Analysis

A barrel measures volume; a tonne measures mass. Moving between them requires the density of the actual liquid on compatible temperature and pressure bases. API gravity offers a familiar petroleum reporting scale, but a quoted API value alone does not turn an operating tank volume, refinery capacity or well observation into a comparable mass measurement.

This guide uses the U.S. Energy Information Administration’s API definition and a published NIST gravimetric-meter example. Its three oil samples, parcel and blend are hypothetical engineering inputs, not reported field data. Calculations retain unrounded intermediate values. The illustration is conceptual, and the chart shows arithmetic derived from the stated inputs rather than a measured assay or forecast.

Start with the reference conditions

EIA defines API gravity through the specific gravity of petroleum at 60 °F relative to water at 60 °F. Specific gravity, SG, is a dimensionless density ratio. The two temperatures matter: 60 °F is approximately 15.56 °C, and an observed density at another temperature is not automatically the same reference quantity. A laboratory report should identify the sample, method, temperature and reporting basis.

Keep the fluid definition with the number. Crude oil, condensate, a refined product and drilling mud are different materials even when their reported density ratios happen to resemble each other. The examples below concern a single petroleum liquid at the stated reference condition. They do not provide a generic correction for water, suspended solids, entrained gas or changing phase composition.

Convert API gravity to specific gravity

The EIA relationship is °API = 141.5 / SG − 131.5. Rearrangement gives SG = 141.5 / (°API + 131.5). Thus higher API gravity means lower reference specific gravity. Water on this relative-density basis has SG = 1 and therefore 10 °API. The formula defines a reporting scale; it does not supply an oil composition or a viscosity.

For the hypothetical 35 °API sample, SG = 141.5 / 166.5 = 0.84984985 after rounding. The 20 and 50 °API samples give 0.93399340 and 0.77961433 respectively. Substituting these unrounded results back into the API equation recovers the inputs. Rounding to 0.85 early may be adequate for a quick estimate but should not replace the original value in a reproducible calculation.

Turn the ratio into a density with units

NIST’s August 2005 article on gravimetric testing of oil meters gives density at 60 °F in pounds per U.S. gallon as SG × 8.337193. Its published 28.3 °API example gives approximately 7.38 lb/US gal, which our calculation reproduces. This is a historical technical reference for the conversion relationship, not a claim that the article replaces current metering requirements.

Using the same coefficient, the 35 °API sample gives 7.08536222 lb/US gal. One petroleum barrel contains 42 U.S. gallons, one avoirdupois pound equals 0.45359237 kg, and one metric tonne is 1,000 kg. EIA and NIST list approximately 0.1589873 m³ per petroleum barrel. Keeping the U.S. gallon explicit avoids substituting the different Imperial gallon.

The illustrative mass equation is m[t] = V[bbl] × 42 × SG × 8.337193 × 0.45359237 / 1000. Dividing the mass per barrel in kilograms by the rounded volume factor gives approximately 849.01 kg/m³ for the 35 °API sample. The displayed SI density is consequently approximate. Retaining extra calculation digits improves reproducibility; it does not create extra precision in the source coefficient or sample measurement.

Calculate a parcel and compare the three cases

Assume a hypothetical 100,000 bbl parcel with its volume already expressed at 60 °F and with a compatible pressure basis. For 35 °API, the equation gives 13,498.24 t after rounding. The reciprocal factor is approximately 7.4084 bbl/t. A generic barrels-per-tonne factor would lose the dependence on liquid density. Applying the factor to a volume measured at an uncorrected tank temperature would introduce a different, unresolved problem.

Holding the reference volume at 100,000 bbl, the 20 °API case gives 14,834.70 t and the 50 °API case gives 12,382.68 t. The accompanying chart compares those calculated masses. This is an isolated density comparison: it does not measure refinery throughput, product yield, crude quality beyond density, or a change in physical production. A daily volume would require its own time denominator before being reported as a mass rate.

Blend density before converting back to API

Consider a second hypothetical calculation: equal reference volumes of the 20 and 50 °API liquids, with volumes assumed additive and no material loss. Average the specific gravities under those assumptions: SGblend = (0.93399340 + 0.77961433) / 2 = 0.85680386. Applying the API formula to the unrounded mean gives 33.64865 °API, or about 33.65 °API.

Averaging the two API numbers directly gives 35 °API, a different answer because the API transformation is nonlinear. The equal-volume assumption is essential; equal masses need a different volume calculation. Real mixtures can also depart from volume additivity. For an actual transfer or blend, use the relevant measured mixture properties and applicable correction method rather than treating this teaching example as a certified blend prediction.

Separate numerical sensitivity from measurement uncertainty

For fixed reference volume and coefficient, mass is proportional to 1 / (°API + 131.5). A first-order relative sensitivity is therefore Δm/m ≈ −ΔAPI / (°API + 131.5). Around 35 °API, a hypothetical ±0.5 °API change corresponds to approximately ±0.3003% mass change when all other inputs remain fixed.

That result is a local sensitivity, not a confidence interval or an instrument accuracy claim. A measurement uncertainty budget must also consider the volume measurement, sampling, temperature determination, calibration and any dependence between inputs. NIST’s meter-testing discussion explains why product data sheets can require assumptions about successive batches. More decimal places in an API-to-SG conversion do not remove those assumptions.

Keep engineering calculations separate from dataset observations

An annual refinery rating in tonnes cannot be converted into barrels per day responsibly without an appropriate density and time basis. Even a correct mass–volume conversion leaves calendar-day capacity, stream-day capacity and actual annual throughput as different quantities. The related refinery-capacity guide explains those denominators. These hypothetical densities are not new capacity values for any Refinery Atlas record.

The same discipline applies to wells. A drilling-fluid density used in hydrostatic-pressure calculations is not a measurement of produced-oil API gravity. A source field’s label, fluid and reference conditions must survive any reuse. Linking technical reading does not establish that two records describe the same physical well, that a field supplies a refinery, or that an unreported fluid property is known.

A reproducible workflow and its limits

Record the original API value and its reference basis; calculate SG without early rounding; apply the explicitly chosen density coefficient; carry units through the barrel, gallon, pound and tonne conversions; and round only the displayed result. A reverse calculation should recover the original API value and parcel volume within the stated rounding. Preserve both source references and the calculation assumptions when sharing the result.

For a real engineering application, establish whether the measured volume and density refer to the same sample, phase and conditions. Obtain appropriate temperature or pressure corrections when they do not. API gravity alone supplies neither those corrections nor a complete crude assay. The examples here are reproducible definitions and arithmetic reviewed on 11 November 2026; they are not operating instructions, a metering certification or a production forecast.