Flash Floods In The Northeast United States: Interactions Between Natural And Human Systems

Abstract

Flash floods are an ongoing threat to life and property in the Northeast US and are generally studied as natural phenomena resulting from extreme weather. This dissertation approaches flash floods from a different perspective, in which I explore not only landscape and atmospheric variables, but also the role of people and the built environment. In the first part, I use a spatial model to predict flash flood reports as a function of landscape variables and human population. I find that flash flood reports are correlated with spatially varying slope, saturated hydraulic conductivity (k sat), and impervious area. The model improves when population is included, indicating a reporting bias toward populated areas. In the second part I assess natural and anthropogenic landscape variables of 180 watersheds affected by flash floods. I create a runoff type classification system based on precipitation rate and the amount (depth) of precipitation resulting from a storm combined with soil properties of k sat and soil moisture deficit in the watershed. This analysis demonstrates that saturation excess runoff, infiltration excess runoff, and poor drainage all contribute to flash floods in the Northeast US. In the third part, I assess one aspect of built infrastructure by calculating the suitability of road culverts to convey runoff from a 5-year design storm. Suitability varies by study watershed and by road ownership with roads owned by New York State being more likely to have suitable culverts than roads owned by county and local governments. I find no significant relationships between culvert suitability and slope, impervious area, or k sat at the watershed scale. At the town level, population is positively correlated with culvert suitability, but median income is not significantly correlated. This work demonstrates that the presence of people and their alterations to the landscape contribute to the incidence of flash floods in the Northeast US.