Machine Learning-based Assessment of Aquifer Depletion and Groundwater Quality Deterioration Risk in the Elwak Aquifer: A Hydrological Assessment of Urbanization Stress and Implications for Managed Aquifer Recharge

The Elwak Aquifer, a critical transboundary sedimentary system spanning negligible parts of Wajir county, Mandera County in Kenya, and extending into Somalia’s Gedo region, sustains a heavily populated pastoralist region in one of the most arid environments in the Horn of Africa. Despite its critical importance, the hydrogeological evolution of the aquifer under urbanization stress and over-abstraction remains poorly quantified, complicated by severe security constraints that prevent systematic cross-border mapping. This study integrates Principal Component Analysis (PCA) with a machine learning (ML) benchmarking framework to model aquifer depletion and water quality risk. Using an 8-variable hydrogeological dataset spanning 33 observation points, we evaluated the latent structure of aquifer vulnerability and compared ten algorithms: Random Forest, Gradient Boosting, Decision Tree, Lasso, Ridge, Support Vector Regressor (SVR), K-Nearest Neighbors (KNN), Gaussian Naive Bayes, Logistic Regression, and Support Vector Classifier (SVC). PCA revealed four dominant orthogonal processes: PC1 represents a spatial hydrogeological gradient driven by distance and salinity; PC2 captures geohydrological vulnerability linking sinkhole density and discharge; PC3 represents anthropogenic pressure and recharge potential; and PC4 isolates institutional funding modalities. The Random Forest regressor achieved exceptional predictive accuracy for depletion risk (, Accuracy = 88.9%), closely followed by Gradient Boosting (). KNN, Naive Bayes, Logistic Regression, and SVC all achieved a robust classification accuracy of 77.8%. We conclude that Elwak’s depletion risk is a catastrophic clash between high population density (urbanization over natural sinkholes) and limited recharge. ML-based diagnostic dashboards can inform targeted climate adaptation funding, including Managed Aquifer Recharge (MAR) and transboundary water governance.

Authors

Publication Details

Journal
American Journal of Science and Advanced Technology
Published
2026-11-12
Primary Topic
Water Quality and Pollution Assessment
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Machine Learning-based Assessment of Aquifer Depletion and Groundwater Quality Deterioration Risk in the Elwak Aquifer: A Hydrological Assessment of Urbanization Stress and Implications for Managed Aquifer Recharge

Geol Anthony Ndungu Gutu, Prof George O. Krhoda, Geol Agatha Njuguna, Meshack Owira Amimo
American Journal of Science and Advanced Technology
Water Quality and Pollution Assessment
article

Machine Learning-based Assessment of Aquifer Depletion and Groundwater Quality Deterioration Risk in the Elwak Aquifer: A Hydrological Assessment of Urbanization Stress and Implications for Managed Aquifer Recharge

Geol Anthony Ndungu Gutu, Prof George O. Krhoda, Geol Agatha Njuguna, Meshack Owira Amimo
article en

Abstract

The Elwak Aquifer, a critical transboundary sedimentary system spanning negligible parts of Wajir county, Mandera County in Kenya, and extending into Somalia’s Gedo region, sustains a heavily populated pastoralist region in one of the most arid environments in the Horn of Africa. Despite its critical importance, the hydrogeological evolution of the aquifer under urbanization stress and over-abstraction remains poorly quantified, complicated by severe security constraints that prevent systematic cross-border mapping. This study integrates Principal Component Analysis (PCA) with a machine learning (ML) benchmarking framework to model aquifer depletion and water quality risk. Using an 8-variable hydrogeological dataset spanning 33 observation points, we evaluated the latent structure of aquifer vulnerability and compared ten algorithms: Random Forest, Gradient Boosting, Decision Tree, Lasso, Ridge, Support Vector Regressor (SVR), K-Nearest Neighbors (KNN), Gaussian Naive Bayes, Logistic Regression, and Support Vector Classifier (SVC). PCA revealed four dominant orthogonal processes: PC1 represents a spatial hydrogeological gradient driven by distance and salinity; PC2 captures geohydrological vulnerability linking sinkhole density and discharge; PC3 represents anthropogenic pressure and recharge potential; and PC4 isolates institutional funding modalities. The Random Forest regressor achieved exceptional predictive accuracy for depletion risk (, Accuracy = 88.9%), closely followed by Gradient Boosting (). KNN, Naive Bayes, Logistic Regression, and SVC all achieved a robust classification accuracy of 77.8%. We conclude that Elwak’s depletion risk is a catastrophic clash between high population density (urbanization over natural sinkholes) and limited recharge. ML-based diagnostic dashboards can inform targeted climate adaptation funding, including Managed Aquifer Recharge (MAR) and transboundary water governance.

American Journal of Science and Advanced Technology
Water Quality and Pollution Assessment
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon