The Shrinking Source: Why Pakistan’s Survival Depends on the Melting Glaciers of Ladakh

Every drop of the Indus that reaches a Pakistani canal began its journey somewhere high, cold and far from our fields. A large share of it starts in the glaciers and snowfields of Ladakh and the Zanskar, Karakoram and Himalayan ranges, and the satellite record now shows that this frozen reservoir is shrinking, slowing and destabilizing faster than most planners in Islamabad, Lahore or Karachi have priced in.

Pakistan has built its economy on a single river system. The Indus basin irrigates the wheat, rice, cotton and sugarcane that feed and clothe a country of more than 240 million people, and it turns the turbines that keep the lights on. Studies put the number of people across the basin who depend on its flows at around 286 million. That dependence is not a footnote to our food security; it is the foundation of it. If the glaciers that regulate the river are failing, then the stability of Pakistan’s farms, cities and power grid is failing with them, even if the water still looks plentiful at Tarbela this November.

This is the argument of this piece: the evidence of long-term glacier loss in Ladakh is no longer ambiguous, the early symptoms are already visible in the way the ice behaves, and Pakistan needs to treat the upper Indus as a national security issue, not as a distant scientific curiosity.

Ladakh sits in a cold, dry rain shadow, yet it has not been spared. Temperature records show an average rise of about 0.02°C a year over the long run, with a sharper climb of roughly 0.21°C to 0.22°C per decade since the late twentieth century. That is in the same range as, or faster than, recent global average warming. Local drivers add to the problem: vehicle emissions from a booming tourism economy and black carbon deposited on snow darken the surface, lower its reflectivity and speed melt.

The glacier-by-glacier record, drawn from satellite imagery and field observation, tells a consistent story across the region.

In the Karzok Range, an inventory of 276 glaciers shows total glacier area fell 25.6% between 1980 and 2023, from 169.4 km² to 125.4 km², with small, low-elevation and steep glaciers the most vulnerable. At the Parkachik Glacier in the Suru valley, satellite records show the annual melt rate from 1999 to 2021 was about six times the 1971–1999 baseline. Across 87 glaciers around Pangong Lake, surface area retreated 6.7% between 1990 and 2019. In the Drass basin, a study of 150 glaciers from 1965 to 2013 recorded widespread marginal losses, with larger glaciers such as Machoi, observed since 1875, carrying the heaviest long-term volume loss.

Two features of these findings deserve attention. First, the loss in the Karzok Range, more than a quarter of glacier area in about four decades, is not a marginal trend. It is the disappearance of a quarter of a water store. Second, the Parkachik result shows acceleration, not just steady decline: melting did not merely continue after 1999, it ran at roughly six times the earlier pace. Acceleration is what turns a manageable problem into a planning emergency.

The newest findings, finalized this year, describe a second symptom that is easy to miss: the ice is physically slowing down. Glaciers in the Zanskar Himalaya moved about 16% more slowly in 2023 than in 1992. A glacier flows because its own weight pushes it downhill. As it thins, that driving stress falls, and the ice crawls. A slower glacier is a thinner, weaker glacier, and one that is further along the road to disintegration.

The thinning itself has accelerated sharply. In the Zanskar region, surface lowering rose from about 0.22 meters a year in 2000–2005 to about 0.57 meters a year in 2015–2020, more than a doubling in a decade and a half.

Meltwater has to go somewhere. Remote sensing of an upper Ladakh catchment shows one moraine-dammed lake growing from 0.03 km² to 0.08 km² between 1969 and 2014, an increase of more than 150%. Such lakes are held back by loose rock and ice, not by engineered concrete, and they are the setup for the most violent hazard the mountains can produce.

The most dangerous misunderstanding about shrinking glaciers is that they mean less water right away. They do not. When a glacier melts faster, rivers briefly run higher, because the stored ice is being spent on top of the annual snowfall. Hydrologists call the top of that curve “peak water.” Before it, the river swells. After it, the river declines, permanently, because the savings account has been drained.

The Zanskar basin is a major feeder of the upper Indus. Its long-term depletion therefore threatens the water, food and energy security of the roughly 286 million people who live downstream. Among the great Asian river basins, the Indus is widely regarded as one of the most dependent on meltwater, because its monsoon rains arrive late and its lower reaches run through an arid plain.

The cruel feature of peak water is that it rewards complacency. A decade of higher summer flows can be read as proof that nothing is wrong, and can justify irrigation expansion and new demand that the river cannot sustain once the curve turns. We should expect that debate in Pakistan, and we should resist it. Nobody can say with precision when each sub-basin will cross its peak, which is exactly why Pakistan needs better measurement now, not later.

Pakistan’s agriculture is one of the most irrigation-dependent in the world. Most of the country receives little rain, so canal water drawn from the Indus and its tributaries is what turns desert into cropland. Farming remains the largest employer in rural Pakistan and a major share of exports flows from it, directly and through textiles. Three pressures from the north converge on that system.

The calendar is shifting. Communities in Ladakh report that mountain snowpacks now melt earlier in the summer. Glacial runoff once sustained crops until late September; streams now regularly run low by mid-August. That is a local observation rather than a basin-wide measurement, but it points at a risk we should take seriously downstream: water arriving earlier and in sharper pulses, then thinning out just when late-season crops such as cotton, rice and the sowing of rabi wheat need it.

The decline will be uneven. A shrinking, earlier-peaking flow does not fall evenly across the basin. Tail-end canal commands in Sindh, which already receive the least reliable supply, will feel a shortfall first and hardest. A system that is already contested between provinces cannot absorb a structural decline in supply without conflict.

The floods will get worse. Glacial lake outburst floods are the sudden face of this slow crisis. In Ladakh, the Nidder flood of 2010 and the Gya flood of 2014 destroyed bridges, houses and generations-old agricultural terraces. Those were small lakes in a sparsely populated valley. A comparable failure higher in the Indus system, in a valley with towns, roads and hydropower, would do far greater damage. Pakistan learned in 2022, when floods submerged a third of the country, how little margin its riverine economy has when the water arrives wrongly.

The ice is in Ladakh, and Ladakh is administered by India. Sovereignty over Jammu and Kashmir, of which Ladakh was part until 2019, remains disputed between India and Pakistan, and nothing in this piece settles that question. But international law does not need it settled to assign responsibility. In its 1971 advisory opinion on Namibia, the International Court of Justice held that effective control of a territory, not the legitimacy of title to it, is what gives rise to a state’s obligations toward others for what happens there. A state that controls the headwaters of a river answers for how it treats them.

Customary international law is the first source of that responsibility, and it binds India whether or not it has signed a given convention. States must not allow activities under their control to cause significant harm across borders, must use shared rivers equitably and reasonably, and must notify and consult downstream neighbours about projects and emergencies. The ICJ confirmed the duty to assess transboundary environmental impact in the Pulp Mills case in 2010. For Ladakh this translates into concrete duties: transparent environmental assessments for dams and roads on the Indus, Suru and Zanskar; timely flood and glacial-lake warnings to Pakistan; and continued exchange of hydrological data.

India is a party to the UN Framework Convention on Climate Change, whose objective is to stabilize greenhouse gas concentrations at a level that prevents dangerous interference with the climate system, and which commits parties to systematic observation, research and cooperation on adaptation. It is a party to the Paris Agreement, which sets the goal of holding warming well below 2°C while pursuing 1.5°C, requires each party to submit and progressively strengthen national targets, and obliges parties to cooperate on adaptation and on averting, minimizing and addressing loss and damage. India’s updated 2022 target is to cut the emissions intensity of its economy by 45% from 2005 levels by 2030, and it has pledged net zero by 2070. For glaciers the temperature goal is what counts, since global modelling finds that ice loss climbs steeply with every additional tenth of a degree.

The International Court of Justice’s 2025 advisory opinion on climate change gives these duties force. It found that states have obligations under customary law to prevent significant harm to the climate system, that the required standard of due diligence is stringent, and that discretion over national climate targets is limited by it. Serious failures can engage state responsibility.

Here is the rewritten text integrated into a smooth, running paragraph: To stabilize the regional climate, Pakistan and India could jointly manage Ladakh by upholding five concrete obligations derived from international environmental principles. First, both nations must cut greenhouse gas emissions and black carbon by pursuing targets consistent with the 1.5°C goal while strictly controlling local soot sources like diesel traffic, biomass burning, and brick kilns that darken the snowpack. Second, they must assess and limit damaging development by conducting transparent environmental impact assessments for infrastructure in glacier zones and honoring Ramsar commitments for vulnerable high-altitude wetlands such as Tso Moriri and Tso Kar. Third, they should systematically observe and share data by monitoring glaciers and exchanging real-time information with downstream communities. Fourth, both countries must warn and prepare for climate hazards by mapping dangerous glacial lakes, setting up early-warning systems, and promptly notifying downstream areas of imminent floods under customary duties of notification and the Sendai Framework. Finally, they must keep the headwaters flowing by respecting the 1960 Indus Waters Treaty, ensuring the western rivers flow freely into Pakistan while strictly limiting upstream infrastructure uses on the Indus, Suru, and Zanskar rivers.

India’s thermal power grid remains heavily reliant on coal, which accounts for approximately 223 to 229 gigawatts—roughly 41 to 43 percent—of the country’s total installed electricity generation capacity of 542 to 555 gigawatts. Because coal facilities operate continuously as base-load power plants, they deliver over 70 percent of the actual electricity consumed. Consequently, their environmental footprint is disproportionately high; emissions from coal combustion generate over 900 million metric tonnes of carbon dioxide annually, contributing roughly 75 to 80 percent of power sector emissions and between 30 and 33 percent of India’s total nationwide anthropogenic greenhouse gas emissions.

These coal emissions drive Himalayan glacier melt through both atmospheric warming and direct pollutant deposition. The massive volume of carbon dioxide released traps thermal radiation in the troposphere, elevating regional temperatures and extending summer melt seasons across High Mountain Asia. Concurrently, coal combustion releases fine aerosolized black carbon and fly ash, which are transported by atmospheric currents from the Indo-Gangetic Plains onto high-altitude snow and ice. When black carbon settles on glacier surfaces, it sharply reduces their albedo, causing the darkened ice to absorb more solar energy and accelerate seasonal ablation. In turn, this accelerated retreat destabilizes seasonal hydrological cycles, increasing short-term risks of glacial lake outburst floods while threatening long-term dry-season water security for major river basins downstream.

The high-density Indian military presence along strategic sectors in Ladakh, such as the Siachen Glacier and Pangong Tso, creates severe environmental stress on high-altitude glacial ecosystems. While global climate change is the main driver of ice loss, military operations accelerate melting through several localized factors. Constant emissions from diesel generators, military transport, and aircraft deposit black carbon and soot on ice surfaces, lowering their reflectivity and increasing solar heat absorption. Decades of deployment have also accumulated vast amounts of non-biodegradable waste, fuel drums, and refuse that alter local thermal dynamics and contaminate vital downstream water sources. Additionally, heavy infrastructure development—including road construction, blasting, and helipads— destabilizes permafrost and generates localized heat, further hastening the degradation of these critical water towers.

What can the United Nations do? Less than Pakistan would like, but more than it is doing. The UN is not a party to the Indus Waters Treaty and has no police power over a river. Its value lies in the tools it holds, and Islamabad should use each of them.

The most useful is the International Court of Justice. The General Assembly can ask the Court, under Article 96 of the Charter, for the legal duties of a state that controls the headwaters of a shared river, including its duties on climate and glacier protection. An opinion is not binding, but the Court’s 2025 climate opinion shows how much weight its reasoning carries in negotiations, domestic courts and diplomacy. A General Assembly majority, not a Security Council veto, would decide whether to ask.

The Secretary-General can offer good offices and mediation support, quietly and without a mandate. The Security Council, where Kashmir has been on the agenda since 1948 and where UNMOGIP still monitors the ceasefire line, can be told under Article 35 that a water dispute is endangering regional peace. Pakistan’s leaders have already described any move against the treaty as a threat to peace, and that is the argument to put before the Council. Realistically, a permanent member’s veto limits what the Council will do, and India has long insisted that its disputes with Pakistan are bilateral, a position rooted in the 1972 Simla Agreement. Pakistan should expect resistance on that point.

The quieter UN work may matter most. The World Meteorological Organization, UNEP and UNESCO can coordinate neutral glacier and snowpack monitoring across the Hindu Kush-Himalaya, building on the UN’s 2025 International Year of Glaciers’ Preservation and its decade of cryospheric science that follows. The UN Economic Commission for Europe’s Water Convention, now open to every UN member, offers a ready-made framework for data exchange and joint flood warning that India and Pakistan could join. The Early Warnings for All initiative can support glacial-lake monitoring and warning downstream. And climate finance channels, from the Green Climate Fund to the loss-and-damage fund, can pay for the adaptation that Pakistan’s farmers will need whatever happens to the treaty.

None of this replaces a decision in New Delhi to honour its obligations as the custodian of the headwaters. But the UN can raise the cost of ignoring it, supply the shared facts a dispute like this lacks, and give two nuclear-armed neighbours a face-saving place to talk. A river that sustains more than a quarter of a billion people deserves that much.

There is a simpler remedy than any court ruling, and it is to take the guns off the ice. Soldiers have faced each other on the heights of Siachen since 1984, and Indian and Chinese forces have confronted each other along the undefined line in eastern Ladakh, most visibly since the deadly clash at Galwan in 2020. The enemy in these places is mostly the mountain itself. Avalanches, frostbite and altitude kill far more soldiers than bullets do, and the 2012 Gayari avalanche alone buried about 140 people in the Siachen sector.

The ice pays too. Garrisons need fuel, helicopters, roads and heated shelters, and every one of them adds soot, heat and waste to a landscape whose glaciers are already thinning. Armies that contest frozen ground are helping to melt it, in a place where the water they stand on feeds the Indus and the lives of hundreds of millions of people downstream.

The proposal is to demilitarize the high-altitude zones of Ladakh, including the Siachen region, in phased, verified steps, and to turn them into a transboundary peace park under joint scientific management, with the aim of World Heritage recognition. The idea of a Siachen peace park has been floated for years by scholars and officials in both India and Pakistan. The precedents are real. The Antarctic Treaty reserved a continent for peaceful purposes and science in 1959. Svalbard has been demilitarized since 1920. The Waterton-Glacier International Peace Park joined Canada and the United States, and is a World Heritage site. After Ecuador and Peru fought over the Cordillera del Cóndor in 1995, their 1998 peace settlement created a protected binational area on that disputed ground. A park does not decide sovereignty. UNESCO’s World Heritage Convention says in Article 11 that listing a property in a disputed territory does not prejudice the rights of the parties, which is what makes it a face-saving vehicle for rivals who cannot yet agree on a line.

The international community should therefore press for it, through the UN General Assembly, UNESCO, the UN Environment Programme and the countries with real leverage in New Delhi, Islamabad and Beijing. India administers most of Ladakh and the heights of Siachen, so its decision is the pivot, and diplomatic attention should concentrate there. Nobody needs to fight over permafrost. They fight over lines, and a park can make the line matter less without erasing anyone’s claim.

The river that built Pakistan is a loan from the mountains, and the lender is now calling it in. Ladakh’s glaciers will not wait for a political consensus. The question for those of us downstream is whether we start measuring, storing, and sharing our water wisely while the Indus still runs strong, or whether we discover its limits in a bad August.