Dr Duarte da Silva on environmental arbitrage: The asset in the flood

Beneath the Witwatersrand, the abandoned gold mines are filling with water — a rising, acid, metal-laden flood that the state spends billions to neutralise and pour away. Above the same ground, one of the world's great cities is running dry, its goldfield towns queuing at boreholes. The water below is treated as a liability. The water above is a desperate shortage. They are, to within a treatment step, the same water. This is a study of environmental arbitrage — the mispricing that lets a thing be a cost and an asset at once — of the obligations reneged that produced the flood, and of the legislation that keeps the asset locked underground while a city stands thirsty over it..By Dr Duarte F da Silva*.Beneath the goldfields of Gauteng, the flooded mine voids of the Witwatersrand discharge acid water by the tens of millions of litres a day, and the state pours billions of rand into neutralising it and sending it down the rivers. Across the same province, Africa's richest city imposes rolling water cuts, its West Rand towns enduring outages of days and weeks at a time, some for the better part of a year, the national supply rationed to the litre. A liability that costs a fortune to discard. A scarcity that costs a fortune to import. And almost no one asking the obvious question: what if the first were the answer to the second? This study calls that gap by its name — environmental arbitrage — and then asks why, decades in, it remains unclaimed..EXHIBIT A · THE MISPRICINGI. Environmental arbitrage, definedAn arbitrage is the act of standing between two prices for the same thing and pocketing the difference. The classic kind lives in markets: a share quoted cheaper in one city than another, a currency mispriced across two screens, value waiting in the gap for whoever notices. Environmental arbitrage is the same move applied to a different ledger. It is the gap between what a thing is worth as a liability and what the identical thing is worth as an asset — a gap opened not by geography or by time, but by accounting and by law. A waste stream carried on the books as a cost to be discarded is, after a defined act of treatment, a product with a market. The arbitrage is the distance between those two valuations, and the discipline is the work of standing in it.Mine water on the Witwatersrand is the cleanest example of environmental arbitrage in the country, and perhaps anywhere. On one side of the ledger it is acid mine drainage — a pollutant the state is legally bound to contain, neutralise and dispose of — the Central and Eastern Basin plants alone rated to lime a hundred and seventy-eight million litres a day, on a long-term bill the state put at ten to twelve billion rand in 2016 — an open-ended charge with no offsetting revenue, a liability that grows with every millimetre the water rises. On the other side of the same ledger, on a highveld that has never had enough water, whose bulk utility is licensed to draw sixteen hundred million cubic metres a year and draws more, that same flow is a scarce raw resource: pump it, treat it to standard, and you have manufactured potable or industrial water in the precise place that is short of it. Nothing physical separates the two valuations. The cubic metre is identical. What separates them is a treatment step, a market, and a thicket of law — and the law is the part this study will argue is the true lock.Why does the gap persist, if it is real? Because liabilities and assets are recorded by different people, in different columns, under different statutes, answering to different departments — and no one is paid to notice that they describe the same water. The polluter wants the cheapest lawful discharge. The state wants the decant contained before it surfaces. The water utility wants more supply from somewhere far away. Each acts rationally inside its own column, and the arbitrage sits unclaimed between them, because claiming it requires standing across all three columns at once, which the law currently forbids. The gap is not an engineering fact. It is an institutional one. And institutional gaps, unlike geological ones, can be closed by decision.The principle generalises far beyond water, which is why it deserves a name. Wherever law or accounting records something as a pure cost — a waste, an effluent, a residue, a closure obligation — there is the possibility that the same thing, treated and re-categorised, is a resource someone needs. Tailings dumps re-mined for the gold left in them; landfill gas burned for power; captured carbon sold; slag turned to cement. Each is the same move: the liability and the asset were always the same material, divided only by a process step and a permission. Acid mine drainage on the Witwatersrand is the starkest case because the liability is so large, the need above it so acute, and the gap between them so purely legal. But the lens is general, and once you have it, you see unclaimed arbitrage everywhere a ledger says cost and a market says want.Nothing physical separates the liability from the asset. The cubic metre is identical. What separates them is a treatment step, a market, and a thicket of law — and the law is the true lock.EXHIBIT B · THE FLOODII. What is rising beneath the WitwatersrandTo understand the asset you must first understand the flood, because the flood is the feedstock. For more than a century the gold mines of the Witwatersrand were kept dry by pumping. Water seeps constantly into deep workings, and for as long as the mines ran, that water was lifted to surface and the levels held down. When the mines closed — basin by basin, as the reef gave out — the pumps stopped. And the water that had been held at bay for a hundred years began, quietly and inexorably, to rise. The abandoned workings of the Witwatersrand are not empty. They are filling, and have been for decades, with a rising underground sea.The water does not stay clean as it rises. The reef and its surrounding rock are rich in pyrite — iron sulphide, fool's gold — and when the rising water and the air in the old workings reach it, the pyrite oxidises and the reaction turns the water to dilute sulphuric acid. That acid then dissolves whatever it touches: iron, manganese, aluminium, and, in the uranium-bearing Witwatersrand, radioactive uranium and its decay products. What rises is therefore not water in any ordinary sense. It is an acidic, metal-loaded, faintly radioactive brine, with a pH measured at about 2 in the Tweelopiespruit below the Western Basin decant and between 2.2 and 2.6 in Robinson Lake, which it drains into — sour enough to corrode steel and kill a stream. This is acid mine drainage, and on the Witwatersrand it gathers in three great underground compartments: the Western Basin below Krugersdorp and Randfontein, the Central Basin below Johannesburg itself, and the Eastern Basin below Springs and Nigel.The Western Basin went first, and it went as a warning. In late August 2002, acid water began flowing out of an abandoned shaft on the old Randfontein workings near Krugersdorp — the first decant, the moment the rising sea reached the surface and spilled. It has not stopped since. The Western Basin has discharged fifteen to twenty million litres a day for years — about a hundred two-litre bottles every second, pouring out of the ground — surging toward sixty million in the wet season. It ran into a stream called the Tweelopiespruit, which it killed; into Robinson Lake, which it poisoned so thoroughly — uranium at sixteen milligrams a litre, forty thousand times the natural background — that the National Nuclear Regulator declared it a radiation area; through the Krugersdorp Game Reserve, where the dams turned rust-orange with iron sludge, the fish died and the lions sickened; and on toward the dolomitic compartments of the Cradle of Humankind, the World Heritage fossil site, and the Sterkfontein caves. The acid attacks the dolomite itself, threatening the stability of the very rock beneath a global heritage landscape.The Central Basin is the largest and the most dangerous, because what sits above it is Johannesburg. After mining stopped, the water there rose at 0.59 of a metre a day, year on year from 2009 — relentless, measurable, a tide with a timetable. Engineers fixed on a number they called the Environmental Critical Level: the depth below surface beneath which the water must be held, because above it the acid would reach the near-surface aquifers, the building foundations, the basements and tunnels of the central city. By 2010 the Central Basin water stood about five hundred metres below surface and was climbing toward that line at better than two hundred metres a year. The Eastern Basin, below Springs and the old Grootvlei mine, told the same story toward the Blesbokspruit, a Ramsar wetland of international standing. Three basins, three rising seas, one chemistry — and a clock running on all of them.How much water is down there? No one knows precisely, because the hydrology of a century of interconnected workings is fiendishly complex — but the scale is vast. The three basins together underlie an area of hundreds of square kilometres, and the voids left by a hundred years of the most intensive deep mining in human history hold water measured, by most estimates, not in millions but in hundreds of millions of cubic metres, recharged continually by rainfall and ingress. This is not a puddle to be pumped out once and forgotten. It is a permanent, self-replenishing underground reservoir — which is precisely what makes it, treated, a permanent water source rather than a one-off clean-up.And the flood does more than poison rivers; it undermines the very ground. Much of the Far West Rand sits on dolomite, a rock that water dissolves into caverns and, when the water table is disturbed, into sinkholes — the catastrophic, sudden collapses that have swallowed homes and machinery on these goldfields before, and that had Carletonville and Khutsong declared a provincial disaster area in August 2025. Acid water moving through dolomite accelerates the process, eating new voids beneath the surface. So the rising mine water threatens three things at once: the rivers it poisons, the heritage rock it destabilises beneath the Cradle of Humankind, and the literal stability of the land that towns are built on. A liability, left to rise, does not stay still. It corrodes outward in every direction.The pumps stopped, and the water that had been held back for a century began to rise. What rises is not water in any ordinary sense — it is acid, metal and uranium, a sour underground tide with a timetable.EXHIBIT C · THE DECANT THAT WAS LET HAPPENIII. Predicted, ignored, and allowed to flowThe flood was not a surprise. It was forecast. Geohydrologists warned through the 1990s — most explicitly in the Water Research Commission's 1995 report on the flooding of the Central and East Rand mines — that once the pumping stopped the Witwatersrand voids would fill and decant poor-quality water to surface, and that it would be acid when it did. The mechanism was understood, the location was roughly known, and the timing was estimable. The warnings were on the record years before 2002. And still, when the water surfaced from the Randfontein shaft, those responsible behaved as though it could not have been foreseen — scrambling to build a containment dam and a power supply after the fact, pumping the first acid into an already-polluted lake, converting an old plant into a treatment works while the stream below them died.Then came the part that turns negligence from a failure of foresight into a failure of will. The Western Basin decanted continuously from 2002, and for years the official response to a flow of up to sixty million litres a day of acid water was to tip lime into the stream — in-stream neutralisation, a chemical fig leaf, treatment in name only. The one plant on the basin could handle around twelve million litres a day against wet-season peaks five times that, and holding the basin below its critical level needed twenty-seven. The water went on entering the environment part-treated or untreated, year after year, through a game reserve and toward a World Heritage Site, while the agencies argued over who would pay. A known, predicted, continuous discharge of poison into a river system was, in practice, managed by hoping it would not get much worse.What finally moved the state to act at scale was not the dead Tweelopiespruit. It was not the rust-coated animals, or the uranium in the borehole water that informal communities were drinking because they had nothing else. What concentrated official minds was the Central Basin — because the Central Basin's rising water threatened the foundations and basements of central Johannesburg and high-value property, and the prospect of acid welling up beneath the economic heart of the country produced, at last, an Inter-Ministerial Committee and an emergency programme. The water rose toward poverty for eight years and produced lime in a stream. It rose toward money and produced a national intervention. Remember that sequence. It tells you what the system treats as an emergency and what it treats as someone else's problem.The negligence is measured in people. Around Johannesburg lie more than two hundred and seventy tailings dumps covering some four hundred square kilometres, most of them unlined, holding billions of tonnes of mine waste laced with the same sulphides and uranium as the water — a hundred thousand tonnes of uranium by one count. The apartheid state placed informal and black settlements hard against these dumps, and the people are still there: a large share of greater Johannesburg's residents live in informal settlements, many on or beside radioactive mine residue, breathing its dust in the dry season and drawing from boreholes the acid water has reached. In parts of the West Rand, families could no longer grow vegetables because the soil itself had turned; borehole water carried uranium; a lake near Randfontein was declared too radioactive for use. This is what the untreated flood does when it surfaces where the poor live.The two halves belong in one sentence. A community on the goldfields, living beside a radioactive dump, drinking uranium-touched borehole water because the municipal taps have run dry — sitting on top of a flooded mine holding, by most estimates, hundreds of millions of cubic metres of water that, treated, would be cleaner than what the rivers now carry. The water is there. The need is there. The technology to join them exists and runs daily a province away. And between the two stands nothing but a body of law and a failure of will. The people drinking poison above the flood are not victims of scarcity. They are victims of an arbitrage no one was permitted, or willing, to claim.Into the abandoned shafts themselves came the zama-zamas — informal miners, many of them former employees of the very mines stripped and closed beneath them, who descend into the unpumped, ungoverned workings to scratch the last gold from the rock. The flooded, decaying shafts have become a lawless underworld with its own violence; bodies have been recovered from these workings, and at Stilfontein, where police sealed an abandoned Buffelsfontein shaft from August 2024 until a court-ordered rescue in January 2025, seventy-eight bodies came up and two hundred and forty-six survivors, among close to two thousand people who surfaced or were brought up over the five-month siege. The same abandonment that produced the flood produced the desperate economy that now dies in it. The reneged obligation has a body count.And it runs against the Constitution itself. South Africa's founding law guarantees everyone the right of access to sufficient water, and the right to an environment that is not harmful to health or wellbeing. A flood of acid water surfacing into the rivers and boreholes of poor communities, beside a city rationing its taps, is the negation of both at once — the environmental right denied below, the water right denied above, by two faces of a single unmanaged resource. The state's duty here is not discretionary. It is constitutional. The flood is not only an economic failure or an environmental one. It is, on the face of it, a standing breach of the two rights the country promised its citizens first.The water rose toward poverty for eight years and produced lime tipped in a stream. It rose toward money and produced a national intervention. That sequence tells you exactly what the system counts as an emergency.EXHIBIT D · THE OBLIGATIONS RENEGEDIV. Who was supposed to pay, and how they leftThere is a principle, old and sound and written into South African law, called polluter pays. The party that profits from extracting the resource bears the cost of cleaning up after it. Mining companies are required to set aside financial provision for rehabilitation and closure, to leave the ground safe and the water managed when the gold runs out. That is the deal society strikes with an industry it allows to hollow out its earth: take the gold, but make good. The flood beneath the Witwatersrand is, in the plainest terms, the physical evidence that the deal was broken — that the making-good was deferred, underfunded, shuffled off, and in many cases simply never done.The renege has a scale. The state today carries a list of roughly six thousand derelict and ownerless mines — sites whose owners have vanished, dissolved or walked away, leaving the public to inherit the hole. The cost of rehabilitating them was put at forty-nine billion rand in 2019; the state spends on the order of a hundred and twenty million a year doing it, which is a four-hundred-year backlog, not a budget. And the cruel irony underneath: the Centre for Environmental Rights has told Parliament that something like sixty billion rand sits in mine-rehabilitation funds across the country, money lodged for closures that may never come, opaque as to who holds it and what it has ever been spent on. Funds exist. The flood continues. The two facts sit side by side and do not touch.The mechanism of the renege has a shape, and the industry knows it well. A large house mines the profitable years, then sells the ageing asset — with its mounting closure liability — to a smaller company. The smaller company works what is left, then sells to a smaller one still, or restructures into a shell, each transfer moving the liability further down the chain toward an entity with less and less substance to meet it. At the bottom of the chain sits a chronically underfunded shell that fails, is liquidated, and the obligation evaporates into an insolvency — leaving the rehabilitation undone and the public holding it. Move the value up the chain and the liability down it until the liability lands on something too weak to be made to pay.Blyvooruitzicht is the case that should be taught in every law school in the country. Once among the richest gold mines in the world, it was abandoned in August 2013, fourteen years ahead of the 2027 closure its own environmental plan had projected, through a collapsed sale and a liquidation. The rehabilitation fund left behind was thirty-five million rand, against a true clean-up cost that its former owner conceded was a multiple of that. Some six thousand people in seven hundred households were left living beside its dust and its tailings. And the pumping that had kept the local water in check was so imperilled that a neighbouring major, AngloGold Ashanti, obtained a court order to take over pumping two of the shafts itself — not out of duty to the public, but to stop the rising water from flooding its own adjacent assets — and in March 2014 the electricity to the mine's last operating shaft was cut off over unpaid power, and it flooded. When a profitable neighbour must go to court for permission to pump a derelict mine's water to protect itself, you are looking at a system in which the obligation to manage the flood has not so much failed as been formally abandoned. Aurora and Mintails and others fill out the same grim catalogue. The flood is not an act of God. It is an act of accounting.If the reneged obligation has a single face, it is the Grootvlei mine at Springs, on the East Rand goldfields, and the company called Aurora — and it is where reneging crossed into the criminal. When the previous owner, Pamodzi Gold, was liquidated in 2009, the operations at Grootvlei and at Orkney were handed to Aurora Empowerment Systems, a company whose directors carried two of the most resonant surnames in the country: Khulubuse Zuma, nephew of the then-president, and Zondwa Mandela, grandson of Nelson Mandela, with the president's own attorney named among the directors for a time, a designation he disputed. The promise was a six-hundred-million-rand investment, job security for thousands, and a future for tired mines. What followed was the opposite of every word of it.Aurora mined and sold gold — some hundred and twenty million rand of it by the union Solidarity's count, a hundred and ten on the liquidators' side — while paying not a single creditor, the proceeds untraceable. It stopped paying the workers; more than five thousand of them went unpaid, some reduced to surviving on charity food parcels. And then it stripped the mines, selling off equipment that did not belong to it — including, fatally, the pumping infrastructure that held the water down. Six newly installed heavy-duty pumps that had kept the Eastern Basin in check were sold for scrap. The department's own officials warned, in real time, that if the pumping at Grootvlei stopped the basin would flood and the acid would decant, and laid criminal charges in May 2010. Two shafts were permanently flooded by January 2010; the last two pumps were taken out in February 2011; the shafts were wrecked, in all likelihood never to work again.That sequence deserves its plain name. A flood this study has called an act of accounting was, at Grootvlei, very nearly an act of intent: the dewatering infrastructure was not merely neglected but physically sold for scrap, by operators who had been warned of exactly what would follow, while five thousand workers went unpaid and the gold revenue vanished. The courts have since said as much in their own register. On 25 June 2015 the Pretoria High Court found the Aurora directors personally liable for some 1.7 billion rand in stripped and destroyed assets, and on 12 May 2016 the Supreme Court of Appeal dismissed their challenge, the trial court having found wilful deception and reckless management. The prosecuting authority has pursued charges against them under the National Water Act itself — unlawful water use, failure to comply with a licence, polluting a water resource, detrimentally affecting it — the criminal provisions of the water law turned against the people who let the basin flood.More than fifteen years on, the Aurora mineworkers have still not been made whole; the water-law prosecution reached the Springs magistrate's court in 2019 and was struck off the roll in February 2020 because the state was not ready; and the flood the looting unleashed beneath the East Rand is now part of the public liability the state pumps and neutralises at the taxpayer's expense. The private actors took the gold and the scrap value of the pumps. The public inherited the flood, the workers inherited the ruin, the rivers inherited the acid. That is the polluter-pays principle turned precisely upside down: the polluter was paid, and everyone else has paid ever since.Aurora is the lurid extreme, but it is not an outlier in kind, only in degree. The same structure — value extracted, liability shed, pumping allowed to fail, the public left holding the flood — recurs across the goldfields in quieter, legal, un-prosecuted forms. It is enabled by a financial-provision regime that was allowed to rot: the official guideline for calculating how much a company must set aside for rehabilitation was written in 2005 and went years without the annual revision it was meant to receive, while the rand lost more than half its value against the dollar, so that the sums held fell ever further short of the true cost. Under-provision was not an accident. It was the predictable result of a guideline no one updated, in a system where care and maintenance became, as the Centre for Environmental Rights has documented, a way of postponing rehabilitation indefinitely.Move the value up the chain and the liability down it, until the liability lands on something too weak to be made to pay. The flood beneath the Witwatersrand is not an act of God. It is an act of accounting.EXHIBIT E · THE COSTINGSV. What it costs to throw the water awayWhen the state finally acted, it acted through the Trans-Caledon Tunnel Authority, directed on 6 April 2011 to install pumping and treatment in all three basins and hold each below its Environmental Critical Level. The engineering is real and the plants are large. The chosen method is high-density sludge treatment: lime is added to neutralise the acid, the dissolved metals and uranium precipitate out as sludge, and the sludge is thickened and disposed of. The Western Basin plant at Krugersdorp was upgraded toward fifty million litres a day; the Central Basin plant at Germiston has run at eighty-two million litres a day since December 2014; and the Eastern Basin plant at Springs, commissioned in June 2016, is among the largest of its kind in the world, built for about a billion rand with a capacity of a hundred and eight million litres a day. Across the basins, the short- and long-term bill for managing the Witwatersrand flood has been put at ten to twelve billion rand, to be shared — in principle — between the state, the mining sector and water users.One detail turns a cost story into an arbitrage story. High-density sludge treatment neutralises. It lifts the pH, strips the metals and the uranium, and renders the water no longer acutely poisonous. What it does not do is remove the salts. The treated water that leaves these plants remains high in sulphate and total dissolved solids — saline, hard, unfit to drink and marginal even for many industrial uses. It is treated to a discharge standard, not to a resource standard. So the billions are spent to take acid water and turn it into salty water, which is then released into the rivers and, in the end, runs down to the sea. The state has built, at vast expense, an elaborate machine for converting one kind of unusable water into another kind of unusable water — and stopping precisely one treatment step short of the asset.That last step — desalination, the removal of the sulphate and the salts to bring the water up to a standard a person can drink or a factory can use — is the expensive one, the technically demanding one, and the one the emergency programme was never scoped to take. It needs membranes, reverse osmosis, energy, and a way to handle the brine and gypsum the process throws off. It roughly doubles the difficulty. And so the official logic has been: neutralise to protect the environment, discharge to be rid of it, and treat the full clean-up to a usable standard as a luxury the emergency cannot afford. Which would be a defensible position — if the same province were not, at that very moment, paying to import water from across an international border because it did not have enough. The cost of the last step looks very different when you set it not against zero, but against the price of the water the city is buying instead.There is a second cost the neutralise-and-discharge approach buries rather than removes: the sludge. High-density sludge treatment does not make the metals and uranium disappear; it concentrates them into a thickened residue that must itself be disposed of somewhere, in perpetuity — a radioactive, metal-rich waste that is its own long-tail liability. And the desalination step that would make the water usable throws off its own stream: a concentrated brine and a mass of gypsum to be managed. None of these is a reason not to treat the water to standard. They are reasons to design the full chain deliberately — to recover the gypsum and the metals as saleable by-products rather than as new dumps, which the better plants already do. The waste of the treatment is itself a candidate for the same arbitrage: liability, correctly handled, becomes product.Against the right comparison, the case turns over. The instinctive objection to full treatment is that desalinating mine water is expensive per litre — and so it is, against the fiction that the alternative is free. The alternative is not free. The province is spending billions to neutralise this water and discard it, and billions more to import replacement water under a mountain range from Lesotho — a scheme whose second phase was costed at eight billion rand at feasibility in 2008, now stands at fifty-three billion, and will not deliver water before 2028. The true comparison for a litre of treated mine water is not zero. It is the all-in cost of the Lesotho litre it would displace, plus the avoided cost of neutralising-and-dumping the mine litre, plus the avoided environmental liability. Counted that way — against what we already spend to throw it away and to buy it back — the expensive last step is not a luxury. It is a saving.The state built, at the cost of billions, an elaborate machine for converting acid water into salty water and pouring it down the rivers — stopping precisely one treatment step short of the asset.EXHIBIT F · THE CITY THAT STANDS THIRSTYVI. The shortage above the floodAbove the underground sea stands a city. Johannesburg was built where there was gold, not where there was water — one of the few great cities on earth founded on neither a river nor a coast, on a high inland plateau that has never had enough. For a century it has imported its water: pumped uphill from distant rivers, and carried under the Drakensberg from the dams of the Lesotho Highlands, one of the largest cross-border water transfer schemes ever built. The engineering that keeps Johannesburg wet is a marvel. And it is failing.Through 2024, 2025 and into 2026 the taps of Gauteng ran dry in rolling waves. Rand Water, the bulk utility, is licensed by the national department to draw sixteen hundred million cubic metres a year from the Integrated Vaal River System, holds a temporary permit to eighteen hundred — topped up by a further two hundred for four months from February 2026 — and has been abstracting above even that — because the three metros routinely demand more than the system is allowed to give; compliance would require Johannesburg alone to cut four hundred and forty-five million litres a day. Reservoirs cannot recover; power cuts knock out the pump stations; pipes burst by the thousand each month in a network starved of maintenance. Forty-two per cent of the water supplied in Gauteng is non-revenue water, and forty-seven per cent nationally — lost to leaks, given away, or stolen — so that for every few litres pumped, one is lost through a broken main and another is never paid for. And the great supposed rescue, the second phase of the Lesotho Highlands scheme, is running nine years late, its additional water not due before 2028. Climatologists at Wits now name a Day Zero drought as the single largest threat to the Gauteng economy.The West Rand is where the flood and the thirst become the same sentence. Krugersdorp, Randfontein, Carletonville, Merafong — this is the goldfield. It is the ground above the Western Basin, the basin that has been decanting fifteen to sixty million litres of mine water a day since 2002. And in 2025 the towns of Merafong went almost a year without reliable tap water, cut back by Rand Water over a municipal debt of R1.4 billion, residents driven to draw from rivers and stormwater drains, supply restored only between 11 and 14 December once the municipality had filed a repayment plan. Fochville organised itself into a water crisis committee and was reported in a court dispute with the municipality over deliveries to Kokosi. Communities on the West Rand goldfields, standing directly above tens of millions of litres a day of water, went thirsty for the better part of a year — because the water beneath them is in the wrong legal column and one treatment step from the wrong quality. The asset was literally flooding the ground under their dry taps.The shortage is not only a matter of volume; it is rot in the delivery. Close to half of the water is lost before anyone pays for it — leaking from unmaintained mains, given away, or stolen — a non-revenue figure that would be a scandal in any well-run system and is, in a water-scarce one, a slow catastrophe. Around the failure has grown an economy of its own: where the pipes fail, water-tanker contracts bloom — a three-year Johannesburg tanker contract of R223 million to R263 million, depending on who counted it, awarded to two politically connected start-ups, declared invalid by the High Court in December 2025 and still running in February 2026 pending appeal — while the Auditor-General reports four billion rand swept out of Johannesburg Water that exists only on paper. Money that should renew the network is diverted; the network decays further; the tankers roll. Scarcity, once it sets in, becomes a business with an interest in its own continuation.What this does to people is the part the figures miss. When the taps fail for days or weeks, water — the most basic of needs — becomes a daily ordeal of queues, buckets and uncertainty, and it falls hardest on those least able to bear it. Businesses close; clinics struggle; the social compact frays, and civic groups form in the vacuum where service should be. The President told Parliament in the 2026 State of the Nation that water is now the single most important issue for many people, from Johannesburg down to the smallest towns, and committed R156 billion over three years to it; by July a National Water Crisis Committee was sitting at the Union Buildings. And all of it unfolds on the same Witwatersrand — including the East Rand goldfields, where the Eastern Basin's water was let loose at Grootvlei — that holds, beneath its dry and angry suburbs, the underground reservoir this study has been describing. The thirst and the flood are not in different countries. They are in the same municipalities, one above the other.Communities on the West Rand goldfields went almost a year without tap water — standing directly above tens of millions of litres a day of mine water. The asset was flooding the ground beneath their dry taps.EXHIBIT G · THE PROOF IT WORKSVII. eMalahleni, and the state's own admissionEverything to this point could be dismissed as a clever way of describing a problem, were it not for one inconvenient fact: the arbitrage has already been claimed, once, and it works. In the coalfields around eMalahleni — the old Witbank, in Mpumalanga — two competing mining houses, Anglo Coal and BHP Billiton's South African coal arm, did in the 2000s exactly what this study argues for. Facing rising mine water on one side and a municipality over-drawing its only dam on the other, they built a plant that takes acid mine drainage and treats it, through neutralisation and multi-stage membranes and reverse osmosis, all the way to the national drinking-water standard, SANS 241. The eMalahleni Water Reclamation Plant was commissioned in 2007. It was the first plant in the world to take acid mine drainage to potable water at scale. It has been running for the better part of two decades — under Thungela Resources since Anglo's coal demerger in 2021 — treating up to fifty million litres a day on a site licensed for sixty, recovering better than ninety-seven per cent of what goes in, and the water it makes is sold to the water-stressed municipality and piped to people's homes, about an eighth of the city's daily need.That precedent disposes of the standard objections one by one. AMD cannot be made drinkable? It is being drunk. The chemistry is too hostile? The same sulphate-and-metals chemistry is beaten daily by a plant that has run for nineteen years. There is no market? The market was a thirsty town next door, exactly as the West Rand and the whole of Gauteng are thirsty now. The case study that described eMalahleni used a phrase this study could take as its thesis: it turned a major liability into a valuable asset. That is environmental arbitrage, built in concrete and membrane, operating in South Africa, on mine water, for going on twenty years. The thing is not hypothetical. It is plumbed in.And the state knows it. The same government that spent a decade tipping lime in streams has, in its own long-term plan for the Witwatersrand flood, conceded the entire argument. The long-term acid-mine-drainage solution, as government itself has framed it, is to treat the water fully — not merely to neutralise it — and feed it into the Vaal River System as usable supply, which would augment Gauteng's water and, in the words of the minister who launched it, defer the need for further costly Lesotho Highlands augmentation beyond Phase 2 for at least another thirty years. In other words, the authorities have formally acknowledged that the flood is a water resource, that treating it to standard would ease the very shortage the city is suffering, and that doing so would save the enormous capital cost of importing more. They have written the arbitrage into their own strategy. They simply have not built it: the desalination phase has sat on hold since about 2019 pending a review of the salt load on the Vaal, and a June 2025 reply to Parliament on the long-term programme offered neither timelines nor costs — nor, indeed, any mention of desalination. What it has put forward instead, in its August 2025 environmental programme, is a packaged reverse-osmosis unit of one million litres a day at each of the Central and Eastern Basin plants — about one per cent, between them, of the flow those plants already treat. Between the admission and the asset stands the law.The prize is not small. The Central and Eastern Basin plants are rated to pump and neutralise a hundred and seventy-eight million litres a day between them — a hundred and six at Springs, seventy-two at Germiston, on the operator's 2025 programme — before the Western Basin is counted, on installed capacity across the three plants of roughly two hundred and thirty to two hundred and forty-five; and the eMalahleni precedent shows the bulk of such a flow can be brought to potable standard at recovery rates above ninety per cent. Two hundred million litres a day is about five per cent of what the three Gauteng metros drew in early 2026, and it is water that originates inside the province, immune to the politics of a cross-border transfer and the delays of a dam nine years behind schedule. The government's own long-term plan made the same point in its own terms: full treatment of the flood could defer the next, multi-billion-rand phase of Lesotho augmentation for thirty years. The asset is not marginal. It is, on the state's own arithmetic, a strategic share of the region's water future, sitting unclaimed beneath it.It turned a major liability into a valuable asset. That is not a slogan — it is the documented result of a plant that has been treating mine water to drinking standard, and selling it to a thirsty town, since 2007.EXHIBIT H · THE LEGISLATIVE CAGEVIII. Why the asset stays lockedIf the water can be made drinkable, the market is gasping next door, and the state itself agrees the flood is a resource — then the only thing left holding the arbitrage shut is the law. And it is. South Africa's water and mining statutes were written to do many things well, but they were never written to permit a liability to be converted into a tradable asset by the party that cleans it. The framework has no category for what this study proposes. Five legal locks hold the asset underground, and each must be named precisely, because vague complaint changes nothing and a specific lock can be specifically opened.The first lock is ownership of the water itself. Under the National Water Act of 1998, water is not private property; the state holds all of it as public trustee of the nation's water resources, and what a user receives is not ownership but an entitlement to use, granted and conditioned by the state. So treated mine water is not simply the property of whoever pumped and cleaned it, to be sold at will. The right to take it, the right to use it, and the right to trade it are not one right but several, each requiring the state's grant. A would-be arbitrageur does not own the very thing whose value they would unlock — and that uncertainty alone is enough to deter the capital a treatment plant requires.The second lock is the water-use licence, and its slowness. Section 21 of the Act lists eleven distinct water uses — taking water, storing it, discharging waste, disposing of water containing waste and removing water found underground among them — and each must be authorised, by licence or, where the thresholds allow, by general authorisation. A scheme that abstracts flooded mine water, treats it, and supplies the result is not one licensed activity but a stack of them, assembled from a regime that is discretionary, chronically slow, and built on an either-or assumption: it pictures the applicant as either a polluter to be permitted a discharge, or a user to be allocated a supply. It has no clean template for an entity that is converting the first into the second. The licensing burden that would be trivial for a conventional waterworks becomes, for a liability-to-asset converter, a years-long thicket — and years of regulatory uncertainty kill the financing of capital-heavy plant.The third lock is the gap between two standards. The quality a discharger must meet to release water safely into a river is one regulatory universe; the quality required to supply water for drinking or industry is another, governed by different instruments and different approvals. The emergency plants were permitted, scoped and funded to hit the discharge standard — neutralise and release. To cross from there to a resource standard is not merely more treatment; it is a move into a different licensing world, with its own approvals to win. The law draws a hard line exactly where the value is, and treats stepping over that line as starting a new and separate regulatory journey rather than completing the one already underway.The fourth lock is the unsettled ownership of the liability — and therefore of the right to profit from solving it. Who, in law, owns the Witwatersrand flood? For derelict and ownerless mines the answer is contested: the original polluters have in many cases dissolved, and the relevant mining legislation does not automatically vest the residual liability in the state simply because a mine is derelict. So the question of who is responsible for the water is unresolved — which means the question of who may legitimately treat it and sell it is also unresolved. An investor cannot build a plant on a liability whose legal owner is undetermined, because the entitlement to the cleaned-up product flows from a responsibility no one has been firmly assigned. Unclear liability is unclear title, and unclear title is unbankable.The fifth lock is the route to market. Bulk water supply in Gauteng runs through what is in practice a state-anchored monopoly: Rand Water abstracts from the Vaal system, treats, and sells on to the municipalities, all within volumes the national department sets. A private producer of reclaimed mine water cannot simply sell into the city; it must sell into or alongside that regulated bulk system, on terms and at a tariff the system controls. Without a defined right to feed treated mine water into the bulk supply at a fair, bankable price, the arbitrageur has manufactured an asset with no permitted buyer. The market that is visibly, painfully short of water is screened off from the producer by the very architecture meant to deliver it.Five locks, not one. The water is not yours to sell; the licence is a thicket; the standard is a cliff; the liability has no owner; the buyer is a monopoly. Each is a clause — and a clause can be rewritten.EXHIBIT I · THE UNLOCKIX. Five locks, five keysName the locks precisely and the keys announce themselves, because every one of these barriers is a legal artefact, not a law of nature — a clause that was written and can be rewritten. The case for doing so is not charity and not green sentiment. It is that the same set of reforms would attack, together, three national crises that are currently being fought separately and expensively: the acid-mine-drainage liability, the Gauteng water shortage, and the dead capital sitting idle in rehabilitation funds. One change to the architecture; three problems addressed; and, done correctly, at no net cost to the fiscus, because the asset pays for the remediation.For the first lock, create the category the law is missing: a defined legal status for remediated mine water as a resource that may be abstracted, treated and supplied under a single, durable entitlement — so that the party who cleans the flood holds a clear, tradable right to the water it produces, conditioned on quality and on the state's custodianship, but bankable. For the second, a combined, fast-tracked licence that bundles the abstraction, treatment and supply of reclaimed mine water into one authorisation with statutory timeframes, ending the years-long stack. For the third, an explicit bridge between the discharge and resource regimes, so that treating beyond the discharge standard toward a usable standard is encouraged and pre-approved within the same licence, rather than treated as a fresh regulatory journey.For the fourth lock, resolve the liability question deliberately: where a mine is derelict and ownerless, let the state, as the party that has in practice inherited the flood, grant a clear right to treat-and-supply to whoever will take on the remediation — converting an unowned liability into an assigned, fundable mandate. And here the sixty billion rand — the last published tally, from 2017 — changes character. Rehabilitation financial provisions, and the residual closure funds, should be permitted to fund treatment to a resource standard, not merely neutralisation — turning money that currently sits frozen against closures that never come into the capital that builds the plants. For the fifth lock, a defined right of feed-in: an obligation on the bulk system to take reclaimed mine water that meets standard, at a fair and published tariff, so that the manufactured asset has a guaranteed, bankable buyer. Five locks, five keys, and the arbitrage is open.What stands on the other side of those five keys is large. The Witwatersrand flood, fully treated, is a standing source of water measured in the low hundreds of millions of litres a day, in the exact province that is rationing supply and importing the balance across an international border. Turning it on would ease the shortage, defer billions in import infrastructure, extinguish a perpetual pollution liability, fund itself out of the value of the water, and put the dead rehabilitation billions to work — all without taking a litre from anyone or building a single new dam. The reforms are specific, the precedent is built and running, and the only thing manufacturing the scarcity is the gap between the statutes. The flood is not the disaster. The disaster is that we are paying to throw it away while we stand in line for the same water clean.The funding should follow the fault. The money to build these plants should come, first, from the provisions the polluters already lodged — the rehabilitation funds that sit frozen against closures that never come — and from the assurances the law already requires of operating mines. The public should not pay twice: once for the gold that was taken, and again to clean the flood that was left. Where the original polluter survives, it should pay; where it has vanished into liquidation, the inherited provisions and a modest levy on the water the asset produces should carry the cost, so that the flood funds its own remediation and the citizen funds neither. The arbitrage, properly designed, is not a call on the fiscus. It is a way to make a century-old debt finally pay itself off.None of this need be attempted everywhere at once, and the right first step is obvious: build where the proof and the need already coincide. The West Rand is the natural pilot. It has the longest-running decant, the Western Basin, discharging continuously since 2002; it has the most acute and visible thirst, in the very towns that went a year without taps; and it has, a province away in the coalfields, a proven template in the eMalahleni plant that can be studied, adapted and transplanted. A single demonstration scheme on the West Rand — abstract the Western Basin flood, treat it through to a resource standard under a combined licence, feed it into the local supply at a published tariff, fund it from the rehabilitation provisions and a levy on the water sold — would prove the whole arbitrage on the ground that needs it most. Prove it once, visibly, and the model travels to the Central and Eastern Basins and beyond. The first plant is not a leap of faith. It is a copy of one already running, placed where the dry taps are.And the moment for it is now, because several clocks have struck at once. The price and scarcity of water in Gauteng have climbed to the point where treated mine water is, for the first time, competitive on any reasonable estimate with the alternatives the province is paying for — a fifty-three-billion-rand dam and a tanker economy. The acid-mine-drainage liability has matured from a future worry into a present, decanting fact the state is already spending billions to manage. The rehabilitation provisions have swollen into tens of billions of rand of capital sitting idle. The technology has been de-risked by nineteen years of operation at eMalahleni. And the Lesotho lifeline is nine years late, leaving a supply gap that has to be filled from somewhere. Each of these on its own is a reason; together they are a convergence. The arbitrage has been available in theory for twenty years. It has become practically irresistible only now — which is exactly the kind of moment that is missed, because it has been coming for so long that everyone stopped expecting it to arrive.There is a harder political truth in the keys, and it should not be dodged. To open the arbitrage is also to confront the reneging head-on — to insist that the provisions be adequate and accessible, that the polluters who survive are made to pay, that derelict liability is assigned rather than left to drift, and that the funds lodged against closure are spent on the flood instead of guarded against a closure that never comes. The same reforms that unlock the asset also, unavoidably, close the escape routes that produced the flood in the first place. That is not a flaw in the proposal. It is its deeper purpose. An arbitrage that turns the liability into an asset only works if someone is finally made responsible for the liability — and making someone responsible is the thing the system has spent a quarter of a century avoiding.One change to the architecture; three crises addressed; and, done right, at no net cost to the state — because the asset pays for its own remediation. The flood is not the disaster. The disaster is throwing it away while we queue for it clean.THE FIVE LOCKSX. The cage, and the keys, on one page.THE VERDICTXI. A city thirsty above its own reservoirTaken together, this is what it amounts to. Beneath the Witwatersrand, a flood that was predicted, then allowed, then half-treated at a cost of billions and poured down the rivers — the physical residue of a polluter-pays bargain that was reneged on, mine by abandoned mine, through a liability-shifting game that left the public holding an underground sea. Above it, a city of millions rationing water to the litre, importing it under a mountain range from another country, its goldfield towns queuing at stormwater drains for the better part of a year. And one treatment step and five legal clauses standing between the two — between the water that is a liability and the water that is an asset, which are, to the cubic metre, the same water.I have spent my life around these goldfields, and I have watched South Africa do this again and again: build something the world could not, then let it go for want of the will to keep it. The flood is that pattern written in water. We pumped this basin dry for a hundred years to reach the gold; we stopped; we let it fill; we let it surface; we spend fortunes to neutralise it and throw it away; and we stand above it, thirsty, buying the same water clean from across a border. The acid mine drainage of the Witwatersrand is not, in the end, a story about chemistry. It is a story about arbitrage left unclaimed — about a value that has been sitting in plain sight, in the gap between two columns of a ledger, while the people who live on top of it go without. The water is there. It has always been there. What is missing is the decision to call it what it is.This study has carried a single idea against its hardest case, and the case has held. Environmental arbitrage is real, and the Witwatersrand flood is its sharpest instance: a thing recorded as a pure liability that is, to the cubic metre, a scarce asset — divided from its own value by a treatment step the state already half-performs, on a hundred and seventy-eight million litres a day of rated capacity in the Central and Eastern Basins alone, and a body of law that has no category for the conversion. The flood was foreseen and allowed; the obligations to manage it were reneged, in places criminally; the proof that the water can be made drinkable has run for nineteen years; the city above is dying of thirst; and the only thing left holding the two apart is a cage of five legal locks, each with a nameable key. Everything required to close the gap exists, except the decision to close it. That is the most damning sentence available, because it is not a sentence about geology or chemistry or money. It is a sentence about will.THE FLOOD BELOW AND THE THIRST ABOVE ARE THE SAME WATER, SEPARATED BY ONE TREATMENT STEP AND FIVE CLAUSES OF LAW.WE PAY BILLIONS TO THROW THIS WATER AWAY, AND BILLIONS MORE TO IMPORT IT CLEAN. THE ARBITRAGE IS THE DIFFERENCE — AND IT IS UNCLAIMED.THE DISASTER IS NOT THE FLOOD. THE DISASTER IS A CITY STANDING THIRSTY ABOVE ITS OWN RESERVOIR..*Dr Duarte F da Silva • Capital Markets Strategist • Managing Director, Northbound Processing • Germiston, September 2026.A STANDALONE STUDY · A NOTEThis is a standalone study in the Gold Papers, written to be read on its own. It takes a single idea — environmental arbitrage, the mispricing that lets one thing be a liability and an asset at once — and tests it against the hardest, most literal case in the country: the acid water rising beneath the Witwatersrand, and the city running dry above it. The flood is documented, the negligence is documented, the thirst is documented, and the proof that mine water can be made drinkable has been running since 2007. The only thing standing between the liability and the asset is a body of law that has no category for turning one into the other. That law was written by people. It can be rewritten by people. The water is waiting..A note on the figuresNorthbound Processing • Germiston, September 2026. The Gold Papers — Paper 21, a standalone study. This article does not constitute legal, environmental or investment advice. Figures are drawn from public sources: the report to the Inter-Ministerial Committee on Acid Mine Drainage (December 2010) and subsequent Department of Water and Sanitation and Trans-Caledon Tunnel Authority material, including the TCTA/Nafasi operational environmental management programme for the Central and Eastern Basins of August 2025 (basin water levels, the Environmental Critical Level, rated treatment capacities of 72 and 106 Mℓ/day, and the packaged reverse-osmosis proposal), the TCTA corporate plan presented to Parliament on 21 April 2026 (Western Basin 35 Mℓ/day) and the department's reply to Parliament NW2519 of 17 June 2025; the Water Research Commission's 1995 report on the flooding of the Central and East Rand mines (WRC 486/1/95); peer-reviewed and technical literature on Witwatersrand AMD chemistry and high-density-sludge treatment (Mine Water and the Environment, 2024); the Federation for a Sustainable Environment's 2010 submission to Parliament (Robinson Lake, tailings inventory); reporting and research on derelict and ownerless mines and mine-closure financial provision (Centre for Environmental Rights, Full Disclosure 2018 and its briefing to Parliament of 25 October 2017; Oxpeckers and News24, May 2017, for the R57–60bn provisions tally; Mail & Guardian, 6 December 2019) and on the Blyvooruitzicht (Lawyers for Human Rights and FIDH, 2017; Bloomberg, March 2014), Aurora (Engelbrecht NO v Zuma, Gauteng Division, Pretoria, 25 June 2015; Supreme Court of Appeal, 12 May 2016; OUTA; TimesLIVE, May 2019) and Stilfontein (January 2025) cases; the eMalahleni Water Reclamation Plant project record (Anglo American / BHP Billiton, commissioned 2007; Thungela Resources regulatory audit, 2024; UNFCCC case study); and contemporary reporting on the Gauteng water crisis, Rand Water's licensed abstraction (EWN, December 2024), non-revenue water (Africa Check), Merafong (The Citizen, 15 December 2025; GroundUp), Johannesburg's tanker contracts (Daily Maverick and amaBhungane, November 2025 – February 2026), the State of the Nation Address of February 2026 and the National Water Crisis Committee of 23 July 2026, and the Lesotho Highlands Water Project (Engineering News, September 2025). The legal analysis refers to the National Water Act 36 of 1998, the Water Services Act 108 of 1997, the Mineral and Petroleum Resources Development Act 28 of 2002 and the financial-provision regulations under the National Environmental Management Act, and the environmental and water rights in sections 24 and 27 of the Constitution. Volumes, costs and liability estimates are drawn from public records and vary across sources; they are indicative and are labelled as such. The interpretation laid on these facts — including the framing of environmental arbitrage — is the author's own.

Dr Duarte da Silva on environmental arbitrage: The asset in the flood

Related Stories

He fought Swimming SA for years. Finally the world body caught up

Da Silva’s Gold Papers: Sibanye's plan to flood the richest reef on earth - at best gold price ever

Hesitant about the stronger Rand? Here's how to structure your hedging strategy — Sebastian Steyn (Sable International)

Dr Duarte da Silva: SA gold mine owners’ windfall funded departure from Reef