Reconciliation isn’t broken — you’re comparing the wrong rock: Why underground mines are solving the wrong problem

A recent industry webinar hosted ahead of the AusIMM International Mining Geology Conference 2026 brought together leading practitioners to tackle one of underground mining’s most persistent challenges: reconciliation.

The session was presented by Rayleen Hargreaves, product owner – reconcile at Datamine, a geologist with more than 30 years of industry experience. She was joined by Ian Glacken, executive consultant at Snowden Optiro and a recognised leader in resource estimation and reconciliation, Tarrant Elkington, general manager at SinoTech Europe with a PhD in mining engineering, and Arja Jewbali, head of global resource management at Newmont, who leads a global team of resource geologists. Together, the panel explored a structured framework for improving how operations compare models, mine performance and plant outcomes.

The discussion centred on a simple but confronting idea.

Reconciliation is not failing because of geology or data.

It is failing because most operations are comparing the wrong things.

The Illusion of Accuracy

Reconciliation reports are typically produced monthly, supported by increasingly sophisticated datasets. Resource models, reserve plans, grade control, production data and plant performance metrics are brought together and compared through established frameworks such as the F-series. On paper, the process appears robust. In practice, it often produces numbers that are difficult to interpret and even harder to trust.

As Hargreaves told attendees:

“We’re very good at reporting differences, but we’re not always as good at explaining where those differences actually come from.”

Variance is observed, reported and often accepted as part of the inherent uncertainty of underground mining. But that assumption is increasingly being challenged.

Apples, Oranges and Ore Bodies

Underground mining introduces complexity that open pit operations largely avoid. Selective sampling, irregular stope geometries, limited drilling density and high short-range grade variability all contribute to uncertainty. However, the most critical issue lies elsewhere.

Different models across the mining value chain are frequently evaluated over different volumes of rock. The resource model represents one interpretation. The reserve incorporates modifying factors such as dilution and ore loss. The mined volume reflects actual extraction geometry. The plant feed is influenced by stockpiles, timing and processing realities.

When these datasets are compared directly, they are not describing the same physical material.

As Hargreaves explained:

“When we compare tonnes, grade or metal… we’re often not comparing like for like.”

The result is a reconciliation outcome that reflects geometry mismatch rather than operational performance. This is the industry’s blind spot.

Introducing Spatial Consistency

The framework presented during the webinar introduces a deceptively simple but powerful concept: the common volume. Rather than allowing each dataset to retain its own spatial definition, all models are evaluated within a single, unified mining footprint that represents the total material mined or effectively removed during a reporting period.

Each dataset is then assessed against that same volume. The result is a like for like comparison that fundamentally changes what reconciliation is measuring.

As Hargreaves put it:

“Now, when we compare… we’re actually comparing like for like, and fundamentally, that changes what reconciliation is telling us.”

This shift moves reconciliation from reporting differences to explaining them.

Beyond the F-Series

The traditional F-series framework remains valuable because it answers a key question about whether operations are delivering to plan. However, on its own, it does not explain why performance deviates.

To address this, the framework extends reconciliation through the R-series, which tracks how modifying factors such as dilution, ore loss and recovery propagate through the system. By linking the resource model to reserve, mine and plant outcomes within the same spatial volume, the R-series provides visibility into how these factors influence final performance.

The result is a clearer understanding of how value is gained or lost as material moves through the mining value chain.

Case Study: Tanami Gold Mine

A central part of the webinar was a case study from Newmont’s Tanami underground operation, a large-scale gold mine characterised by narrow, high-grade ore bodies mined using long-hole stoping.

Tanami presents a complex reconciliation environment, with significant planned dilution and multiple datasets across the value chain. Under traditional approaches, variability in tonnes and grade could be difficult to interpret, and it was often unclear whether differences reflected performance issues or inconsistencies in how volumes were defined.

When recalculated using a common volume, the picture changed.

Tonnes increased and grade decreased relative to the resource model, which might typically suggest underperformance. However, the R-series analysis showed that these changes were the direct result of planned dilution being applied consistently through the system. Contained metal remained aligned with expectations.

As Hargreaves noted:

“The real value… was not in changing the result, but in making that result transparent, explainable and defensible.”

The operation had not changed. The understanding had.

From Reporting to Understanding

For mining professionals, this distinction is critical. Reconciliation has long been treated as a reporting exercise, often driven by compliance or governance requirements, with the focus on producing numbers rather than interrogating their meaning.

That mindset limits its value.

Ian Glacken was direct in his assessment during the session:

“Reconciliation… is often treated as a compliance thing, when in fact it’s there to tell a story and understand what’s actually happening.”

Unlocking that value requires integration across disciplines. Geology, mining, surveying and processing teams must align their data and assumptions. Without that alignment, reconciliation remains fragmented and prone to misinterpretation.

As Glacken added:

“Reconciliation does tend to be siloed.”

Persistent Industry Challenges

The webinar discussion also highlighted several persistent challenges. Stockpiles continue to distort reconciliation outcomes, particularly where blending obscures the relationship between mined material and plant feed. Timing delays between extraction and processing introduce further complexity.

Data management remains another constraint. Many operations still rely heavily on spreadsheets, increasing the risk of error and limiting efficiency.

Hargreaves was blunt on this point:

“Getting that data together… is so time consuming. It’s prone with error… we need to get it out of Excel.”

These are not technical limitations. They are operational ones.

A Shift in Perspective

The key takeaway from the session is straightforward. Most reconciliation issues are not caused by poor geology or insufficient data. They are caused by invalid comparisons.

As Hargreaves concluded:

“The biggest source of error… is not always the geology, it’s how we’re comparing the model.”

Before interpreting any reconciliation result, operations should ask a single question: are we comparing the same material?

If the answer is no, the conclusions are unreliable. If the answer is yes, reconciliation becomes one of the most powerful diagnostic tools in the mining value chain.

The Future of Reconciliation

As underground mining continues to expand, the complexity of reconciliation will only increase. Frameworks that prioritise spatial consistency and system-level understanding will become standard practice.

The opportunity is clear. Reconciliation can move beyond compliance and become a driver of operational performance.

But only if the industry stops comparing apples with oranges and starts comparing like with like.

As noted during the session, the webinar was based on a paper included in the proceedings of AusIMM International Mining Geology Conference 2026, taking place in Brisbane on 21–22 April. It offered a preview of the technical depth expected at the event, while reinforcing the value of in-person collaboration across the industry.