When the Mountain Fell on the Powerhouse: Chamoli 2021 and the Unfinished Protection of Himalayan Hydropower

B.A.P-I Retrospective, 2020 to 2024. This paper looks back at 2021 and the lesson it holds for the protection of India's critical infrastructure.

1. Introduction

On the morning of 7 February 2021, a mass of rock and glacier ice broke away from the north face of Ronti Peak in the upper catchment of the Rishiganga in Chamoli district, Uttarakhand, and within minutes the valleys below had been converted into a corridor of destruction. The functional 13.2 MW Rishiganga small hydro project was washed away, and the 520 MW Tapovan project of NTPC on the Dhauliganga, then under construction, was struck by the flood that followed (MHA, 2021, Feb. 9). More than 200 people were killed or went missing, and the great majority of them were workers at the two project sites (Shugar et al., 2021, Jun. 10; PSI, 2021, Jun. 10).

Five years on, Chamoli is remembered in public discussion largely as a "glacier burst", a phrase that the scientific investigation did not support and that tends to obscure the lesson. The event was a hazard cascade: a slope failure at altitude that turned into a highly mobile debris flow and then met infrastructure placed squarely in its path. This paper revisits what the investigations found and what has changed since, and argues that the siting, warning and protection of Himalayan hydropower are matters of national security rather than of project engineering alone.

2. What the Investigations Found

The most complete reconstruction came from an international team of fifty-three scientists, Indian and foreign, who combined satellite imagery, seismic records, numerical modelling and eyewitness video. Their paper in Science concluded that roughly 27 million cubic metres of rock and glacier ice had collapsed from Ronti Peak, that the avalanche transformed into a very large and mobile debris flow, and that this flow carried boulders more than 20 metres across and scoured valley walls up to 220 metres above the valley floor (Shugar et al., 2021, Jun. 10). The mass fell from about 5,500 metres to 1,800 metres, at speeds the team estimated at between 205 and 216 kilometres per hour, and about one fifth of it was glacier ice (Temblor, 2021, Jul. 16).

Three findings deserve particular attention for so far as protection is concerned. First, the failure was not without precursors. Satellite images from 2015 to 2021 showed a crack on Ronti Peak that had begun to widen by 2018, and earlier ice masses had fallen from the same ridge in recent years and struck the valley just above the Rishiganga facility (Temblor, 2021, Jul. 16; PSI, 2021, Jun. 10). Second, the event was visible to instruments as it happened: distant seismic stations recorded both the detachment and the impact on the valley floor 58 seconds later (Temblor, 2021, Jul. 16). Third, the two hydropower plants lay 15 and 26 kilometres from the source, and eyewitness videos showed how little warning the people in the path of the flow received (Temblor, 2021, Jul. 16).

The cause, in other words, was not a lake that burst. It was a mountainside that had been visibly preparing to fail, in a valley with a recorded history of mass movement, above two plants for which no site-specific warning of a slope failure then existed. One of the authors observed that the scale of the event "certainly argues for avoiding development in these areas" (PSI, 2021, Jun. 10).

3. The State Response and the Human Cost

The national machinery for response worked broadly as designed. The Home Minister's statement to Parliament recorded that the National Crisis Management Committee met under the Cabinet Secretary at 1630 hours on the day itself, that 450 personnel of the Indo-Tibetan Border Police, five National Disaster Response Force teams, eight Army teams, an Indian Navy diving team and five Indian Air Force helicopters had been deployed (MHA, 2021, Feb. 9). Some 25 to 35 people were believed trapped in one NTPC tunnel. Of the 197 people then missing, 139 were from the NTPC project, 46 from the Rishiganga project and 12 were villagers (MHA, 2021, Feb. 9).

The search inside the Tapovan tunnel, choked with silt and debris, became the defining image of the weeks that followed. By May 2021 the official toll stood at 83 dead, with 121 people still missing (Deccan Herald, 2021, May 14). What comes through from all of this is a response capacity that mobilised at national scale within hours; the reform agenda lies earlier in the chain, in moving workers off the valley floor before such a flow arrives.

4. The Project and Its Recovery

Tapovan-Vishnugad, a 4 x 130 MW project of NTPC, had originally been scheduled for commissioning in 2012-13 against an original cost of Rs 2,978.48 crore; by September 2021 the Central Electricity Authority recorded an anticipated commissioning in 2024-25 and a latest cost of Rs 5,867.38 crore, and identified the flash flood of 7 February 2021 as the event that had held up construction (CEA, 2021, Sep. 30). The official response to the damage was methodical. Expert teams from the Central Water Commission, the Central Electricity Authority, the Geological Survey of India, the Central Soil and Materials Research Station and NTPC were constituted to assess the damage and plan restoration, restoration work began on their recommendations, and the Authority recorded that a warning system had been made operational at the site (CEA, 2021, Sep. 30).

That combination of multi-agency assessment and on-site warning is the right model, and India has the institutional depth to make what was done at Tapovan after the event the standard for every high Himalayan project before it.

5. Siting, Warning and Protection after Chamoli

In hindsight, three things have moved. The first is a sharper national position on siting in the upper Ganga basin. In May 2026 the Ministry of Environment, Forest and Climate Change told the Supreme Court that no new hydropower projects should be permitted in the Alaknanda and Bhagirathi basins beyond seven projects, among them Tapovan-Vishnugad, then reported at 75 per cent construction; a committee constituted by the Court had rejected seven further projects as lying in the path of glacial lake outburst floods (Mishra, 2026, May 21). The second is a funded programme for glacial hazards. The National Glacial Lake Outburst Flood Risk Mitigation Project carries an outlay of Rs 150 crore across Arunachal Pradesh, Himachal Pradesh, Sikkim and Uttarakhand, with a component for monitoring and early warning; the Wadia Institute of Himalayan Geology has identified 1,266 lakes in the glaciated terrain of Uttarakhand alone (PIB, 2025, Apr. 1). The third, and most recent, is instrumentation in Chamoli itself: in October 2026 the State announced a pilot warning station at Vasundhara Tal with water-level, tilt and avalanche sensors reporting by satellite to state and scientific dashboards (ThePrint, 2026, Oct. 13).

These are substantial steps, and they open a natural next stage of reform. The national effort is framed around glacial lakes, while Chamoli was a rock and ice avalanche with no lake involved. A warning architecture designed for lake outbursts will not, on its own, detect a slope that has been cracking for three years. What the Chamoli record shows is that the signal was available in two forms, as slow deformation visible from orbit and as a seismic signature at the moment of failure, and that neither was connected to any alarm at the project sites. Keeping this into consideration, the protection of Himalayan dams requires a hazard model that treats slope instability, ice loss and lake formation as one system, and a warning chain that ends not in a dashboard but in a siren at the adit mouth and a drilled evacuation route for every shift underground.

There is also the question of who owns the risk. A hydropower plant in a high valley is at once an energy asset, a dam that can harm communities downstream and a workplace for hundreds of contract labourers; these dimensions fall under different ministries and agencies, and none is charged with assessing the asset whole in its terrain. That fragmentation is precisely the condition which the Bharat National Resilience Index was proposed to measure, by scoring assets across hazard exposure, continuity and governance domains rather than within a single sectoral silo (Dash, 2025, Oct. 28).

6. Make Every Himalayan Dam a Monitored Asset: Recommendations

The lesson of 2021 is not that hydropower has no place in the Himalaya; it is that a plant placed in a valley of known mass movement must carry its protection with it from the day of approval. Three recommendations follow. First, every high Himalayan hydropower project, existing or proposed, should hold a basin-scale hazard assessment covering slope instability and ice avalanches as well as glacial lakes, refreshed with satellite deformation data, as a condition of approval. Second, the warning effort now under way for glacial lakes should be widened into a multi-hazard network for the upper catchments of operating projects, with alerts tied by protocol to the project control room and to automatic evacuation of tunnels and riverbed works. Third, these assets should be brought under a single resilience audit, independent of the developer, that examines the plant, its workforce and its downstream reach together and reports to a national authority.

Chamoli was a disaster that the mountain announced in advance and that instruments recorded in real time. India already possesses the satellite capability, the scientific institutions and the disaster response forces to act on such signals. Once siting, warning and protection are governed as one problem, the country will be able to protect its hydropower workers as surely as it has learnt to rescue them.