Proposed RAPID UK model THC intercomparison project

Human activities may significantly increase the risk of rapid climate change in the coming decades by inducing a possible slowdown of the Atlantic Ocean thermohaline circulation (THC), a fundamental controlling factor on the European climate owing to its crucial role in transporting northward the excess of heat imparted in the equatorial regions. The ultimate objective of the NERC funded RAPID programme is to quantify the probability and magnitude of this potential future rapid climate change, and the uncertainties in these estimates. The main tools for making projections of anthropogenic climate change in the coming century are coupled atmosphere-ocean general circulation models (AOGCMs), as well as Earth Models of Intermediate Complexity (EMICs). Such models, however, exhibit widely divergent responses to a CO2 increase, which may range from no response of the THC to an almost complete disappearance. Such differences in the THC response imply large uncertainties in the character and magnitude of climate change over coming decades, especially in the north Atlantic and Europe. To understand how to quantify and possibly reduce uncertainties in model predictions, one may usefully distinguish between three distinct categories: 1) Sensitivity to poorly known initial conditions, due to the inherently chaotic nature of the climate system; 2) Uncertainties in the forcing scenarios; 3) Sensitivity to the parameter values and functional form of parametrisations of subgridscale processes, as well as structural uncertainties associated with model bias, resulting from models being an imperfect representation of reality. Among these, only parametric and structural uncertainties (3) can be reduced over time by continuous physical refinements, while the best that can be achieved for (1) and (2) is to quantify their importance. RAPID round 2 has funded a project on “understanding uncertainty in simulations of THC-related rapid climate change” (NE/C509366/1) with the aim of addressing (3), by comparing climate models of a range of complexity and design, run under similar scenarios, to identify and quantify the physical reasons for different predictions among them, and thereby to get insights into how to reduce parametric and structural uncertainties on predictions of climate change. Part of the funding is to coordinate a UK model THC intercomparison project. The present document sets out a plan for this and outlines the resources requested. This proposed UK project is complementary and provides added value to the existing CMIP coordinated THC experiments, which have already been carried out by several international groups. This CMIP programme not being specially funded, however, it has been short on effort, only allowing for a cursory analysis of the results. The purpose of the UK project will be to construct its own hierarchy of coupled ocean/atmosphere models, and to conduct a much more thorough analysis of their results with the above objectives. To achieve its objectives, the present project brings together the expertise of a critical mass of UK experts in ocean modelling, ocean theory, and statistics, which will be combined with the existing complementary expertise of the international participants in the CMIP project.