High magnitude floods in the late Holocene: insights from muds and models

Abstract:

Riverine floods rank among the costliest and most frequently occurring natural disasters in the United States, with large uncertainties in near- and long-term projections of flood hazard. In this talk, I will describe a series of projects that integrate instrumental datasets, paleoclimate and paleoflood records, climate reanalysis, and climate model output to examine the response of river systems to climate variability and change. Over inter-annual and decadal time-scales, we show how flood occurrence on the Mississippi River and its tributaries is highly sensitive to major modes of climate variability centered in the Pacific (ENSO, PDO) and Atlantic Oceans (NAO, AMO), providing a means to improve seasonal flood forecasts. We also examine the response of midcontinental North American rivers to the relatively warm and dry Medieval period (ca. AD 1000–1300) – a partial analogue for projected warming – to show that flood hazard was reduced across the Mississippi River basin at this time. These findings support model projections of reduced Mississippi River discharge in response to greenhouse forcing, but do not incorporate changes in river and floodplain management that are also critical to mitigating flood risk today and in the future.