Search

The response of coasts to climate change is an uncertain threat to the infrastructure and inhabitants in these areas. Rock coasts are ubiquitous landscapes found along coastlines worldwide and are often sites of important infrastructure, which are vulnerable to coastal erosion and cliff retreat. Furthermore, the stability of rock coasts is threatened by future climate change, particularly the influence of accelerating sea level rise and increased storminess. Constrains need to be placed on past erosion rates at rock coasts across millennial timescales in order to better forecast how future cliff retreat rates will respond to climate change. The ability to quantify coastal cliff retreat rates across millennial timescales has been transformed with the application of exposure dating across shore platforms using cosmogenic radionuclide (CRN) analysis. Previous studies have successfully quantified long-term cliff retreat rates across the Holocene for a range of rock coast sites; however, previous work has been limited by the use of a simplistic geometric-based model to interpret CRN concentrations. Considerable potential remains in the use of CRN analysis along rock coasts. In this thesis, first, I develop a routine that is the first application of a process-based model to interpret CRN concentrations. Next, I apply this technique to a range of rock coast sites across the UK to quantify long-term, transient cliff retreat rates for the past and also make forecasts for the future to better understand the response of rock coasts to climate change. A multi-objective optimisation routine is developed in order to optimise a process-based model to both measured topographic and CRN concentration data for two sandstone sites in the UK. The results highlight that the multi-objective optimisation routine improves model results by reducing equifinality to constrain uncertainties and produce consistent trends in transient long-term cliff retreat rates. This approach is employed to quantify long-term, transient cliff retreat rates for the past 7,000 years at two sandstone rock coast sites. These long-term cliff retreat rates clearly link to the rate of relative sea level rise at both sites. I also use the process-based model to decode insights into the key erosional processes involved in rock coast evolution. Next, I use the optimised parameters from the process-based model results to make forecasts of future cliff retreat rates to the year 2100 using future sea level rise scenarios. At both sandstone sites, cliff retreat forecasts are likely to accelerate by 3-7 times present-day rates, to maximum retreat rates of 22-30 cm yr-1, using the current trajectory of relative sea level rise. Comparisons to cliff retreat rates quantified for the past reveal that these projected cliff retreat rates are unprecedented in the last 3-5 millennia for both sites. These results highlight that even historically stable rock coasts are highly sensitive to sea level rise and need to be included in future planning. Finally, I apply the newly developed optimisation routine to four new chalk rock coast sites on the south coast of England to quantify long-term cliff retreat rates across the late-Holocene. My results identify a recent acceleration in cliff retreat rates, and such rapid cliff retreat rates last occurred 5300-6800 years ago. Furthermore, the optimised, process-based model detects significantly different dominant factors controlling erosion between the sandstone and chalk rock coast sites. Model results for the chalk sites also suggest that heterogeneous lithology and beach material play an important role in the long-term erosion at these sites, and advocates for the inclusion of these factors into future process-based rock coast evolution models. Overall, the results presented in this thesis highlight the knowledge gained from implementing a process-based model to interpret CRN concentrations and quantify transient cliff retreat rates across millennial timescales. Through the methods that I develop, the expected relationship between long-term cliff retreat rates and the rate of sea level rise is supported with empirical data for the first time. Multi-objective optimisation improves model equifinality, which allows for more accurate and well constrained past and future cliff retreat rate forecasts. Results clearly demonstrate the power and utility of these methods to better understand rock coast response to climate change and make more reliable forecasts of cliff retreat rates at coastlines worldwide.