In a warmer world, the hydrological cycle will change in intensity and in its geographic behaviour. This, in turn, will change patterns of river flood and the risk associated with them. The Last Interglacial (LIG; 125,000 years ago) is the most recent instance of climate warmer than today - especially in the high northern latitudes-, sea level was higher, ice sheets were smaller and monsoons were stronger. We use daily output from multi-century LIG simulations of an ensemble of paleoclimate models, and study how global precipitation patterns and extremes de­viate from the preindustrial climate. We validate these results by comparing them with the first compilation, to our knowledge, of global LIG precipitation patterns. Successively, we use the daily temperature and precipitation from the paleoclimate models to drive two global hydrological models (PCR-GLOBWB and CWATM), and simulate river discharges at 5-30’ resolution. With this, we force a hydrodynamic model, CaMa-Flood, and produce floods maps for different return periods. At the end of this model cascade, we look into what would happen if a climate similar to the LIG were to materialize in the coming decades: we combine the flood maps with maps of exposure through vulnerability relationships, and to calculate the risk that floods may pose to future people and assets.

Coastal risks are increasing due to the warming of the climate, resulting in rising mean sea levels and changes in storminess. Projections of future coastal flooding rely on global climate models based on greenhouse gas scenarios with inherent large uncertainties. The past warm climate of the Last Interglacial (LIG, ~127,000 years ago) is considered a partial analogue of a future warmer world. Therefore, understanding how coastal systems were affected by changes in atmospheric and relative sea levels during the LIG can inform us about possible future changes. In this contribution we will analyze extreme sea levels and coastal flooding during the LIG. The analysis is based on the hydrodynamic Global Tide and Surge Model (GTSM; Muis et al., 2016, doi: 10.1038/ncomms11969). To simulate storm surges during the LIG GTSM will be forced by 6-hourly wind and surface pressure fields from LIG simulations of IPCC-type climate models. Due to non-linear effects, tides and surge levels will be influenced by changes in mean sea level. Therefore, a key input variable is map of regional mean sea levels during LIG. However, there is still considerable uncertainty on sea level high-stands and regional patterns during the LIG. Using output from a Glacial Isostatic Adjustment model (GIA), we will model tides and surges for a set of plausible scenarios of relative sea levels and assess sensitivities.

We investigate hydrology during a past climate slightly warmer than the present: the Last Interglacial (LIG). With daily output of pre-industrial and LIG simulations from eight new climate models we force hydrological model PCR-GLOBWB, and in turn hydrodynamic model CaMa-Flood. Compared to pre-industrial, annual mean LIG runoff, discharge, and 100-year flood volume are considerably larger in the Northern Hemisphere, by 14%, 25% and 82%, respectively. Anomalies are negative in the Southern Hemisphere. In some boreal regions, LIG runoff and discharge are lower despite higher precipitation, due the higher temperatures and evaporation. LIG discharge is much higher for the Niger, Congo, Nile, Ganges, Irrawaddy, Pearl, and lower for the Mississippi, Saint Lawrence, Amazon, Paraná, Orange, Zambesi, Danube, Ob. Discharge is seasonally postponed in tropical rivers affected by monsoon changes. Results agree with published proxies on the sign of discharge anomaly in 15 of 23 sites where comparison is possible.