The snow line nobody prices

In April 2026, ICIMOD published its annual snow update for the Hindu Kush Himalaya. Snow persistence across the winter of November 2025 to March 2026 came in at 27.8% below the long-term average — the fourth consecutive below-normal year, and a deeper deficit than the record of 23.6% set the year before. Ten of twelve basins were below normal.

The finding was reported widely and correctly, as an environmental story. It was not reported as what it also is: a revision to next season's cash flow for every generator, irrigator and agricultural lender operating downstream of a snow-fed catchment.

That gap between how the Himalayan cryosphere is measured and how it is valued is the subject of this analysis. The measurement is world-class. The valuation does not exist.

Two different clocks

Public discussion tends to merge glaciers and snow into a single narrative of melting mountains. Financially they are opposites.

Glacier mass is a stock. It depletes over decades, and it matters to anyone holding a thirty-year concession, a long-tenor loan, or a sovereign development plan. ICIMOD's 2026 glacier assessment maps 63,761 glaciers across roughly 55,782 square kilometres, with about 78% of that ice area concentrated in the 4,500 to 6,000 metre band — the elevation most exposed to warming. A system that concentrated is not a diversified portfolio.

Seasonal snow is a flow. It arrives and departs within a year, contributing roughly 23% of annual runoff regionally, and it determines whether a plant meets its lean-season generation target or a farmer gets a spring irrigation. It is a working capital variable.

Treating the two as one produces analysis useful to nobody. Anyone holding a twenty-year asset needs the stock series. Anyone financing a crop cycle needs the flow series. The 2026 data illustrates why the distinction is not academic: while the regional aggregate hit a record low, the Ganges basin recorded snow persistence 16.3% above normal. A regional headline would have erased that. So would a blended index.

The exposure is already known. The mapping is not.

TERI's assessment places roughly 52% of India's hydropower generation and 33% of its thermal generation on rivers originating in the Himalaya. That figure circulates widely.

What does not exist anywhere in public data is the next step: which capacity sits in which basin, and how that basin is doing. Installed capacity is published by state. Cryosphere condition is published by basin. Nobody has joined the two tables.

The consequence is that the single most important physical variable governing a majority of Indian hydropower cannot be looked up against the assets it governs. Not because it is secret — because nobody has done the join.

Sikkim, October 2023

The reason this matters is not hypothetical.

On 3 October 2023, a permafrost landslide fell into South Lhonak Lake at roughly 5,200 metres in Sikkim. The moraine dam breached. The resulting flood travelled some 385 kilometres down the Teesta and into Bangladesh, eroding an estimated 270 million cubic metres of sediment along the way, according to reconstruction published in Science.

It destroyed the 60-metre dam of the 1,200 MW Teesta-III project — Sikkim's largest hydropower asset, held by a state government enterprise. Downstream, sluice gates at Teesta-V were damaged. Casualty figures differ across sources; Science records 55 dead and 74 missing, while contemporaneous reporting cited approximately 40 dead and 76 missing. Around 88,400 people were directly affected, with 16 roads and 33 bridges damaged.

Within months, the state had agreed to sell its majority stake in the project to its minority partner.

Three things about that sequence deserve more attention than they have received. A single cryosphere event destroyed a gigawatt-scale asset outright rather than degrading it. The loss landed on a sub-sovereign balance sheet and forced a change of ownership. And environmental review for projects in that basin had considered outburst-flood risk in submissions without translating that consideration into design or risk transfer.

None of that is a warning about the future. It is a description of something that has already happened, once, in a region with many comparable exposures — and nobody can currently say how many, or where.

Not a Himalayan problem

It would be a mistake to frame this as regional exceptionalism.

Published global assessment finds roughly 15 million people exposed to potential glacial lake outburst flood impacts worldwide. High Mountain Asia is the most exposed region, with about a million people living within ten kilometres of a glacial lake. The population at direct risk has risen by around 3.2 million — some 27% — since 2000. The same research flags the Andes as carrying broadly comparable potential impact with substantially less published study behind it.

This is a global asset class with identifiable peers. The Alps have decades of instrumented hydrological finance. The Andes are under-analysed. High Mountain Asia is data-rich and analysis-poor.

That last category is where institutions get built.

What we propose to build

HEF is constructing a Basin Meltwater Exposure Index: a basin-resolved register that sets cryosphere condition against the generation capacity, irrigated area and downstream population sitting beneath it, alongside the institutional capacity to absorb a loss.

Three design commitments, stated here so they can be held against us.

The index will report at basin and sub-basin level and will never publish a single regional score, because a single regional score would have hidden the 2026 Ganges result.

Stock and flow will be reported separately and never blended, because they answer different questions for different balance sheets.

Where a figure cannot be sourced, the index will say so rather than estimate it. Insurance penetration on Himalayan generation assets, for example, does not appear to be published anywhere in consolidated form. We will attempt to obtain it; if we cannot, the gap will be published as a gap. The shape of what is missing is itself a finding.

Why this is an argument, not an accusation

The 16th Finance Commission's report, submitted to Parliament in February 2026, assigns 10% weight to forest and ecology in the horizontal devolution formula, and revised that criterion to include open forests and to reward growth in forest cover. It also set a disaster management corpus of ₹2,04,401 crore, with 90:10 centre–state cost sharing for Himalayan and north-eastern states.

Himalayan states have argued for years that snowfields and cold-desert areas above the tree line perform an ecological and economic function that forest cover does not capture — that these are the surfaces which generate the water.

At present that argument is made rhetorically. There is no index to point at.

Building one is not a critique of anybody's policy. It is the construction of an evidence base that a finance ministry can carry into a room — domestically before the next Finance Commission, and internationally wherever adaptation finance is negotiated. An independently maintained index, built from published inputs with a transparent method, is a stronger instrument than the same claim advanced by the claimant.

The measurement already exists. What is missing is the arithmetic that turns it into money. That is a solvable problem, and it is the one we intend to solve.