From Risk to Resilience:
What It Will Actually Take

We know which lakes are dangerous. We have the satellite data. We have the science. So why are communities still being caught off guard? Why are people still dying?

On 19 March 2026 — World Glacier Day — the third session of the Glacier Dialogues series convened as a UN side event co-organised with UNESCO and the World Meteorological Organization (WMO). Glacier Dialogues is TVGF's monthly convening series, and this session was designed to move the conversation forward: not to re-establish that the crisis is real, but to ask honestly what it will actually take to meet it. Four voices on glacier risk in India were brought together with a clear intention — not just to describe the problem, but to argue about what will fix it.

The Risk Is Not Just Rising — It Is Changing Shape

Anurag Maloo opened the first round with a provocation that set the tone: "If the glaciers did not melt, we would have no Ganga, no Yamuna. So we need glaciers to melt — just not like this. But right now, that same melt is killing people. How do we explain that paradox to communities whose water depends on the very thing that threatens them?" It is a question that reframes the problem entirely — glacier loss is not simply an environmental crisis but a water security paradox playing out in real time across 2 billion lives.

Dr. Ashim Sattar opened his response with a deceptively simple analogy. Think of glaciers as a banking system — savings accounts of frozen water that release funds slowly, predictably, seasonally. For generations, that bank was reliable. What we are witnessing now is not just a bank that is losing money. It is a bank that has become erratic.

The shift he described is not primarily about frequency — scientific studies do not yet show a confirmed trend in the number of Glacial Lake Outburst Floods (GLOFs). What is changing is magnitude. The South Lhonak disaster released approximately 50 million cubic metres of water, which then eroded debris orders of magnitude larger as it moved downstream. The Chamoli disaster destroyed two hydropower projects. The Dharali floods of 2025 laid waste to an old landslide fan on which hotels had been built.

What these events share is a pattern: isolated hazards have given way to cascading ones. A rockfall triggers an avalanche. The avalanche triggers a GLOF. The GLOF triggers a flash flood. The flood carries debris that dams the river, which then fails catastrophically downstream. These are no longer single events — they are chain reactions in an increasingly destabilised system.

There is also a category of hazard that is receiving insufficient attention even within scientific circles: supraglacial lakes — bodies of water that form on the surface of a glacier through differential melting. These lakes bypass standard monitoring frameworks entirely. They can drain catastrophically through the glacier itself without any external trigger, leaving no warning window. The 2024 Rasua floods — which caused transboundary impact across the Nepal-India border — originated from exactly this kind of supraglacial formation. The category is poorly mapped, and current monitoring infrastructure was not designed to catch it.

A related gap was raised during the audience Q&A: water pocket outburst floods (WPOFs), distinct from standard GLOF events, are not currently covered under the National Glacier Risk Mitigation Programme (NGRMP). This is acknowledged as a gap to be addressed in future programme phases — but the hazard exists now.

Dr. Aparna Shukla added the governance layer: India's national frameworks for glacier hazards are evolving, but they remain structurally underprepared for the pace of change. The core bottleneck is not political will but meteorological data. Without reliable, high-altitude weather observations — which deteriorate or fail routinely in Himalayan terrain — hazard assessments remain incomplete. Mission Mausam, the Ministry of Earth Sciences' newly launched initiative, aims to extend the automated weather station network significantly across Himalayan regions. This is a necessary step, and its scale and speed will determine how much it matters.

She also highlighted a promising development: the BIMSTEC Himalayan Science Council, being led by the Ministry of Earth Sciences, would create a shared research and advisory structure across Bangladesh, Myanmar, Sri Lanka, Thailand, Bhutan, India, and Nepal — mirroring the successful BIMSTEC Weather and Climate Centre. Transboundary hazards require transboundary intelligence infrastructure.

An active Indo-Swiss collaboration is also currently live: a joint call for research proposals between the Ministry of Earth Sciences and the Swiss National Science Foundation (SNSF) focuses specifically on mountain hazards. Researchers working on Himalayan risk were publicly invited to contribute — a direct bridge between Indian institutional science and global cryosphere research capability.

The Honest Governance Reckoning

Safi Ahsan Rizvi opened with the kind of candour that is rare at policy forums. He refused the easy narrative that government is the villain. "The science is not ready. The research is not ready. The mitigation programs are not ready. The communities are not ready. The state governments are not ready. Everybody is not ready." But crucially, he added — beginnings have been made, and they are good beginnings.

Safi Ahsan Rizvi then laid out what those beginnings look like. The Committee on Disaster Risk Reduction (CODRR), a multi-institutional platform, identified 7,500 glacial lakes of one hectare or more. Through deliberation with the Geological Survey of India (GSI), Central Water Commission (CWC), Swiss Development Corporation (SDC), scientific institutions, and NDMA itself, that list was narrowed to 200 lakes declared at risk. Forty of those lakes were physically visited in the summers of 2024 and 2025 — an exercise requiring helicopter access, acclimatisation, and extreme logistical precision.

The NGRMP, cleared by the Government of India, is now active across four states: Himachal Pradesh, Uttarakhand, Sikkim, and Arunachal Pradesh — with Jammu & Kashmir and Ladakh running parallel programmes from their own funds. The programme operates on five pillars:

The Finance Commission architecture has also quietly shifted in a meaningful way. The 15th Finance Commission moved disaster management funding away from a model where 65% went to relief and response, replacing it with a four-window structure where 60% goes to pre-disaster activities and long-term recovery. The 16th Finance Commission has raised the total national allocation further, to ₹2.83 lakh crore over five years. These are structural changes, not announcements — and they matter for whether the NGRMP's harder interventions can actually be sustained over a decade.

Safi Ahsan Rizvi also pushed back against the systemic pessimism that sometimes enters conversations about India's capacity to manage complex, long-horizon challenges. His counter was concrete: look at what the India Meteorological Department (IMD) did after the 1999 Odisha super cyclone, which killed thousands with virtually no early warning. Twenty-five years later, IMD provides seven-day cone-of-uncertainty cyclone forecasts, enabling district magistrates to evacuate coastal villages before landfall. That transformation was the result of sustained institutional investment over two decades. The same trajectory is possible — and in some ways already underway — for glacier hazard management.

But Safi Ahsan Rizvi was equally pointed about what is not working. Anurag Maloo pressed him directly on the asymmetry that haunts disaster response in India: a terrorism event killing 50 people commands 200 times the media coverage and political resources of a glacier disaster killing thousands. The ecosystem of media verticals, security analysts, and government communications built around national security over decades simply does not exist for climate adaptation — and building it requires the same kind of deliberate, long-term investment.

He made a direct and provocative call: India's large philanthropic foundations — which collectively spend ₹5,000–6,000 crore annually — allocate only 2–3% to disaster management. Of the approximately ₹900 crore that does reach the sector, the overwhelming share goes to capacity building rather than to the harder, more expensive work of technical mitigation. He called for corporate philanthropy and CSR funding to shift from climate mitigation (where significant global effort already exists) toward climate adaptation — specifically to the twelve nationally recognised disaster categories, prominent among which is GLOF.

Antony Joh Moothedan: From Lake to Last Mile

Antony Joh Moothedan reframed the entire problem usefully: a GLOF is not a glacial lake problem. It is a river basin problem that begins at a glacial lake. The real questions are not "is this lake dangerous?" but "if this lake fails, what happens next?" — how fast does the wave travel, where does it amplify, what critical infrastructure lies in its path?

This is the contribution of hydrological basin-scale modelling: not to predict the precise moment of failure, but to build a library of scenarios — large breach, small breach, gradual release, sudden outburst, monsoon context, dry season context — so that decision-makers can prepare for a range of consequences rather than a single anticipated event. The three parameters that matter most in translating lake failure to downstream impact are lake volume, breach width, and time of breach formation. These determine peak discharge, inundation area, flood depth, velocity, and — most critically — arrival time at downstream settlements.

The Intelligence That Lives in Mountain Communities

The third round may have been the most consequential — and it began not with data, but with a story Anurag Maloo chose to place at the centre of the conversation. It is the kind of story that formal policy forums rarely allow: specific, human, and inconvenient to the assumption that technology is always the answer.

In Gilgit-Baltistan, a local shepherd with no satellite and no model noticed something wrong with a glacier lake above his valley — a change in the water, in the animals, in some quality of the landscape he had watched for decades. He raised the alarm. His village evacuated. Hundreds survived a severe GLOF event. Meanwhile, millions were lost to corruption in GLOF projects in the same region. The contrast, Anurag Maloo pointed out, tells us something critical: value in glacier risk management does not always live where we expect to find it — and the most important early warning systems are sometimes the ones that cannot be switched off in a Himalayan winter.

This is not anecdote. This is data infrastructure that we have not yet learned to formalise.

Dr. Ashim Sattar brought a striking research finding to the table: in the South Lhonak disaster, 36% of early warning communication came through mobile phones and WhatsApp — relatives texting relatives, observers upstream alerting people downstream. Community members near the Indo-Tibetan Border Police (ITBP) post noticed the river level rising and passed the signal along. The system worked not because of sensors but because of people.

The practical challenge is making this observation credible within formal scientific pipelines. Dr. Ashim Sattar described a path: map the hazard zones scientifically, communicate them to communities in plain language, train community members to observe specific indicators — water colour, turbidity, sediment load, unusual sounds, animal behaviour at high altitudes, and notably the foul smell that often precedes a GLOF event as disturbed sediment is released — and create standardised reporting mechanisms that feed into early warning systems.

He also named something that institutions frequently overlook: many Himalayan communities hold spiritual and cultural relationships with glacial lakes. These lakes are not simply water bodies — they are sacred presences. Any engagement framework that arrives with instruments and warnings but no acknowledgement of this relationship will fail. Trust is built by entering respectfully, not by arrival with expertise alone.

Schools were identified as a particularly underdeveloped entry point. GLOF evacuation drills are entirely absent from mountain school curricula, despite being standard practice for earthquakes and fires. Children are the most reliable vector for community preparedness — and the absence of glacier-specific drills in the highest-risk districts is a gap that could be addressed at relatively low cost.

Vulnerability surveys — asking about elder care needs, communication access, education levels, disability — allow scientific hazard data and community social data to be combined into household-level risk scores that drive targeted preparedness.

What emerged clearly is that community participation is not a supplement to technical systems. It is the layer that makes technical systems survivable. Antony Joh Moothedan noted the three reasons early warnings fail to produce evacuation: people don't understand the warning, people don't trust the warning, or people receive no instruction on how to act. The third failure is the most preventable — and it is a failure of the gap between science and community, not a failure of the science itself.

What a Community-Led GLOF Adaptation Framework Actually Requires

Dr. Aparna Shukla laid out six non-negotiable building blocks for any village-level adaptation framework that is designed to hold — not just look credible on paper:

Safi Ahsan Rizvi reinforced a point that cuts across all six: communities should not be designed for as beneficiaries. They should be designed with as participants. The distinction matters enormously. When early warning systems are installed by external teams without community understanding or buy-in, they fail — not technically, but socially. When communities build the system alongside scientists and administrators, the system is trusted. And trusted systems get used.

He also pushed back against the instinct to create a new dedicated institution for glacier hazards. The Disaster Management Act already includes GLOF as one of twelve nationally recognised disasters. NDMA, SDMA, CODRR, and the network of state agencies already exist. The problem is not architecture — it is coordination speed, fund utilisation timelines, and the political urgency (or lack thereof) behind implementation.

One exchange from the Q&A session crystallised the systemic gap clearly. A researcher from a national policy institute described reviewing adaptation documents for Himalayan states and finding almost nothing on glaciers — discussions of agriculture, forests, biodiversity, but not the cryosphere. This is the Orphaned Crisis in action: glacier preparedness falling between the mandates of climate mitigation, disaster response, and development finance — belonging fully to none of them.

A related and equally urgent gap was raised regarding long-term water system adaptation. Current National Adaptation Plans and State Action Plans on Climate Change do not adequately address what happens when the smaller glaciers disappear — not just as a GLOF risk, but as a systemic water failure for agriculture, drinking water, and energy across Himalayan states. Technologies such as ice stupas in Ladakh and experimental glacier shielding with reflective materials are being explored but have not been adopted at government programme scale. The questions of downstream water quality, community consent, and ecological impact remain unresolved. A dedicated working group focused on this longer horizon — beyond immediate disaster scenarios — was identified as an urgent need.

A final accountability question surfaced repeatedly: who actually owns community communication and early warning coordination at the district level? The Disaster Management Act and the institutional architecture of NDMA, SDMA, and DDMA provide the legal framework. But in practice, responsibility at the last mile remains diffuse. No single district-level actor feels fully accountable for ensuring that communities understand, trust, and respond to warnings. Naming this gap — and closing it — is as important as installing any sensor.

If You Were Designing a 10-Year Roadmap — Where Do You Begin Today?

Each panelist was asked for a single, non-negotiable first step. Their answers, taken together, are a sequence:

The Knowledge Exists. What Is Missing Is the Bridge.

What made this session different from most conversations about glacier risk was not what was said — much of the science is already published. It was who said it, together, in real time, willing to name failures alongside progress. And it was the quality of the questions that held the conversation together — questions shaped by someone who has spent time inside a glacier, not just studying one.

A former NDMA advisor acknowledging that systems-level readiness is not yet achieved. A Ministry of Earth Sciences scientist naming the gaps in research-to-policy translation. A glaciologist describing how a WhatsApp message network saved more lives than any instrument. A hydrological modeller reframing the entire hazard from "dangerous lake" to "what happens downstream, and when?" These are not abstract insights — they are the direct outputs of a conversation designed to move from description to diagnosis.

In his closing remarks, Anurag Maloo drew together what the 90 minutes had made undeniable:

That framing — the preparedness gap as a bridge problem, not a knowledge problem — is one Anurag Maloo has held since his survival on Annapurna. It is the animating logic behind The Voice of Glaciers Foundation: not to produce more reports, not to add to the body of published science, but to build the infrastructure of translation, trust, and last-mile connection that turns what science already knows into what communities can actually survive on.

Glacier Dialogue #4, already in planning, will move into deliberately different territory — connecting glaciers through memory, art, and storytelling. This is an intentional choice. Anurag Maloo has long held that preparedness is not only technical. It is also about whether people feel the glaciers belong to them — whether the loss registers as personal, whether the grief is named, and whether that grief becomes the kind of motivation that outlasts any single policy window.

The glaciers aren't waiting. Neither should we.