Dundee Precious Metals - Tsumeb Smelter: 3D Groundwater Flow … · 2019. 7. 2. · non-reactive transport modelling does not consider adsorption, precipitation and retardation processes,

DUNDEE PRECIOUS METALS -

TSUMEB SMELTER: 3D GROUNDWATER FLOW AND CONTAMINANT TRANSPORT

Prepared for:

Dundee Precious Metals Tsumeb (Pty) Limited

SLR Project No.: 733.04040.00010 Revision No. 1: Month/Year: March 2018

Dundee Precious Metals Tsumeb (Pty) Limited Dundee Precious Metals - Tsumeb Smelter: 3D Groundwater Flow and Contaminant Transport File name: Appendix E Addendum -

DPMT_Groundwater_flow_and_transport_model_Report_v1.1

SLR Project No.: 733.04040.00010

Month/Year: March 2018

i

DOCUMENT INFORMATION

Title Dundee Precious Metals - Tsumeb Smelter: 3D Groundwater Flow and Contaminant Transport

Project Manager Arnold Bittner

Project Manager e-mail [email protected]

Author Markus Zingelmann, Winnie Kambinda

Reviewer Arnold Bittner

Keywords Keywords

Status Final

Authority Reference No

SLR Project No 733.04040.00010

DOCUMENT REVISION RECORD

Rev No. Issue Date Description Issued By

Revision No. 0 January 2018 Client Draft report issued to client AB

Revision No. 1 March 2018 Final Report AB

BASIS OF REPORT

This document has been prepared by an SLR Group company with reasonable skill, care and diligence, and taking account of the manpower, timescales and resources devoted to it by agreement with Dundee Precious Metals Tsumeb (Pty) Limited part or all of the services it has been appointed by the Client to carry out. It is subject to the terms and conditions of that appointment.

SLR shall not be liable for the use of or reliance on any information, advice, recommendations and opinions in this document for any purpose by any person other than the Client. Reliance may be granted to a third party only in the event that SLR and the third party have executed a reliance agreement or collateral warranty.

Information reported herein may be based on the interpretation of public domain data collected by SLR, and/or information supplied by the Client and/or its other advisors and associates. These data have been accepted in good faith as being accurate and valid.

SLR disclaims any responsibility to the Client and others in respect of any matters outside the agreed scope of the work.

The copyright and intellectual property in all drawings, reports, specifications, bills of quantities, calculations and other information set out in this report remain vested in SLR unless the terms of appointment state otherwise.

This document may contain information of a specialised and/or highly technical nature and the Client is advised to seek clarification on any elements which may be unclear to it.

Information, advice, recommendations and opinions in this document should only be relied upon in the context of the whole document and any documents referenced explicitly herein and should then only be used within the context of the appointment.



SLR Project No.: 733.04040.00010


ii

EXECUTIVE SUMMARY

The Environmental Impact Assessment (EIA) for the proposed Dundee smelter expansion required a ground- and surface water study to investigate the likely impacts that additional infrastructure could have on the environment and for these results to be included in a cumulative site impact assessment. From the Groundwater and Surface water Study (SLR 2016), it is been concluded that the planned smelter expansion will have a relatively small impact on the groundwater quantity, but potentially a cumulative negative impact on the groundwater quality. However, there is already a negative impact on the groundwater quality existing, due to significantly intensive smelter operations during the past 100 years.

For this purpose, a reliable estimation of the groundwater flow and contaminant transport was needed. In line with the Environmental and Social Impact Assessment (ESIA) Amendment Process for the Proposed Tsumeb Smelter Upgrade and Optimisation Project, a 3D groundwater flow and transport model has been setup to meet these requirements. This model provides firstly the results for regional and local groundwater flow conditions and secondly a realistic and reliable assessment of potential groundwater contaminant transport.

The model has been setup using the finite-element based software FEFLOW 7 with a steady-state flow and a coupled transient transport solution. Applied hydraulic conductivity values range between 0.002 m/d within the Basement-Complex (aquiclude) and 55 m/d in the fractured and karstified Hüttenberg-Formation (dolomite aquifer). In addition, since site specific porosities are not available, default values according to empirical investigations and groundwater models developed in similar environments were assumed. Porosity values range between 0.1% in the low permeable basement and up to 8% in the fractured dolomites of the Tsumeb-Subgroup (e.g. Hüttenberg-Formation). Regional and local faults and fracture systems have been included as discrete-features into the finite-element model, assuming that they are acting as groundwater pathways. Groundwater recharge has been taken into consideration and varies from 0.01% to 10% of mean annual rainfall (i.e. recharge ranges from 1 to 55 mm/a). For the model calibration existing water level records from a previous hydrocensus (GCS 2016) and further groundwater level readings were used. The calibrated model results meet requirements for a reliable and robust flow model solution (see Reilly & Harbaugh 1996) with an absolute model error of 2.51%. In general, the groundwater flow pattern is comparable to the existing regional flow (see GKW & BICON Namibia 2003).

Based on the calibrated flow solution, groundwater contaminant transport has been simulated. The transient non-reactive transport modelling does not consider adsorption, precipitation and retardation processes, which could further reduce the transport of contaminants. No specific source concentration was modelled and the plumes are illustrated in percentages of the relative source concentration. This is a worst case assumption as in reality seepage concentration will decline over time due to the above mentioned processes. The potential pollution scenarios are showing the predicted situation after 10 years, 25 years, 100 years, and 200 years after the initial contamination started from the relevant source areas. The following results can be summarised:

After 10 years, the plume is predicted to spread mainly towards the general flow direction and concentrations decrease rapidly over a short distance.

After 25 years, the plume is predicted to spread mainly in the general flow direction to the north of the Tsumeb Smelter site. After approximately 800 m, the contaminant concentrations drop to general background concentrations.

After 100 years, at a maximum distance of around 3.2 km from the origin concentrations drop to below 5% of the initial concentration. Although, the plume is spreading further towards the northeast, concentrations do not change, in general (this scenario represents most likely the current situation).

After 200 years, there is no significant change of the plume extent, although minor changes of the concentrations in the close vicinity of the source can be seen. At this stage, a kind of equilibrium has been reached, where reduction processes and dispersion cancel each other out. The hydraulic barrier, represented by the Maieberg Formation limestone and shale, potentially also retards the pollution plume to the north and into the irrigation farming area.



SLR Project No.: 733.04040.00010


iii

CONTENTS

EXECUTIVE SUMMARY ................................................................................................................................ II

1. INTRODUCTION ................................................................................................................................. 1

1.1 BACKGROUND................................................................................................................................................ 1

1.2 MODEL OBJECTIVES ....................................................................................................................................... 2

1.3 SOFTWARE AND MODEL FUNCTION .............................................................................................................. 2

2. GENERAL SETTINGS ............................................................................................................................ 2

2.1 FOOTPRINT OF THE TSUMEB SMELTER ......................................................................................................... 2

2.2 TOPOGRAPHY AND HYDROLOGY ................................................................................................................... 3

2.3 GEOLOGY AND HYDROGEOLOGY ................................................................................................................... 4

2.1 GROUNDWATER QUALITY AND CONTAMINANTS .......................................................................................... 7

3. CONCEPTUAL MODEL ......................................................................................................................... 8

3.1 PREVIOUS MODEL SOLUTIONS ...................................................................................................................... 8

3.2 INPUT DATA ................................................................................................................................................... 8

3.2.1 DIGITAL ELEVATION MODEL ...................................................................................................................................................................... 8

3.2.2 RAINFALL AND RECHARGE ......................................................................................................................................................................... 8

3.1 DELINEATION OF THE MODEL AREA .............................................................................................................. 9

4. NUMERICAL GROUNDWATER FLOW AND MASS TRANSPORT MODEL ................................................ 11

4.1 MODEL DISCRETISATION ............................................................................................................................. 11

4.1.1 HORIZONTAL MESH .................................................................................................................................................................................. 11

4.1.2 VERTICAL DISCRETISATION ....................................................................................................................................................................... 12

4.2 AQUIFER CHARACTERISTIC .......................................................................................................................... 13

4.2.1 HYDRAULIC PROPERTIES .......................................................................................................................................................................... 13

4.2.2 FLOW BOUNDARY CONDITIONS .............................................................................................................................................................. 14

4.2.3 DISCRETE-FEATURE ELEMENTS AND ABSTRACTION WELLS ................................................................................................................... 14

4.2.4 CALIBRATED GROUNDWATER RECHARGE............................................................................................................................................... 15

4.2.5 TRANSPORT BOUNDARY CONDITIONS .................................................................................................................................................... 16

4.3 STEADY-STATE FLOW CALIBRATION AND RESULTS ...................................................................................... 17

4.3.1 WATER LEVELS AND FLOW STATISTICS ................................................................................................................................................... 17

4.3.2 WATER BUDGET RESULTS ........................................................................................................................................................................ 19

4.3.3 PARTICLE TRACKING AND STREAMLINES ................................................................................................................................................ 19

4.4 TRANSIENT TRANSPORT CALIBRATION ........................................................................................................ 20

4.4.1 NON-REACTIVE TRANSPORT MODEL ....................................................................................................................................................... 20

4.4.2 SIMULATED TIME STEPS ........................................................................................................................................................................... 20

5. PREDICTIVE SIMULATION ................................................................................................................. 21

5.1 TRANSPORT SCENARIO NO. 1 (10 YEARS) .................................................................................................... 21

5.2 TRANSPORT SCENARIO NO. 2 (25 YEARS) .................................................................................................... 21

5.3 TRANSPORT SCENARIO NO. 3 (100 YEARS) .................................................................................................. 22

5.4 TRANSPORT SCENARIO NO. 4 (200 YEARS) .................................................................................................. 23

6. MODEL COMPARISON ...................................................................................................................... 23

7. CONCLUSIONS AND RECOMMENDATIONS ........................................................................................ 24

8. REFERENCES .................................................................................................................................... 26



SLR Project No.: 733.04040.00010


iv

APPENDICES

APPENDIX A: OVERVIEW MAP (MAP SCALE 1:75,000) ................................................................................. 27 APPENDIX B: GROUNDWATER CONTOURS, STEADY-STATE (MAP SCALE 1:75,000) .................................... 28 APPENDIX C: PARTICLE TRACKING RESULTS (MAP SCALE 1:50,000) ............................................................ 29 APPENDIX D: POTENTIAL PLUME AFTER 10YEARS (MAP SCALE 1:25,000) ................................................... 30 APPENDIX E: POTENTIAL PLUME AFTER 25YEARS (MAP SCALE 1:25,000) ................................................... 31 APPENDIX F: POTENTIAL PLUME AFTER 100YEARS (MAP SCALE 1:25,000) ................................................. 32 APPENDIX G: POTENTIAL PLUME AFTER 200YEARS (MAP SCALE 1:25,000) ................................................. 33 APPENDIX H: OBSERVED WATER LEVELS VS. COMPUTED HEADS (STATISTICS) ........................................... 34 APPENDIX I: IMPACT OF POTENTIAL POLLUTION ........................................................................................ 35

LIST OF TABLES

TABLE 3-1: INITIAL GROUNDWATER RECHARGE VALUES .................................................................................. 9 TABLE 4-1: VERTICAL DISCRETISATION AND NUMERICAL PARAMETERS ........................................................ 12 TABLE 4-2: HYDRAULIC PROPERTIES TAKEN FROM THE TSUMEB GROUNDWATER STUDY ............................ 13 TABLE 4-3: STEADY-STATE CALIBRATION STATISTICS ...................................................................................... 17

LIST OF FIGURES

FIGURE 1-1: LOCATION OF INFRASTRUCTURE ASSOCIATED TO THE TSUMEB SMELTER .............................. 1 FIGURE 2-1: TOPOGRAPHY AND HYDROLOGY OF THE INVESTIGATION AREA .............................................. 3 FIGURE 2-2: GEOLOGICAL MAP (SOURCE: SLR, RE-ADAPTED AFTER THE TCL GEOLOGY MAP OF 1974) ..... 4 FIGURE 2-3: GEOLOGICAL MAP AND LOCATION OF THE CROSS SECTION .................................................... 6 FIGURE 2-4: SW-NE CROSS SECTION THROUGH THE TSUMEB TOWN AND SMELTER SITE AREA ................ 6 FIGURE 3-1: DELINEATED MODEL AREA AND HYDROGEOLOGICAL UNITS ................................................. 10 FIGURE 4-1: FEFLOW SUPERMESH AND CORRESPONDING FINITE-ELEMENT-MESH ................................. 11 FIGURE 4-2: 3D-FEFLOW-MODEL WITH VERTICAL RESOLUTION AND HORIZONTAL MESH ....................... 12 FIGURE 4-3: BOUNDARY CONDITIONS, HYDRAULIC CONDUCTIVITY AND DISCRETE-FEATURES ................ 14 FIGURE 4-4: RECHARGE WITHIN THE MODEL AREA .................................................................................... 16 FIGURE 4-5: SCATTER PLOT OF COMPUTED VS. OBSERVED HEADS............................................................ 18 FIGURE 4-6: SIMULATED STEADY-STATE GROUNDWATER CONTOURS, MODEL-AREA (RIGHT PICTURE) AND TSUMEB SMELTER PROJECT AREA (LEFT PICTURE) ....................................................................................... 18 FIGURE 4-7: WATER BALANCE OF THE CALIBRATED STEADY-STATE MODEL.............................................. 19 FIGURE 4-8: SIMULATED TIME-STEPS FOR THE TRANSIENT TRANSPORT SOLUTION ................................. 20 FIGURE 5-1: INITIAL STARTING CONCENTRATIONS (LEFT PICTURE) AND POTENTIAL PLUME AFTER 10YEARS (RIGHT PICTURE), FOR DETAILED LEGEND SEE MAPS IN APPENDICES ................................................... 21 FIGURE 5-2: INITIAL STARTING CONCENTRATIONS (LEFT PICTURE) AND POTENTIAL PLUME AFTER 25YEARS (RIGHT PICTURE), FOR DETAILED LEGEND SEE MAPS IN APPENDICES ................................................... 22 FIGURE 5-3: INITIAL STARTING CONCENTRATIONS (LEFT PICTURE) AND POTENTIAL PLUME AFTER 100YEARS (RIGHT PICTURE), FOR DETAILED LEGEND SEE MAPS IN APPENDICES ................................................. 22 FIGURE 7-1: PROPOSED ADDITIONAL MONITORING BOREHOLES ........................................................................ 25



SLR Project No.: 733.04040.00010


v

ACRONYMS AND ABBREVIATIONS

Acronym / Abbreviation Definition

Beak Beak Consultants GmbH

BC Boundary condition

GCS GCS Environmental Engineering (Pty) Limited

DPMT Dundee Precious Metals Tsumeb (Pty) Limited

EPL Exclusive Prospecting License

ESIA Environmental and Social Impact Assessment

EW Electro winning

HRU Hydrological research Unit

k Coefficient of hydraulic conductivity

km Kilometre

m amsl Metres above mean sea-level

MAP Mean Annual Precipitation

ML Mining License

mm Millimetre

NCS Namibia Custom Smelters (Pty) Limited

SLR SLR Environmental Consulting (Namibia) Pty Limited

Dundee Precious Metals Tsumeb (Pty) LimitedDundee Precious Metals - Tsumeb Smelter: 3D Groundwater Flow and Contaminant Transport File name: Appendix E Addendum -


SLR Project No.: 733.04040.00010


Page 1

1. INTRODUCTION

1.1 BACKGROUND

SLR Environmental Consulting (Namibia) (Pty) Limited (SLR) together with Beak Consultants GmbH (BEAK) has been appointed by Dundee Precious Metals Tsumeb (Pty) Limited to develop a 3D groundwater flow and transport model in line with the ESIA (SLR 2017) Amendment Process for the Proposed Tsumeb Smelter Upgrade and Optimisation Project. This model should firstly provide results for the regional and local groundwater flow conditions and secondly a realistic and reliable assessment of potential groundwater contaminant transport. The smelter is located on the outskirts of Tsumeb in the Oshikoto Region of Namibia, approximately 2 km northeast of the Tsumeb town centre (see Figure 1-1). Constructed in the early 1960s, the smelter is one of only five commercial-scale smelters in Africa capable of processing concentrates with a high arsenic content.



SLR Project No.: 733.04040.00010


Page 2

FIGURE 1-1: LOCATION OF INFRASTRUCTURE ASSOCIATED TO THE TSUMEB SMELTER

1.2 MODEL OBJECTIVES

Due to the significant potential environmental impacts associated with the general operations of a smelter of this nature coupled with the ongoing public interest in the facility, a full ESIA process was carried out.

The deliverables detailed in the project are:

Review of the existing information and model results.

Realistic image and overview of the local and/or regional groundwater flow conditions.

Prediction scenarios for contaminant transport or respectively a potential pollution plume.

A report detailing groundwater flow results, transport solutions, and predictive simulations.

1.3 SOFTWARE AND MODEL FUNCTION

For this project it has been decided to use the modelling software FEFLOW 7.0 (Finite-Element-Subsurface-Flow-System, DHI-Wasy GmbH, 2017), which is based on a finite-element solution. This means that the model area or domain is represented by a number of nodes and elements. Hydraulic properties are either assigned to these nodes or elements depending on the nature of the input data. An equation is developed for each node, based on the surrounding nodes. A series of iterations are then run to solve the resulting matrix problem, and the model is said to have “converged” when errors reduce to within an acceptable range. For the current project the so-called SAMG-solver, an algebraic-multigrid for symmetric and asymmetric matrix, has been chosen. According to a widely accepted agreement (by scientist and associated professional), a flow model is said to be calibrated within an acceptable range when errors reduce to < 5%.

2. GENERAL SETTINGS

2.1 FOOTPRINT OF THE TSUMEB SMELTER

Figure 1-1 above shows the infrastructure associated with the Tsumeb Smelter. The following infrastructure and site areas are assumed to be potentially relevant with regard to groundwater and contaminant transport:

Active tailings dam,

Eastern tailings dam,

Evaporation ponds,

DPMT smelter site,

Sewerage works (managed by the Tsumeb Municipality),

Legacy slag (tailings stockpile),

Arsenic and calcine legacy site,

Blast furnace slag stockpile,



SLR Project No.: 733.04040.00010


Page 3

Hazardous waste disposal site ,

Shaft No. 1.

2.2 TOPOGRAPHY AND HYDROLOGY

Tsumeb is located on the eastern side of the Etosha Basin catchment, which is an endoreic drainage system where runoff flows from the north via the Cuvelai Ephemeral River system into the Etosha Pan from where it then evaporates. The area around Tsumeb is predominantly karstic, due to the dissolution of soluble base rock mainly dolomite and limestone in this area, characterised by underground drainage systems with sink holes and caves. It is observed that areas proximal to Tsumeb are of high elevation, up to 1550 m amsl, in comparison to areas further north of the town where elevation drops to 1150 m amsl (see Figure 2-1). Due to the geology of the area, there is no well-defined drainage pattern in the Tsumeb-Grootfontein area, but rather many small individual drainage systems, dependant on the local geology. No major surface runoff is observed due to rapid intake of water by the karstic features in most areas.



SLR Project No.: 733.04040.00010


Page 4

FIGURE 2-1: TOPOGRAPHY AND HYDROLOGY OF THE INVESTIGATION AREA

2.3 GEOLOGY AND HYDROGEOLOGY

The area is situated in the Otavi Mountainland (OML). As part of the ESIA (SLR 2017) a detailed introduction of the geology, stratigraphy and tectonic situation is given. Figure 2-2 shows the geological situation of the investigation area with Quaternary cover.

FIGURE 2-2: GEOLOGICAL MAP (SOURCE: SLR, RE-ADAPTED AFTER THE TCL GEOLOGY MAP OF 1974)



SLR Project No.: 733.04040.00010


Page 5

In hydrogeological terms, relevant groundwater flow is restricted to the dolomites of the Tsumeb Subgroup and minor to the other fractured but less permeable hard rocks. All geological formations acting either as aquifer, aquitard or aquiclude are summarised in stratigraphical order:

The Nosib Group unconformably overlies the Basement Complex. It consists of the Nabis and Varianto formations. The environment of deposition progressively developed from predominantly fluvial to marine when finer grained shales were deposited (Kamona & Günzel, 2006). Aquifers of the Nosib Group are generally of moderate potential.

The Otavi Group consists of Abenab and the Tsumeb Subgroups which are unconformably overlying the Nosib Group and the Basement Complex (Hedberg, 1979). The Group’s sediments form major aquifers in the area utilised for bulk water supply. The Tsumeb Subgroup, is subdivided into 8 litho-zones (T1 to T8) from the clastic Ghaub Formation to the carbonate dominant Maieberg, Elandshoek as well as the Hüttenberg Formations elaborated on below:

o The Ghaub Formation, referred to as T1, is a glacio-marine tillite with lenses of dolomite and schist. The Maieberg Formation is a platform slope, deep water deposit and overlies the Ghaub Formation. The lower Maieberg Formation (T2) consists of slump brecciated and laminated carbonate and argillaceous sediments. The upper Maieberg Formation (T3) comprises bedded and finely laminated carbonates.

o The Elandshoek Formation conformably overlies the Maieberg Formation. It covers most of the northern limb of the Otavi Valley north of Kombat Mine. The lower Elandshoek Formation (T4) comprises of massive dolomite and is responsible for the rugged geomorphologic terrain of the northern limb of the Otavi Valley. The brecciation is generally intensive and therefore T4 is regarded as an important aquifer (Van der Merwe, 1986). The upper Elandshoek Formation (T5) is fairly thin and not easily distinguishable from T4.

o The Hüttenberg Formation marks the change from the deep sea environment observed in the Elandshoek Formation to shallow lagoon shelves. The high potential aquifer consists of a grey bedded basal dolomite, stromatolite rich (T6), overlain by two upper units, a massive dark and bedded dolomite with chert and with phyllite (T7) and T8 is marked by pisolite and oolite.

The Mulden Group is characterised by the Kombat Formation in the southern part of the OML, which consists of a siliciclastic molasses (poorly graded phyllite, arkose, argillite and siltstone) deposited syntectonically during the early stage of the Damara Orogeny, and the Tschudi Formation (Arkose and feldspathic sandstone) in the northern part of the OML, and is separated from the Tsumeb Subgroup by an angular disconformity.

Tertiary and Quaternary cover is mainly composed by sediments of the Kalahari Supergroup form a thin cover in most areas around Tsumeb; it can be thicker to the north and northwest of the town. The Kalahari sediments are represented by calcrete deposits around Otavi and they are locally important shallow aquifers.

A stratigraphic column for the OML is shown in detail in SLR 2016, with the regional geology depicted in. The town of Tsumeb lies on the northern edge of the OML and the dolomites of the Otavi Group characterise the area. The sandstones and shales of the Mulden Group have been preserved in the Tschudi Syncline which



SLR Project No.: 733.04040.00010


Page 6

extends in an east-west direction and is the representative geology of most of the area covered by the town. This syncline has been depicted within the GCS 2016 report (see Figure 2-3 and Figure 2-4).

FIGURE 2-3: GEOLOGICAL MAP AND LOCATION OF THE CROSS SECTION

FIGURE 2-4: SW-NE CROSS SECTION THROUGH THE TSUMEB TOWN AND SMELTER SITE AREA

The town falls within the Tsumeb Sub-basin of the Cuvelai-Etosha Basin. Groundwater occurs in the Tsumeb Karst Aquifer (TKA) with the Mulden Group shale and sandstone acting as an aquiclude. The Smelter site is located on the Elandshoek and Hüttenberg Formation, both are fractured and locally karstified dolomitic aquifers, in an ESE-WNW sloping valley formed as part of a synclinal/anticlinal structure. The groundwater is expected to move along fractured tectonic structures like fold axes, pressure relief joints, faults or on lithological contact zones (see Figure 2-2 above). The average natural groundwater levels in Tsumeb are at approximately 1 210 m amsl (60 m below the land surface in the town area) with little seasonal fluctuation in the levels.

The Tsumeb Mine was operational from 1907 to 1996, temporarily closed until 2000 then recommissioned for a short period. (GCS, July 2013). Dewatering occurred at Shaft No.1 (approximately 1 600 m deep) at the old Tsumeb Mine, south-west of the smelter. In 2000, water was pumped from the shaft for mining of mineral specimens from the upper levels of the mine (approximately 250 m below ground surface) at a rate of 350 m3/hr (WSP Walmsley, 2004). It is understood that during actual mining operations until 1996, the water was pumped from a much greater depth.



SLR Project No.: 733.04040.00010


Page 7

The natural groundwater flow from Tsumeb is in a northerly direction. The dolomite of the Hüttenberg Formation shows high transmissivity and it is estimated that water migrates at a rate of approximately 1.08 m/day or 360 m/annum (GCS, 2013), although secondary fracture flow in the area may result in localised acceleration of the groundwater flow rates. The high elevations areas are considered as recharge zone for the groundwater. Noteworthy, groundwater flow is predominantly happening within the Tsumeb Subgroup. The Kalahari aquifer is acting with its thin layered sand and calcrete as perched water table.

2.1 GROUNDWATER QUALITY AND CONTAMINANTS

Based on a hydrocensus undertaken in November 2012, GCS (2013) summarises the background groundwater quality as follows:

High calcium, magnesium, bicarbonate water is encountered as expected from dolomitic water.

pH values between 6.9 and 7.4 were measured in the hydrocensus boreholes. pH across boreholes is stable.

Elevated concentrations, above Namibian drinking water standards, of sulphate (SO4), arsenic (As) and molybdenum (Mo) were measured at the boreholes situated on the smelter site.

Elevated iron (Fe) is also observed in all boreholes and is probably a result of the mineralogical composition of the rock.

It is important to view the groundwater quality monitoring results against some background values for the larger karst region, specifically when looking at arsenic pollution. Data from studies that have focused on the wider area indicates elevated arsenic concentrations in areas not previously affected by mining which may be reflective of naturally high background arsenic levels in the geological formations (see SLR 2016 and SLR 2017).

Results of DPMT’s July 2015 groundwater sampling round showed that only the Calcine- and Return Boreholes had arsenic concentrations exceeding the Namibian Guideline values for drinking water. All other boreholes had concentrations falling within the Group B or better quality drinking water according to the Guideline. According to the outcome of the groundwater model update in 2016 (see GCS 2016), there are a couple of elevated species and elements in the on-site boreholes. The following list summarises the important elements with significantly elevated concentrations:

Elevated sodium (Na) and sulphate, which is probably originating from the As / Calcine legacy slag site,

Elevated Fe in all on-site boreholes,

Elevated Cd near the Calcine legacy slag site,

High concentration of As and Mo in all on-site boreholes, with peak concentrations near the Calcine legacy slag site.

However, potential contaminant sources currently related to the site include more than 100 years of smelter operations with 2 tailings dams and various storage areas of slag and calcine. It appears that runoff from the site has, occasionally, accumulated away from the site, acting as an additional (secondary) source of groundwater contamination. Due to the fact, that there are no monitoring boreholes located within the potential down gradient flow field to the North of the Tsumeb Smelter Site, it is not possible to delineate a potential pollution plume based on monitoring data. Thus, all on-site boreholes can only give an estimate of the peak concentrations within the site area. Therefore, contaminant transport has to be carried out as conservative worst-case scenarios.



SLR Project No.: 733.04040.00010


Page 8

3. CONCEPTUAL MODEL

3.1 PREVIOUS MODEL SOLUTIONS

In 2013, GCS Environmental Engineering (Pty) Limited (GCS) carried out a hydrogeological study which included a groundwater contaminant transport model (GCS 2013). This model was established to provide a management and risk assessment tool. The ESIA Report (SLR 2017) underlined that the GCS groundwater model for DPMT was comparably simplistic and can be regarded as a low confidence- high level model. The outcome of the GCS model of 2016 and the extent of the pollution plume was partly influenced by erroneous laboratory data, showing elevated arsenic concentrations in production boreholes on farm Mannheim. Later sampling rounds and most recent sampling by DWAF (DWAF, 2017) shows that this is not the case and arsenic concentrations in sampled boreholes on farm Mannheim are below detection limit. The ESIA (SLR 2017) recommended developing an improved model that accounts in more detail for.

complex geology of the area, including the Mannheim Dome with Maieberg Formation as possible hydraulic barrier;

information from new boreholes (still to be drilled) to provide water level, geological, geophysical and chemical data for more accurate predictions on plume migration and

the impact of groundwater abstraction.

Such a model should build on the existing model for the area considered during the Tsumeb Ground Water Study (TGWS) in 2003 (GKW CONSULT / BICON, 2003). For that study, a regional numerical groundwater flow model was set up for the TKA (IV Synclinorium of the OML). The TGWS model showed more realistic and reliable results of the regional groundwater flow patterns. Nevertheless, the model developed under this study and an updated version from 2016 (GCS 2016) represent an important basis for the new FEFLOW-model developed under the recommendation of the ESIA Report (SLR, 2017).

3.2 INPUT DATA

3.2.1 Digital Elevation Model

A digital elevation model of the project area derived from SRTM (Shuttle Radar Topographic Mission, JARVIS et al, 2008) data was used as a general terrain reference. The SRTM 90 m DEM’s have a resolution of 90 m at the equator, and are provided in mosaicked 5 deg x 5 deg tiles for easy download and use. All are produced from a seamless dataset to allow easy mosaicking. These are available in both ArcInfo ASCII and GeoTiff format to facilitate their ease of use in a variety of image processing and GIS applications.

3.2.2 Rainfall and Recharge

Aquifer recharge resulting from rainfall events is dependent on the temporal and spatial variation in precipitation as well as the host rock surface and subsurface susceptibility in terms of infiltration and storage. By virtue of its surface and subsurface hydrogeological conditions, the dolomite outcrops are potentially a good recharge area. Hence it is also susceptible in terms of pollution. Due to the fact that most of the precipitation in the region is in the form of thunderstorms, it is likely that some of the rain water infiltrates into open fractures and faults to finally reach the groundwater table before evaporating or flowing into the next river bed. Indirect recharge due to flood events in ephemeral rivers is assumed to be of minor importance in the project area.

A large range of recharge rates were reported by several investigations within Namibia and the study area itself (see Bardenhagen 2007). Values vary from 0.01% to 10% of mean annual rainfall, accordingly. Tsumeb has an annual average rainfall of 520 mm (between 500 and 600 mm in the area) with most of the rainfall occurring in the summer months (October to April). Approximately two thirds of the rainfall occurs in the months of



SLR Project No.: 733.04040.00010


Page 9

January, February and March, with the highest number of productive rainfall days (i.e. days with rainfall of 10 mm and more) registered in January and February. According to the Tsumeb-Groundwater-Study (Bardenhagen 2007) groundwater recharge is extremely variable in the area (see Table 3-1).

TABLE 3-1: INITIAL GROUNDWATER RECHARGE VALUES

Formation description Type Assigned recharge value

Dolomite covered by a thin layer of Kalahari calcrete or sand

Aquifer Approximately 1% of the long-term average annual precipitation

Dolomite Outcrop Area Aquifer 10% of the mean annual precipitation at the central and eastern ranges and 6% of the mean annual precipitation towards the west

Schist, mixtite and phyllite Aquitard Less than 0.1% of the annual precipitation

Basement sub outcrop areas Aquitard 0.01% of the annual precipitation

Dolomite outcrops are showing relatively high groundwater recharge, whereas basement sub outcrops are significantly lower. Wherever dolomite and/or other rock types are overlain by Kalahari calcrete or sandy layers, lower recharge values have been observed.

3.1 DELINEATION OF THE MODEL AREA

To delineate the model area all information that potentially is controlling the regional and local groundwater flow has been gathered and depicted in Figure 3-1. Due to the absence of a proper natural drainage system the surface water catchments are hardly significant. Furthermore, they are not necessarily depicting the groundwater flow pattern since the flow within the relevant dolomites is restricted to fractures and fissures. Groundwater contours are taken from the so called Tsumeb Groundwater Study (see GKW GmbH & BICON 2003; Bittner 2004). They indicate a general flow direction from the southern basin margin towards the basin centre in the North (see Figure 3-1).

Taken the groundwater flow (indicated by the existing groundwater contours) and the surface water catchments into account it becomes clear that there is a natural western and eastern boundary limiting the groundwater flow to the West and the East. This boundary is most likely caused by runoff and thus resulting recharge. However, a northern and/or southern limitation of the model area neither be derived from the surface water catchments nor the regional groundwater flow pattern. In this case the geological formations are striking E-W. Hence, distribution boundaries of geological and/or hydrostratigraphical units can be taken as basis for a northern and southern model boundary. Thus, a model area has been delineated (see Figure 3-1) with the boundaries described above. A separate description of the numerical implementation is given in chapter 4.2.2.



SLR Project No.: 733.04040.00010


Page 10

FIGURE 3-1: DELINEATED MODEL AREA AND HYDROGEOLOGICAL UNITS



SLR Project No.: 733.04040.00010


Page 11

4. NUMERICAL GROUNDWATER FLOW AND MASS TRANSPORT MODEL

4.1 MODEL DISCRETISATION

4.1.1 Horizontal Mesh

Various input data have been used to define and create a finite-element-mesh. The following input information and spatial data were taken as a basis for the super mesh (i.e. all input polygons and lines and point data):

Potentially contaminated areas and groundwater relevant smelter infrastructure,

Monitoring boreholes and abstraction wells,

Tectonic structures (i.e. faults and fractures), dykes and lineaments,

Ephemeral rivers and drainage system,

Geological units and formation.

One of the biggest advantages of a finite-element solution compared to a finite-difference solution is the possibility to setup a triangular mesh or to work in a totally unstructured mesh. Thus, linear structures can be considered realistically. For this project a regular mesh based on triangles has been developed (see Figure 4-1).

FIGURE 4-1: FEFLOW SUPERMESH AND CORRESPONDING FINITE-ELEMENT-MESH

The numerical groundwater flow model consists of 418,458 elements and 281,192 nodes in 3 layers. The model area is covers an area of 697.9 km².



SLR Project No.: 733.04040.00010


Page 12

4.1.2 Vertical Discretisation

Based on the finite-element mesh, the model has been expanded to fully 3-D by the inclusion of layers and elevation data. The triangles of the horizontal mesh were changed to triangular prisms (see Figure 4-2).

Following the hydro-stratigraphy of the investigation area (see chapter 2.3) the vertical resolution and discretisation has been summarised in Table 4-1.

TABLE 4-1: VERTICAL DISCRETISATION AND NUMERICAL PARAMETERS

Hydro-stratigraphic Unit

Numerical Layer No.

No. of Zones

Type Status Average Thickness

Weathered zone 1 5 Aquifer / Aquitard Phreatic ~ 10 m

Tsumeb subgroup 2 5 Aquifer / Aquitard Confined 10 – 1000 m

Basement complex 3 1 Aquiclude Confined 300 m

FIGURE 4-2: 3D-FEFLOW-MODEL WITH VERTICAL RESOLUTION AND HORIZONTAL MESH



SLR Project No.: 733.04040.00010


Page 13

4.2 AQUIFER CHARACTERISTIC

4.2.1 Hydraulic Properties

Based on a variety of borehole information (see Bittner 2004), hydraulic- and pump-tests (see Bittner 2004, GKW & BICON 2003), and previous model solutions (see GCS 2013, GKW & BICON 2003) a detailed picture of the k-values can be given (see Table 4-2).

TABLE 4-2: HYDRAULIC PROPERTIES TAKEN FROM THE TSUMEB GROUNDWATER STUDY

Hydro-stratigraphic Unit k [m/d] Storage

Coefficient [-]

Porosity

[%] Characterisation Description

Tschudi-Formation

(Mulden-Group) 0.032 3.0 X 10

-04 1 Aquitard

Fractured aquitard. Fault zone acting

as a vertical conduit within low

permeable fractured rock. Evaluation

suggests bilinear rather than linear

flow field indicating a finite

conductive vertical fracture zone.

Hüttenberg-Formation

(Tsumeb-Subgroup) 55.65 1.6 X 10

-03 8 Aquifer Homogeneously fractured aquifer

Elandshoek-Formation

(Tsumeb-Subgroup) 12.60 7.0 X 10

-04 6 Aquifer Homogeneously fractured aquifer

Maieberg-Formation

(Tsumeb-Formation) 1.23 9.0 X 10

-05 2.8

Aquitard or

Aquifer

Fractured aquifer. Fault zone acts as

a vertical conduit. Length of

hydraulically active zone appears to

be much smaller than the length of

the lineament determined from

geophysical data.

Gauss-Formation

(Abenab-Subgroup) 4.66 6.0 X∙ 10

-04 3.8 Aquifer

Fractured aquifer. Either

homogeneously fractured rock or

finite conductive fracture zone with

considerable storage properties

Basement-Complex 0.002 0 0.1 Aquiclude Virtually impermeable rock

Although there is hydrogeological information available, it is heterogeneously distributed in the area. Given the lack of information on extensive hydrogeological characteristics especially for porosity, a literature review on hydraulic properties of rock types in the project area has been completed. In this case, guidelines for groundwater modelling approaches (see Barnett et al 2012, Reilly & Harbaugh 1996, and Robertson Geo-consultants & SRK 2012) recommend using average hydraulic property values for hydro-stratigraphic units instead of an interpolated hydraulic property map. Figure 4-3 shows the model k-values.

Since k-values, storage coefficient and porosity represent average values, these hydraulic properties have not been changed during the calibration process. Furthermore, it is believed that these hydraulic properties are the only known fixed values in the area. For the model calibration only groundwater recharge (for groundwater flow), advection, longitudinal and transversal dispersion (for the transport solution) have been changed within a reasonable range (i.e. ± 10% of the initial values).



SLR Project No.: 733.04040.00010


Page 14

4.2.2 Flow Boundary Conditions

For most boundary conditions it is of capital importance to use them as little as possible. 1st kind boundary conditions (constant head or Dirichlet-BC) create an inflow or outflow, which easily over or underestimates the water budget. When applying 2nd kind boundary conditions (fluid-flux or Neumann-BC) a realistic estimation or proof of the in- and outflowing groundwater volume is needed. 3rd kind boundary conditions (fluid-transfer or Cauchy-BC) typically represent rivers or other water bodies. For this particular numerical groundwater model only the northern boundary shall be constrained with a constant head boundary to simulate the outflow of the project area. Firstly this assures a minimum impact on the available water volume, secondly it avoids an over or underestimation of the groundwater volume as the only source for groundwater within the model area is recharge from rainfall.

The boundary condition has been assigned according to the delineation of the model area (see Section 3.1 and Figure 4-3) with the groundwater heads of the existing regional model (see Bäumle 2003).

FIGURE 4-3: BOUNDARY CONDITIONS, HYDRAULIC CONDUCTIVITY AND DISCRETE-FEATURES

4.2.3 Discrete-Feature Elements and Abstraction Wells

Discrete Features can be inserted in the model domain to represent high-conductivity features that can be approximated as one- or two-dimensional, such as:

Conductive faults, fractures and fissures

Horizontally screened wells

Shafts, tunnels, pipes and drains

Discrete Features are additional finite elements of a lower dimension than the basic finite-element mesh, i.e. one-dimensional finite elements in 2D models and one- or two-dimensional discrete features in 3D models. Discrete Features can be placed along edges or faces of the mesh elements or they connect two arbitrary nodes.



SLR Project No.: 733.04040.00010


Page 15

For the Tsumeb model, two different types of discrete-feature have been incorporated (see Figure 4-3):

Regional faults and fractures were included as conductive lineaments as part of the Tsumeb-Subgroup,

The Tsumeb Shaft No. 1 has been included as shaft-feature-element.

Information and estimations of the conductivity of faults and fractures within the predominantly dolomitic Tsumeb-Subgroup has been taken from Bardenhagen (2007), GKW Consult & BICON Namibia (2003), and Bittner (2004) where available. Thus, the conductivity of these features is elevated compared to the rocks in the vicinity by a factor of 1.

The shaft has been incorporated as open space with a constant discharge rate of 320 m³/h, contributing to the Tsumeb water supply. Since 2000, for operation of mine, 300 m³/h to 340 m³/h or 2.63 Mm³/a to 2.98 Mm³/a, respectively, gets pumped out of the mine shaft. Water is pumped out of mine shaft from 235 - 250 m below surface. No drop of groundwater level in observation boreholes neither in the area of Tsumeb town nor in a radius of 20 km has been observed. Additionally, other abstraction wells are contributing to the Tsumeb Municipality water supply.

The current abstraction can be summarised as follows:

Tsumeb area:

o 1.83 Mm³/a are delivered to the Smelter,

o 1.67 Mm³/a are used by the Tsumeb Municipality for public water supply,

Irrigation / farm water supply:

o 2.03 Mm³/a are used by irrigation farms

o 0.15 Mm³/a are used by other farms.

All in all, this leads to an overall abstraction volume of 5.68 Mm³/a.

4.2.4 Calibrated Groundwater Recharge

Table 3-1 above lists the initial groundwater recharge rates. During the calibration processes these values were adjusted and modified within a certain limit (< 15%). Figure 4-4 is showing the groundwater recharge after the terminated calibration process, it is observed that highest recharge is in the south of the model area.



SLR Project No.: 733.04040.00010


Page 16

FIGURE 4-4: RECHARGE WITHIN THE MODEL AREA

4.2.5 Transport Boundary Conditions

According to the flow boundaries, separate boundary conditions for the solute transport solution are needed. In this case, a conservative approach has been chosen. Porosity values of 0.1% - 8% were used in the bedrock units. The longitudinal dispersivity (DL) is assumed to range between 30 and 60 m. A DL of 60 m was used in the simulations discussed below. For the longitudinal, transversal and vertical dispersivity (DL, DT, DV) a reasonable ratio was specified with 100:10:1 (Kinzelbach et al. 1995).



SLR Project No.: 733.04040.00010


Page 17

4.3 STEADY-STATE FLOW CALIBRATION AND RESULTS

4.3.1 Water Levels and Flow Statistics

Table 4-3 lists all calibration targets used and shows residuals and calibration statistics for the steady state flow solution. The outcome underlines the model performance, robustness and reliability of prediction scenarios.

A couple of statistical errors are commonly used to describe numerical model performances. However, the most substantial statistical error to determine the calibration quality is considered to be the relative model error (Ē) and the coefficient of determination (CD). According to good practice standards (see chapter 1.3 and Barnett et al. 2012) the CD should be close to 1 and the Ē for good calibration <10% and for excellent calibration <5%. Apart from this, the residual mean is giving an indication whether the computed heads are plotting too high (i.e. water levels were simulated higher than observed ones) or the water levels are plotting too low (i.e. heads were simulated lower than original ones).

TABLE 4-3: STEADY-STATE CALIBRATION STATISTICS

Item Abbrevation Statistics

Number of observation boreholes 41

Residual mean RM 1.90 m

Minimum residual Min -3.57 m

Maximum residual Max 7.35 m

Range in target values Range 72.00 m

Relative model error Ē 2.51%

Coefficient of determination CD 0.93

Root mean squared RMS 2.99

Figure 4-5 displays the scatterplot of residuals giving a graphical representation of the goodness of fit. The scatterplot shows that there is no systematic error in the spatial differences between computed and observed groundwater heads.

No expectable fit has been achieved for borehole WW33382, which has been excluded from the statistics. This residual is possibly due to lack of information, complex geology, unknown geometries, and hydraulic parameters. Furthermore, elevation data for this borehole is non-explicit, with values ranging from 1,266 m amsl (topo information from Google) and 1,232 m amsl (topo information from DEM, SRTM-dataset). However, there are a few computed heads showing minor deviations. Since these values are lying within an acceptable limit no further action is required.



SLR Project No.: 733.04040.00010


Page 18

FIGURE 4-5: SCATTER PLOT OF COMPUTED VS. OBSERVED HEADS

The results of the steady-state calibration process are shown in Figure 4-6. In general, groundwater heads indicate a flow from South to North. Where ever groundwater is migrating in low permeable hydrogeological units the gradient is relatively steep. This can be observed within the Tschudi-Formation (Mulden-Group) in the area of the Tsumeb town centre (see Figure 4-6). Just to the North, in the surroundings of the Tsumeb-Smelter, better permeable units (especially the Huettenberg-Formation) are causing lower gradients.

In general, there is neither any impact due to water abstraction from farm boreholes visible, nor any impact from the Tsumeb shaft water abstraction. Due to the high permeable rocks (fractured dolomites, partially karst formations), no drop in groundwater levels is recorded (see chapter 4.2.3).



SLR Project No.: 733.04040.00010


Page 19

FIGURE 4-6: SIMULATED STEADY-STATE GROUNDWATER CONTOURS, MODEL-AREA (RIGHT PICTURE) AND TSUMEB SMELTER PROJECT AREA (LEFT PICTURE)

A detailed map of the modelled groundwater contours can be seen in the appendices (see Appendix B).

4.3.2 Water Budget Results

The major controlling parameter for the water budget in the area is groundwater recharge from rainfall with an average daily volume of 41,161 m³ (this is equal to 15 Mm³/a). Furthermore, it is the only source for water inflow into the system (see Figure 4-7). A volume of approximately 15,561 m³/d (which is approx. 5.68 Mm³/a) has been simulated as abstraction for farming and irrigation purposes and from the Tsumeb Municipality. It is believed that the amount of abstracted groundwater is a realistic estimation, since it is based on the Integrated Resources Management Plan of 2010 and the prediction for 2015-2020.

FIGURE 4-7: WATER BALANCE OF THE CALIBRATED STEADY-STATE MODEL

Apart from groundwater abstraction, most of the groundwater volume is leaving the model area as outflow in northern directions, following the regional groundwater flow pattern (see Dirichlet BC, constant head boundary).

The overall groundwater budget is lower compared to the previous model solution (see GCS 2013). All in all, the volumes are reasonable, following the results of the Tsumeb Groundwater Study (see Bäumle 2003).



SLR Project No.: 733.04040.00010


Page 20

4.3.3 Particle Tracking and Streamlines

Particle tracking in combination has been applied to evaluate the flow paths. Starting locations were set at potential contamination sites (see Figure 1-1 and Appendix A). From each location a number of path lines indicate the flow field. Results have been depicted in Appendix C.

4.4 TRANSIENT TRANSPORT CALIBRATION

4.4.1 Non-Reactive Transport Model

In the transient non-reactive transport modelling only advection, longitudinal and transversal dispersion were considered. Adsorption, precipitation and retardation were not considered and the solute was treated as a conservative tracer. Hence processes which could reduce transport of contaminants were not modelled. In addition, since site specific porosities are not available, default values according to empirical investigations and groundwater models in similar environments were assumed. No specific source concentration was modelled and the plumes are illustrated in percentages of the relative source concentration. This is a worst case assumption as in reality seepage concentration will decline over time due to retardation.

Since no observation borehole is located within the pollution plume, there is no reliable calibration point available.

4.4.2 Simulated Time Steps

Since the model has been setup as and transient transport in a steady-state flow field, time steps for the contaminant transport solution had to be incorporated (see Figure 4-8). All in all, 7 time steps were set, simulating an observation period of 100 a.



SLR Project No.: 733.04040.00010


Page 21

FIGURE 4-8: SIMULATED TIME-STEPS FOR THE TRANSIENT TRANSPORT SOLUTION

The spreading of the potential plume within the weathered zone (layer 1) and layer 2 (predominantly dolomites of the Tsumeb-Subgroup) is illustrated by percentages of the source concentration at three different times, namely 10, 25, and 100 years. These time steps represent the following predictive scenarios.

Step-NoSimulated

Time [a]

Simulated

Time [d]

1 1 365

2 5 1825

3 10 3650

4 25 9125

5 50 18250

6 75 27375

7 100 36500



SLR Project No.: 733.04040.00010


Page 22

5. PREDICTIVE SIMULATION

5.1 TRANSPORT SCENARIO NO. 1 (10 YEARS)

The 1st scenario represents the situation after 10 years. After 10 years, the plume is predicted to spread mainly towards the general flow direction. Concentrations decrease rapidly. At a maximum distance of around 150 m from the origin concentrations drop to below 5% of the initial concentration (see Figure 5-1).

FIGURE 5-1: INITIAL STARTING CONCENTRATIONS (LEFT PICTURE) AND POTENTIAL PLUME AFTER 10YEARS (RIGHT PICTURE), FOR DETAILED LEGEND SEE MAPS IN APPENDICES

During the first years the plume is leaching through the weathered zone, which obviously needs some time. After 10 years a first plume has been developed. Appendix D shows that the plume extends outside of the DPMT property boundary after 10 years with less than 5 % of original concentration


The 2nd scenario represents the situation after 25 years. After 25 years, the plume is predicted to spread mainly towards the general flow direction. Compared to the situation 10years after start, pollution plumes were predicted from all potential sources. Figure 5-2 shows that concentrations are below general background concentrations at a distance of 800 m from the potential contaminated sites.

Appendix A shows the potential pollution plume of the 2nd predictive scenario, the plume extends outside of the DPMT property boundary after 10 years with less 5-10% of original concentration.



SLR Project No.: 733.04040.00010


Page 23



The 3rd scenario represents the situation after 100 years (see Figure 5-3). At a distance of approximately 3.2 km from the origin, concentrations drop to below 5% of the initial concentration. Although, the plume is spreading further into North-East, concentrations exceeding 10% of the initial concentration do not change at all. At this stage, the plume seems to have reached a kind of equilibrium, since no significant change is predicted beyond 100 years from the start (see comparison between Figure 5-2 and Figure 5-3). Appendix F shows that the plume extends outside of the DPMT property boundary with concentrations exceeding 10% of original concentration.



SLR Project No.: 733.04040.00010


Page 24



A last 4th prediction scenario has been calculated (see Appendix G) at 200 years after initial start. However, there is no significant change between the 3rd and the 4th scenario observed. After 200 years, there is no significant change of the plume length visible, although minor changes of the concentrations in the close vicinity of the source can be seen. At this stage, a kind of equilibrium has been reached, where reduction processes and dispersion cancel each other out. The Maieberg Formation aquitard, located between the smelter in the south and the irrigation farms in the north, forms a hydraulic barrier as indicated by the narrowing groundwater contour lines. In this regard, the barrier is expected to slow down groundwater flow towards the irrigations areas (see Appendix I). Further to the west of the plume; a groundwater divide is observed and outlined, this implies that potential contamination from the DPMT may not be transported to the west towards the sewage works and more importantly towards the water supply boreholes managed by the Tsumeb Municipality (see Appendix I).

6. MODEL COMPARISON

With regard to the results, there are a couple of differences between the current FEFLOW-based groundwater flow model solution and the previous GCS-model. The following should be noted:

For the GCS-model, porosity-values above 20% have been used. These values are far too high. Results from the Tsumeb Groundwater Study (e.g. GKW & BICON 2003) and general literature are indicating porosities below 10%.

Longitudinal dispersivity values of the GCS model were adjusted, so that the current situation (with erroneous elevated arsenic concentrations of irrigation farm monitoring boreholes) were reproduced and simulated. Thus, dispersivity values of the previous model are supposed to be significantly higher and not realistic.

In contrast with the GCS-model, the current SLR 2018 model aims to build on the existing models for the area considered during the Tsumeb Ground Water Study (TGWS) in 2003 (GKW CONSULT / BICON, 2003) and also using the most detailed geological map of the area in the form of the old TCL map of 1974 (digitised as part of the TGWS in 2003). The digital map was made available to DWAF for further use.

The SLR 2018 model includes the hazardous waste disposal facility as a potentially unlined/leaky facility. A potential pollution plume was simulated for best practice purposes. However it should be noted that the facility is an engineered lined facility and therefore not expected to be a source of groundwater pollution within the first at least 25 years.

The SLR 2018 model produced a potential plume and does not quantify any concentrations in areas where the plume extends due to uncertainties that result from lack of information in those areas.

The two models are rather incomparable due to the above mentioned points (although the groundwater flow direction and flow path of the potential pollution plume are similar at least for the higher concentrated parts of the plume). It is observed that the basis of the GCS-model resulted in an overestimation of the plume length (for low arsenic concentrations) even when it comes to a conservative approach. Furthermore, since no observation wells are located in the area of the expected plume, a reliable simulation and re-production of the plume is not possible. The SLR 2018 model is more conservative and the resulting potential plume is observed not extend into the farming area.



SLR Project No.: 733.04040.00010


Page 25

7. CONCLUSIONS AND RECOMMENDATIONS

As a result of the simulated groundwater flow as a steady-state solution and the predicted contaminant transport as a time-varying function, further conclusions can be drawn and recommendations can be given.

The calibrated groundwater flow depicts a mixture of water level records measured at different dates. Hence, a single reference date measurement, which takes all relevant monitoring boreholes and farming boreholes into account, will improve the model adjustment and its performance.

Groundwater abstraction does not have any impact on groundwater levels in the area, although the real abstraction rates (e.g. from farming wells) might be significantly higher.

Predicted contaminant transport is showing high concentrations within the Tsumeb Smelter area, especially for arsenic.

Although no monitoring borehole outside the site area is proving contamination, simulated potential plumes are indicating that a pollution plume has developed well outside the site area.

According to the transport modelling results, it is assumed that the current (year 2018) plume does not extend further north than shown in the 100 years scenario (Appendix F).

In fact, processes which could reduce the transport of contaminants (adsorption, precipitation and retardation) were not considered.

Results of the transient transport are indicating an equilibrate situation of the potential plume.

Models are dynamic decision making tools that should continually be re-evaluated with new information. This allows for high level of confidence in the results from models. The following recommendations are arising from the processing and interpretation of the groundwater flow and contaminant transport results:

A revision and update of the groundwater flow should be pursued once consistent water level records (reference date measurement) are available.

It is of capital importance to drill new monitoring boreholes, especially in the area of the expected pollution plume. Eleven (11) monitoring borehole are proposed with preliminary locations shown in Figure 7-1 below.

The newly drilled boreholes should be sampled and monitored on a regular basis and incorporated into the groundwater monitoring network, as recommended in the ESIA (SLR, 2017).

After evaluation of the monitoring results, the numerical flow and contaminant transport model should be updated and revised by expanding it into a reactive transport solution model, where adsorption, precipitation and retardation processes are incorporated. The calibrated reactive transport model should result in a more realistic simulation of the plume extent.



SLR Project No.: 733.04040.00010


Page 26

FIGURE 7-1: PROPOSED ADDITIONAL MONITORING BOREHOLES

UNSIGNED COPY

Markus Zingelmann

(Report Author)

Winnie Kambinda

(Report Co-author)

Arnold Bittner

(Project Manager and Reviewer)



SLR Project No.: 733.04040.00010


8. REFERENCES

JARVIS, A., H.I. REUTER, A. NELSON & E. GUEVARA (2008): HOLE-FILLED SRTM FOR THE GLOBE VERSION 4, AVAILABLE FROM THE CGIAR-CSI SRTM 90M DATABASE (HTTP://SRTM.CSI.CGIAR.ORG).

SLR ENVIRNONMENTAL CONSULTING NAMIBIA (PTY) LIMITED (2017): ESIA AMENDMENT PROCESS FOR THE PROPOSED TSUMEB SMELTER UPGRADE AND OPTIMISATION PROJECT: ENVIRONMENTAL IMPACT ASSESSMENT REPORT. REPORT FOR DUNDEE PRECIOUS METALS INC. TSUMEB. SLR PROJECT NO.: 734.04040.00008.

SLR ENVIRNONMENTAL CONSULTING NAMIBIA (PTY) LIMITED (2016): ESIA AMENDMENT PROCESS FOR THE PROPOSED TSUMEB SMELTER UPGRADE AND OPTIMISATION PROJECT: ENVIRONMENTAL IMPACT ASSESSMENT REPORT. APPENDIX E: GROUNDWATER AND SURFACE WATER ASSESSMENT. REPORT FOR DUNDEE PRECIOUS METALS INC. TSUMEB. REPORT NO.: 2016-WG28.

GCS ENVIRONMENTAL ENGINEERING (PTY) LIMITED (2013): NCS GROUNDWATER FLOW AND TRANSPORT MODEL. REPORT FOR NAMIBIA CUSTOM SMELTERS (PTY) LTD.

GCS ENVIRONMENTAL ENGINEERING (PTY) LIMITED (2016): TSUMEB SMELTER GROUNDWATER MODEL UPDATE. REPORT FOR DUNDEE PRECIOUS METALS – TSUMEB SMELTER.

GKW CONSULT GMBH & BICON NAMIBIA (2003): TSUMEB GROUNDWATER STUDY. MAIN REPORT. PREPARED FOR DEPARTMENT OF WATER AFFAIRS, WINDHOEK.

BITTNER, A., A. MARGANE, R. BAEUMLE, F. SCHILDKNECHT, M. MUUNDJUA, W. METZGER (2004): TECHNICAL COOPERATION PROJECT INVESTIGATION OF GROUNDWATER RESOURCES AND AIRBORNE-GEOPHYSICAL INVESTIGATION OF SELECTED MINERAL TARGETS IN NAMIBIA. GROUNDWATER INVESTIGATIONS IN THE OSHIVELO REGION. DOCUMENTATION COMPENDIUM ON THE 2004 DRILLING CAMPAIGN. PREPARED FOR DEPARTMENT OF WATER AFFAIRS, WINDHOEK.

BITTNER, A., A. MARGANE, R. BAEUMLE, F. SCHILDKNECHT, M. MUUNDJUA, W. METZGER (2005): TECHNICAL COOPERATION PROJECT INVESTIGATION OF GROUNDWATER RESOURCES AND AIRBORNE-GEOPHYSICAL INVESTIGATION OF SELECTED MINERAL TARGETS IN NAMIBIA. GROUNDWATER INVESTIGATIONS IN THE OSHIVELO REGION. MAIN HYDROGEOLOGICAL REPORT. PREPARED FOR DEPARTMENT OF WATER AFFAIRS, WINDHOEK.

KLOCK, H. (2001): HYDROGEOLOGY OF THE KALAHARI IN NORTH-EASTER NAMIBIA WITH SPECIAL EMPHASIS ON GROUNDWATER RECHARGE, FLOW MODELLING AND HYDROCHEMISTRY. DISSERTATION, JULIUS-MAXIMILIANS-UNIVERSITY OF WÜRZBURG.

DHI – WASY GMBH (2018): FEFLOW 7.0. FINITE ELEMENT SIMULATION SYSTEM FOR SUBSURFACE FLOW AND TRANSPORT PROCESSES. DHI-WASY GMBH.

DEPARTMENT OF WATER AFFAIRS (1991): THE WATER ACT (ACT 54 OF 1956), “THE NAMIBIAN NATIONAL WATER QUALITY STANDARDS”, DEPARTMENT OF WATER AFFAIRS, MINISTRY OF AGRICULTURE, WATER AND RURAL DEVELOPMENT, GOVERNMENT OF THE REPUBLIC OF NAMIBIA.

MENDELSOHN, J., A. JARVIS, C. ROBERTS & T. ROBERTS (2002) : “ATLAS OF NAMIBIA: A PORTRAIT OF THE LAND AND ITS PEOPLE". DAVID PHILIP PUBLISHERS, CAPE TOWN, RSA.

BARNETT B., L.R. TOWNLEY, V. POST, R.E. EVANS, R.J. HUNT, L. PEETERS, S. RICHARDSON, A.D. WERNER, A. KNAPTON & A. BORONKAY (2012): AUSTRALIAN GROUNDWATER MODELLING GUIDELINES. PREPARED BY SINCLAIR KNIGHT MERZ AND NATIONAL CENTRE FOR GROUNDWATER RESEARCH AND TRAINING FOR AUSTRALIAN GOVERNMENT - NATIONAL WATER COMMISION.

REILLY, T. E., A.W. HARBAUGH (1996): GUIDELINES FOR EVALUATING GROUND-WATER FLOW MODELS. OFFICE OF GROUND WATER, USGS. TECHNICAL MEMORANDUM NO. 96.04

ROBERTSON GEOCONSULTANTS INC. & SRK CONSULTING (CANADA) INC. (2012): GUIDELINES FOR GROUNDWATER MODELLING TO ASSESS IMPACTS OF PROPOSED NATURAL RESOURCE DEVELOPMENT ACTIVITIES. PREPARED FOR WATER PROTECTION & SUSTAINABILITY BRANCH, MINISTRY OF ENVIRONMENT, BRITISH COLUMBIA/CANADA.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF RECLAMATION (1995): GROUND WATER MANUAL. A GUIDE FOR INVESTIGATION, DEVELOPMENT, AND MANAGEMENT OF GROUND-WATER RESOURCES.

BARDENHAGEN (2007): PLATVELD AQUIFER STUDY. CONCEPTUAL GROUNDWATER MODELLING REPORT. DEPARTMENT OF WATER AFFAIRS. S&P.

http://www.cgiar-csi.org/2010/03/108/uot;http:/srtm.csi.cgiar.org



SLR Project No.: 733.04040.00010


KINZELBACH ET AL. (1995): GRUNDWASSERMODELLIERUNG; EINE EINFUEHRUNG MIT UEBUNGEN; GEBRUEDER BORNTRAEGER, BERLIN, STUTTGART, 238 PP.



SLR Project No.: 733.04040.00010


APPENDIX A: OVERVIEW MAP (MAP SCALE 1:75,000)



SLR Project No.: 733.04040.00010


APPENDIX B: GROUNDWATER CONTOURS, STEADY-STATE (MAP SCALE 1:75,000)



SLR Project No.: 733.04040.00010


APPENDIX C: PARTICLE TRACKING RESULTS (MAP SCALE 1:50,000)



SLR Project No.: 733.04040.00010


APPENDIX D: POTENTIAL PLUME AFTER 10YEARS (MAP SCALE 1:25,000)



SLR Project No.: 733.04040.00010


APPENDIX E: POTENTIAL PLUME AFTER 25YEARS (MAP SCALE 1:25,000)



SLR Project No.: 733.04040.00010


APPENDIX F: POTENTIAL PLUME AFTER 100YEARS (MAP SCALE 1:25,000)



SLR Project No.: 733.04040.00010


APPENDIX G: POTENTIAL PLUME AFTER 200YEARS (MAP SCALE 1:25,000)



SLR Project No.: 733.04040.00010


APPENDIX H: OBSERVED WATER LEVELS VS. COMPUTED HEADS (STATISTICS)



SLR Project No.: 733.04040.00010


APPENDIX I: IMPACT OF POTENTIAL POLLUTION

AFRICAN OFFICES

South Africa

CAPE TOWN

T: +27 21 461 1118

FOURWAYS

T: +27 11 467 0945

SOMERSET WEST

T: +27 21 851 3348

Namibia

WINDHOEK

T: + 264 61 231 287

SWAKOPMUND

T: + 264 64 402 317

Documents

Dundee Precious Metals - Tsumeb Smelter: 3D Groundwater Flow … · 2019. 7. 2. · non-reactive transport modelling does not consider adsorption, precipitation and retardation processes,