Past and future changes of temperature and precipitation variability in climate model projections and transient simulations of the Last Deglaciation

Abstract

Reliable projections of future climate change are vital for mitigation and adaptation efforts. Such efforts require not only projections of mean changes but of changes in variability, too, since those directly affect the occurrence of extremes. The evaluation of climate models regarding their ability to simulate expected changes in variability of temperature and precipitation relies on the comparison of observations with simulations of past and present-day climate. As such, studying past periods of warming furthers the understanding of the climate system and its projected changes. However, the response of the climate system to forcings depends on the background state. Thus, understanding how insights from studies of the past transfer to future projections and the limitations of this transfer is vital.Here, we present an analysis of temperature and precipitation variability in transient simulations of the Last Deglaciation and projected future climate. To this end, we analyze how the distributions of temperature and precipitation change as exemplified by the moments of the distribution, i.e. variance, skewness and kurtosis. We identify trends in the projections and compare them to results for the Last Deglaciation and present commonalities and differences between the responses in these climate states. We further present how these changes relate to differences in the background state, forcings, and the timescales on which these forcings act as well as the limitations imposed by these differences. Based on this analysis of the state-dependency of variability and its change with a warming mean state, we present conclusions on how past climates can inform and support studies of future climate change.