Sustainable Groundwater Exploration
AI-powered geophysical methods to locate clean water in data-scarce regions across Africa and beyond.
The Challenge: Finding Water Where Data Is Scarce
In rural Africa, failed water boreholes represent wasted resources and missed opportunities for communities in need. Traditional geophysical surveys are expensive and time-consuming. My research applies AI to AMT and CSAMT data to dramatically improve the probability of drilling success.
The Workflow
ERT & Seismic Investigation — multi-method survey for deep aquifer delineation
Open-Source Tools for a Global Challenge
Research impact requires accessible tools. I developed pyCSAMT and WATex — open-source Python packages that put state-of-the-art geophysical inversion and AI-assisted hydrogeology workflows in the hands of scientists and practitioners worldwide.
Open science is not just about sharing papers — it's about sharing the tools that make the science reproducible and applicable.
Abandoned dry borehole · Côte d'Ivoire — each failure costs thousands of dollars and leaves communities without water
Field Evidence
Survey Operations on the Ground
Côte d'Ivoire · 2019–2022 · From reconnaissance to clean water delivery
Pre-Survey Reconnaissance
Terrain inspection and site assessment before deploying geophysical equipment — critical for adapting survey geometry to local conditions.
Electrode Array Deployment
Mounting and connecting ERT electrodes along a profile line. Precise spacing is essential for accurate resistivity inversion.
Water Quality Inspection
On-site water sampling and quality checks at a newly completed borehole — verifying the water is safe before community handover.
Borehole Drilling Operations
Rotary drilling at a target identified by the AI-assisted geophysical model. Survey-guided siting reduced the likelihood of a dry borehole outcome.
Functional Water Point
Post-drilling inspection of a completed water point. Each successful installation serves hundreds of people who previously had no reliable access to clean water.
MADF — Detecting Toxic Leachate with AI
Beyond locating fresh groundwater, protecting it from contamination is equally critical. This line of work applies a Multifaceted Anomaly Detection Framework (MADF) to electrical resistivity tomography (ERT) data to map and delineate toxic leachate plumes escaping from failed landfill liners.
When Landfills Fail: Toxic Leachate in the Subsurface
Anti-seepage membrane failures in landfills release toxic lixiviate — a mix of organic contaminants, heavy metals, and pathogens — into the surrounding soil and groundwater. Traditional ERT inversion locates anomalies roughly, but cannot reliably delineate excavation boundaries with the precision remediation teams need.
Majority-Vote Funnel: Three Detectors, One Ground Truth
MADF runs three unsupervised anomaly detectors — Isolation Forest (IF), One-Class SVM (OC-SVM), and Local Outlier Factor (LOF) — on resistivity features extracted from ERT sections. A majority-vote funnel fuses their binary outputs into a single confirmed anomaly mask (B_MADF), suppressing false positives from any single detector while preserving true leachate signatures.
MADF majority-vote funnel: IF + OC-SVM + LOF → confirmed binary leachate truth (B_MADF)
From ERT Sections to Excavation Perimeter
Validated on two profiles (G1040 and G2030) at an active landfill site, MADF output achieved a Youden index of J = 0.095–0.052 — a substantial improvement over standard inversion (J = 0.027–0.046). The delineated leachate zone matched the confirmed membrane tear at 15 m depth, enabling engineers to define a precise excavation perimeter.
Traditional ERT inversion vs. MADF AI output on profiles G1040 and G2030
Confirmed leakage zone and recommended excavation perimeter derived from MADF
Impact
Key Research Outcomes
Peer-reviewed papers in Water Resources Research, Geophysical Prospecting, Computers & Geosciences, and more.
Demonstrated success rate across field campaigns in Côte d'Ivoire using AI-assisted site selection.
watex and pyCSAMT released as installable Python packages with full documentation and tutorials.
WATER4ALL for Africa launched to scale AI-assisted groundwater exploration across the continent.
Publications
Related Work
A mixture learning strategy for predicting aquifer permeability coefficient K
Kouadio, K. L.; Liu, J.; Liu, W.; Liu, R.
A novel approach for water reservoir mapping using controlled-source audio-frequency magnetotelluric in Xingning area, Hunan Province, China
Kouadio, K. L.; Liu, R.; Malory, A. O.; Liu, W.; Liu, C.
Ensemble Learning Paradigms for Flow-Rate Prediction Boosting
Kouadio, K. L.; Liu, J.; Kouamelan, S. K.; Liu, R.
watex: machine learning research in water exploration
Kouadio, K. L.; Liu, J.; Liu, R.
Groundwater Flow-Rate Prediction from Geo-Electrical Features using Support Vector Machines
Kouadio, K. L.; Loukou, N. K.; Coulibaly, D.; Mi, B.; Kouamelan, S. K.; Gnoleba, S. P. D.; Zhang, H.; Xia, J.
pyCSAMT: An alternative Python toolbox for groundwater exploration using controlled-source audio-frequency magnetotelluric
Kouadio, K. L.; Liu, R.; Mi, B.; Liu, C.