The land/ocean temperature contrast in natural variability

In global warming scenarios, global land surface temperatures (T_land) warm with greater amplitude than sea surface temperatures (SSTs), leading to a land/ocean warming temperature contrast. This land/ocean contrast is not only due to the different heat capacities of the land and ocean as it exists for transient and equilibrium scenarios. Similarly, the interannual variability of T_land is larger than the covariant interannual SST variability, leading to a land/ocean temperature contrast in natural variability. This work investigates the land/ocean temperature contrast in natural variability based on observations, coupled global model simulations, atmospheric global model simulations with different SST forcings, and using idealised models with simplified geometry or processes. The land/ocean temperature contrast in interannual variability is found to exist in observations and models to a varying extent in global, tropical and extra-tropical bands. There is agreement between models and observations in the tropics but not the extra-tropics. Causality in the land-ocean relationship is explored with modelling experiments forced with prescribed SSTs, where an amplification of the imposed SST variability is seen over land. The amplification of T_land to tropical SST anomalies is due to the enhanced upper level atmospheric warming that corresponds with tropical moist convection over oceans leading to upper level temperature variations that are larger in amplitude than the source SST anomalies. This mechanism is similar to that proposed for explaining the equilibrium global warming land/ocean warming contrast. The tropospheric structure is studied with single and two column models. It is found that realistic values of the land/ocean contrast can be simulated with the simplified models, and the atmospheric structure is similar to the mean tropical response in global climate models. However on regional scales the simple models fail to represent the magnitude or patterns of the global response. The regional response is then explored with the Globally Resolved Energy Balance (GREB) model. This model represents the circulation, clouds and soil moisture as boundary conditions. Perturbation experiments allow for attribution of the regional response of a complex coupled climate model to these boundary conditions.