A Framework for Integrated Surface–Subsurface Geochemical Hazard Modeling

Thesis Abstract

This study addresses the escalating geochemical hazards arising from coupled surface and subsurface processes in mineralized and industrially impacted environments, where decoupled monitoring has hindered accurate risk assessment and mitigation planning. The aim is to develop an integrated framework for surface–subsurface geochemical hazard modeling that synthesizes hydrological dynamics, geochemical reactivity, and transport processes to predict contaminant mobilization, fate, and exposure risk under varied climate and land-use scenarios. Specific objectives include (i) to formulate a conceptual model that links surface runoff, infiltration, and subsurface flow with mineral dissolution/precipitation, adsorption–desorption, and redox transformations; (ii) to operationalize a coupled modeling framework incorporating physically based hydrology (Richards’ equation and Darcy-scale flow), reactive transport (HYDRUS-1D/2D coupled with PHREEQC), and geochemical databases (PHREEQC databases and MINTEQ) to simulate spatiotemporal hazard trajectories; (iii) to calibrate and validate the framework using multi-source datasets from a representative mining-impacted watershed containing 312 monitoring wells, 24 surface water gauges, and 60 soil profiles; (iv) to assess sensitivity and uncertainty through global variance-based methods (Sobol’ indices) and Bayesian calibration; (v) to evaluate hazard indicators (contaminant concentration exceedance probabilities, plume migration times, and redox-front advancement) under three climate scenarios and land-use change projections; and (vi) to develop decision-support tools and risk metrics for stakeholders. The methodology adopts a pragmatic mixed-methods research design that integrates quantitative simulations with qualitative stakeholder inputs to refine model assumptions. The population includes hydrologic and geochemical processes at the watershed scale, while the sample comprises spatially distributed observations from 68 geochemical boreholes, 11 rainfall-runoff stations, and 9 meteorological stations over a 10-year observation window. Data collection instruments encompass high-resolution spectroscopic analyses (ICP-MS forTrace elements, ICP-OES for major ions), mineralogical characterization (XRD), field geophysics for subsurface property estimation (Electrical Resistivity Tomography and Ground-Penetrating Radar), as well as continuous loggers for water level, pH, redox potential, and dissolved oxygen. Model implementation follows a three-tier coupling strategy (1) a surface hydrological component based on a distributed rainfall–runoff model; (2) a subsurface transport and reactive-chemistry component leveraging HYDRUS-1D/2D and PHREEQC for speciation and mineral equilibria; (3) a data assimilation layer applying Kalman-based updating and Bayesian inference to reconcile observations with predictions. Model validation employs split-sample testing, with 70% of data for calibration and 30% for validation, and performance metrics including Nash–Sutcliffe efficiency, Kling-Gupta efficiency, RMSE, and receiver operating characteristic (ROC) curves for hazard exceedance. Key expected findings include improved predictive skill for contaminant plumes and redox fronts through integrated coupling, revealing nonlinearity and hysteresis in surface–subsurface exchange under transient rainfall events. The study anticipates identifying critical thresholds in hydraulic conductivity and mineral phases that govern abrupt shifts in hazard trajectories, and quantifying the relative influence of climate variability versus land-use change on hazard retardation or acceleration. The contribution to knowledge lies in advancing a transferable, theory-informed modeling framework that unifies surface hydrology, geochemical reactivity, and transport physics within a single coherent platform, enabling scenario-based risk forecasting for complex geochemical hazards. The work builds on and extends theoretical underpinnings from reaction-transport theory (Wheeler and Guinot), surface–subsurface coupling (Bear and McCuen adaptations), and uncertainty quantification (Sobol’ and Bayesian updating), explicitly integrating them into a practical workflow with transparent parameter estimation and data provenance. The main conclusion expected is that an integrated surface–subsurface geochemical hazard framework significantly reduces predictive uncertainty relative to decoupled approaches and enhances decision-support capability for hazard mitigation, remediation prioritization, and land-management planning. Recommendations include embedding the framework in regional environmental monitoring programs, expanding the database of mineral-water–organic matter interactions to improve reactive transport accuracy, adopting adaptive monitoring strategies guided by uncertainty-trajectory analyses, and transferring the methodology to other settings with analogous geochemical risk profiles.