Rodney Garrard examines how geoscientists’ roles must now extend beyond site characterisation into uncertainty governance and institutional realism, and whether alternative concepts such as deep borehole disposal may offer proportionate options alongside large, mined repositories.
Abstract
Nuclear energy has re-entered the global policy and investment arena, driven by concerns around energy security, system-scale decarbonisation, and grid stability. Yet its long-term credibility remains contingent not on reactor performance, but on how societies manage nuclear refuse - the terminal by-product of the fuel cycle (This article uses ‘nuclear waste’ in its broad, encompassing sense, consistent with common usage. By contrast, ‘spent fuel’ reflects a more sanitised, policy-laden term that preserves the notion of residual value via potential reprocessing. Therefore, in material terms, the term ‘refuse’ more directly captures this finality, chain-terminally, but it is not standardised in IAEA / NEA usage).
This article argues that while deep geological disposal is underpinned by decades of conservative subsurface science and a mature safety philosophy, the dominant risks facing new high-level disposal programmes have migrated away from geology and into institutional, financial, and societal domains.
Drawing on experience across geoscience, the upstream oil and gas sector, nuclear refuse management, and insurance, the article contends that precisely because geological concepts and safety cases are now robust, governance quality and cost realism have become the primary determinants of programme success. Weak institutional design, blurred accountability, and persistent optimism bias in cost estimation increasingly threaten otherwise sound disposal strategies.
The article examines how geoscientists’ roles must now extend beyond site characterisation into uncertainty governance and institutional realism, and whether alternative concepts such as deep borehole disposal may, in specific national contexts, offer proportionate options alongside large, mined repositories (Figure 1). Ultimately, it argues that the credibility of nuclear energy will be determined less by subsurface performance than by societies’ willingness to confront uncertainty, cost, and institutional limits with intellectual honesty.
Introduction: Credibility lives in the back end
Nuclear energy has returned to the policy and investment conversation with a momentum few would have predicted a decade ago. Energy security, ‘net-zero’ ambitions, and the physics of grid stability have realigned political incentives in nuclear’s favour. Yet this resurgence remains fragile. Nuclear does not lose credibility in the reactor hall; it loses it in the refuse debate.
That reality is neither new nor controversial. What has changed is where the dominant uncertainty now resides.
This article builds on a recent EAGE presentation, June 8, 2026 in Aberdeen, and on five years spent working directly in the nuclear refuse sector, following an earlier career in the E&P industry. Across this period – including involvement with geological disposal programmes, waste packaging strategies, and long-term safety cases – one conclusion has become increasingly clear: the scientific and geological foundations of deep disposal are stronger than ever, but institutional confidence has not kept pace.
Four connected arguments are developed. First, deep geological disposal represents a scientifically mature and conservatively engineered concept, even though construction and operation of high-level waste (HLW) repositories at scale remain first-of-a-kind (FOAK). Second, the dominant risks to disposal programmes are now institutional rather than geological. Third, geoscientists – particularly those trained in upstream subsurface disciplines – have an under-recognised role in restoring credibility by insisting on disciplined realism around uncertainty, time, and cost. Finally, while large national repositories remain essential for major nuclear nations, alternative concepts such as deep borehole disposal deserve serious, non-ideological evaluation where inventory scale and national context justify it.
From conceptual improvisation to engineered conservatism
The early history of nuclear refuse disposal reflects both ingenuity and overreach. Proposals to dispose of radioactive material beneath ice sheets, in ocean trenches, or via space launch now read as artefacts from a period when subsurface realism was limited and governance expectations minimal.
More damaging than conceptual dead ends were attempts to reuse existing subsurface infrastructure without sufficient geological or institutional scrutiny. The repurposing of salt mines at Lyons (Kansas) and later at Asse in Germany demonstrated how compromised host geology, inadequate monitoring, and opaque decision-making can permanently erode public trust. These experiences continue to shape contemporary debates.
From these failures emerged the modern consensus: purpose-built deep geological repositories, developed explicitly around a multi-barrier containment philosophy. This approach – now broadly adopted across nuclear nations – combines conservative host-rock selection with engineered barriers and evolutionary safety cases designed to err deliberately on the pessimistic side of uncertainty.
Importantly, progress is no longer theoretical. Finland’s Onkalo repository is entering operations. Sweden has issued construction licences. France, Switzerland, and Canada have advanced siting, engineering design, and predictive modelling programmes rooted in extensive subsurface characterisation. The central question has shifted from whether geological disposal can work to how it should be governed, financed, and sustained over multigenerational timescales.
Where geoscience is strong — and why it Is no longer sufficient
From a geoscientific perspective, the degree of convergence is striking. Clay formations, evaporites, and crystalline basement dominate host-rock selection. Disposal depths typically range from 400 to 900 metres – below active hydrological circulation yet above conditions that introduce unnecessary construction risk. Characterisation emphasises hydrogeology, geochemistry, fracture systems, and long-term geomechanical stability.
For geoscientists trained in E&P, this territory is familiar. Many of the tools – core analysis, geophysics, fracture modelling, and probabilistic uncertainty methods – transfer directly. In some respects, the standards of evidence demanded by nuclear refuse programmes exceed those of conventional energy projects, precisely because assumptions must withstand scrutiny over geological timescales rather than project lifetimes.
What changes fundamentally is the objective function. There is no production upside. No operational optimisation for performance gain. Conservatism is not a preference; it is the design philosophy.
Yet the absence of production upside does not remove optimisation pressures – it displaces them. Disposal programmes are subject to a different, often more pernicious form of optimisation: one driven by cost containment under the implicit assumption of safety neutrality. The expectation to ‘spend less without affecting the risk profile’ generates design churn, defers decisions, and obscures uncertainty rather than reducing it. This dynamic feeds directly into the systematic optimism bias observed in disposal cost and schedule estimates.
Cost realism, institutional risk, and silent erosion of credibility
Independent analysis by the Oxford Global Projects Group (2018), applying reference-class forecasting across nuclear megaprojects including waste facilities, reveals a consistent pattern. Nuclear refuse programmes exhibit cost-overrun behaviour closer to nuclear power plants than to conventional underground mining.
At a 50% confidence level (P50), cost uplifts of approximately 67% are required. At an 80% confidence level (P80), required uplifts approach or exceed 200%. Schedule slippage and cost escalation are tightly coupled. Yet official repository cost estimates are routinely presented at confidence levels closer to P20–P30, while public and political discourse implicitly assumes P80 certainty.
When costs rise – as they almost inevitably do – the geological disposal concept is often blamed, despite no deterioration in the subsurface case. This is not merely a financial problem. Repository safety cases depend on durable institutions: long-term funding, regulatory continuity, monitoring commitments, and knowledge transfer across generations. Persistent cost underestimation therefore becomes a technical risk in its own right, directly undermining institutional credibility.
Here, geological robustness is silently eroded above ground by failures in cost realism and financial stewardship.
Social licence: Realism over reassurance
At the timescales relevant to nuclear refuse disposal, social and political change is inevitable. Geological disposal is therefore not a technocratic outcome delivered by science alone, but a social choice informed by it.
Public resistance has rarely hinged on fracture permeability or diffusion coefficients. It is driven by distrust – shaped by historical opacity, perceived minimisation of uncertainty, and repeated episodes of deferred accountability. The long legacies of Asse and Morsleben continue to frame debates, particularly in Europe.
Transparency about uncertainty does not weaken the safety case; it strengthens legitimacy. But transparency alone is insufficient. Disclosure must be accompanied by explanation, contextualisation, and institutional willingness to educate rather than reassure. Without this, transparency risks being dismissed, misunderstood, or buried – reinforcing rather than repairing distrust.
Right-sizing solutions: Where deep boreholes may fit
This raises an uncomfortable but necessary question: does every nuclear nation require a large, mined geological repository (Figure 1)?
For countries with extensive nuclear fleets and significant legacy inventories, the answer remains yes. But for nations with smaller inventories – or new entrants deploying next generation small modular reactors – the economic, institutional, and social burden of a full-scale repository may be disproportionate.
In such contexts, deep borehole disposal merits serious evaluation. Concepts such as those developed by Deep Isolation propose emplacement several kilometres into crystalline basement, relying on depth, geochemical conditions, and engineered containment.
From a geoscience perspective, these environments are not simple. Fractured crystalline basement is neither impermeable nor hydraulically inert, and such concepts place a greater burden on canister integrity and societal acceptance of effective irretrievability. Deep boreholes are not a panacea, nor a substitute for mined repositories, but they may offer modularity, scalability, and reduced megaproject exposure where national circumstances justify them.
A broader role for geoscience
What ultimately unites mined repositories and deep boreholes is not geology alone, but intellectual discipline. Both demand honest articulation of what is known, what remains uncertain, and how uncertainty is bounded.
Geoscientists are trained for this. They operate professionally within heterogeneity, incomplete data, and probabilistic outcomes. Those skills are now as critical to institutional credibility and governance as they are to site characterisation.
Conclusion
Nuclear energy’s return is real, but its credibility will not be secured by reactor announcements alone. It will be determined by whether societies believe the refuse problem is being handled conservatively, transparently, and proportionately.
Geological disposal works – not because it is simple, but because it is robust. That robustness must now be reflected in governance structures, funding models, and institutional endurance. Where appropriate, alternative concepts such as deep borehole disposal should be assessed pragmatically rather than dismissed ideologically.
The irony is stark. The science is ready. Governance must now catch up.
If nuclear is to scale sustainably this century, it will do so not by minimising the refuse problem, but by confronting it openly – grounded in geoscience, disciplined by data, and framed with institutional humility.
Epilogue: Who pays for certainty?
A final, unresolved question remains. Cost overruns in nuclear refuse disposal are not a hypothetical risk; they are a near certainty. Unlike power generation projects, disposal offers no upside against which overruns can be absorbed. There is no revenue stream, no strategic optionality, and no sunk-cost logic compelling continued investment.
National waste funds, where they exist, are typically actuarially optimistic by construction and politically constrained by design. When costs exceed expectations, the burden is deferred – implicitly – to future taxpayers. This transfer is rarely acknowledged explicitly, yet it is central to public trust.
If geological disposal is to retain credibility, societies must confront not only how risk is contained underground, but how cost risk is allocated above it. A repository whose funding model cannot tolerate realistic overruns is not conservatively designed – regardless of its geology.
References
- Brady, P.V. et al. [2009]. Deep Borehole Disposal of High‑Level Radioactive Waste. Sandia National Laboratories, SAND2009‑4401.
- Flyvbjerg, B. [2014]. What You Should Know About Megaprojects and Why: An Overview. Project Management Journal, 45(2).
- International Atomic Energy Agency (IAEA). [2011]. Geological Disposal Facilities for Radioactive Waste: Specific Safety Guide SSG‑14.
- OECD Nuclear Energy Agency (NEA). [2020]. The Safety Case for Deep Geological Disposal of Radioactive Waste.
- Oxford Global Projects. [2018]. Survival of the Unfittest: Why Nuclear Energy Is Not the Solution to Climate Change (Cost and Schedule Analytics Dataset). (Cited here for reference‑class forecasting of nuclear and waste megaprojects.)
- Posiva Oy. [2021]. Safety Case for the Disposal of Spent Nuclear Fuel at Olkiluoto. TURVA‑2020 Main Report.
- SKB [2011]. Long‑Term Safety for the Final Repository for Spent Nuclear Fuel at Forsmark. Svensk Kärnbränslehantering AB, TR‑11‑01.
