Identifying confined aquifers using EOF data-driven approach and MODFLOW simulation
Year: 2025
Type: article
Abstract: • This study offering a fast and efficient approach for aquitard localization. • Deep aquifer uncertainty can be reduced by integrating EOF. • Inconsistent patterns in the same area indicate effects from clay layers distribution. • EOF analysis effectively distinguish confined or unconfined aquifer. Traditionally, groundwater studies have heavily relied on numerical simulation software to model regional groundwater flow. However, the accuracy and effectiveness of these simulation depends on well-defined boundary conditions and reliable hydrogeological parameters, such as hydraulic conductivity. In reality, these conditions are often uncertain and not as deterministic as simulations typically assume, especially in the deeper aquifer. Conversely, data-driven approach could infer hydrogeological conditions from groundwater observations. However, it often lacks a clear connection to their underlying mechanisms. This study first constructed an experimental control case to explore groundwater behavior under different geological condition. The experiments confirm that the lower hydraulic conductivity of confined aquifers restricts groundwater flow, leading to predominantly horizontal flow. Based on these findings, this study integrates a widely used groundwater numerical simulation model, Modular Three-Dimensional Groundwater Flow Model (MODFLOW) with Empirical Orthogonal Function (EOF) which extracts dominant spatial-temporal variation patterns from simulated groundwater levels. The results demonstrate that EOF not only effectively distinguishes geological conditions, particularly between confined and unconfined aquifers, in both ideal case and real case in the Minxiong Industrial Zone, Taiwan, but also reveals how regions with similar hydraulic conductivity display synchronized groundwater level changes, while areas with larger contrasts exhibit delayed responses. These findings indicate that analyzing groundwater fluctuations with EOF can successfully identify subsurface geological conditions, such as the presence of aquitards, providing a rapid and effective approach for aquitard localization.
Source: Results in Engineering
Authors: Huan-Sheng Lin, Jui-Yun Hsieh, Hsin-Wen Pai, Yuan-Chien Lin
Institution: National Central University
Language: English
Subfield: Environmental Engineering
Field: Environmental Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,940
Awards: Geological Uncertainty and Societal Risk: the Perspectives of Engineering, Environment, and Geohazards( I ), Combining multi-spectral satellite imagery and machine learning methods to establish automatic identification system of landslides and factor analysis of typhoon events(2/5), Adding Value and Applications of Formosat and Triton Satellite Gnss Data( I ), Integration of Formosat Satellite Data for Healthy Exposure Construction in Taiwan( I ), Combining Multi-Spectral Satellite Imagery and Machine Learning Methods to Establish Automatic Identification System of Landslides and Factor Analysis of Typhoon Events( I ), Combining Multi-Spectral Satellite Imagery and Machine Learning Methods to Establish Automatic Identification System of Landslides and Factor Analysis of Typhoon Events( Iv ), Combining Multi-Spectral Satellite Imagery and Machine Learning Methods to Establish Automatic Identification System of Landslides and Factor Analysis of Typhoon Events( III ), Air Pollution Data Mining Analysis and Diffusion Simulation in Petrochemical Plants, Study of Disaster Information Integration between Central Government and Local Government - Hsin-Chu County( II ), TaiWan AI-based Management Strategy, TWAI, for Earth and Environmental Disasters(1/2)
OpenAlex ID: W4415219787