Establishing flood thresholds for sea level rise impact communication

Area of Science:

• Environmental Science
• Coastal Geomorphology
• Climate Change Science

Background:

• Sea level rise (SLR) significantly alters coastal flood patterns, complicating flood risk management, particularly in regions lacking tide gauge data.
• Existing methods for estimating high tide flood (HTF) thresholds and SLR rates are often limited to specific locations, leaving large coastal areas unmonitored.
• Effective monitoring of SLR and associated flood risks is crucial for coastal communities worldwide.

Purpose of the Study:

• To develop and validate a machine learning (ML) system for estimating SLR and HTF thresholds at a fine spatial resolution (10 km) along U.S. coastlines.
• To provide a method for assessing SLR and flood risks in ungauged coastal areas, complementing traditional monitoring techniques.
• To enhance community awareness and support adaptation planning by documenting chronic HTF signals.

Main Methods:

• Implementation of a novel high tide flood (HTF) thresholding system utilizing machine learning (ML) algorithms.
• Training and validation of the ML system using historical data from the National Oceanic and Atmospheric Administration (NOAA) tide gauge network.
• Estimation of SLR and HTF thresholds at a 10 km spatial resolution across U.S. coastlines.

Main Results:

• The ML system demonstrated strong performance, achieving an average Kling-Gupta Efficiency (KGE) of 0.77 in estimating SLR and HTF thresholds.
• The system successfully provided spatially distributed estimates of SLR and HTF thresholds, including for ungauged coastal segments.
• The results highlight the potential for ML to offer detailed insights into localized flood risks.

Conclusions:

• The proposed ML-based system offers a promising approach for monitoring SLR and HTF thresholds at local scales, particularly in data-scarce regions.
• The findings underscore the value of applying ML techniques to improve flood risk management and coastal adaptation strategies.
• This research encourages the broader adoption of ML for generating spatially continuous environmental data crucial for climate change adaptation.