When Glacial Lakes Burst 

High in the mountains, glacier lakes can create floods that arrive without warning. 

Thousands of lakes sit quietly behind walls of ice, rock and debris. Framed by retreating glaciers, they might look innocent – even serene – to those unaware of the dangers they harbour. Because when these natural dams break, they pave the way for sudden, violent, and hard to predict disasters. These events, known as glacial lake outburst floods, or GLOFs, are becoming one of climate change’s most dangerous and least understood threats. 

A landmark global study published in 2023, “Glacial lake outburst floods threaten millions globally” estimates that 15 million people worldwide are currently exposed to potential GLOF impacts, making this no longer a remote, high-altitude problem but a growing global risk. About 90 million people live in river basins that contain glacial lakes, and roughly one in six of them, 16.6%, live in zones considered directly exposed to a possible outburst flood. As glaciers melt and new lakes form, the danger is spreading faster than governments can respond.

What exactly is a GLOF? 

“GLOF is an acronym for glacial lake outburst flood,” explains Céline Walker to REVOLVE, a doctoral researcher at Friedrich-Alexander-Universität, working on glacier-related hazards. It refers to sudden floods that occur when a lake associated with a glacier, whether in front of it, on top of it, or dammed by ice, rock, or moraine – the accumulated debris carried and deposited by a glacier – rapidly releases its water. “When that natural dam fails,” she says, “the lake can drain very quickly, producing a powerful flood downstream.” 

Walker says that this unpredictability is precisely what makes GLOFs so dangerous. “That doesn’t mean we can predict when a GLOF will happen,” she explains, “but we can identify where they could occur.” 

Not just a Himalayan problem 

Such events are not confined to one part of the world. They have been recorded across continents, with particularly high concentrations in the Himalayas and the Andes. But Europe is not immune. “GLOFs have also occurred in the European Alps in the past,” Walker notes, adding that when they happen, they pose serious threats not only to people, but also to infrastructure and ecosystems downstream. “That’s why they require a lot of precaution and careful management.” According to research published in 2016, Glacier floods have directly caused at least 393 deaths in the European Alps. 

However, the geographical risk of GLOFs is very uneven. High Mountain Asia, home to the Himalayas, Karakoram, and Hindu Kush ranges, accounts for 62% of the world’s exposed population, around 9.3 million people. Four countries alone, India, Pakistan, Peru, and China, make up more than half of global exposure.

Walker cautions, however, against thinking of vulnerability as a purely geographic issue. “A region with glacial lakes is not necessarily vulnerable if there are no people or infrastructure nearby,” she explains. GLOFs do occur in remote areas where no one is affected, but such events rarely attract attention. “Risk only exists when people, infrastructure, or ecosystems are exposed,” she says. In that sense, vulnerability is shaped as much by where people live and how societies are organised as by the presence of glaciers themselves. 

Minutes that mean life or death 

When it comes to GLOFs, distance can mean the difference between survival and catastrophe. Around one million people live within 10 kilometres of a glacial lake, where flood waves can arrive with almost no warning. An even smaller group, about 300,000 people, or just 2% of those exposed, live within five kilometres, yet they face the highest mortality risk because of the sudden onset and force of GLOFs. Once a moraine or ice dam collapses, escape is often impossible. 

Early warning systems can buy critical minutes, but they are never foolproof. “A warning system does not always guarantee an evacuation in time,” Walker says. “But if there’s no early warning system, then there is no warning at all.” 

Lessons from past disasters 

History shows that when these risks materialise, they can have deadly consequences. In June 2013, Chorabari Lake, perched nearly 4,000 metres above sea level in India’s Uttarakhand state, burst after days of intense rainfall and rapid snowmelt. In just five to ten minutes, an estimated 262 million litres of water thundered downstream, devastating Kedarnath and killing thousands. The flood then surged down the Mandakini River, destroying settlements far beyond the immediate surroundings of the lake. Today, with its embankment destroyed. Chorabari Lake is almost completely dry. And yet flood-control walls are still being built to protect Kedarnath from a threat that scientists say can no longer come from that direction. The case has become a symbol of what experts warn against: misreading hazards and investing in the wrong solutions. 

Globally, glacier floods are neither rare nor new. A long-term assessment of 1,348 glacier outburst floods across 20 countries over the past 1,000 years shows that more than 12,000 deaths worldwide have been directly attributed to these events. The deadliest impacts have occurred in Central Asia and South America, while countries such as Peru, Nepal, and India have suffered disproportionately high damage despite fewer recorded floods. The pattern underscores a key lesson: exposure and vulnerability often matter more than frequency alone.

In other words, GLOFs become disasters when they intersect with people, infrastructure, and weak governance. “The impacts really depend on where a GLOF occurs,” Walker says, “and a lot comes down to the financial and institutional capacity of a region.” In poorer or less-prepared areas, communities often lack the resources to monitor risks, respond quickly, or recover after disaster strikes. 

For Walker, technology alone is not enough. “It’s really important that the local community is informed about the risks,” she says. “When an alarm goes off, people need to know how to react.” 

Being prepared makes the difference 

She points to a recent example in Blatten, Switzerland, where a massive collapse of ice and rock buried an entire village in a narrow mountain valley. Authorities feared that ice-debris-mixture could dam water and form a new lake, raising the risk of another flood downstream. In that case, however, extensive monitoring systems and emergency resources were already in place. “There were substantial systems in place beforehand,” Walker explains, allowing authorities to act quickly. 

Still, that level of preparedness is not common. In many regions, GLOFs occur without warning or mitigation. A historic example is Lake Palcacocha in Peru, which burst in 1941 and destroyed parts of the city of Huaraz. Over time, Walker notes, settlements have expanded into areas once avoided because of natural hazards. “What we often see is that settlements were not originally built close to these zones,” she says, “but they have become increasingly populated. When a GLOF occurs there, people can lose their homes, their livelihoods, and sometimes their lives.” 

The Peruvian case has also become emblematic of a broader debate about climate responsibility. Scientists and local communities have sought to hold major emitters accountable for accelerating glacier melt and increasing flood risks. A high-profile lawsuit against German energy company RWE argued that its emissions contributed to glacier retreat and heightened the risk of future GLOFs near Lake Palcacocha. The court ultimately rejected the claim, ruling that RWE’s contribution to climate change was small enough not to establish direct liability. However, the case remains historically significant. As Walker notes, the court nonetheless recognised that “a company could, in principle, be held liable for its contribution to climate change impacts.” The ruling marked a legal milestone, illustrating how glacier hazards are increasingly entangled with questions of climate justice and accountability. 

When it comes to GLOFs, the climate crisis is tilting the balance further. New research published by an international team of scientists projects that by 2100, the world could lose 25–29% of glacier mass under low-emission scenarios, and up to 46–54% under high-emission pathways. As glaciers retreat, they leave behind new terrain where new lakes can form, particularly in the Andes and High Mountain Asia, regions already struggling with limited monitoring capacity and scarce resources. 

The Andes, now identified as the second most dangerous GLOF hotspot globally, have experienced glacial lake growth of 93% since 1990, yet receive less than 8% of global GLOF research attention. The mismatch between risk and research investment raises concerns that some of the world’s most vulnerable regions are being left without the data needed to prevent future disasters. 

What prevention can look like 

Against this backdrop, prevention is no longer optional – it’s urgent. One of the clearest examples of effective risk reduction comes from Northern Pakistan, where a US$7.6 million pilot project supported by the Adaptation Fund and UNDP, ran between 2011 and 2015. The project combined early warning systems, infrastructure, policy reform, and community engagement. It installed weather stations, sensors, and community-managed early warning systems, revised disaster management laws to include GLOFs, and trained local disaster committees in some of the world’s most glacier-dense terrain. 

The results were striking. Evaluators later described the project as highly cost-effective, and its success led to a much larger Green Climate Fund–backed scale-up, now covering 10 districts and building more than 250 protective structures. The initiative demonstrated that combining science, governance, and local participation can significantly reduce disaster risk, even in some of the most challenging environments on Earth. 

But experts increasingly warn that engineering alone is not enough. The same briefing stresses that effective GLOF risk reduction must address hazard, exposure, and vulnerability together, combining early warning systems, land-use planning, strong governance, and local knowledge. Structural measures such as lake drainage or dam reinforcement often provide only short- to medium-term protection, especially as climate change accelerates glacier melt. 

The most resilient strategies, the briefing argues, are those that treat GLOFs as part of a multi-hazard reality, alongside landslides, floods, and extreme rainfall.  

“Risk is defined by impact and vulnerability,” Walker reiterates. “If there are no people, there is no risk, only a natural process.” That is why countries like Iceland, despite facing GLOFs, often experience far fewer consequences than densely populated mountain regions elsewhere. 

From the CryoSCOPE perspective, the goal is simple but urgent: reduce blind spots before disaster strikes. “There are still many places where we don’t quite know yet whether a GLOF could happen,” Walker warns. 

The science is clear, and the stakes are rising as millions of people live under the shadow of destructive or even deadly GLOFs. As glaciers shrink and mountain populations grow, GLOFs are shifting from rare disasters to predictable risks. Whether they become future tragedies, or largely preventable events, will depend on how quickly governments, scientists, and communities act on the warning signs already written into the ice.