Stochastic Parameterizations and the Climate Response to External Forcing: An Experiment With EC‐Earth

Global climate models are a powerful tool to study the current climate and its future projections. Applying those models in high resolution is computationally demanding. A cheaper alternative to grid refinement is the use of stochastic physics schemes in which the unresolved subgrid processes are parameterized by introducing a term of randomness into the physical parameterizations. In this work, we study the impacts of including stochastic physics schemes in the atmospheric component of the ECEarth climate model by analyzing longterm coupled simulations from 1850 to 2160. It was found that the global temperature increase is lower when the stochastic physics is included, which implies reduced climate sensitivity. However, with the melting of Arctic sea ice, the opposite occurs. An increase of highclouds due to the skewed interaction between stochastic perturbations and the process of condensation accelerates the increase in temperature in the simulations with stochastic physics.