Prediction of the long-term climate response in a coupled climate model using response theory

This contribution provides new insights on the application of the Ruelle Response Theory (RRT) for the analysis of long-term evolution of the climate. (Lucarini et al., 2017 J. Stat. Phys.; Ragone et al., 2016 Clim. Dyn.) successfully applied the theory to an intermediate complexity atmospheric model in order to predict the global and local response of the climate system to increases in the CO2 concentration. We now investigate the oceanic scales of the response by looking at the long-term evolution of a low-resolution coupled model. We consider two ensembles, each composed of 20 members. Each member is a 2000 year simulation performed with the Max Planck Institute Earth System Model (MPI-ESM v1.1) in its coarse resolution version. The first ensemble, namely “predictor” consists of 20 runs with an initial abrupt CO2 concentrations doubling (2xCO2). The second ensemble, namely the “predictand”, consists of runs where CO2 is increased by 1% per year until 2xCO2 threshold is reached (1pctCO2), and then the concentration is kept constant. Results show that the evolution of a number of observables relevant to the climate projections are well captured by the convolution of the observable-dependent Green's function, obtained with the 2xCO2 ensemble, and the time modulation of the predictand's forcing. Remarkably, the prediction is not only able to reproduce the slow relaxation of the system to the equilibrium, but also the transient response of the system to the forcing.