Kathmandu's earthquake nightmare
By analysing fault lines in Nepal scientists hope better to predict the next big one. The problem is finding them, as Kate Ravilious discovers.
The Mahendra highway marks the line where the mighty Himalayan mountain range starts to rear up from the pancake-flat Terai plains. Slicing Nepal in two, it runs for more than 1,000 kilometres east to west. Back in the 1960s it was seen as a convenient place to construct a highway. The ground was flat and it didn’t cross valuable agricultural land. Little did the road engineers realise that they were following the fault line responsible for Nepal’s largest earthquakes.
Just outside the small town of Bardibas, 100 kilometres south-east of Nepal’s capital city, Kathmandu, the highway is busy. Horns blare, engines roar and lorries laden with gravel trundle past. It is hot, the sun is glaring down and a thick choking orange dust fills the air. But Paul Tapponnier is not distracted by these discomforts. His eyes scan methodically across a crumbly rock-face a short distance back from the road. Occasionally he pokes a penknife in between the cobbles and scrapes a small heap of rubble onto his trowel. Intent on his task, the silver-haired, bespectacled Frenchman sweeps away the dust and pulls out his magnifying lens to scrutinise the larger lumps. Most of the time he flings them away dismissively, but every now and then his face crinkles into a smile of victory and he shouts for a nail and a foil envelope. With great care he slides a tiny black flake, often no bigger than a sesame seed, into the envelope. As soon as it is safely in its package he spears a paper label with the nail and bangs it into the hole the little flake came from.
The fragments that Tapponnier, a geologist based at the Earth Observatory of Singapore, is searching for are pieces of charcoal – the remnants of long-dead trees. “To us this charcoal is like gold – and a key part of understanding the biggest earthquakes in Nepal,” he says.
Earthquakes are a fact of life in Nepal. India is slamming into Asia at a rate of four centimetres a year and the strain that accumulates in the tectonic plates periodically releases itself in the form of earthquakes. This collision of continents has forced up the vast Himalayan mountain chain as the Indian plate slides beneath Asia. Tremors of magnitude 4 or 5 happen more than 10 times every year. But the real worry is the “great” earthquakes – magnitude 8 or more – that occur every century or so as strain in subsections of the fault up to 200 km long suddenly and violently releases.
'Personally, I’m terrified. I’ve worked in the aftermath of some of the world’s biggest earthquakes ... but this is going to be far worse.'
The last great earthquake – a magnitude 8.4 – occurred in 1934, razing around a quarter of Kathmandu and killing nearly 17,000 people across both India and Nepal. It was caused when the eastern segment of the fault lurched to life (see map). Prior to that a great earthquake in 1833 devastated the Kathmandu valley. People are now feeling exceedingly jittery about the next one.
“Personally, I’m terrified. I’ve worked in the aftermath of some of the world’s biggest earthquakes – Haiti, Bam, Kashmir and Gujarat – but this is going to be far worse,” says Moira Reddick, coordinator of the Nepal Risk Reduction Consortium, based in Kathmandu. “It gives me nightmares. I have a disaster preparedness kit at the bottom of my garden – a shovel to dig people out, water, tinned food, a battery radio and so on. Other people think I’m crazy but I’m envisaging that I might have to house 30 people in my garden.”
Since 1934, the population of Nepal has risen fivefold, and a rural exodus has swelled Kathmandu’s population more than sevenfold. Today these city dwellers are living in one of the riskiest earthquake locations in the world.
Surrounded by steep mountains, the 25-kilometre-wide Kathmandu valley is a geographical freak, the only flat and fertile land for miles around. But the ancient lake sediments that made it such an attractive place to settle are also what makes Kathmandu so vulnerable to earthquakes. Seismic waves can’t pass as quickly through soft sediments as they can through solid rock, bottling up energy there and shaking the ground all the harder.
“They magnify the shaking of an earthquake as much as eight times,” explains Lok Bijaya Adhikari, a seismologist at the National Seismological Centre in Kathmandu. On a geological map of the Kathmandu valley Adhikari points out the wobbliest place of all – the thick layer of sand, silt and mud underneath downtown Kathmandu.
Thamel, a district at the heart of tourism in Kathmandu, is made up of a network of courtyards connected by passages only as wide as a person’s outstretched arms. Throngs of people surge through these narrow channels, bikes and scooters brushing past them as looping electric phone and internet cables dangle precariously overhead. The density of humanity in this place is astounding – 60,000 people to each square kilometre. Houses, businesses and shops tower up on either side of the passageways – eight, nine, 10 storeys high, cantilevering inwards as they go up, so that the sky above is just a thin blue line.
Often the first two storeys of these buildings are traditional brick and wood, some still bearing fissured scars from the 1934 quake. But weighing down on top of these dainty constructions are another five or so storeys made of modern concrete and steel. “This is very dangerous, as the lower mortar part is more flexible while the upper concrete floors are quite rigid,” explains Bijay Upadhyay, one of the directors of the National Society for Earthquake Technology in Nepal.
To make matters worse many of the houses have been divided vertically, with each skinny slice (sometimes only two metres wide) inherited by different sons in a family. And sometimes a son will decide to knock his slice down and rebuild from scratch. “You end up with different floor heights between the buildings, which means they no longer brace each other during an earthquake,” says Upadhyay.
'There will be a stampede of people trying to escape and it
will be completely blocked by rubble.'
Following Upadhyay’s nimble footsteps, I duck through a passageway just 1.5 metres high and pop out into a tiny tranquil courtyard. Suspicious residents peer out of their windows as Upadhyay points out the uneven bricks and wonky wooden window frames in a house that obviously survived the 1934 quake.
“People whose houses survived the 1934 quake have faith in them, but they forget about ageing and lack of maintenance, and they don’t realise that adding extra concrete storeys and continually slicing the houses is making them weaker,” says Upadhyay, concern wrinkling his brow. And even if their houses do survive they are going to struggle to get to a safe open space. “The only exit from this courtyard is through that tiny passage.”
Meanwhile, in the alleyways of Thamel, Upadhyay envisages cantilevered architecture collapsing and air conditioning units, electricity transformers and masonry raining down. “There will be a stampede of people trying to escape and it will be completely blocked by rubble,” he says.
The story is similar throughout the Kathmandu valley, most of which is now built up with high-rise concrete blocks. To make things worse, the next earthquake will sever Kathmandu’s contact with the outside world. “There will be no vehicle access to Kathmandu for months, as landslips and bridge failures will block the three roads into the valley. Meanwhile cracks across the airport runway are likely to prevent planes landing for at least 72 hours,” says Reddick.
“And communications will be knocked out because all the transmitters are currently situated on the tallest and most vulnerable buildings,” she adds.
Conservative estimates suggest a repeat of the 1934 earthquake today could kill more than 100,000 people and leave millions injured and homeless. Which is why Tapponnier and his colleagues are working against the clock, trying to track down the ruptures created by previous great earthquakes. In the same way the cracks in an old house can tell you how the foundations are moving, the rocks straddling a rupture can tell the story of how often that segment of fault moves, and give some idea of when it might judder into action again. Armed with this information, cities like Kathmandu can better focus their efforts to make themselves more resilient to the next big quake.
Usually it is easy to find the rupture caused by a whopping great magnitude 8 earthquake – but not in Nepal. Monsoon rains wash soils down the hillsides and dense jungle covers much of the land, making it difficult to pinpoint a rupture’s origin. Despite decades of searching no one had ever managed to find the rupture from the 1934 quake, leading many geologists to conclude that Nepal’s earthquakes are “blind” – the movement occurring deep underground but never quite slicing all the way up to the surface.
But Tapponnier was never convinced by that theory. “For me the idea of a ‘blind’ great earthquake is a paradox. We don’t find them in other parts of the world so the theory seems very strange.” It was a tantalising problem, the solution to which had potentially tremendous rewards for Nepal. “If a fault is blind then there is very little you can do except sit and wait for the earthquake – but if you know exactly where the earthquake tore the ground you can really start to prepare and reduce risk,” he says. In 2006, after a case of high-altitude pulmonary oedema made it dangerous for Tapponnier to continue working on the Tibetan plateau, he decided to descend to the plains of Nepal and try to solve the “blind” earthquake puzzle.
Pulling up a satellite image on his laptop he shows me the method he has developed to track down earthquake ruptures. “It is only since 2001 that we have been able to do this. Before then the resolution wasn’t good enough.” Zooming in, his finger traces the Sir Khola River, a 30-metre-wide, arrow-straight gash in the landscape which, a raging torrent during monsoons, carries water from the Himalayan foothills to the Terai plains. But suddenly his finger jumps sideways, tracing out the path of a large zigzag in the river’s path. “It is an anomaly in the landscape and a big clue that something unusual has happened here,” he says.
Field trips soon proved Tapponnier right. He now knows that this kink in the river’s path sits right at the fault where India meets Asia. Each time this section of the Indian plate jolts forward the land to the north of the fault is pushed up, creating a ridge that the river gets caught behind. The Sir Khola is forced to flow east or west for a while, until it can cut its way through to resume its southward journey (see graphic).
After a bumpy jeep ride I walk up the cobble-strewn Sir Khola riverbed (thankfully dry at the moment) with one of Tapponnier’s colleagues, Som Sapkota, a geologist from the Department of Mines and Geology in Kathmandu. Geckos scuttle away into the undergrowth and a band of small children temporarily abandon their goats to follow us. Approaching the downstream end of the dog-leg in the river Sapkota strides over to a seemingly unimpressive section of the riverbank rising 30 metres above us. “Just here the river has done all the hard work for us,” he explains. It has incised a cross-section right through the fault.
To the untrained eye the escarpment looks like no more than a jumble of loose rocks, though there is a distinct diagonal line ramping up through it dividing fine yellowish sediments from coarse orangey layers of river cobbles. “That is the fault,” says Sapkota, pointing at the diagonal line that marks the point where the Indian plate is forced to angle down under Asia. His eyes twinkle as he recalls the excitement he felt when he and Tapponnier discovered it in 2008. Sapkota points out how the fault line forms the diagonal of a distinctive “Z” shape. The top and bottom lines of the “Z” were once a continuous bed of gravel, but in 1934 the earthquake sliced through and left the gravel to the north hanging five metres above the gravel to the south. Charcoal fragments found in the bed have been carbon-dated up to 1934, showing that the river was depositing them until the earthquake occurred.
“Some of the local people have strong memories of this earthquake. One elderly man told me that he was 16 years old and was out cutting grass when the quake came. He clutched a tree as the ground shook and watched as his house collapsed. Luckily his family were all okay, but six people in his village died,” Sapkota recalls.
A couple of metres beneath the distinctive 1934 gravel layer Sapkota points out a more orange-coloured gravel layer. It too is truncated by the fault and then re-emerges some 10 metres up the river bank – another giant “Z” in the escarpment. “We’ve dated this movement to the 1255 great earthquake which is recorded in Nepalese history as having devastated Kathmandu and which mortally wounded Nepalese King Abhaya Malla,” Sapkota tells me.
Interestingly there is no sign of any other earthquakes in between these two, despite historical records indicating that great earthquakes have occurred every century or so. “We’re seeing a gap of 679 years between earthquakes on this section of fault. We think other historic earthquakes, like the 1833 earthquake, must have been different beasts, which broke different segments of fault,” says Tapponnier.
Over their past five field seasons Sapkota and Tapponnier have dug several trenches into the foothills of this region, trying to trace the extent of this surface rupture and to go back further into the quake history of the fault. Their dedication is impressive as they work obsessively in the baking heat until darkness falls. Lunch breaks seem an inconvenience. Other geologists have come and gone, but these two, like a pair of scent hounds, are determined to sniff out the entire fault. Sometimes they disagree and forcefully state their cases, but even when they argue there is a sense of deep underlying respect.
To date, their findings indicate that each time it ruptures this segment of fault cracks open a line across the landscape around 150 km long. But because it creates such huge “steps” in the landscape each time it moves – five metres or more of vertical movement – Sapkota and Tapponnier have yet to dig deep enough to find the rocks sliced by any earthquake prior to 1255.
However, this year a stroll upstream along the dog-leg of the Sir Khola revealed a clue they hadn’t fully appreciated before. Along the top of the river bank Tapponnier and Sapkota point to three wide U-shaped channels hidden by trees and long grass. “We call this one the ‘mad buffalo’ channel because our attention was drawn to it when a buffalo came charging out of it,” says Tapponnier with a grin, pointing to the first of the “U”s. The channel lies at about six metres above our heads, while the next channel is around 10 metres and the last, further downstream, about 14 metres high.
The two men suspect that each channel represents a place where the river forced through the newly uplifted land. They have collected little fragments of charcoal from the base of each channel. The dates from the charcoal will reveal when that ancient vegetation died – perhaps smothered by earthquake-triggered bank slides – indicating the timing of the earthquakes that shifted that channel upwards. The lowest channel is most likely the path the river followed up until the 1934 quake. The 10-metre-high channel is probably where the river flowed up until the 1255 quake. “And if this segment of fault moves roughly every 700 years, then we’d expect the highest channel to have a date in the 6th century perhaps,” says Tapponnier.
Back down by the noisy roadside near Bardibas the team is hoping to go back even further in time with the aid of sophisticated LIDAR (Light Detection and Ranging) technology. Perched on top of a tripod alongside the road a sleek silver cylinder slowly rotates through 360° sending out 70,000 pulses of laser light every second. The resulting reflections are used to build a high-resolution three-dimensional picture of the surface. Once the images are taken to the laboratory they will be processed to remove trees, shrubs and grass, as well as the odd inquisitive child who walked across the LIDAR field of view, leaving just the bare earth surface.
“With the LIDAR images we can clearly see a series of terraces, going like stairs up this hillside. Each step may correspond to one earthquake,” explains Tapponnier. The team has found a sequence of seven earthquake terraces nearby giving them more than 4,000 years of earthquake history. They are working hard to decode that right now. Meanwhile, a seismic survey of the region is helping the scientists to determine what the fault looks like underground and the size of area it is capable of shaking.
Based on the findings so far, it looks like the Bardibas eastern segment of fault will remain locked for another 600 years or so, which means Nepalese people can relax for the time being – right? Unfortunately not. Although the scientists are reasonably confident that there will not be a repeat of the 1934 earthquake any time soon they are concerned about great quakes occurring on neighbouring segments of the fault.
“The place I’m worried about now is central Nepal. I suspect that this central segment of thrust produced the great earthquake of the 14th century, which damaged much of Kathmandu. If that segment has a consistent recurrence time then the next great quake could be due there very soon,” says Tapponnier.
It sometimes helps to stand back and get some perspective when trying to trace the path of Nepal’s great earthquake fault, something a new satellite could help with. On 3 April 2014 a Russian Soyuz rocket blasted off from Europe’s Spaceport in French Guiana carrying the satellite Sentinel-1A. “This satellite is going to transform our understanding of continental deformation and enable us to map which parts of the Earth are under the greatest strain,” says Richard Walters, a geologist at the University of Leeds, UK, who is poised to analyse Sentinel-1A’s first readings.
Once Sentinel 1A’s twin, Sentinel 1B, is aloft – it is due to launch in 2015 – the pair of satellites will gather radar images of virtually every point on Earth every six days. By scouring the images for changes between each flyby Walters and his colleagues will be able to see the way the Earth fidgets, spotting twitches of just a few millimetres per year.
Until now scientists have had to wait more than a month for a satellite to do a repeat flyby. “This length of time limited our ability to see Earth movements because other changes, like people ploughing their fields or leaves coming out on trees, obscured the signal,” explains Walters, who is also a member of a partnership project called Earthquakes without Frontiers. With the Sentinel twins, “we’ll be able to see exactly where the ground is being stretched apart or sheared,” he says.
In recent years, devastating examples of continental interior faults in action have included the Bam earthquake in Iran in 2003, Muzzafarabad in Pakistan in 2005 and Wenchuan in China in 2008. Altogether these earthquakes killed 175,000 people – around one third of the local populations. Since 1900, earthquakes on continental faults have killed twice as many people as earthquakes on ocean-continent boundaries.
Nepal is high on the list of countries that concern the Earthquakes without Frontiers team, and Walters is keen to assess strain in the Earth’s crust there as soon as possible. Tapponnier and Sapkota agree the Sentinel satellites are going to bring interesting additional insight, although they maintain that there is no substitute for tracing the fault along the ground. Equipped with all this information, the people of Nepal will be better able to prioritise their actions to increase their resilience to great earthquakes.
But already, throughout the Kathmandu valley, local people are learning about the vulnerability of their neighbourhoods by joining earthquake walks. And increasingly they are taking responsibility for themselves by setting up local emergency response groups. Meanwhile the government of Nepal organises an earthquake safety day every January, including shake-table demonstrations of the damage a quake can do and street-theatre performances to graphically illustrate the aftermath. All are helping to raise awareness.
While in Kathmandu education is working, down on the Mahendra highway few seem to realise that they are living next to – or even right on top of – one of the most dangerous faults in the world. Every year the road becomes busier and the towns along it more populated. “In places like California you are not allowed to build on an active fault – no building can survive that kind of movement. But here the highway follows the fault and the buildings straddle it,” says Sapkota.
Because of the way earthquake faults shape the landscape and control drainage and water availability, people are unwittingly attracted to live near them and build infrastructure along them. But, as other countries demonstrate, great earthquakes do not have to have catastrophic consequences. By understanding the beast that they are living with people have the power to change their destiny. In Nepal the beast has been glimpsed but can people change their destiny before it awakes?
The author's trip to Nepal was supported by a journalism fellowship from the European Geosciences Union.