EAGER: Satellite-Based Evaluation of US Flood-Mitigation Infrastructure Performance

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

The objective of this EArly-concept Grants for Exploratory Research (EAGER) project is to support research focused on developing and piloting a novel method for faster and more accurate assessment of flood mitigation infrastructure performance. While billions of public dollars have been invested in levees, detention basins, channel realignments, and other flood-mitigation infrastructures, nationwide assessment of their performance is lacking. Few researchers have attempted to merge infrastructure records with continent-scale satellite observations and replace anecdotal, piecemeal evaluation with satellite-verified report cards. The research looks to advance knowledge by generating observed flood data to measure if and where such infrastructures reduce flooding in the US. The research trains undergraduates and postdocs to produce data science tools and prepares them for artificial intelligence, remote sensing, and other technical career opportunities. Current floodplain maps underrepresent flood risk, which most research attributes to failure to produce frequent updates or incorporate different flood drivers. Yet, the “undermapping” phenomena could also be due to reductions in regulatory floodplains though the Letters of Map Revision (LOMR) policies. While traditional assessment relies on expensive physics-based flood models and site investigations, this research project seeks to develop and pilot an observation-based method that enables faster and more accurate assessment of flood risk reduction. Two key questions look to be addressed: 1) where flood mitigation infrastructures are effective at reducing flooding in the US and 2) where new opportunities exist for optimal investment. The research project will analyze locations where flood maps have been modified due to LOMR policies and will look to build a tempo-spatial database of flood events from multiple satellite observations from 2001 to 2025 using machine learning. The database will then augmented with layout of flood mitigation infrastructures. All code, time series, satellite flood layers, and flood exposure change analytics will be released as open-source, enabling engineers, planners, insurers, and researchers to prioritize investments, calibrate models, and design innovative solutions. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.