Yucca Mountain: a million years of certainty

In their book "Useless Arithmetic: Why Environmental Scientists Can’t Predict the Future" Pilkey and Pilkey-Jarvis (2007) several case studies are presented showing that quantitative mathematical models used by policy makers and government administrators to form environmental policies are seriously flawed. The issue taken by the authors is not that modelling is useless, but rather that quantitative modelling should be used with caution in relation to environmental systems.

One of the several interesting examples discussed in the book concerns the Yucca Mountain repository for radioactive waste disposal. The environmental impact of various engineering activities 50 years into the future is calculated even more frequently than cost-benefit ratios are. The mother of all environmental impact predictions is the required assurance of 10,000 years of safety from the Yucca Mountain repository of the nation’s radioactive waste, required by a regulatory standard set by EPA. Billions of dollars have been spent at Yucca Mountain on the unrealistic goal of predicting what the climate and groundwater flow will be thousands of years from now. The American judiciary apparently is even more clueless than the scientists of the Department of Energy who are charged with proving the safety of Yucca Mountain—recently a federal court decreed that the prediction must cover 300,000 to 1 million years! The New York Times quotes an incredulous bartender in Las Vegas as saying, “The earth might not even be here a million years from now.” The disappearance of the earth is perhaps not likely, but certainly over the next several hundred thousand years there will be two or three ice ages, the sea level will fall and rise by hundreds of feet, and Yucca Mountain will experience major changes in climate, perhaps an earthquake or two, maybe even a volcanic eruption. Undying faith in mathematics stilled the voice of scientific caution and skepticism that should have warned Congress and the judiciary that the predictive requirements they established for a repository at Yucca Mountain were impossible to achieve.

There was a huge mismatch between that need for long term certainty and the actual state of knowledge. A very large model called TSPA (for total system performance assessment) is used to guarantee the safe containment of the waste. TSPA is Composed of 286 sub-models. TSPA (like any other model) relies on assumptions – a crucial one being the low permeability of the geological formation and hence the long time needed for the water to percolate from the desert surface to the level of the underground disposal. Major limitations exist to the quantitative modeling approach taken by the US-DOE’s TSPA model. The analysts face radical uncertainty and ignorance of uncontrolled conditions of very long term unknown and indeterminate future.

A crucial parameter in the model was the percolation flux, decribing the rate of transport from rain water into the rocks that cover the underground repositry. The TSPA model assumed a percolation flux of 0.5 mm per year (expert guess). Elevated levels of Chlorine-36 isotope in faults (a global marker originating from the atomic bomb tests on the Bikini Atol in the 1950's) uncovered by tunnel boring reveald that the true percolation flux had been at least 3000 mm per year over the past 50 yr, otherwise Clorine-36 could not have been found that deep inside the rocks. The confidence of the stakeholders in TSPA was not helped when evidence was produced which could lead to an upward revision of 4 orders of magnitude of this parameter.

Pilkey and Pilkey conclude that the reliance on mathematical models has done tangible damage to our society in many ways. Bureaucrats that don't understand their limitations often use modeled predictions. Agencies that depend upon project approvals for their very survival (such as the U.S. Army Corps of Engineers) can and frequently do find ways to adjust models to come up with correct answers that will ensure project funding. Most damaging of all is the unquestioning acceptance of the models by the public because they are assured that the modeled predictions are the state-of-the-art way to go.

All this highlights the key importance of openness about uncertainty and high standards for good scientific practice and quality control in the science policy interface. The new field of Knowlegde Quality Assessment aims to provide the concepts, tools and approaches to achieve this.

Reference
Orrin H. Pilkey, Linda Pilkey-Jarvis (2007) Useless arithmetic: why environmental scientists can't predict the future. Columbia University Press