Uncertainty-Aware Deep Learning Framework for Forecasting Coastal Water Level in Virginia Beach

Hasan, Md Mahmudul; Schram, Malachi; Katragadda, Sridhar; McSpadden, Diana; Udomvisawakul, Alisa N.; Richter, Heather; Liu, Frank

Uncertainty-Aware Deep Learning Framework for Forecasting Coastal Water Level in Virginia Beach (Papers Track)

Md Mahmudul Hasan (Thomas Jefferson National Accelerator Facility); Malachi Schram (Thomas Jefferson National Accelerator Facility); Sridhar Katragadda (City of Virginia Beach); Diana McSpadden (Thomas Jefferson National Accelerator Facility); Alisa N. Udomvisawakul (City of Virginia Beach); Heather Richter (Old Dominion University); Frank Liu (Old Dominion University)

Paper PDF Poster File Cite

Time-series Analysis Cities & Urban Planning Disaster Management and Relief Extreme Weather Oceans & Marine Systems Data Mining Uncertainty Quantification & Robustness

Abstract

Coastal areas like Virginia Beach, USA, are increasingly vulnerable to flooding. To mitigate the impact of flooding, it is crucial for the City of Virginia Beach to have reliable 72-hour-ahead (3 days) forecasts of water levels at key gauge locations. To support this effort, several sensors have been installed throughout the city to monitor water levels and other environmental parameters such as wind speed, precipitation, and atmospheric pressure. Leveraging sensor data from one of these locations, we developed an uncertainty-aware deep learning model to forecast water levels. We employed deep quantile regression (DQR) to quantify variability in the predictions and examined the performance of three different model architectures. In addition to exclusively including historical data, we investigated the improvement wind forecasts provide to the accuracy of 72-hour-ahead water level predictions. The results show a twelvefold improvement in the flood forecast for a real flooding event.