Session: 06-02-02 Coastal Engineering - II

Submission Number: 175158

Multi-Fidelity Surrogate Models for Wave Attenuation From Mangrove Forests 

Efficient prediction of wave attenuation by mangrove forests is essential for enhancing coastal resilience and can be further advanced through digital twin applications. Mangrove root systems reduce wave energy through drag forces, turbulence generation, and energy dissipation within their dense and irregular structures. These processes protect shorelines, limit erosion, and reduce the impact of storm surges on nearby communities. However, quantifying the hydrodynamic behavior of mangroves remains a major challenge because it involves complex interactions between unsteady wave motion and porous vegetation flow fields.

High-fidelity computational models are capable of resolving these nonlinear flow interactions, but they require significant computational time and resources. This limitation restricts their use in uncertainty quantification, design optimization, and real-time coastal management. To address this issue, we developed a multi-fidelity surrogate modeling framework that integrates high-fidelity numerical simulations, intermediate-fidelity potential flow models, and laboratory experiments. The framework combines the theoretical basis of Gaussian process regression with the scalability of deep learning methods to emulate nonlinear hydrodynamics while quantifying uncertainty. Experimental data from artificial mangrove models are used to calibrate and validate the surrogate model across different fidelity levels.

The results show that the surrogate model accurately predicts wave energy dissipation, drag forces, and turbulence intensity with much lower computational cost than full numerical simulations. This approach enables rapid scenario testing, supports probabilistic reliability assessment, and provides a flexible platform for digital twins of natural coastal systems. The outcomes of this study can guide the design and implementation of effective nature-based coastal defenses and contribute to sustainable shoreline management in hurricane-vulnerable regions.

Presenting Author: Taemin Heo Hongik University

Presenting Author Biography: Dr. Taemin Heo is an Assistant Professor in the Department of Civil and Environmental Engineering at Hongik University, Seoul. He received his Ph.D. in Civil, Architectural, and Environmental Engineering from the University of Texas at Austin, and previously held postdoctoral positions at the Computational Visualization Center (UT Austin) and the MIT Energy Initiative. His research focuses on climate-resilient infrastructure, offshore renewable systems, and data-driven structural reliability. He integrates machine learning and uncertainty quantification to improve the design and operation of sustainable and low-carbon energy structures.

Authors:

Taemin Heo Hongik University
Wen-Huai Tsao McNeese State University
Ding Peng Liu Independent Researcher
Christopher Kees Louisiana State University
Lance Manuel The University of Texas at Austin