Conceptual deconstruction of the simulated precipitation response to climate change

Abstract

State-of-the-art climate change projections of the CMIP5 simulations suggest a fairly complex pattern of global precipitation changes, with regions of reduced and enhanced precipitation. Conceptual understanding of these projected precipitation changes is difficult if only based on coupled general circulation model (CGCM) simulations, due to the complexity of these models. In this study we describe a simple deconstruction of the ensemble mean CMIP5 projections based on sensitivity simulations with the globally resolved energy balance (GREB) model. In a series of sensitivity experiments we force the GREB model with four different CMIP5 ensemble mean changes in: surface temperature, evaporation and the vertical atmospheric velocities mean and its standard deviation. The resulting response in the precipitation of the GREB model is very close to the CMIP5 ensemble mean response, suggesting that the precipitation changes can be well represented by a linear combination of these four forcings. The results further provide good insights into the drivers of precipitation change. The GREB model suggests that not one forcing alone can be seen as the main driver, but only the combination of all four changes results in the complex response pattern. However, the dominant forcings are the changes in the large-scale circulation, rather than the pure thermodynamic warming effect. Here, it is interesting to note that changes in high-frequency atmospheric variability of vertical air motion (weather), that are partly independent of the changes in the mean circulation, have a control on the pattern of the time-mean global precipitation changes. The approach presented here provides a powerful basis on which the hydrological cycles of CGCM simulations can be analysed.