2-9 A Machine Learning Approach to Predicting Groundwater Levels in Coastal Agricultural Landscapes Experiencing Effects of Saltwater Intrusion

Session: Coastal Hydrogeology in an Age of Rising Seas

Presenting Author:

Authors:

Abstract:

Rising sea levels are altering groundwater dynamics in coastal regions, challenging the sustainability of freshwater resources for agriculture and increasing vulnerability to flooding. In this study, we explore the use of machine learning to identify key hydrologic drivers of groundwater level variability and to inform efficient monitoring strategies in data-scarce coastal settings. Using time-series data from a coastal field site experiencing the effects of saltwater intrusion, we trained Gradient Boosting and Support Vector Regression models to predict groundwater levels based on past groundwater levels, precipitation, soil water content, and surface water levels. Predictor influence analysis revealed that recent groundwater and surface water levels consistently dominated model performance, while direct precipitation inputs and soil moisture contributed less. Notably, models using only precipitation and surface water levels as inputs, supplemented with historical groundwater level data for training, maintained high predictive performance. These findings reflect the strong lateral hydraulic connectivity typical of low-lying coastal systems. The ability to predict groundwater levels using only precipitation and surface water level inputs demonstrates the potential for simplified, resource-efficient monitoring strategies in data-scarce agricultural settings.  Such approaches offer scalable, cost-effective tools for adaptive water management in vulnerable coastal agricultural landscapes that are experiencing environmental change.

Geological Society of America Abstracts with Program. Vol. 57, No. 6, 2025

doi: 10.1130/abs/2025AM-7501

© Copyright 2025 The Geological Society of America (GSA), all rights reserved.