. Most fuel disputes that end in a temperature explanation begin with someone reasonably assuming that litres in equals litres out, because for a solid or a countable object that assumption holds.
02The published numbers
This is not an estimate anyone has to make. Measurement Canada publishes volume correction factors for diesel: 1.0000 at 15 °C, 0.9958 at 20 °C and 0.9873 at 30 °C. Reading down that list gives the working figure of roughly 0.00085 per degree Celsius, and putting it on a real tank makes the size concrete, about 6.8 litres on a 400-litre tank across a 20-degree temperature swing. Note the direction. The factor is below 1.0000 above the base temperature, because the correction converts a warm, expanded volume back to what it would occupy at the reference temperature. So warm fuel measured at ambient reads high in litres and corrects downward. Nearly seven litres is not a rounding error on a fleet report, and it is comfortably inside the range where an alert rule would be justified in firing.
03Why the Gulf makes this bigger, not smaller
The correction is proportional to the temperature difference from the reference, not to how hot the fuel is in absolute terms, which means what matters is the daily swing rather than the peak. A tank sitting in direct sun on a Gulf afternoon and cooling under a clear desert night sees a large swing, and both halves of that cycle appear on the chart: fuel apparently appears during the heat and apparently disappears overnight. A tank kept in shade or buried sees a much smaller swing and much less apparent movement, which is why two sites with identical fuel and identical sensors can produce very different amounts of unexplained variance. The practical consequence is that a threshold tuned at one site is not transferable to another with a different thermal exposure. The same alert rule generates noise at the exposed site and silence at the sheltered one.
04The standard behind the numbers
The correction is codified rather than customary. Petroleum volumes are corrected to a standard reference temperature under ASTM D1250 / API MPMS Chapter 11.1, and the standard carries two metric base temperatures, 15 °C and 20 °C, with 20 °C incorporated in addition to the original 15 °C. The computation order is specified as well: 60 °F is computed first, and the metric output is then corrected to 15 °C or 20 °C. Diesel falls inside a named commodity group, Fuel Oils, covering approximately −10 °API to 37 °API. This matters practically because two correction figures for the same fuel at the same temperature can differ legitimately if they were computed against different base temperatures. When a supplier's corrected figure and yours disagree, the base temperature is the first thing to compare rather than the last.
05It is not one coefficient
The 0.00085 per °C figure is a working approximation for diesel around ambient, and it is worth knowing that the standard itself does not treat expansion as a single constant. In API MPMS 11.1 the thermal expansion coefficient is a density-dependent correlation, expressed as α60 = K0/ρ² + K1/ρ + K2 with K-coefficients assigned per commodity group. Two consequences follow. First, a denser or lighter product expands at a measurably different rate, so a coefficient borrowed from gasoline or a heavy fuel oil does not apply to diesel. Second, a blended or off-specification product sits somewhere the correlation was not built for, which means the correction becomes an estimate rather than a standard result. For everyday fleet reconciliation the single figure is sufficient. For a commercial dispute over a large delivery it is not, and that is when the density has to be measured rather than assumed.
06The correction is a required step
Measurement Canada does not present the correction as good practice; it instructs that the uncompensated meter reading be multiplied by the volume correction factor for the measured temperature. That framing changes how a fleet ought to treat the reconciliation. It means an uncorrected comparison between a delivery note and a tank reading is not a slightly imprecise comparison, it is an incomplete calculation, one term is missing. It also means the temperature at the moment of measurement is part of the measurement, not metadata. A fuel record carrying litres without the temperature they were measured at cannot be corrected afterwards, so the reading and the temperature have to be captured together. Where a sensor reports its own internal temperature, that value is the one worth logging alongside the level rather than an ambient figure taken elsewhere on site.
07Separating expansion from theft
Once the mechanism is understood, expansion and theft look different on a chart despite affecting the same number. Thermal change is slow, smooth and reversible: it tracks the daily temperature curve, and litres that vanished overnight come back the next afternoon. Theft is a step, a fast, one-directional drop that does not recover. The shapes are distinguishable, but only if the sampling is fine enough to show the shape at all. The OMNICOMM listing states a response time under one second, so the sensor is not the limit; the record interval configured on the tracker is. A single daily reading collapses both patterns into one unexplained delta and makes them indistinguishable. Two readings taken at the same hour each day, together with the temperature at each, is the cheapest arrangement that lets the two causes be told apart at all.
| Signal property | Thermal expansion | Siphon or drain |
|---|---|---|
| Rate of change | Slow, following the daily temperature curve | Fast, a step in the trace |
| Direction | Both ways, up in the heat, down at night | One way only |
| Reverses on its own | Yes, on the next cycle | No |
| Size on a 400 L tank | ≈6.8 L across a 20 °C swing | Whatever was removed |
| Correlates with tank temperature | Yes | No |
| Removed by applying a VCF | Yes, that is what the factor is for | No |
Figures and sources
- Diesel volume correction factors per API MPMS Ch. 11.1, and the apparent change they imply: VCF 1.0000 at 15 °C, 0.9958 at 20 °C, 0.9873 at 30 °C ≈ 0.00085 per °C (≈6.8 L on a 400 L tank over 20 °C) (Source: Measurement Canada (ISED), Volume correction factors, diesel fuel, as of 2026-08-31)
- Measurement Canada instruction that temperature correction is a required reconciliation step: multiply the uncompensated meter reading by the VCF for the measured temperature (Source: Measurement Canada (ISED), Volume correction factors, diesel fuel (PDF), as of 2026-08-31)
- ASTM D1250 / API MPMS 11.1 base temperatures and the required computation order: 60 °F computed first; metric output then corrected to 15 °C or 20 °C (Source: API MPMS Chapter 11.1 Addendum 1, as of 2026-08-31)
- ASTM D1250 metric base-temperature scope: 20 °C incorporated in addition to 15 °C (Source: ASTM D1250 official scope, as of 2026-08-31)
- The API MPMS 11.1 commodity group that governs diesel, by API gravity band: Fuel Oils: approximately -10 °API to 37 °API (Source: ASTM D1250 / API MPMS Ch. 11.1 scope, as of 2026-08-31)
- The thermal expansion coefficient in API MPMS 11.1 is a density-dependent correlation, not one constant: α60 = K0/ρ² + K1/ρ + K2, with K-coefficients per commodity group (Source: NFOGM / North Sea Flow Measurement Workshop, API/IP density referral constants, as of 2026-08-31)
- OMNICOMM response time and operating temperature range, the sensor side of a temperature-aware record: < 1 second; −40 °C to +85 °C; ±0.5% accuracy (Source: ENTEK.AI product catalogue, as of 2026-08-31)
- Teltonika record-acquisition triggers, the shape of a change is visible only if records are taken during it: four simultaneous methods: time (Min Period), distance, angle, speed delta; zero disables each (Source: Teltonika Telematics Wiki, FMB120 Data acquisition settings, as of 2026-08-31)
Questions
How many litres can temperature actually move?
Published correction factors give roughly 0.00085 per °C for diesel, about 6.8 litres on a 400-litre tank across a 20-degree swing. The factors themselves run 1.0000 at 15 °C, 0.9958 at 20 °C and 0.9873 at 30 °C.
Is the fuel actually lost when the volume drops?
No. The mass stays the same; only the volume it occupies changes. Mass is what an engine burns and what a supplier delivered. Volume is the convenient thing to measure and the wrong thing to treat as conserved across a temperature change.
Which reference temperature should I use?
Whichever your counterparty uses, and confirm it. ASTM D1250 carries two metric base temperatures, 15 °C and 20 °C, with 20 °C incorporated in addition to the original. Two correct corrections against different bases give different litres, so compare the base before disputing the number.
Can I use one expansion coefficient for every fuel?
Not rigorously. API MPMS 11.1 treats it as a density-dependent correlation, α60 = K0/ρ² + K1/ρ + K2, with K-coefficients per commodity group; diesel sits in Fuel Oils at roughly −10 to 37 °API. The single figure is fine for daily reconciliation, not for a commercial dispute.