Predicting flood damage using the flood peak ratio and Giovanni Flooded Fraction

Area of Science:

• Environmental Science
• Data Science
• Civil Engineering

Background:

• Accurate flood hazard intensity understanding is crucial for predicting infrastructure reliability and losses.
• Current methods like hydrodynamic modeling and static flood maps have significant limitations in speed and variability.
• There is a need for rapid, spatially-resolved flood damage prediction methods.

Purpose of the Study:

• To develop an integrated, data-driven approach for rapid flood damage prediction.
• To utilize emerging flood intensity heuristics: Flood Peak Ratio (FPR) and NASA's Giovanni Flooded Fraction (GFF).
• To create predictive models at nationwide and statewide levels for the contiguous United States.

Main Methods:

• Utilized National Flood Insurance Program (NFIP) flood claim data as a proxy for flood damage.
• Incorporated flood exposure variables: regional slope and population.
• Applied statistical learning methods to generate predictive models.
• Conducted variable importance analysis to assess the significance of FPR and GFF.

Main Results:

• FPR and GFF data were found to be significant predictors of flood damage.
• State-level models demonstrated higher performance than the nationwide model.
• The data-driven approach proved effective at regional levels.

Conclusions:

• The integrated data-driven approach using FPR and GFF offers a promising alternative for rapid flood damage prediction.
• The method's reliance on accessible data and computational speed makes it a valuable tool.
• Regional-level modeling enhances the accuracy and effectiveness of flood damage predictions.