Robust climate attribution of modern floods needs palaeoflood science

Abstract The basis for attribution assessments of current extreme weather and climatic events such as droughts and floods is that the record of such events is of sufficient length to be able to compare the occurrence and severity of recent events with those in the past. If this assumption holds, then the magnitude and frequency of extreme hydrological events in the current anthropogenically forced climate can be compared with those in the past under an unforced climate. Attempts to attribute recent floods to anthropogenically-forced climate change have been made, but we argue that such assessments have failed to correctly analyse the true frequency and magnitude of past floods, when anthropogenic Greenhouse Gas (GHG) forcing was low. In this paper we use well-dated, multi-millennial and multi-centennial length records of large floods from multiple sites across Western and Southwestern Europe that demonstrate past floods were occasionally of much higher magnitudes than those of the present-day, and that attribution studies are presently unable to claim that human-created greenhouse gas emissions have increased flood magnitude. We show that flood magnitude was significantly higher before the 20th century, despite there being a negligible greenhouse gas contribution from humans, which means that natural variability might be significantly higher than assumed by climate modellers. This has profound implications for flood planning and climate adaptation policy, as many recent floods cannot be viewed as unprecedented, even in the historical record.