Critical data gap to be filled: a lack of local information on sea level rise and flooding thresholds ‎pose a serious challenge to coastal communities along the ungauged stretches of the US shorelines

This study utilizes machine learning to predict sea level rise rates and high tide flooding thresholds at high-resolution along the US coastlines, aiding decision-makers in planning for future chronic impacts of high tide flooding.
Published in Earth & Environment and Statistics
Like

Share this post

Choose a social network to share with, or copy the shortened URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

High tide flooding (HTF), also known as sunny day or nuisance flooding, poses challenges for coastal communities due to elevated sea levels, leading to business disruptions and infrastructure degradation. Despite minimal immediate damage, the cumulative impacts of HTF over time can destabilize structures. Monitoring and assessing these impacts are critical, especially with the discernible increase in HTF events attributed to sea level rise (SLR) in a warming climate.

Our knowledge of SLR and HTF relies on information from installed gauges. To establish HTF thresholds for coastal communities, two main components are used: local tidal records and a flood monitoring system. Tidal records offer insights into baseline water levels and tidal patterns. However, to accurately identify nuisance flooding impacts, a local flood monitoring system is essential. By analyzing data from this system, coastal communities can determine specific water levels beyond which the nuisance flooding impacts are expected. This is a crucial information for setting HTF thresholds.

However, challenges arise with these components. The first challenge relies in the fact that tide gauges are sparse and unevenly distributed. In the United States, if the coastline is divided into 10 km sections, 75% of coastal communities do not have tide gauges nearby (i.e. within a 10 km radius). Hence, decision-makers often rely on assumptions and information from the nearest tide gauge, which may be over a hundred kilometers away, posing accuracy and uncertainty challenges. This approach assumes consistent characteristics between the target ungauged point and the nearest official threshold location, which may not hold true across different coastal areas. Another challenge arises when, despite an installed tide gauge, the region lacks a reliable flood monitoring system. These challenges highlight the decision-makers’ and stakeholders’ need for high-resolution data of the SLR rates and HTF thresholds.

Sweet et al. 1 proposed a regression approach to estimate HTF thresholds (above mean lower low water (MLLW) datum) at ungauged basins by considering tidal range as the sole independent variable. However, proposing the HTF threshold above MLLW could obscure the variability of the actual flooding threshold above mean higher high water (MHHW), especially in regions with larger tidal ranges. MHHW serves as a more suitable reference datum for flood analysis, as it represents the expected high tide level for coastal communities on a regular basis. A more comprehensive to consider multiple physically relevant variables would be insightful.

In this study, we utilized machine learning (ML) algorithms to characterize the non-linear interactions between the components affecting target variables (SLR rates and HTF thresholds).  Moreover, the input features of RF do not share similar characteristics all over our case study (West, Gulf, and Atlantic Coastlines). Hence, we implemented K-means clustering on the input features to achieve similar clusters on which one RF algorithm could be developed and trained.

The final results represented a promising comparison (Kling-Gupta efficiency (KGE) of 0.77) between the official thresholds provided by NOAA and the values predicted by the developed RF algorithms. The high-resolution SLR rates and HTF thresholds are demonstrated in Figure 1a and Figure 1b. Please, check the full article for details on the methodological details and the validation process.

Figure 1. Spatially distributed estimates of HTF thresholds and SLR rates: a) SLR rates; b) HTF thresholds and NOAA official thresholds in larger points, the box plots show the median, 0.25, and 0.75 quantiles, maximum, and minimum of each dataset.

Reference:

  1. Sweet, W. V., Dusek, G., Obeysekera, J. & Marra, J. Patterns and Projections of High Tide Flooding Along the U.S. Coastline Using a Common Impact Threshold. 44 https://tidesandcurrents.noaa.gov/publications/techrpt86_PaP_of_HTFlooding.pdf ‎‎(2018).

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Subscribe to the Topic

Climate Change
Physical Sciences > Earth and Environmental Sciences > Earth Sciences > Climate Sciences > Climate Change
Coastal Sciences
Physical Sciences > Earth and Environmental Sciences > Earth Sciences > Earth System Sciences > Coastal Sciences
Ocean Sciences
Physical Sciences > Earth and Environmental Sciences > Earth Sciences > Ocean Sciences
Machine Learning
Mathematics and Computing > Statistics > Statistics and Computing > Machine Learning

Related Collections

With collections, you can get published faster and increase your visibility.

Materials and devices for separation, sensing, and protection

In this Collection, the editors of Nature Communications and Communications Materials welcome the submission of primary research articles that highlight the development and application of functional materials in the areas of separation, sensing, and protection.

Publishing Model: Open Access

Deadline: Jun 30, 2024

Applied Sciences

This collection highlights research and commentary in applied science. The range of topics is large, spanning all scientific disciplines, with the unifying factor being the goal to turn scientific knowledge into positive benefits for society.

Publishing Model: Open Access

Deadline: Ongoing