38 Geoscience and Hydrogeology in the AI Era: From Predictive Models to Real-Time Applications
This session offers a unique platform to showcase advancements in Artificial Intelligence (AI), Machine Learning (ML), and physics-informed methodologies, focusing on topics in hydrogeology, subsurface flow and transport, real-time geophysical inversion, critical minerals, and energy applications.
Sunday, 19 October 2025
1:30 PM - 5:30 PM
HBGCC, 210AB
An Efficient Surrogate-based Multi-Objective Optimisation Framework Using Sequential Sampling Strategy and Evolutionary Algorithm for Sustainable Island Groundwater Management under Recharge Change and Sea-level Rise
Paper number: 38-5
Program: Technical Sessions
Time: 2:50 PM – 3:05 PM
Authors
-
Yipeng ZhangOklahoma State University, Stillwater, Oklahoma State
Abstract
To efficiently and effectively obtain optimal pumping schemes (OPS) in coastal regions, striking a balance between sustainability, economic costs, and seawater intrusion (SWI) control under recharge change and sea-level rise (SLR), this study develops an efficient offline surrogate-based multi-objective optimisation framework (OFSMOOF). In this framework, SEAWAT is employed to simulate aquifer response under pumping to produce training data. Gaussian Process (GP) emulation is applied to construct model surrogates, predicting management objective and constraint values. A sequential sampling strategy within the iterative process is proposed to select training points, where new sampling locations are found by selecting from a large network of candidate points, the one that has the maximum distance from the closest available training point. Evolutionary algorithm, continuous ant colony optimisation, is employed to search OPS across the input space according to the GP model predictions instead of calling SEAWAT simulations. This study focuses on a two-objective groundwater pumping optimisation considering recharge change and SLR, formulated on a simplified three-dimensional island aquifer, using hydrogeological conditions of the San Salvador Island (Bahamas). GP models can provide associated uncertainties in predictions by conducting repeated Monte Carlo simulations using these GP models, so it is possible to ascertain the probability of Pareto optimality for each pumping scheme. OPS given by OFSMOOF are characterised by the Pareto-optimal probabilities and validated by SEAWAT simulations. Results reveal that the proposed OFSMOOF can efficiently produce trustworthy OPS for sustainable island groundwater management under recharge change and SLR.
Geological Society of America Abstracts with Programs. Vol. 57, No. 6
doi: 10.1130/abs/2025AM-413188
© Copyright 2025 The Geological Society of America (GSA), all rights reserved. Permission is hereby granted to the author(s) of this abstract to reproduce and distribute it freely, for noncommercial purposes. Permission is hereby granted to any individual scientist to download a single copy of this electronic file and reproduce up to 20 paper copies for noncommercial purposes advancing science and education, including classroom use, providing all reproductions include the complete content shown here, including the author information. All other forms of reproduction and/or transmittal are prohibited without written permission from GSA Copyright Permissions.