A gold operation estimates the gold in the ground long before it is mined, estimates it again more closely just before mining, measures it again when the ore is fed to the mill, and finally weighs and assays the gold bars it sells. Each of these figures is an estimate made with different tools and subject to different errors. Reconciliation aligns them, period after period, so that the business knows which estimates can be relied on and where gold is being gained or lost.
Gold does not pass through a processing plant instantly. Ore fed to the mill today may not become a gold bar for several days, because the dissolved gold must load onto carbon, the carbon must be stripped, and the gold must be electrowon and smelted. Gold anywhere along that path at the end of the month is gold in circuit (GIC). It is real gold, it belongs in the month's production, and it is readily overlooked.
A complete reconciliation establishes whether the gold adds up from the resource model to the gold bar. This article covers the whole chain, the factors used to measure agreement, the plant balance, and the element most often mishandled: gold in circuit and the way the production schedule raises and lowers it.
The reconciliation chain
Gold is estimated or measured at several points between the orebody and the refinery, and each point produces its own tonnes, grade and contained gold for the same period:
| Stage | Basis |
|---|---|
| Resource or reserve model | The long-term estimate built from widely spaced exploration drilling |
| Grade control model | The short-term estimate built from close-spaced drilling just before mining, which determines what is sent to the mill as ore and what goes to waste |
| Mined and hauled | Truck counts and surveys of the material that left the pit, by destination |
| Stockpiles | Ore placed on and reclaimed from run-of-mine (ROM) stockpiles, with surveyed volumes and estimated grades |
| Mill feed | Weighed on a belt weightometer and sampled, usually after the primary crusher or at the mill feed conveyor |
| Plant outputs | Gold poured, gold remaining in the tailings, and the change in gold in circuit |
| Refinery outturn | The refinery's final weighed and assayed gold credited to the company: the most accurate figure in the chain, against which all others are ultimately tested |
F1, F2 and F3: the reconciliation factors
Most Australian operations express agreement between stages as ratios, usually called F-factors after the framework popularised by Morley. Each factor is calculated separately for tonnes, grade and contained gold:
where
- how well the resource model predicted what grade control found, for the same mined volume
- how well grade control predicted what the mill received
- the overall agreement between the resource model and the mill
A factor of 1.00 indicates perfect agreement, and a factor above 1.00 indicates that the later stage found more than the earlier stage predicted.
Worked example
Illustrative figures for one quarter, all for the same mined volume:
| Source | Tonnes | Grade (g/t) | Gold (kg) |
|---|---|---|---|
| Resource model | 260,000 | 1.70 | 442 |
| Grade control model | 280,000 | 1.65 | 462 |
| Mill (reconciled) | 275,000 | 1.72 | 473 |
| Factor | Tonnes | Grade | Gold |
|---|---|---|---|
| F1 (grade control ÷ resource) | 1.077 | 0.971 | 1.045 |
| F2 (mill ÷ grade control) | 0.982 | 1.042 | 1.024 |
| F3 (mill ÷ resource) | 1.058 | 1.012 | 1.070 |
Grade control found 7.7% more ore tonnes than the resource model at a slightly lower grade, and therefore 4.5% more gold. The mill then received slightly less tonnage than grade control reported but at a higher grade, a further 2.4% more gold. Overall, the mill received 7% more gold than the resource model predicted. A pattern of this kind repeated over several quarters would prompt a review of the resource model by the geologists; a single quarter is not sufficient grounds for action.
The mine call factor
In South African underground gold mining, the equivalent measure is the mine call factor: the gold recovered at the mill plus the gold in the residue, expressed as a percentage of the gold that the mine's own sampling and measurement indicate was sent. It addresses the same question as F2 and F3 in a single figure.
The plant gold balance
Within the plant, every gram of gold that enters must be accounted for: it has been poured, it remains in the circuit, or it has left in the tailings:
where
- gold fed to the plant: dry tonnes × head grade (g)
- gold poured into bars during the period (g)
- change in gold in circuit: closing GIC minus opening GIC (g)
- gold leaving in the tailings solids: dry tonnes × residue grade (g)
- dissolved gold leaving in the tailings solution (g)
The two recovery figures of interest are then:
where
- plant recovery including the change in gold in circuit: the true recovery for the period
- recovery based on poured gold alone, which is distorted whenever GIC changes
The difference between these two expressions is the subject of the remainder of this article. If GIC changes during the month and is not measured, the poured-only recovery will be in error, sometimes by several percent.
Because every figure in the balance carries measurement error, the balance will rarely close exactly. Closing it correctly is a statistical task, covered in mass balance data reconciliation.
Gold in circuit
In a carbon-in-leach (CIL) or carbon-in-pulp (CIP) plant, gold in circuit is held in several locations:
| Location | Gold held |
|---|---|
| Solution in the leach tanks | Dissolved gold not yet adsorbed onto carbon |
| Carbon in the tanks | Gold loaded onto the activated carbon in each tank, usually the largest single component |
| Loaded carbon | Carbon removed from the circuit and awaiting elution, or in the elution column |
| Elution and electrowinning | Gold in the eluate and plated on electrowinning cathodes |
| Gold room | Cathode sludge and other material not yet smelted into bars |
| Gravity circuit | Concentrate awaiting intensive leaching, and gold in the intensive leach solution |
| Mill circuit | Coarse gold trapped behind liners, in sumps and in spillage, which SRK describes as the "miscellaneous" part of the inventory |
Measurement of gold in circuit
A GIC survey is a snapshot taken at the exact cut-off time for the period. The two main calculations are simple:
where
- volume of solution in the tank (m³)
- gold in solution (g/m³, which is the same as mg/L)
- mass of carbon in the tank (t)
- gold loading on the carbon (g/t)
- volume of slurry in the tank (m³)
- carbon concentration in the slurry (g/L)
Each tank is sampled for solution grade and for carbon concentration and gold loading, and tank volumes are known from the design. Loaded carbon in bins and columns is weighed or estimated by volume and sampled. Elution, electrowinning and gold room inventories are measured from solution volumes, assays and weighed sludge. An illustrative survey for a mid-sized CIL plant follows:
| Location | Basis (illustrative) | Gold (kg) |
|---|---|---|
| Solution in CIL tanks | 6,900 m³ of solution at 0.30 g/m³ | 2.1 |
| Carbon in CIL tanks | 9,000 m³ of slurry at 10 g/L carbon = 90 t, at 300 g/t | 27.0 |
| Loaded carbon in elution | 4 t at 3,000 g/t | 12.0 |
| Eluate and electrowinning | Solution and cathode estimate | 6.5 |
| Gold room, unsmelted | Weighed sludge × assay | 4.0 |
| Gravity concentrate and intensive leach | Concentrate mass × assay | 3.6 |
| Total | 55.2 |
In this example the circuit holds 55 kg of gold, roughly four days of production. Gold in the mill circuit is excluded because it is very difficult to measure. This is a common and defensible choice, provided it is documented and applied consistently every month.
Scheduling effects on gold in circuit
Gold in circuit is not a fixed quantity. It rises and falls with the plant feed and with the timing of the downstream steps, and the mine schedule, the plant maintenance schedule and the elution and smelting schedule all affect it. The examples below quantify these effects for one illustrative plant, milling 8,333 t/d (250,000 t in a 30-day month) at a normal head grade of 1.8 g/t, with a true plant recovery of 92%, so that about 13.8 kg of gold is recovered from the ore each day. Gold takes four days on average to travel from the leach to a poured bar, so the circuit normally holds four days of recovered gold, 55.2 kg.
Example 1: a high-grade campaign at month-end
The mine schedule sends high-grade ore, at 3.0 g/t instead of 1.8 g/t, for the last five days of the month. The plant recovers it at the usual 92%, but the additional gold is still on carbon and in elution at the cut-off. Gold fed in the month is 500 kg and gold recovered from the ore is 460 kg. Gold poured is 423.2 kg, because the last few days of high-grade gold have not reached the gold room, and gold in circuit rises from 55.2 kg to 92.0 kg, an increase of 36.8 kg.
Poured gold divided by gold fed gives 84.6%. Including the increase in GIC gives the true recovery of 92.0%. Without a GIC survey, the plant would report a recovery fall of 7 percentage points that did not occur, followed by an apparent increase the next month when that gold is poured.
Example 2: a delayed elution
One elution strips 6 t of loaded carbon at 3,000 g/t, or 18 kg of gold. If the elution circuit is down for maintenance and two strips are deferred from the last days of month A into month B, 36 kg of gold moves from month A's pour to month B's. With 450 kg fed and 414 kg truly recovered each month:
| Month | Gold poured | Poured ÷ fed | Change in GIC | True recovery |
|---|---|---|---|---|
| A (strips delayed) | 378 kg | 84.0% | +36 kg | 92.0% |
| B (strips caught up) | 450 kg | 100.0% | −36 kg | 92.0% |
A recovery of 100% is a reliable indication that GIC has been drawn down, not that the plant has become perfect.
Example 3: a shutdown at month-end
A planned two-day mill reline is scheduled for the last two days of the month. Feed stops, but elution and smelting continue, so the plant continues to pour gold already in the circuit. Gold fed is 420 kg over 28 days and gold recovered from that ore is 386.4 kg. Gold poured is 414 kg, and GIC falls from 55.2 kg to 27.6 kg.
Poured gold divided by gold fed gives 98.6%, against a true 92.0%. The reverse occurs the following month as the circuit refills.
Example 4: blend and ore-type changes
Scheduling also changes the ore itself. A change to harder ore lowers throughput, and a change to slower-leaching or preg-robbing ore alters the amount of gold in solution and on carbon at any time. A change in carbon concentration or carbon advance rate alters the quantity of carbon, and therefore of gold, held in the tanks. None of these are errors. They are real changes in inventory, and they distort the month only if they are not measured.
Managing scheduling effects
GIC should be surveyed at every reporting cut-off, with the same method, locations and time each month, and recovery reported as gold poured plus the change in GIC. Poured-only recovery may be shown alongside it if management requires it, but should be clearly labelled.
When the mine plans a high-grade campaign or a stockpile reclaim near month-end, the plant should be informed, so that elution and smelting capacity can be planned and the GIC change anticipated. Where possible, elution and smelting should continue at a steady rate across the month-end rather than being accelerated or deferred around the cut-off. Rolling three-month figures should be reviewed alongside single months, because timing effects largely cancel over a quarter while real changes in recovery remain.
The reconciliation procedure
- Fix a single cut-off date and time for every measurement: mining, stockpiles, mill feed, GIC survey and gold room.
- Collect the mining data: grade control tonnes and grades for the material mined, by destination, and truck or survey tonnes.
- Update the stockpiles: opening balance plus additions less reclaims, checked against a survey of stockpile volume and an agreed density.
- Close the plant balance from mill feed tonnes and grade, tailings solids and solution, gold poured and the change in GIC, reconciling the measurements so that the balance closes, as set out in mass balance data reconciliation.
- Compare poured bar weights and assays with the refinery outturn when it is received, and carry any difference forward.
- Calculate F1, F2 and F3 for tonnes, grade and gold, for the period and as a rolling trend.
- Investigate any result outside the agreed limits. The variance that triggers an investigation should be agreed in advance, and a structured approach used, such as the metallurgical investigation method.
- Report and sign off, recording the data, the adjustments, their approval and the explanations. The AMIRA P754 Code of Practice expects metal accounting to be transparent and auditable.
Common errors
The most common errors are different cut-off times for the mine, the plant and the gold room; the absence of a GIC survey, or a survey that omits the elution and gold room inventory; comparison of wet tonnes with dry tonnes, or use of an out-of-date moisture or density figure; and stockpile grades that are never checked against the grade the mill received when they were reclaimed.
Forcing the figures to agree by adjusting whichever is easiest, instead of finding the cause, defeats the purpose of reconciliation, as does judging a single month rather than the trend.
Saldora holds the daily plant figures, the month-end reconciliation and the change history in one place, so that each month's adjustments and their reasons remain on record. Theseus can assist with a reconciliation variance by listing likely causes and the evidence for each.
Glossary
- Reconciliation
- Comparing what one stage of the operation estimated with what the next stage measured, and explaining the difference.
- Resource model
- The long-term estimate of tonnes and grade in the ground, built from exploration drilling.
- Grade control model
- The short-term estimate made from close-spaced drilling just before mining, used to decide what is ore and what is waste.
- ROM stockpile
- Run-of-mine stockpile: ore stored between the pit and the crusher.
- F1, F2, F3
- Ratios comparing grade control with the resource model (F1), the mill with grade control (F2), and the mill with the resource model (F3).
- Mine call factor
- Gold recovered plus gold in residue, as a percentage of the gold the mine's measurements say was sent to the plant.
- Gold in circuit (GIC)
- Gold recovered from the ore but not yet poured into bars, held in solution, on carbon, in elution, electrowinning and the gold room.
- Elution
- Stripping gold off loaded carbon with a hot caustic cyanide solution so it can be recovered by electrowinning.
- Electrowinning
- Plating gold out of solution onto cathodes using an electric current.
- Refinery outturn
- The refinery's final statement of the gold and silver it credits from each shipment of bars.
- Cut-off time
- The single date and time at which every measurement for a reporting period is taken.
Frequently asked questions
Why does reported recovery vary so much from month to month?
Usually because gold in circuit changed and was not counted. Gold recovered near the end of one month is poured in the next. Including the change in GIC removes most of the variation.
How often should gold in circuit be measured?
At every reporting cut-off, which for most operations is monthly, and always with the same method and locations.
What is a good F3 factor?
There is no single correct value. Each operation agrees its own limits based on its geology and measurement quality, and monitors the trend over several periods rather than a single month.
Can recovery exceed 100%?
Not in reality. A reported recovery above 100% means that gold in circuit was drawn down or that the feed grade was understated.
Should gold behind mill liners be counted in GIC?
It is real, but very difficult to measure. Many operations exclude it and state that they do. The important point is to apply the same rule every month and to estimate it when the mill is relined.
Discuss your plant with a metallurgist
Tiegel is a Perth metallurgy company. We help with testwork, plant optimisation, metal accounting and investigations, and we build the Tiegel Suite software.
Talk to a metallurgistReferences and further reading
- Reconciliation principles for the mining industry. AusIMM, 35th APCOM Symposium (2011). Link
- Ebrahimi, A. (2025). Reconciliation explained and solutions recommended. SRK Consulting. Link
- SRK Consulting. Gold-in-circuit inventory measurement. Link
- Applying the AMIRA P754 Code of Practice for Metal Accounting. AusIMM MetPlant 2008. Link
- Mine call factor definition, from a Sibanye Gold pre-listing statement. Law Insider. Link
Worked examples and charts marked illustrative use invented figures, not data from any operation.