Research paper
Conceptual hydrogeological and numerical groundwater flow modelling around the Moab Khutsong deep gold mine, South Africa.

https://doi.org/10.1016/j.gsd.2019.100266Get rights and content

Highlights

  • Fact based and comprehensive conceptual groundwater flow model is a pre-requisite to a successful groundwater flow numerical modeling.

  • Environmental isotope and hydrochemical data are used to improve conceptual groundwater flow model.

  • Numerical models are essential tools to evaluate water flows even in highly disturbed flow regimes.

  • The study underscores the significance of integrating various datasets to improve groundwater flow numerical modeling.

Abstract

Prediction of response to fissure water pumping from a shaft on the surrounding areas is conducted with a numerical model, which was refined using environmental stable isotope (ESI), Tritium and hydrochemical datasets. The numerical model domain covers the local carbonate rock units mainly because it is known to be the main source of groundwater.

Based on the hydrochemical dataset of the fissure water pumped at the shaft, strong signature of Na–Cl type water originating from the seepage of a waste rock dump as well as polluted water from mine offices, loading/unloading and parking lot areas, is evident. The environmental isotope data demonstrated that fissure water intercepted at the shaft has strong footprint of surface water and polluted groundwater from nearby areas. The 30-year numerical model prediction indicates that pumping at a rate of 2500 m3/day or more would cause the radius of influence to expand to areas where the central tailings storage facility is located and could possibly draw highly acidic water. Interception of acid rock drainage could cause sever corrosion of pump infrastructure, enhance dolomite dissolution and local ground subsidence as well as possibly compromise the structural integrity of the central tailings dam.

It is recommended that the current groundwater monitoring network be revised for taking proactive action. The study highlights the relevance of integrating various sets of data to improve understanding of the conceptual flow before any numerical modelling is attempted.

Introduction

Proper management of groundwater in underground mining is one of the most critical aspects of mining economics (Lines, 1985; Rubio and Loca, 1993; Wolkersdorfer, 2008; Woldeyohannes et al., 2015 a and b). Failure to have a thorough understanding of groundwater flow dynamics could make or break the profitability of the overall mining operation due to exceedingly high human, operational and material costs (Rubio and Loca, 1993; Balasubramaniam and Panda, 2004). Wasteful and expensive dewatering schemes can be avoided with proper, economically sound as well as environmentally sustainable design of mine water management on the basis of clear and sound understanding of groundwater storage and flow direction as well as flow rate. In many occasions, the conventional groundwater flow modelling approaches may not be strictly applicable to design effective groundwater management in mining conditions owing to the complexity of groundwater flow system (Woldeyohannes and Webb, 2015).

Groundwater management in underground mines has another layer of difficulty because of the substantial infrastructure development within and in the vicinity of the mine and the considerable change of the groundwater flow regime relative to natural conditions (Hodgson et al., 2001). Conventional hydrological tools designed to characterize groundwater flow even in undisturbed conditions mostly fall short of representing real conditions in a satisfactory manner (Reilly and Habough, 2004; Ji-Chun and Xian-Xi, 2013). Mining activities including construction of tunnels and haulages, massive hauling of rock debris from underground, construction of tailings storage facilities, waste rock dumps, evaporation dams and installation of associated metallurgical plants on surface, causes significant change in groundwater recharge, flow and storage conditions. Mining operations further complicate conceptualization of groundwater and surface water flow regimes with conventional hydrologic tools (Rubio and Loca, 1993). It is almost unconceivable that any viable numerical modelling can be developed without a fact based and reliable conceptual model (Zhau and Li, 2011; Singhal and Goyal, 2011). It has been proven that the uncertainty caused by mining operations exceed all the known uncertainty reduction methods including parameter estimation, calibration, sensitivity analyses as well as invoking additional conceptual model recalibration tools (Stauffer, 2005; Fu and Gomez-Hernandez, 2009: Keating et al., 2010; Vrugt et al., 2014; Mengistu et al., 2015).

The application of environmental stable isotope finger printing method has been widely used in combination with other methods such as geochemical foot printing and tracer testing to evaluate various hydrological processes, which include identification of recharge area, qualitative groundwater pathway and mixing as well as water rock interaction (Abyie et al., 2011; Jaunat et al., 2012; Chesson et al., 2014; Mengistu et al., 2015). Data from environmental stable isotope composition of precipitation waters show close relationship with a number of environmental parameters, including source of moisture, surface air temperature, amount and seasonality of precipitation, and recharge altitude. The relationship between climate and mean annual stable isotope contents of precipitation (Craig, 1961; Clark and Fritz, 1997; Dotsika et al., 2010) provides significant insights into paleoclimatic conditions and the underlying pre-existing evaporation and condensation processes.

Environmental stable isotopes of water (δ18O and δ2H)) are commonly used to develop conceptual groundwater flow path, identify recharge/discharge areas, mixing process, salinization process of groundwater (Pulido-Bosch et al. 1997; Larsen et al., 2001; Huang and Chen, 2012; Schofielda and Jankowski, 2004). Various processes in the hydrologic cycle including evaporation, condensation, recharge, mixing and water-rock interaction cause fractionation of Hydrogen and Oxygen isotopes, thereby modifying the equilibrium of isotopic ratios from baseline condition (Aravena, 1995; Mazor, 1997). Thus, by systematically sampling, measuring and mapping the isotopic ratio of various water sources, it is possible to understand the process that the water was subjected (Dotsika et al., 2010). Successful quantification of evaporation process can also be done using the extent to which stable isotope ratios Deuterium and Oxygen-18 deviate from the defined local or global meteoric water line values on a catchment level provided that long-term, spatially representative and detail stable isotope and meteorological data are available (Gibson et al., 1993). In a scenario where considerable mixing of surface water from mining and metallurgical processes and fissure water from shaft dewatering activities is expected, qualitative extent of mixing of surface water – groundwater can be estimated from measured Environmental Stable Isotope (ESI) data (Mengistu et al., 2015).

Hydrochemical data of major ions and cations can be interpreted using mixing models and water rock interaction processes to classify water types. On the basis of these various classes of water chemistry, possible flow paths through which the water may have circulated could be identified (Woldeyohannes et al., 2015a, 2015b). The hydrochemical data is proven to be indicative of percent mixing thereby showing groundwater migration in a defined direction as well as vertical migration of water, if any.

Improving predictive model result using various data sets has shown increasing acceptance and has been advocated to be one way of limiting uncertainties (Xu et al., 2012; Mengistu et al., 2015). In this study, the use of hydrochemical, environmental stable isotope (ESI) and radioactive isotope data are used to refine the understanding of conceptual groundwater flow. Information on groundwater flow paths, groundwater recharge conditions, groundwater residence times and evaporative processes can be retrieved and integrated with the initial conceptual groundwater model for refinement, making it a novel approach.

Numerical modelling is a powerful tool in providing improved understanding of groundwater flow and information on relevant groundwater management parameters provided that scientifically sound and robust conceptual model is available even in the most stressed site conditions such as in mining areas (Konikow and Bredehoeft, 1974; Robson, 1974; Owen et al., 1996; Gvirtzman et al., 1997; Mills et al., 2002; Leake et al., 2005; Mengistu et al., 2015). The most important aspect of using numerical models as a tool for various management purposes lies on addressing the issue of how well the real site condition is captured as closely as possible (Voss, 2011a, 2011b; Merz, 2012; Woldeyohannes et al. 2015a, 2015b). One school of thought advocates, developing and implementing generalized groundwater guidelines as an overarching document and modify it to fit the specific site conditions (Johnson, 2010; Barnett et al., 2012). Others argue that the approach of using groundwater numerical modelling guidelines doesn't bring much traction because of the fact that the real world is too complex to be fairly represented with guidelines (Reilly and Habough, 2004). In some instances analytical and empirical models calibrated to fit the specific site are used in conjunction with water balance data (Woldeyohannes and Webb, 2015). The most common approach to the majority of the scientific community is that uncertainties associated with anisotropy and heterogeneity of aquifer parameters in real field conditions had to be minimized as much as possible through various model result validations (Swanson and Bahir, 2004; Ji-Chun and Xian-Xi, 2013; Sarkar et al., 2015). However, the use of various data sets to refine a conceptual model to reduce model result uncertainties has not been employed. Therefore, in this study, refining of conceptual model, a pre-requisite for improving numerical modelling through minimizing uncertainties is implemented by integrating hydrochemical and environmental isotope information.

Section snippets

Study area

The study area is situated about 200 km southwest of Johannesburg City, near the boundary of Northwest and the northern part of Free State provinces, South Africa. The study area covers approximately 96 km2 surrounding an active mining Shaft, located 3 Km South of the Vaal River with a waste rock dump very close to the shaft (Fig. 1). The mine started stoping from the shaft in 2003 reaching full production in 2010 reaching in depth to over 4 km below groundwater level (DWA, 2006). The main

Geological and hydrogeological setting

The geology of the area is mainly characterized by the Archaean Malmani Dolomite Group of the Transvaal Supergroup (3.2 Ga), composed of carbonates and arenacious rocks, exposed mainly in the northern half of the project area covering roughly 30% of the overall surface geology and referred as Transvaal dolomite the surface geology map (Fig. 2). There is a subordinate cover of the Karoo fluvio-lacustrine sequences (Paleozoic era), which becomes more important in the south quarter of the study

Methodology

Numerical groundwater modelling was executed by Groundwater Modelling System (GMS 9.2) based on MODFLOW 2005, a broad graphical user environment developed by Aquaveo, LLC in Provo, Utah. GMS is compatible with GIS based graphical pre-processing tools to automate and streamline the modelling process for conceptualizing and performing groundwater flow simulation. The entire GMS system consists of a graphical user interface (the GMS program) and a number of modelling codes (MODFLOW, MT3DMS, etc.),

Environmental stable isotope

Twelve samples were taken from the Vaal River, seepage water from a return water dam, shallow aquifer, deep aquifer and very deep shaft water intercepted around 900 m bgl during summer season (June–July 2012) (Fig. 4). Four boreholes are upstream of the shaft although the majority of boreholes cluster around the shaft whereas three boreholes are situated around the main tailings storage facility (Fig. 4).

The environmental stable isotope data of deuterium ranges from most depleted −49‰ in a

Conclusions

The conceptual groundwater flow model built on the basis of existing borehole logs and detail deep mine map has been verified using hydrochemical and environmental isotope data and Tritium data (δ18O, δ2H, 3H). All these datasets demonstrate that at the current time, shaft fissure water has a strong signature of Na–Cl type water presumably originating from surface mine infrastructure such as the local mine office and loading/unloading as well as parking areas. The shaft fissure water also shows

Acknowledgement

The Authors would like to acknowledge Mr. Robert White of Basic Bed Rock Strata, Mr. Charles Human and Mr. Joel Mallan of AngloGold Ashanti Mine, Mr. Ugo Nzota, Ms. Thato Kgari and Mr. Yazeed van Wyk of the Council for Geoscience as well as staff of the Northwest Regional office of the Department of Mines and Energy.

References (70)

  • R. Aravena

    Isotope hydrology and geochemistry of Northern Chile groundwaters

    Bulletin. Inst. Fr. EstudesAndines

    (1995)
  • Y. Balasubramaniam et al.

    Tackling water management in mining

    AWA Water Journal

    (2004)
  • B. Barnett et al.

    Australian groundwater modelling guidelines

    Waterlines

    (2012)
  • I. Clark et al.

    Environmental Isotopes in Hydrogeology

    (1997)
  • N.A. Clay

    The Geology of Malmani Dolomite Subgroup in the Carlettoville Area

    (1981)
  • H. Craig

    Isotopic variations in meteoric waters

    Science

    (1961)
  • (2006)
  • P.G. Erikson et al.

    The transvaal Supergroup and its precursors

  • ESRI
  • A.G. Fox et al.

    Evaluation of a stream aquifer analysis test for deriving reach-scale streambed conductance

    American Society of Agricultural and Biological Engineers

    (2011)
  • J. Fu et al.

    Uncertainty assessment and data worth in groundwater flow and mass transport modeling using a blocking Markov chain Monte Carlo method

    J. Hydrol.

    (2009)
  • M. Gedeon et al.

    Sensitivity of analyses of a combined groundwater flow and solute transport model using local-grid refinement: a case history

    Math. Geosci.

    (2012)
  • J.J. Gibson et al.

    Estimating evaporation using stable isotopes: quantitative results and sensitivity analysis for two catchments in northern Canada

    Nordic Hydrology

    (1993)
  • H. Gvirtzman et al.

    Hydrogeological modeling of the saline hot springsat the Sea of Galilee

    Israel. Water Resour Res

    (1997)
  • A.W. Harbaugh et al.

    MODFLOW-2000; the U.S. Geological Survey Modular Ground-Water Model – User Guide to Modularization Concepts and the Ground-Water Flow Process

    (2000)
  • C. Harris et al.

    O- AND H-Isotope record of Cape Town rainfall from 1996 to 2008, and its application to recharge studies of table mountain groundwater, South Africa

    S. Afr. J. Geol.

    (2010)
  • X. He et al.

    Analyzing the effects of geological and parameter uncertainty on prediction of groundwater head and travel time

    Hydrol. Earth Syst. Sci.

    (2013)
  • F.D.I. Hodgson et al.

    Prediction Techniques and Preventative Measures Relating to the Post-operational Impact of Underground Mines on the Quality and Quantity of Groundwater Resources. Water Research Commission, Report No. 699/1/01

    (2001)
  • P. Huang et al.

    Recharge sources and hydrogeochemical evolution of groundwater in the coal-mining district of Jiazuo, China

    Hydrogeol. J.

    (2012)
  • W. Ji-Chun et al.

    Review of the uncertainty analysis of groundwater numerical simulation

    Chin. Sci. Bull.

    (2013)
  • J. Johnson

    Framework to Effectively Quantify and Communicate Groundwater Model Uncertainty to Management and Clients

    (2010)
  • U. Kafri et al.

    Hydrogeology of the Malmani dolomite in the klip river and natalspruit basins, South Africa

    Environ. Geol.

    (1989)
  • H.E. Keating et al.

    Optimization and uncertianity assessment of strongly non-linear groundwater models with high parameter dimensionality

    Water Resour. Res.

    (2010)
  • L.F. Konikow et al.

    Modeling flow and chemical quality changes in an irrigated stream aquifer system

    Water Res.

    (1974)
  • S.A. Leake et al.

    Numerical Groundwater Change Model of the C Aquifer and Effects of Groundwater Withdrawals on Stream Depletion in Elected Reaches of Clear Creek, Chevelon Creek, and the Little Colorado River, North-eastern Arizona

    (2005)
  • Cited by (22)

    • Characterizing groundwater flow in a former uranium mine (Bertholène, France): Present status and future considerations

      2022, Journal of Hydrology: Regional Studies
      Citation Excerpt :

      For this purpose, accurate monitoring of the hydrogeochemical variables (water level, discharges, pH, electrical conductivity and environmental tracers) is required to trace and identify potential water contamination (Wolkersdorfer et al., 2020). Data analysis is then used to improve the understanding of the groundwater flows and is essential before any attempt at numerical modeling (Mengistu et al., 2019). 3D-modeling based on the monitoring and analysis of climatic and hydrogeological data also enables the functioning of groundwater flows at the catchment scale, and in particular of the TSF, to be better understood.

    • Integrated modeling of hydrological processes and groundwater recharge based on land use land cover, and climate changes: A systematic review

      2022, Environmental Advances
      Citation Excerpt :

      The results revealed that changing land cover increased groundwater recharge. Mengistu et al. (2019), employed SWAT and HSPF models to analyze hydrological responses to changes in LULC and management strategies and identified that the change in LULC pattern modified the rainfall-runoff relationship and significantly increased runoff. Existing information on LULC's long-term consequences on hydrology has been revealed by the use of hydrological models (Fletcher et al., 2013).

    • Insights into the hydrogeological framework of the NW Himalayan Karewas (India)

      2021, Environmental Challenges
      Citation Excerpt :

      The surface topography of the study area was derived from 90 x 90 m SRTM DEM (Fig. 4). Spatially distributed borehole lithologs are widely used for creating hydrogeological framework models (Du Bui et al., 2011; Adam and Appiah-Adjei, 2019; Mengistu et al., 2019; Ezekwe and Oji, 2019) and these provide a fairly good idea of aquifer location and thickness. Borehole lithologs corresponding to different locations across the study area were provided by DGM, J&K.

    View all citing articles on Scopus
    View full text