During the past 70 years, Peruvian engineers virtually eliminated the risks posed by glacial lake floods. But climate change and a political blind eye are increasing the dangers once again
By Jane Palmer
In 1941, an ice avalanche triggered what is known as a glacial lake outburst flood (GLOF) from Lake Palcacocha. The lake is located at 4,562-meter altitude, below the Pucaranra and Palcaraju summits of the Cordillera Blanca mountain range of the Peruvian Andes. The flood killed an estimated 4,000 people in Huaraz, a third of its population at the time. Such GLOFs have plagued glaciated mountain ranges in places such as Canada, Italy, Nepal, and Bhutan, but the world’s most fatal GLOFs have occurred in Peru.
Two years after the Lake Palcacocha disaster, Peruvian geologist Jorge Broggi deduced that the hazards developed after the end of the Little Ice Age, around 1850. The glacial lakes began to form as the climate naturally warmed, and as they grew, the stress on their natural dams of rocks and debris, or moraines, which sometimes have a core of melting ice, also grew. In 1951, when the Peruvian government realized the full extent of the problem, it launched the Control Commission of Cordillera Blanca Lakes to assess the situation and mitigate lakes that posed a threat. Over the next several decades, led by this commission or other official bodies, Peru drained or built dams at almost 40 lakes in the Cordillera Blanca.
Thanks to those early solutions and continuing vigilance and engineering upgrades, there have been no fatal GLOFs since the mitigation work began. “We worked for 35 years and we solved the problem in that time,” said glaciologist Benjamín Morales Arnao, executive president of the National Institute for Research in Glaciers and Mountain Ecosystems (INAIGEM).
Climate change is creating new threats
In the past few decades, however, the increasingly warming climate has caused the volume of the glacial lakes in Cordillera Blanca to expand rapidly. When scientists from Peru’s Natural Resources Management Office conducted bathymetric surveys of Lake Palcacocha, they found that in 2009 it contained 17 million cubic meters of water, 34 times the volume of water it had held when the dams were completed in 1974.
Consequently, glacial lake hazard mitigation needs to evolve. “While the early efforts of glacial lake mitigation focused on reducing the dangers posed by erodible moraines and high lake levels, anthropogenic climate change is now challenging researchers to take a broader perspective,” said glaciologist John Reynolds, who manages a geohazards consultancy company based in the United Kingdom.
That perspective encompasses another growing hazard caused by melting glaciers: water scarcity. Future glacial shrinkage will limit water resources for both urban and rural residents, noted César Portocarrero Rodríguez, an engineer and consultant who was a leader in Peru’s early efforts to lower glacial lake dangers, particularly those carried out by the glaciology unit of Electroperú, a state electric power utility created in 1969. He now believes that GLOF mitigation infrastructures need to evolve, not just to prevent floods but also to retain the water that glaciers are increasingly shedding. “We have to combine our security work and water resources work by building small dams on many, many lakes to store enough water for the future,” he explained.
Mitigation at Lake Palcacocha
Alerted to glacial lake flood dangers by the 1941 GLOF, Peruvians first addressed Lake Palcacocha, which still posed a significant threat to the residents of Huaraz. Water scarcity was not yet widely recognized as a companion threat to the region. The Lake Palcacocha GLOF had cut into the lake’s moraine, so engineers placed two pipes into this notch to drain the lake and prevent the water levels from rising again. They topped the pipes with an additional 8 meters of earthen wall that formed a barrier against waves caused by rockfalls or avalanches into the lake.
These safety measures worked until 1970, when a magnitude 7.9 earthquake cracked the earthen dam, which had not been compacted because of the difficulty of transporting the necessary equipment to such a remote location. The engineers removed the broken dam and pipes and installed a 122-centimeter-diameter steel drainpipe and 7 meters of a more robust wall above it.
The strategy proved sufficient when in 2003, an avalanche of ice into the lake triggered a large wave but not a GLOF. By draining glacial lakes and reinforcing dams around the region, the engineers likely prevented many such catastrophes, said Portocarrero, who directed 18 of Peru’s projects to date to reduce glacial lake dangers. “We were pioneers in this kind of work, because we learned the formula: how to work and what needed to be done first, second, and third,” he said.
The work itself was risky. Each day, workers and engineers would dig out a portion of the terminal moraine so that they could insert the pipes or reinforce it. At nightfall, they would shore up the excavation with sandbags. “They wouldn’t sleep at night for fear of an avalanche causing waves that would have easily breached the now weakened moraine,” said Alton Byers, a mountain geographer at the University of Colorado Boulder who has been involved in GLOF monitoring and mitigation in Peru and Nepal. “But they still did it, and they achieved remarkable engineering feats.”
Disaster averted at Carhuaz
The mitigation work continued for several decades. Although proven to be successful, political will at the national level to fund the program began to weaken in the 1980s. In 1989, local leaders secured additional financial support from the British and Austrian governments to install siphons to lower the water level of Lake 513, which threatened downstream Carhuaz, a district estimated to have more than 14,000 residents. In 1991, these measures prevented a catastrophe when an ice avalanche triggered a minor GLOF that caused no casualties. But, to the engineers’ concern, the flood did wash out the lake’s moraine, exposing a large rock bar that held back the lake.
The engineers made a plan to reduce the water level by excavating a single tunnel through the rock bar. To check on the viability of that solution, they sought the input of Reynolds, who had conducted the first assessment of the Peruvian glacial lakes in 1988 for the glaciology and hydrology unit of Electroperú Energy and Mines. Funded by the British government, he and a colleague flew out, traveled on horseback to the site, and worked with the on-site team to devise a system of four tunnels, each of which could safely draw the lake level down 5 meters at a time. “A single tunnel could have caused the whole rock bar to fail, and they could have had a massive disaster on their hands,” Reynolds told Eos.
For the next 18 years, the system at Lake 513 withstood several avalanches from Mount Hualcán. Then, in 2010, a large avalanche generated a wave towering 25 meters above the lake surface that breached the dam. The deluge destroyed some bridges, farmland, and houses downstream but caused no fatalities. “It was a tremendous success story,” Reynolds said. “And it made us realize that taking an engineering and a glaciological approach to the problem can have a significant effect.”
Transferring knowledge
In Peru, engineers and scientists have “figured out how to find all the glacial lakes, how to evaluate lake stability, and then how to drain and dam the most dangerous ones so they didn’t produce GLOFs,” explained Mark Carey, a science historian at the University of Oregon in Eugene. Moreover, they achieved all of those advances with relatively small budgets and limited resources, doing work that was frequently in remote and challenging environments. For those reasons, the knowledge they’d gleaned proved useful to other countries that were grappling with similar limitations and hazards.
In the 1990s, the Nepalese government appointed Reynolds as its technical adviser on glacial hazards. He developed glacial hazard training courses in Nepal, Bhutan, and Chile and adapted the lake-lowering siphon design used for Lake 513 at Tsho Rolpa Lake in Nepal. (Meanwhile, in Peru, the national government closed Electroperú’s glaciology unit in 1996 or 1997 as part of privatizing the electric power industry. Some aspects of the unit continued in other government and private organizations).
Since then, through at least 2013, when funding ended, the transfer of know-how from the Peruvian pioneers to Nepal continued by means of international workshops on glacial flooding and risk management sponsored by the U.S. Agency for International Development (USAID), as well as by site visits in both countries. The visits seemed to have “helped the Nepalese have more confidence that the type of work they were proposing to do would be successful—which it was,” said engineer and professor emeritus Daene McKinney of the University of Texas at Austin, who has comanaged workshops with Byers through the High Mountains Adaptation Partnership. Portocarrero, who attended a 2011 workshop, also went to Nepal several times, funded by the United Nations, to work on draining the lower Imja Lake.
An evolving hazard
Although GLOF mitigation specialists around the world have made progress in the past few decades, it has become apparent to them in recent years that they face more complex challenges than ever before.
The breadth of climate change effects on glacial lake hazards has become clearer. On account of the huge rockfall that triggered the 2010 GLOF at Lake 513 and other incidents, scientists are concerned that warming temperatures are destabilizing the slopes that surround glacial lakes. Anthropogenic climate change is causing some of the lower-altitude permafrost belts to thaw for part of the year, effectively melting the “glue” that holds some slopes in place. “This is starting to cause more rock and ice avalanches and landslides,” Reynolds said. “So these are part and parcel of the triggering mechanisms that we need to be looking at in relation to how a glacial lake might become destabilized.”
In the past few years, Reynolds has evaluated the many factors of glacial hazards in the Upper Indus Basin of Pakistan, in Tibet, and in Kashmir. He believes it is necessary to start evaluating all the geological hazards within a particular catchment rather than focusing solely on the lakes, via an approach he terms an integrated geohazards assessment. In 2017, the World Bank adopted the approach and incorporated it into some of its guidelines for dam safety and funding of hydropower projects.
Progress stalled at Palcacocha
In Peru, the glaciology unit was briefly resuscitated in 2001, but barely a year later, then president Alejandro Toledo effectively shifted responsibility and funding for disaster prevention to regional governments ill equipped to handle either.
For GLOF mitigation at Palcacocha, that political shift has brought setbacks. When the lake’s high water levels became apparent in 2009, rather than implement a plan to lower water levels by 15 meters, the regional government chose a cheaper, but only temporary, option to reduce risk. In 2011, the government installed six tubes in the lake to siphon off 7 million cubic meters of water, lowering the water level by only a few meters.
By leaving the lake so full, Peruvians aren’t building on decades of hard-won experience, Carey said. “The main lesson from the past was simply doing the work at a low cost, with a more comprehensive approach to lower the lake level and build a security dam for long-term protection,” he noted. “It’s now all about politics and not about engineering or scientific analysis of the lake.”
Prevention turns to warning
For Peruvians who worked on the early mitigation projects, the priority is to gain the funds to drain Palcacocha. “This is not an earthquake we are talking about; we know how to prevent this hazard,” Morales said.
Instead, concerned citizens of downstream Huaraz, with a population now estimated to exceed 120,000, have pushed for the next best thing: an early warning system. In April 2018, regional authorities pledged nearly 5 million Peruvian soles (approximately $1.5 million) for a system that will connect sensors that measure soil vibrations and water levels in Lake Palcacocha to sirens in the city. The notice, if it works, should give the residents of Huaraz about 30 minutes to seek safety from the flood. The shift in priority is a retreat for a region that more than half a century ago, pioneered the engineering to prevent such catastrophes. Compared to stalled efforts to drain more of the lake’s water, there’s at least some hope that the city can dodge another major GLOF disaster.