The Indus doesn’t read maps

Pakistan is debating a new constitutional order as if water will adjust afterwards. It will not.

In 1947, the Radcliffe line was drawn across Punjab while its canal system remained one machine. When the standstill arrangements expired in the spring of 1948, supplies from headworks now on the Indian side were interrupted. A temporary Inter-Dominion settlement followed in May, but the lesson was already clear: a political boundary could turn hydraulic dependence into leverage.

It took more than a decade — negotiations, World Bank good offices and international finance — to convert that emergency into the Indus Waters Treaty of 1960. The treaty broadly assigned the Ravi, Beas and Sutlej to India and the Indus, Jhelum and Chenab to Pakistan. But it did not leave Pakistan’s canal economy where it found it. Bahawalpur’s Sutlej Valley works had drawn from Sulemanki and Islam; the Lower Bari Doab Canal ran from Balloki through the Okara–Sahiwal belt toward Khanewal. After the treaty, such commands were kept alive less by the eastern rivers than by replacement works: Mangla and Tarbela, and link canals carrying western-river water into commands whose historical supplies had been allocated away.

That history should make us humble. The last time a major line was drawn across this basin before the water was settled, Pakistan needed outside mediation, replacement storage and a generation of works to recover. Today we are debating not simply new provinces but a different architecture of the federation — more units, possible federal territories, stronger local governments and a stronger executive centre — as if the river will simply adapt.

I should declare an interest. I have served three terms from South Punjab constituencies and one as Punjab’s irrigation minister. Likewise, I have spent much of my political life arguing that the south is failed by a government seated in Lahore. I have no sentimental attachment to the four-province map. But I know what it costs when a political map is drawn first and water is discussed later.

No final draft has settled the question. That is precisely why the water question must be asked now, while the geometry is still fluid.

Article 239 is often quoted as a barrier. It is better understood as leverage. A constitutional amendment requires two-thirds of the total membership of each House; a bill altering the limits of a province must also be passed by the affected provincial assembly by two-thirds of its total membership before presidential assent. But Article 239 protects boundaries, not hydrology. It can stop a unilateral federal map; it cannot, by itself, settle entitlement, rotation, telemetry, Kotri or shared headworks. Consent given before those are settled is consent given blind.

The divisional template is tempting because it asks no one to invent anything. The districts exist. The divisional offices exist. The names exist. But administrative geography and hydraulic geography are not the same thing. Districts follow revenue boundaries. Irrigation follows rivers, doabs, barrages and commands. Where those logics diverge, a river becomes a boundary; and when a river becomes a boundary, every barrage acquires a left-bank government and a right-bank government.

The reported redesign changes the question again. A stronger centre — through a stronger presidency, governors, chief ministers drawn into federal decision-making or empowered districts — may claim that it can coordinate a more fragmented federation. Perhaps. But coordination is not settlement. A centre strong enough to order a canal closure is not necessarily a centre bound by transparent entitlements, telemetry, drought rules, audit and downstream obligations. Power created before law is not a water settlement.

Stand at Sukkur and the problem becomes visible: one barrage, seven canals, right-bank commands toward Shikarpur and Larkana, left-bank commands reaching Khairpur, Nawabshah, Sanghar and Badin. Guddu already lives the interprovincial reality: standing in Sindh, feeding canals that irrigate parts of Balochistan. Trimmu feeds systems whose districts would not sit together under many circulating maps. Kotri sits low in the system while helping supply Karachi through the Kalri Baghar feeder and Keenjhar. If Karachi becomes a federal territory, the city does not leave the Indus; the federation becomes its claimant while the works that feed it remain embedded in another jurisdiction. Gwadar teaches the opposite lesson: a port can be designated for strategy, but it cannot import a river.

The post-1960 links were built because borders and water did not coincide: Chashma–Jhelum, Taunsa–Panjnad, Qad irabad–Balloki, Balloki–Sulemanki, Trim mu–Sidhnai. If a link sits in one unit and its command lies in another, the questions multiply: who operates the gate, who pays for maintenance, who controls telemetry, who bears a closure, who has standing when the other government objects? Beneath the canals lies an even less obedient system: groundwater. A province can license tube-wells; it cannot stop an aquifer from crossing the boundary.

Consider what four provinces have failed to complete. The 1991 Water Apportionment Accord recognised the need for storage; recognition has not become sufficient storage. It recognised the importance of flows below Kotri and left the quantity to further study; the delta has continued to suffer. Kachhi shows that building a canal is not the same as commanding land. The Chashma Right Bank story shows that older gravity works and a delayed lift-cum-gravity project are not the same as delivered water. These are failures among four governments. Anyone who says many more governments will coordinate better carries the burden of proof.

The deepest division is not Punjab against Sindh. It is upstream against downstream. Splitting provinces may create more upstream claimants without creating a stronger downstream voice. Sindh’s representation could rise while its coherence as the lower riparian weakens. Those are not the same thing; the second matters more. That is why those of us who want South Punjab must be honest: a new capital, secretariat, and flag are worth little if the new province inherits canals without securing the institutions, headworks and links that supply them.

When water runs short, the politics hardens. Losses are often measured as a residual: what entered, minus what left, minus what was withdrawn. Seepage, evaporation, inefficiency, and theft are bundled together because measurement and audit are weak. Nor can shortage be shared by cutting every canal by the same percentage. Below usable discharge, canals lose command and tails fail. So the system runs on rotation: one command takes water at a workable flow while another waits, then the order reverses.

Rotation requires an authority trusted to say no — repeatedly, in a drought, to governments with different mandates. Picture the first dry year after notification. Upstream units protect standing crops. Downstream units invoke entitlements. A federal city demands lifeline supply. The delta demands flow. A rotation order closes a canal that now answers to another government. The matter reaches court after sowing has been decided and lost. You cannot litigate a hydrograph. A crop has no lawyer.

None of this means the present arrangement is satisfactory. Large provinces do last-mile water badly. Watercourse maintenance, abiana collection, outlet tampering and tail-end distress can benefit from government closer to the farmer. Muzaffargarh may be failed by Lahore much as Badin is failed by Karachi. Delivery can devolve; basin allocation cannot.

So the test is not whether Pakistan reforms its federation. It is what is settled before reform becomes irreversible. No province bill, federal-territory bill or district-government scheme that changes control of an Indus command should reach a final vote until five schedules are written into law: the entitlement each unit inherits and how it is calculated; the shared headworks, barrages, links and reservoirs, and who operates and pays for them; the drought rotation rules that bind every unit; the telemetry, audit and public-data standard that ends competing numbers; and the lifeline and downstream obligations for cities, federal territories and the flow below Kotri. These schedules should be prepared through a parliamentary process with IRSA and the provincial irrigation departments, approved alongside the constitutional changes, and made enforceable on a crop-cycle timetable.

Article 239 should be used to exact that price. No assembly should give its two-thirds until the schedules are ratified; and once ratified, those schedules should not be changeable by a simple later majority. The principle is simple: no new political boundary across an Indus command without a prior water settlement for everything that boundary will divide.

The present buzz will fade. Another map will replace it. But the river will remain. I want South Punjab to have its province; I have spent much of my political life fighting for it. But 1947 taught us the price of drawing the line first and asking about water later. A province that cannot say where its water comes from is not yet a province. It is a flag waiting for a crisis.

Copyright Business Recorder, 2026