Assessing future ice-sheet variability for long-term safety of deep geological repositories

Abstract. In many regions considered for deep geological repositories (DGR) for nuclear waste, repeated glaciations could occur within the time frame relevant to their long-term post-closure safety (up to 1 million years; Myr). Ice sheets can affect the long-term safety of a DGR by elevating the hydrostatic pressure at DGR depth, altering groundwater flow and chemistry, influencing the frequency and severity of earthquakes, and causing surface bedrock erosion. Therefore, DGR safety assessments must account for uncertainties in future ice-sheet variability, including the timing, frequency and duration of ice sheets at the DGR site. Using coupled ice sheet-climate models to constrain uncertainties in ice-sheet variability over the next 1 Myr is not feasible due to the long timescales involved and substantial computational requirements. Instead, we propose a simplified methodology to assess future ice-sheet variability at potential DGR sites using reconstructions of past ice sheets and global simulations of future climate change. The simulations are conducted using a conceptual climate model driven by changes in insolation and atmospheric CO2 concentrations resulting from anthropogenic emissions. The model is calibrated with 500 000 years (500 kyr) of climate proxy data inferred from deep-ocean sediments. Applying this methodology to the planned Swedish DGR site intended for disposal of spent nuclear fuel in Forsmark suggests that the onset of the next glaciation at the site will not occur until 100 kyr after present, even in the absence of anthropogenic CO2 emissions. However, anthropogenic emissions have the potential to delay the next glaciation in Forsmark by several hundred thousand additional years. Following the initial glaciation, our results suggest that the frequency and duration of subsequent glaciations in Forsmark resemble those observed over the last 800 kyr. Considering uncertainties in anthropogenic emissions and future climate evolution, a wide range of possible future glacial developments is identified. At the extremes of this uncertainty range – developments with a low likelihood of occurrence but relevant for evaluating the robustness of the DGR – we find that Forsmark could experience ice-sheet coverage for nearly half of the next 1 Myr or remain almost entirely ice-free throughout this period. The proposed methodology is easy to implement and applicable to any potential DGR site with a recorded history of glaciations.