AUTHOR=Kopsiaftis George , Kaselimi Maria , Protopapadakis Eftychios , Voulodimos Athanasios , Doulamis Anastasios , Doulamis Nikolaos , Mantoglou Aristotelis 

TITLE=Performance comparison of physics-based and machine learning assisted multi-fidelity methods for the management of coastal aquifer systems

JOURNAL=Frontiers in Water

VOLUME=Volume 5 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/water/articles/10.3389/frwa.2023.1195029

DOI=10.3389/frwa.2023.1195029

ISSN=2624-9375

ABSTRACT=In this work we investigate the performance of various lower-fidelity (LF) models of seawater intrusion (SWI) in coastal aquifer management problems. The variable density model is considered as the high-fidelity (HF) model and a pumping optimization framework is applied on a hypothetical coastal aquifer system in order to calculate the optimal pumping rates which are used as a benchmark for the LF approaches. The examined LF models could be classified in two categories: 1) physics-based models, which include several widely used variations of the sharp-interface (SI) approximation and 2) machine learning (ML) assisted models, which aim to improve the efficiency of the SI approach. The Random Forest (RF) method was utilized to create a spatially adaptive correction factor for the original SI model, which improves its accuracy without compromising its efficiency as a LF model. Both the original SI and ML assisted model are then tested in a single-fidelity optimization method. The optimal pumping rated which were calculated using the ML-based SI model sufficiently approximate the solution from the VD model. The ML assisted approximation seems to be a promising surrogate for the HF-VD model and could be utilized in multi-fidelity groundwater management frameworks.