The massive devastation seen in Kedarnath (July 2013) was the impact of a cascading hydro-met disaster, amplified by the failure of the Chorabari glacial lake. The next massive such glacial lake outburst flood (GLOF) was of Sikkim's South Lhonak glacial lake (3 October 2023), which devastated the
Teesta valley biosphere and took out a 1200mw dam, putting Sikkim's economy back by nearly a decade. India's
NGRMP (National GLOF Risk Mitigation Programme) is an effort to address these challenges.
The years 2023, 2024, 2025 and 2026 are on track to be the four hottest years on record. And it's getting hotter, by each decimal of a centigrade - risking reaching 2°C above the pre-industrial baseline by mid-century under current emission trajectories. Apart from rising heat making it close to impossible to survive for humans, animals, crops and infrastructure, it will exacerbate glacial melt. As a result, glacial lakes are already increasing in size. The main cause of GLOFs is slope failure, often from ice or rock avalanches creating impulse waves, seepage or piping leading to dam failure, overtopping from heavy precipitation or snowmelt and earthquake-induced slope collapse.
The National Programme
This rising risk is being addressed by the ₹150 crore NGRMP wherein Himachal Pradesh, Uttarakhand, Sikkim and Arunachal Pradesh are to receive funds in three instalments.
NDMA identified 195 at-risk glacial lakes amongst the 7,500 identified by ISRO for urgent mitigation measures under this programme.
In order to coordinate, NDMA created an online-cum-seminar format called CoDRR (Committee on Disaster Risk Reduction) for periodic micro-meetings, one for each disaster - the GLOF committee being the first. It brought together all relevant entities, including central ministries, states and their departments, scientific institutions, international experts and institutions, academicians and practitioners. No standing committee, no fixed agenda, no compulsory attendance, but all voluntary and participative.
There was no boss, just an NDMA advisor as convenor and coordinator. The first two meetings drew polite attendance, but the periodicity, continuity, simplicity and action-oriented approach of CoDRRs got us all talking. One-to-one institutional relationships were incubated, and states began asking for help from whichever institution they found suitable. At a ripe stage, NDMA began to de-captain itself from the leadership role, as the states were where all work had to be done and they had begun to take ownership of this next-gen problem.
The Science
For more than a decade, our understanding of glacial lakes has been dominated by satellite imagery and geospatial models. Landsat, Sentinel‑2 and SAR data have allowed us to build inventories, track lake expansion and identify susceptible basins using multi‑criteria weighting schemes. The remote‑sensing phase was necessary, but it was always only the first step. Sikkim took the lead, with a dedicated cross-institutional team, while HP and J&K were not too far behind. Uttarakhand, Ladakh and Arunachal Pradesh had capacity issues, which are slowly being firmed up. The results have just about begun to come in. In the first summer of 2024, nearly 40 at-risk lakes, classified as A-risk were expeditioned, led by the states.
Led by officials of the Sikkim government, nine lakes were studied over several seasons using bathymetry (volume of water), electrical resistivity tomography (ERT), detailed morphometry and hydro‑meteorological monitoring. An unmanned survey vessel mapped depth profiles and volumes. ERT revealed saturated moraine zones and buried ice. Field visits and DEM (digital elevation models) analysis captured dam geometry, avalanche sites, freeboard (gap between lake level and peak dam height) and outlet conditions. AWS (automated weather monitoring stations) at Shako Cho and South Lhonak lakes tracked temperature, precipitation and lake‑level changes in real time.
Two conclusions from that work should unsettle our comfortable assumptions. First, lakes that look similar from space can have radically different hazard profiles. La Chho and Lachung Khangtse, for example, have nearly identical surface areas but differ by an order of magnitude in volume because their basins are shaped differently. Second, deep water immediately in front of a calving glacier, as at Khangchung Chho, implies a fundamentally different impulse‑wave and moraine‑stress risk than a shallow, spatula‑shaped basin like Tikip.
Shako Cho: Where Mitigation Stops Being a Slogan
Shako Cho has become a kind of touchstone in both the publication and my public writing. It is a high‑risk moraine‑dammed lake at around 5,000m, with no visible outlet, seepage through a saturated dam, rising volumes under glacier melt and limited freeboard. On paper, that description could belong to many Himalayan lakes. But this lake forces us to confront what mitigation actually requires. It reveals that the relationship between lake level and stored volume is non‑linear, governed by the unique bathymetric profile. A sudden 20m drawdown, the paper shows, would remove about 11 million cubic metres of water and cause devastation downstream. The South Lhonak GLOF drawdown was > 50mcm and took out an entire dam and the Teesta valley.
The Strategy
As a nation clearly concerned about climate change, we must be willing to invest in multi‑year, high‑altitude lake‑lowering, using solar pump systems, siphoning and flow-through retention structures to break the force and volume of GLOF catastrophes. In order to achieve this, we need to stop serial processing risk reduction ie first the expedition, then the studies, then the ideation on mitigation, then the actual mitigation and capacity building of communities. The same can be done in parallel, as much as possible. Conduct the expedition, studies, deployment of AWS and ideation on mitigation on the first trip itself, in the first available summer when the lake is accessible. Prepare the DPRs etc in the winter and implement the more complex EWS (early warning system) and mitigation measures in the next summer. On a war footing, this can be done, each state on its own. The technology, the scientists, the contractors - all exist.
The
AWS at Shako Cho, for example, has already captured shifts in winter temperature and ice‑cover duration that, while still emerging, are clearly worrying signals in a high‑risk basin. If we had insisted on 'completing' hazard mapping before allowing any monitoring to be installed, those early cues would have been lost. In a warming Himalayas, the GLOF programme has to behave like a learning system, rather than like a static project with a fixed Gantt chart as some would have us believe.
(Safi Ahsan Rizvi is a retired 1989-batch IPS officer with over three decades of experience in national security, counter-terror financing, and geo-strategy, including nearly 18 years at the ministry of home affairs and a UN peacekeeping posting in Kosovo. Most recently, he advised the national disaster management authority (NDMA) on disaster risk reduction and risk finance until January 2026.)
