HISTORICAL AND PALEO PERSPECTIVES ON THE ORIGIN, TIMING, AND SPATIAL DISTRIBUTION OF EXTREME FLOODING IN THE NORTHEASTERN UNITED STATES

Large magnitude floods are by their very nature infrequent events. Nonetheless, the potential for significant damage to property and infrastructure as well as the threat to human lives associated with these events underscores the importance of understanding the factors which influence their occurrence and frequency. Historical records are typically inadequate for accurately constraining the recurrence interval or identifying trends in the occurrence of the most extreme events. In contrast, natural geologic archives extending beyond the historical period can provide important constraints on the occurrence of extreme events in the past. This study synthesizes existing reconstructions of past flood occurrence from lacustrine archives and evaluates their significance in the context of a new analysis of the timing and spatial distribution of extreme flood events recorded in historical USGS stream gage records from throughout the region. We have identified flood events on individual rivers with magnitudes that exceed those of 10, 25, 50, and 100 year recurrence intervals. Over the past 100 years there have been over 60 unique hydroclimatic events that have impacted multiple watersheds simultaneously and produced floods with magnitudes on all affected rivers that exceeded the estimated magnitude of 10 year recurrence interval floods. While spring snow melt is the dominant source of peak annual flows, late season flooding associated with tropical and extratropical cyclones forms an increasingly significant source of flooding as larger flood magnitudes are considered. Additionally, the most extreme flood events consistently impact large regions of the northeastern United States with floods occurring simultaneously in multiple watersheds spanning 1000s of square kilometers. Patterns in the historical records highlight the regional significance of extreme flooding and support the interpretation of paleo-flood reconstructions from individual lake sites as being regionally important. Similarities among coastal overwash reconstructions from Massachusetts and our paleoflood reconstruction from an upland site in central Vermont suggest that changes in tropical cyclone frequency are a significant driver of changes in flood frequency over the past 1,000 to 2,000 years.