Changes in Precipitation Extremes under Two Climate-Change Scenarios

Variations in the spatiotemporal patterns of extreme precipitation are one of the most societally relevant but least-understood features of climate change. This study uses a 50-km-resolution global climate model (GCM) to force the atmosphere with three cases of sea surface temperatures (SST): a control utilizing the 1980-2000 climatological SST, a perturbation case responding to a nearly uniform SST increase of 2 K from the control, and a second perturbation case corresponding to a “realistic” warming as simulated with a fully coupled GCM from the IPCC A1B emissions scenario. Averages and three categories of extremes (comprising the greatest 10%, 5%, and 1% of daily precipitation values) are computed for the globe, for six 30-degree zonal bands, and then separately over the same domains for continental and oceanic areas, with a particular focus on the Northern Hemisphere mid-latitudes during their winter. The basic results -- of general increases in extremes, and of greater precipitation overall in the higher latitudes -- are in line with previous studies. More-involved calculations reveal that the plus-2K case sees a greater increase over the control in the zonally averaged mid-latitude 99th-percentile daily precipitation threshold, despite a weakening of the large-scale temperature gradient relative to both the control and the realistic-warming cases. Indeed, there is a larger extreme-precipitation response in the plus-2K warming scenario by almost all measures. Meanwhile, extreme values and trends are examined over land and over oceans, and the land-ocean differences are found to be statistically significant. Eddy kinetic energy and changes in circulation patterns are also found to be largely insufficient to explain these results. By way of explanation, the influence of changes in local surface temperature is discussed, especially in their correlation with empirically derived fractional increases in extreme precipitation. The contributions of larger-scale zonal factors, such as changes in regional P-E, are also examined.