Projecting long-term past and future climate change within the context of post-closure performance assessments for disposal of radioactive waste

Abstract

Multi-millennial transient simulations of climate changes have various important applications, such as for investigating key past geologic events for which high resolution proxy data are available, or for projecting the long-term impacts of future climate evolution on the performance of geological repositories for radioactive waste disposal. However, due to the high computational requirements of current General Circulation Models (GCMs), long-term simulations can generally only be performed with less complex models and/or at lower spatial resolution. In this thesis, two tools are developed which can be used to rapidly project long-term (>1000 years) changes in climate and one of its primary forcings – atmospheric CO2 concentration. A multi-exponential analysis is performed on a series of atmospheric CO2 decay curves predicted by an Earth system model, producing an empirical response function that can rapidly project the long-term response of atmospheric CO2 to any plausible cumulative CO2 release. With increasing total emissions, the buffering and CO2 uptake by the ocean on anthropogenic timescales is found to progressively saturate, whilst CO2 uptake due to carbonate weathering processes progressively increases in importance to compensate. In contrast, the uptake timescale and relative importance of the silicate feedback is almost unchanged. Long-term “continuous” projections of climate evolution are then produced using a statistical emulator calibrated using GCM simulations with varying orbital configurations and atmospheric CO2 concentrations. The emulator is applied to model climate in the late Pliocene and over the next 1 million years. For both periods, emulated climate varies on an approximately precessional timescale, with evidence of increased obliquity response at times. Uncertainties include the emulator’s inability to represent true transient changes in the climate system, and its limited range of fixed ice sheet configurations. The climate data produced is suitable for use in modelling performed for a post-closure safety assessment for a radioactive waste repository.

Date of Award	16 Jan 2018
Original language	English
Awarding Institution	University of Bristol
Supervisor	Dan Lunt (Supervisor), Andy J Ridgwell (Supervisor) & Mike Thorne (Supervisor)