1. Understanding AMD from a Business Model Perspective Public Debate on AMD 29 August 2013 by Anthony Turton Water Stewardship Council Trust Image © Anthony Turton 2011

2. Emerging Trends and Issues Expect fast rollout of infrastructure in all gold mining basins now that AMD is better understood at technical/engineering level – Who Pays for What? Two Business Models underpinning AMD mitigation – Legacy Model (simple but costly) – PPP Model (complex cost/benefit matrix) Two Process Models are implied but latent – Centralized Model (Water Boards providing potable quality water to ALL users at a flat tariff) – Dual Steam Model (off-take agreements for non-potable AMD treated water to industrial users at different tariffs)

3. Gold Production Life Cycle Work for the GDARD MRA Reclamation Project shows the existence of three peaks in gold. Source: Hartnady ‘s specialist report in GDARD MRA Reclamation Project First peak was shallow gold driven by surface resources that were easily accessible. Third peak is about the recovery of gold from tailings dams driven by metallurgical engineering. The Legacy Model follows this trajectory because it assumes that mining has ended The PPP Model follows this trajectory because it assumes that the business case for mining can be reinvented Second peak was deep gold driven by ventilation engineering and rock mechanics.

4. Legacy Model AMD is the result of irresponsible mining AMD is generated mostly in the void AMD will be a perpetual problem – dumps remain Void remains open forever – Zama Zama/crime issue Uranium remains a hazard – dust fallout Service provider will be a contracted third party Taxpayer will foot the entire bill forever Rehabilitation is not part of the process Mine companies inevitably become insolvent Water has to be treated to potable standards Economics favours Centralized Process Model Known OPEX cost to taxpayer = ≥R 12.00 m³ (neutralized water only) (best available data – might change)

5. PPP Model AMD can be mitigated by responsible Closure Mining AMD is mostly generated on surface AMD is temporary until dumps are rehabilitated Void can be closed out with emerging best practice Uranium can be sequestered (possible subsidy?) Service provider is restructured mining companies Revenues from Closure Mining will subsidize rehabilitation Rehabilitation drives the process – land is the end product Mine companies become rehabilitation companies Water can be treated to industrial standards Either Centralized or Dual Stream Model Known OPEX cost to taxpayer = ±R 4.00 m³ (neutralized water only) (best available data – might change)

6. Centralized Process Model State owns all water resources State has the sole right to treat and distribute it Energy analogy = Eskom All water treated to one standard (potable) One price based on average national cost Treated AMD will become drinking water Reconciliation of demand/supply at national level Monopolistic so price insensitive Consistent with current laws and policy

7. Dual Stream Process Model State is custodian of all water resources Partnerships for treatment and distribution Energy analogy = Independent Power Producers Water treated to different standards Varied price based on local treatment cost and end-users requirement AMD does not have to become drinking water Reconciliation of demand/supply at local level Non-monopolistic so price sensitive Will need some legal / policy reform

8. Centralized Process Model Dual Stream Process Model Legacy Business Model PPP Business Model Mining is the problem State to contract third party service provider (E-Tolls Model) Taxpayer to pay in perpetuity Landscape not rehabilitated AMD becomes drinking water HDS Storage facilities needed Cost = ≥R12.00 m³ (neutralized) Closure Mining is the solution Mines are service provider Mining revenues cover majority of cost (partnership) Landscape is rehabilitated AMD does not become drinking water HDS Storage facilities not needed Cost = ±R4.00 m³ (neutralized) Economics do not stack up for this combination Process can support this combination HDS Storage facilities not necessarily needed depending on process selected Economics supports this combination Process supports this combination HDS Storage facilities not needed dependimg on process selected

9. Rehabilitated Land is the result of the PPP Model Source: Tang & Watkins (2012) The Big Picture Post-Mining

10. Flow Pathway “A” Genesis of AMD Flow Pathway “B” Gatekeeper Flow Pathway “C” Aquatic Ecosystems Flow Pathway “D” Underground Fallout of Uraniferous dust Rainfall Decant Ingress © AR Turton 2013 Streamflow Ingress Wind AMD is complex so we need carefully designed interventions at specific parts of the overall cycle Misdiagnosis of the cause-effect relationships will cause massive investment into the wrong intervention and an angry taxpaying public over time

11. Flow Pathway “A” Mine Dump Rainfall Dust Wind Evaporation © AR Turton 2013 Flow Pathway “A” Genesis of acidification Scale - Localized in the dump Mechanism - acid rain falls on alkaline tailings generating hydrogen Acid leaches Uranium Wind blows dust over a larger geographic area Remedy: Policy on coal combustion Rehabilitate to prevent ingress and dust Remove dump if possible Uranium sequestration © AR Turton 2013

12. Flow Pathway “B” Gateway to alternative pathways Scale – Quaternary catchment Mechanism – mostly hydraulic Either AMD seeps from the base of the dump into the void Or AMD flows off the dump into adjacent wetlands Remedy: Rehabilitate to prevent slumping and erosion Remove dump if possible Uranium sequestration Flow Pathway “B” Gatekeeper Dump Surface Underground © AR Turton 2013

13. Flow Pathway “C” Aquatic Ecosystems Scale - Catchment and beyond Mechanism – hydraulic mobilization of dust deposition and dump erosion by rainfall BUT also precipitation in rivers Yellow Boy as manifestation Metals accumulate in wetlands Remedy: Rehabilitate dumps to prevent wind and water erosion Wetland rehabilitation to sequester metals and trap sediment Ingress control into void Neutralize decant as temporary measure Uranium sequestration Flow Pathway “C” Aquatic Ecosystems Dust Rainfall Sediment from Dumps Streamflow © AR Turton 2013

14. Flow Pathway “D” Underground Scale – Mining basin and multiple catchments Mechanism – Ingress from surface hydraulic flows Acid mobilizes heavy metals and oxidises pyrite in pillars Bacteria (Thiobacillus feroxidans) convert Pyrite to additional acid Remedy: Ingress control from surface Pillar removal Closure of void by paste backfill Drawdown of water level Neutralize as temporary measure to prevent Uranium mobilization © AR Turton 2013 Ingress from Dump Flow Pathway “D” Underground Ingress from rivers Decant to rivers

15. Image © Anthony Turton 2013 Thank You

16. Tailings Management AMD Management Rainfall as a Variable Deposited at high pH (>10) Solutions: (1) legal reform of waste classification, (2) develop MRA policy, (3) remove/consolidate dumps Lowers pH and triggers release of Hydrogen – genesis of acidification Acidifying (pH <3) Solutions: (1) neutralize pH in void, (2) close ingress points, (3) remove pillars, (4) close void with tailings, (5) rehabilitate surface © AR Turton 2013 Falls on dumps Attacks pillars mobilizing metals and pyrite Decants at lowest shaft Becomes AMD in riversLeaches in situ Uranium Concentrates in the crust Becomes dust Dumps classified as “spoil” Eroded by water Ingress into the void via multiple points Falls on land Solutions: energy/water/food policy reform - WEF Nexus

17. Rainfall pH < 3 Ingress into void Tailings pH ˃ 10 Land © AR Turton 2013 Wind erosion mobilizes dust Fallout over terrestrial ecosystems Acid leaches in situ uranium Uranium concentrates in crust Hydraulic flows into wetlands AMD flows in rivers Acid leaches pyrite and metals from pillars Decants at lowest shaft Flow Pathway “A” Flow Pathway “B” Wind mobilization Flow Pathway “C” Flow Pathway “D” Source: SAWEF Paradigm Shifter