Advancing wave overtopping prediction at eco-engineered Seawalls: Integrating laboratory experiments and machine learning
M. A. Habib, Soroush Abolfathi, John O’Sullivan, Paul R. Brooks, M. Salauddin
Abstract
Eco-engineering interventions, such as vertipools, rockpools, and textured tiles are increasingly integrated into traditional seawalls to enhance biodiversity. However, their influence on wave overtopping remains underexplored. This study, for the first time, investigates the hydraulic performance of textured tiles designed to mimic natural topography in reducing mean wave overtopping discharge. A series of controlled laboratory-scale tests were conducted in a wave flume, where eco-engineered tiles were retrofitted onto a vertical seawall. Mean overtopping discharge and the spatial distribution of overtopping water were measured under both swell and storm representative sea conditions. The results revealed that the presence of eco-engineered tiles reduced mean wave overtopping discharge, spatial distribution of overtopping hazards zone, and proportion of overtopping waves by up to 75 %, 67 % and 55 %, respectively, compared to a conventional vertical seawall. Experimental datasets were utilised to develop machine learning-based predictive models for wave overtopping assessment at eco-engineered seawall. Gradient Boosted Decision Trees (GBDT) outperformed other algorithms with r 2 , R , RMSE and RAE values of 0.92, 0.96, 0.000226 and 0.21, respectively. Furthermore, a predictive equation was derived using Genetic Programming method to facilitate simplified wave overtopping estimation for eco-engineered Seawalls. These findings provide valuable insights for design and operation of ecologically enhanced critical coastal protection infrastructures to combat extreme climatic events and coastal flooding. • Experimental Investigation of wave overtopping at eco-engineered vertical seawalls. • Eco-engineered seawalls achieved a maximum of 75 % reduction in wave overtopping. • The travelling distance of waves behind the eco-engineered wall was reduced up to 67 %. • Predictive performance of five ML Algorithms are assessed for eco-engineered seawalls. • Novel predictive equation is proposed for overtopping at eco-engineered seawalls.