The Cryosphere as a Financial Layer

Executive summary

1. The Himalayan cryosphere is one of the most intensively monitored physical systems in Asia and one of the least intensively priced. Glacier extent, mass balance and snow persistence are published annually by credible institutions. Almost none of that output is expressed in the units used by finance ministries, regulators, lenders or insurers.

2. The 2026 baseline is materially worse than the 2025 baseline on both the stock and the flow measure. ICIMOD's HKH Glacier Outlook 2026 maps 63,761 glaciers across roughly 55,782 sq km, with approximately 78% of that glacier area sitting between 4,500 and 6,000 metres — the elevation band most exposed to elevation-dependent warming. Separately, ICIMOD's HKH Snow Update 2026 records snow persistence for November 2025–March 2026 at 27.8% below the long-term average, the fourth consecutive below-normal year and a deeper deficit than the previous record of 23.6% set in 2025.

3. These two datasets measure different things and matter to different balance sheets. Glacier mass is a long-duration buffer that affects asset life and dry-season firm capacity. Seasonal snow is a within-year flow variable that affects next season's generation, irrigation and working capital. Conflating them — as most public commentary does — produces analysis that is useless to both the ten-year investor and the one-year borrower.

4. The exposure on the Indian side is concentrated and quantifiable in principle. TERI's assessment places roughly 52% of India's hydropower generation and 33% of its thermal generation on rivers originating in the Himalaya. No public dataset maps that generation capacity against basin-level cryosphere condition.

5. The region has already recorded a total loss of a utility-scale asset to a cryosphere event. The October 2023 South Lhonak glacial lake outburst flood destroyed the 1,200 MW Teesta-III project in Sikkim. The state-owned developer subsequently agreed to sell its majority stake. This is the clearest available demonstration that cryosphere risk in the Himalaya is not a distant scenario but a realised, ownership-changing financial event.

6. The instrument HEF proposes is a Basin Meltwater Exposure Index (MEI) — a repeatable, annually refreshed composite that maps cryosphere condition against installed generation, irrigated command area, and downstream population and asset exposure at basin and sub-basin scale. Section 6 sets out the methodology; the accompanying tracker specification (HEF-TR-01) sets out the data architecture.

7. This work is constructive, not adversarial. A defensible basin exposure index is the evidence base a state or national government requires when arguing for adaptation finance, disaster-transfer weightings and concessional capital in multilateral settings. It is more useful as an argument than as an accusation.

1. The framing problem

There are two distinct literatures on the Himalayan cryosphere and they do not speak to each other.

The first is glaciological and hydrological. It is technically strong, internationally peer-reviewed and updated on a reliable cycle. It produces numbers about ice volume, snow persistence, mass balance and runoff contribution.

The second is financial and fiscal. It concerns installed capacity, plant load factors, power purchase agreements, royalty and free-power schedules, state revenue projections, disaster relief corpora and devolution formulae. It produces numbers about rupees.

Between the two sits an unpopulated space. When a hydropower concession is valued, the cryosphere enters as a hydrology annexe rather than as a risk factor with a distribution. When a state projects revenue from water-based levies, the projection is built on installed capacity and tariff, not on basin condition. When a devolution formula assigns weight to forest and ecology, snowfields and areas above the tree line have historically not been counted in the same way as forest cover — a point Himalayan states have pressed repeatedly.

The consequence is not that anyone is wrong. It is that the region's largest physical asset is absent from the region's financial arithmetic.

The purpose of this paper is narrow: to argue that the missing layer is constructible, to specify what it would contain, and to identify precisely which inputs already exist and which HEF would have to build.

2. The 2026 baseline

2.1 The stock measure: glaciers

ICIMOD's HKH Glacier Outlook 2026 is the most recent regional synthesis. Its headline parameters:

Two features of this dataset matter for financial use.

First, the elevation concentration is the risk concentration. A system with 78% of its ice in a single 1,500-metre band is not diversified. Warming that shifts the freezing level upward acts on most of the stock simultaneously rather than on a marginal fraction.

Second, the basin shares are not proportional to the economic exposure they underwrite. The Ganga and Brahmaputra basins hold a minority of the region's glaciers but carry a substantial share of downstream Indian generation, irrigation and population. Glacier count is not a proxy for economic weight, and any index that treats it as one will mislead.

2.2 The flow measure: seasonal snow

ICIMOD's HKH Snow Update 2026 covers a different variable on a different clock. Snow persistence — the fraction of time snow remains on the ground after snowfall — is measured across the November–March window.

The Ganges basin result is the analytically interesting one. In the year the regional aggregate hit a record low, the Ganges basin recorded a positive anomaly. A regional headline number would have concealed that entirely.

This is the single strongest argument for basin-level rather than regional reporting, and it is the design principle behind the index proposed in Section 6. It also carries a warning: a positive single-year anomaly in one basin is not a trend reversal. The prior year saw the Ganga basin at a 23-year low. High interannual variance around a declining mean is precisely the signature that makes seasonal snow a working-capital risk rather than a slow structural one.

2.3 Why the two measures must be kept separate

An index that averages the two produces a number that answers nobody's question.

3. Transmission channels

The cryosphere reaches financial statements through four identifiable channels. Each is measurable. None is currently measured as a channel.

Channel 1 — Generation assets

Himalaya-origin rivers underwrite a majority of India's hydropower and a substantial minority of its thermal generation, per TERI's assessment (approximately 52% and 33% respectively). Cryosphere condition affects these assets through three mechanisms: total annual water availability, seasonal distribution of that availability, and catastrophic physical risk (Section 4).

The financially relevant variable is not annual mean flow. It is firm capacity in the lean season — the period when snow and ice melt, rather than monsoon rainfall, dominate runoff. That is exactly the period the snow persistence deficit hits first.

Channel 2 — Revenue instruments

States have attempted to monetise the resource directly. Himachal Pradesh's Water Cess on Hydropower Generation Act, 2023 was struck down by the Himachal Pradesh High Court as beyond state legislative competence; the Supreme Court subsequently stayed the refund direction. The state had projected annual revenue in the order of ₹2,000 crore from over 170 hydel projects. Uttarakhand, Jammu & Kashmir and Sikkim have pursued comparable levies, over Union Power Ministry objections.

The relevance here is not the legal merits. It is that a revenue line of that scale, contingent on an unresolved constitutional question, sits on top of a physical resource whose future yield is not modelled. That is a compound uncertainty and it is not disclosed as one.

Channel 3 — Fiscal transfers

The 16th Finance Commission's report, submitted to Parliament on 1 February 2026, sets the transfer architecture for 2026–31. Relevant parameters:

• Vertical devolution retained at 41% of the divisible pool.

• Horizontal formula: income distance 42.5%, population (2011) 17.5%, demographic performance 10%, area 10%, forest and ecology 10%, and a new 10% weight for contribution to GDP.

• The forest criterion was reworked to include open forests, with density-based weighting, and to reward increases in forest area between 2015 and 2023.

• Disaster management corpus of ₹2,04,401 crore for state disaster funds, with 90:10 centre–state cost sharing for north-eastern and Himalayan states; the Centre's share is ₹1,55,916 crore.

Himalayan states had argued for snowfields, alpine pasture and cold-desert areas above the tree line to be recognised within the ecology weighting on the grounds that these are the region's water-generating surfaces.

The analytical point for HEF is that the ecology weighting is a proxy variable and the cryosphere is the underlying asset it is proxying for. A defensible index of cryosphere contribution is the missing evidence for that argument — and it will still be missing when the 17th Finance Commission convenes unless someone builds it in the interim.

Channel 4 — Insurance, credit and contingent liability

This is the least-documented channel and probably the largest. It comprises project insurance on hydro and transmission assets, sovereign and sub-sovereign contingent liability where the developer is a state enterprise, agricultural credit exposure in snow-fed irrigation commands, and the fiscal cost of post-event reconstruction.

Public data on insured versus uninsured exposure to cryosphere hazard in the Indian Himalaya is, as far as this paper's review has established, not published in any consolidated form. This is stated as a gap, not as a finding.

4. Case study: Teesta-III

The October 2023 South Lhonak event is the reference case because it is the only Himalayan cryosphere event in the recent record that produced a documented, utility-scale, ownership-changing financial loss.

Sequence. On 3 October 2023 a permafrost landslide entered South Lhonak Lake at approximately 5,200 m in Sikkim, breaching the moraine dam and initiating a glacial lake outburst flood. The resulting flow travelled some 385 km down the Teesta and into Bangladesh, and was compounded by concurrent heavy rainfall.

Physical scale. Published reconstruction in Science estimates roughly 270 million cubic metres of sediment eroded. Hydrodynamic modelling of the event (IIT Ropar / IIT Roorkee) reconstructs peak discharge of approximately 7,355 m³/s at Chungthang and approximately 8,282 m³/s at Gazoldoba.

Asset loss. The 60-metre-high dam of the 1,200 MW Teesta-III project — Sikkim's largest hydropower asset, developed by a state government enterprise — was destroyed. Sluice gates at the downstream Teesta-V station were also damaged.

Human and infrastructure impact. Casualty figures vary across sources: Science reports 55 deaths and 74 missing in Sikkim; earlier contemporaneous reporting cited approximately 40 dead and 76 missing. Around 88,400 people were reported directly affected, with two government buildings, 16 roadways and 33 bridges damaged.

Financial consequence. The state subsequently agreed to sell its majority stake in the project to its minority partner, subject to competition-regulator approval.

Why this case matters analytically. Three things are established by it that were previously treated as modelled possibilities:

1. A single cryosphere event can destroy a gigawatt-scale asset outright rather than degrade it.

2. The loss propagates to sovereign and sub-sovereign balance sheets where the developer is a state entity, and can force a change of ownership.

3. Environmental review processes for projects in the affected basin had considered GLOF risk in submissions and did not translate that consideration into a design or risk-transfer response.

The event is therefore not usable as a tail-risk anecdote. It is a base-rate observation, and the correct question is how many comparable exposures exist and where.

That question is currently unanswerable from public data. That is the gap this paper proposes to close.

5. The global comparison

Framing the Himalaya as a special case weakens the analysis. The correct frame is that the Himalaya is the largest and most exposed instance of a global class.

Published global assessment finds approximately 15 million people worldwide exposed to potential glacial lake outburst flood impacts, with High Mountain Asia the most exposed region and roughly one million people living within 10 km of a glacial lake. The exposed population at direct risk has risen by about 3.2 million (27%) since 2000. The same literature flags the Andes as a region of broadly comparable potential impact but substantially less published research.

This comparison does real work for HEF:

• It converts the subject from a regional grievance into a global asset class with identifiable peers.

• It gives the index external validity — an MEI methodology that works for the Ganga should be portable to the Rímac or the Rhône.

• It positions HEF against a genuine gap. The Alps have decades of instrumented hydrological finance; the Andes are under-analysed; High Mountain Asia is data-rich and analysis-poor. The third of those is where an institution can be founded.

6. Proposed instrument: the Basin Meltwater Exposure Index

6.1 Design principles

1. Basin and sub-basin resolution, never regional aggregate. The 2026 Ganges anomaly proves the aggregate misleads.

2. Stock and flow reported separately, never blended into one score.

3. Exposure weighted by economic mass, not by ice mass.

4. Every input traceable to a named, dated, publicly citable source. No proprietary black box. The credibility of the instrument is its methodology, not its cleverness.

5. Annual refresh on a fixed calendar, aligned to the publication cycle of the underlying cryosphere datasets.

6. Published gaps. Where a component cannot be sourced, the index reports it as unavailable rather than estimating it.

6.2 Structure

The MEI is a composite of three sub-indices, reported separately and never summed into a single headline figure:

A. Cryosphere Condition Sub-index (CC)

• Glacier area and change, by basin

• Share of glacier area in the 4,500–6,000 m band

• Snow persistence anomaly, current season and rolling multi-year

• Glacial lake count and area change, where published

B. Economic Mass Sub-index (EM)

• Installed hydropower capacity, by basin (operational, under construction, allotted)

• Thermal generation dependent on basin water

• Irrigated command area

• Population and built asset value in modelled runout and floodplain zones

C. Institutional Absorption Sub-index (IA)

• Disaster fund allocation and cost-sharing terms applicable to the state

• Insurance penetration on generation and transmission assets, where disclosable

• Early warning system coverage in the basin

• Ownership structure of exposed generation assets (state, central, private) — determining where a loss lands

Reported output: a matrix, not a ranking. Each basin receives three scores and a stated data-completeness percentage.

6.3 What exists and what HEF must build

The two bolded construction items are the moat. Everything else is aggregation of published material, which is useful but replicable. Capacity-to-basin mapping and asset-value-in-runout-zone are the components that would make the MEI citable and difficult to duplicate.

7. Policy implications

Stated in the register in which they are most useful to the institutions that would act on them.

For finance ministries and Finance Commissions. The ecology weighting in the horizontal devolution formula is currently proxied by forest cover. The underlying claim advanced by Himalayan states is that water-generating surfaces above the tree line have ecological and economic value that forest cover does not capture. That claim is presently argued rhetorically. An MEI would let it be argued numerically, which is a materially stronger position ahead of the 17th Finance Commission.

For power sector planners and regulators. Long-tenor hydropower concessions are underwritten by hydrological assumptions with a stated design life. Where those assumptions predate the recent snow persistence record, they warrant revision. A basin-level condition series makes that revision routine rather than exceptional.

For disaster management authorities. The 90:10 cost-sharing pattern for Himalayan states recognises differential cost disability. It does not currently differentiate within the Himalayan states by basin-level hazard exposure. An MEI would permit that differentiation.

For India's external negotiating position. Claims for adaptation finance are strongest when they rest on an independently constructed, methodologically transparent evidence base rather than on advocacy. An index maintained by an independent Indian institution, using internationally published inputs, is a more durable negotiating asset than the same argument made by a claimant.

8. Limitations, conflicts and disclosures

Source hierarchy. This paper draws its primary cryosphere parameters from ICIMOD, its energy-dependence parameter from TERI, its GLOF reconstruction from peer-reviewed journal literature, and its fiscal parameters from the 16th Finance Commission report and PRS Legislative Research. Secondary press reporting has been used only where it is the sole available source for a policy decision, and is identified as such.

Known conflicts in the record.

• Casualty figures for the October 2023 Sikkim event differ across sources (55 dead / 74 missing per Science; approximately 40 dead / 76 missing in contemporaneous reporting). Both are reported above; neither is adopted as authoritative.

• Basin nomenclature differs between ICIMOD's twelve-basin framework and Indian administrative basin definitions. Any MEI build must publish a concordance table before publishing values.

Not addressed in this paper. Transboundary flow allocation, upstream infrastructure in jurisdictions outside the HEF scope, and attribution of specific events to specific emissions pathways. These are excluded deliberately; the paper's argument does not depend on them and their inclusion would substitute a contested debate for a constructible one.

Data currency. All cryosphere figures are as of the 2026 publication cycle. Fiscal figures reflect the 16th Finance Commission report as submitted 1 February 2026. Readers should verify currency before citation.