Transposing flood risk from extreme rainfall events: A case study of Hurricane Harvey

Abstract

Hurricane Harvey produced unprecedented flooding that altered flood frequency statistics near Houston, Texas. While Harvey could have made landfall elsewhere along the Gulf coast, traditional flood frequency methodologies only consider the risk of Harvey in the region that it hit. This may be a shortcoming of flood frequency methodologies as the intensity of Harvey was greater due to climate change; therefore, Harvey may be more indicative of future hurricanes than other historical observations. To that end, this study investigates what effect Harvey would have had on flood frequency statistics if it had made landfall elsewhere. To do so, a Monte Carlo simulation was used to shift Harvey's rainfall within alternative landfall locations. This rainfall was then applied to synthetic unit hydrographs to estimate peak flows that were applied in Log Pearson III and Regional Flood Frequency Analyses. Log Pearson III analyses with simulated Harvey streamflows produced median 100-year peak flows that were 17%–66% higher than analyses that only used historical records. A regional flood frequency analysis in the central coastal geomorphologic region of Texas showed that predictive equations, based upon basin area and shape factor, had an average increase of 30.2% in the 100-year peak discharge.