The analysis, published in Frontiers in Nuclear Engineering by Claudio Pescatore of Linnaeus University in Sweden, does not dispute the standard regulatory distinction between normal evolution and human intrusion. Normal evolution is the expected long-term behaviour of the disposal system, modelled in exhaustive detail: canister degradation, waste-form dissolution, groundwater transport, sorption, decay, dilution and biosphere transfer. Human intrusion, by contrast, is treated as a stylised future human action, because no assessment can credibly forecast the motives, technologies, institutions or collective memory of societies thousands of years from now. International guidance from the IAEA and ICRP deliberately discourages speculative forecasting and instead permits stylised intrusion cases based on present-day knowledge. That caution, Pescatore argues, is fully justified — but it can obscure a protection-relevant comparison that safety cases currently do not make explicitly.

The core of the proposal is a single quantity: P(D ≥ D), the probability that a given pathway produces a dose equal to or greater than a specified meaningful-dose threshold D at least once during the assessment period. The threshold could be a regulatory dose constraint such as 0.3 millisieverts per year, an ICRP reference level, or a lower diagnostic benchmark chosen to mark protection relevance. By expressing both normal evolution and human intrusion in this common probability language, the framework allows the two pathways to be compared on a protection basis rather than merely catalogued as separate scenario categories. The central claim is architectural: compliance under normal evolution, the cumulative probability of boundary bypass, the conditional consequence after physical intersection, and the likelihood of meaningful dose are four distinct quantities that should never be collapsed into a single plausibility judgement.

The distinction matters because the two pathways behave very differently as the dose threshold changes. Normal evolution produces receptor doses only after release, transport and dilution have acted, so the resulting doses are comparatively low; its conditional exceedance probability falls steeply as D* rises and reaches zero once the threshold exceeds the maximum dose the attenuated pathway can deliver. Human intrusion behaves in the opposite way. Once waste or contaminated material is physically intersected and brought toward the surface, it creates a concentrated, largely unattenuated source, so the conditional exceedance probability may remain high across a broad range of thresholds. Raise the threshold high enough and only intrusion remains capable of exceeding it; lower it and normal evolution comes back into contention. The choice of threshold is therefore not a neutral calibration performed before the comparison — it enters the comparison itself and can determine which pathway dominates.

The radionuclide in question sharpens the asymmetry further. For strongly retained actinides such as plutonium-239, normal evolution requires an entire series of retention steps — canister failure, fuel dissolution, mobilisation, transport through buffer and geosphere, sorption, biosphere uptake — that the repository is designed precisely to make ineffective. Intrusion does not weaken that sequence; it removes it, delivering plutonium-bearing fuel fragments, cuttings or contaminated slurry directly into an accessible exposure configuration. For mobile long-lived nuclides such as iodine-129 or chlorine-36, normal evolution remains a credible release route, but intrusion still bypasses the delay, dilution and dispersion that separate a failed canister from the biosphere. As Pescatore puts it compactly: for plutonium, intrusion matters because it bypasses retention; for iodine-129, it matters because it bypasses delay and dilution.

The Finnish TURVA-2012 construction-licence review illustrates how this plays out in practice. The Radiation and Nuclear Safety Authority STUK reported that inadvertent human intrusion was the only disturbance scenario producing a dose within the 10,000-year dose window of the assessment. That finding does not make intrusion predictable, but it shows that a pathway treated as stylised and analytically separate may nonetheless be the more likely route by which repository material reconnects with dose-relevant exposure during the very period in which human dose is explicitly assessed. Posiva’s later FHA-OLA analysis reinforces the point, evaluating both immediate driller exposure and a construction-worker exposure from contaminated landfill material after canister-intersecting drilling, with very high conditional doses for spent-fuel material brought to the surface.

Time adds a further dimension that conventional assessments tend to understate. Boundary-bypass opportunity does not attenuate with time the way release and transport do; it accumulates. Using a Poisson representation with an annual hit rate B, the expected number of canister intersections over a period T is the product of rate and time, and the probability of at least one intersection grows accordingly — it can never decrease as the period lengthens. Sweden’s SR-97 assessment, one of the earliest to quantify this, combined a repository-area probability, a conditional probability of canister penetration and an assumed deep-borehole rate, and reported a cumulative probability of 1 for drilling through a canister at some time over 100,000 years under an illustrative rate of 200 deep boreholes per year in relevant crystalline basement. The US Waste Isolation Pilot Plant provided a regulatory precedent for combining a stipulated drilling rate with cumulative exceedance displays, and expert studies cited in the paper estimated roughly 1 to 10 percent cumulative probability of human intrusion over 10,000 years.

Pescatore is careful about the limits of such rate calculations. They translate a specified annual hit rate into a cumulative probability without pretending to forecast future society, and four cautions apply. Physically distinct intrusion modes — shallow excavation, deep borehole, resource extraction — should each carry their own mode-specific rates, intersection probabilities and conditional consequences, because aggregating them discards exactly the information needed for protection-relevant comparison. The rate itself is not stationary: memory loss, reinterpretation and institutional discontinuity are time-dependent, and the stable social configuration that would make a constant rate defensible is the thing least likely to persist. Land-use and technological change act on both the rate and the consequence at once. And a rate calculation says nothing about the conditional dose after intersection or the likelihood that it exceeds a threshold. Notably, even the Nordic KAN 1.3 work on transferring information to future generations — a study about memory preservation — has itself largely fallen into oblivion within a few decades.

The practical upshot is that a safety case can be complete as a compliance demonstration while remaining incomplete as a protection claim. A calculated dose below 0.3 millisieverts per year under normal evolution over a million years is an important result, but it is a statement about one pathway, not evidence that normal evolution is the dominant route to meaningful exposure. The paper proposes four checks that regulators and developers can apply without any speculative forecasting: report the cumulative boundary-bypass probability separately from the normal-evolution dose calculation; compare conditional exceedance probabilities by radionuclide class, since the two pathways are dominated by different nuclides at different epochs; identify the limiting exposure pathway explicitly as a finding rather than letting the most fully modelled pathway dominate by default; and name the intrusion-created source configuration — breached canister, drill core, cuttings, slurry, contaminated soil or landfill-like mixture — because that configuration governs which radionuclides matter, on what timescale, and under which exposure-situation interpretation.

The framework extends beyond spent fuel. Posiva’s assessment of both its roughly 430-metre-deep spent-fuel repository and its roughly 180-metre-deep low- and intermediate-level waste repository shows the same analytical structure applies wherever isolation depends partly on depth and intrusion can create an accessible exposure configuration. Intrusion can even make radionuclides dose-relevant that barely feature in groundwater assessments — silver-108m appears among contributors to dose maxima in Posiva’s drilling and landfill cases despite not being a principal nuclide in normal-evolution calculations. The deeper message is a challenge to the field: compliance, robustness and protection may not always coincide, and confronting that possibility requires expressing both pathways in a common probability language under clearly stated assumptions. Only then can safety cases answer not just whether the repository complies, but which route to meaningful dose deserves the most protection attention.

Subject of Research: Comparing human intrusion and normal evolution as competing exposure pathways in geological disposal safety cases

Article Title: Meaningful dose in geological disposal safety cases: human intrusion and normal evolution as competing exposure pathways

Article References: Pescatore, C. (2026). Meaningful dose in geological disposal safety cases: human intrusion and normal evolution as competing exposure pathways. Frontiers in Nuclear Engineering, 5, Article 1921673. https://doi.org/10.3389/fnuen.2026.1921673

Image Credits: AI Generated

DOI: 10.3389/fnuen.2026.1921673

Keywords: geological disposal, radioactive waste, human intrusion, normal evolution, meaningful dose, safety case, pathway efficiency, Poisson process, plutonium-239, iodine-129, TURVA-2012, nuclear engineering