Skip to main content

Sarwidi Versus the Volcano

An engineer’s radical idea to help people survive an eruption: stay put
  • Author:
  • Updated:

THE WORLD"S ONLY LOW-COST VOLCANO BUNKER is not a high-tech affair. Called the Rulinda—the name is a portmanteau of the Indonesian words for “emergency-protection room”—the bunker is a brick box that can, in theory at least, shelter nine adults from superheated ash for up to an hour. It features walls two feet thick, a door customized to face away from the relevant volcano’s particular blast pattern, and ground-hugging breathing holes that block toxic ash while admitting (mostly) healthy air. Any half-decent mason can throw up a Rulinda in a day or two for about $320. The science in the walls, however, comes from a civil engineer in his mid-40s with a doctorate in seismic-resistant design.

Only three Rulindas exist, all within a few miles of each other, near the town of Kaliurang, Java. A former Dutch hill station in the middle of Indonesia, Kaliurang sits halfway up a volcano named Merapi, where eruptions have killed some 335 people and displaced one million or more in the last 20 years.

The town is also home to the Rulinda’s inventor, who goes by Doctor Sarwidi. (Like many Javanese, Sarwidi does not use a surname.) Sarwidi is convinced that building more Rulindas will save lives—though the idea of staying put when a volcano erupts flies in the face of decades of official disaster-mitigation practice. And common sense.

Buddhist statues are covered in ash after the eruption of Mount Merapi in Central Java, Indonesia. (Photo by Kemal Jufri)


Sarwidi began the project after an ash cloud from Merapi killed a childhood friend in 1994, just 500 meters from the doctor’s own house. The news reached Sarwidi abroad. He was in Troy, New York, finishing a graduate degree in civil engineering at Rensselaer Polytechnic Institute, specializing in seismic work—the study of building things to survive earthquakes.

The memory is still potent enough to derail the grammar in Sarwidi’s otherwise fluent English. “My mother is there! My father is there!” says Sarwidi, a short, stocky man sitting at a teak table in his engineering-department office at the Islamic University of Indonesia in Kaliurang, on a clear day last March. The university campus is entirely within Merapi’s blast zone, just five miles from the bubbling caldera that has been erupting continually since 1992. The cone coughs out smoke daily.

While still in New York, Sarwidi heard a tale that inspired him to create the Rulinda. According to villagers, in the 1994 eruption, nine people had gone outside their Kaliurang homes to gawk at the spectacle when a “pyroclastic” cloud—one composed of superheated ash—came roiling down the mountain, catching them by surprise. The mass of ash and debris smashed into Kaliurang’s outskirts “faster than a tornado,” Sarwidi says.

Burning volcanic materials cover the interior of a house at a village badly hit by pyroclastic flows from the Mount Merapi eruption. (Photo by Kemal Jufri)


It overwhelmed the first eight people. But the ninth, the story goes, heard his landline phone ringing, and ran inside to answer it. A finger of the debris caught the edge of his cement home’s door, slamming it shut and sealing him in a pocket of cool, breathable air. He survived.

The story gave Sarwidi his line of investigation: Could a simple Javanese brick structure resist an ash cloud? He decided to figure out how to build a structure to survive Merapi’s moods—or those of any volcano.

When a volcano like Merapi erupts, it hurls toxic ash and rocks into the air, forming clouds that can fly as fast as 300 miles per hour and travel as far as 20 miles before smashing back to Earth. Lava isn’t the biggest problem in most eruptions; pyroclastic clouds caused the casualties in famous blasts like Mount Saint Helens and Pompeii.

Protective bunkers have been tried from time to time in other regions. In Japan, officials have experimented with a design that leaves a wall open to the downward slope, and even on Merapi the Indonesian government built a cement bunker—which failed. “The best sort of designs I’ve seen are buried, and surrounded by a slipstream in the surrounding slope,” says Ray Cas, a volcanologist at Australia’s Monash University. “If it’s just a box, it would have to be of incredible strength to survive.”

“The best thing, of course, is to evacuate the people,” says Sarwidi. But that’s easier said than done. People hate to leave their homes, even in the face of death. When they do, they tend to trickle back too soon: after a week or two in a temporary camp, people start to worry about their untended homes and livestock left behind.

Then there’s the issue of scale. Four million people live within 30 miles of Merapi. In its most recent eruption, in 2010, the volcano blasted debris as far as the outskirts of Yogyakarta, a nearby city with a population roughly the size of Seattle’s. Merapi’s preeruption rumblings sent some 300,000 people fleeing to temporary camps and shelters, where they remained for more than two months. Factoring in undocumented evacuees—people who left town to take cover with family elsewhere on Java—Sarwidi estimates there could have been one million people displaced in 2010. Even so, when the mountain finally blew, at least 273 people died of asphyxiation or burns. Sarwidi says his design can’t prevent the disruption of so many lives, but it may be able to lower the death toll.

Bad design, however, could raise it. During a 2006 explosion, a pyroclastic cloud hit three miles from Kaliurang. Seeing it coming, two young men ran into the government-built cement bunker and slammed shut the iron hatch. Emergency crews later found the door open and the two men dead inside.

The government had installed the door backward, to open in; the cloud blew the door off its hinges and incinerated the men inside. Had the door opened out so that it rested on its frame when closed, the force would have distributed across the door and the pressure from the cloud would have kept it shut.

Sarwidi cautions that even a better design could still carry risks. Installing the Rulinda could inspire false confidence—think of the nuclear-fallout shelters of the 1950s. That could make it more difficult for the government to evacuate people, he explains.

EFFECTIVE EVACUATION requires good predictions of where the deadly clouds will fly, which can be difficult. In most disaster-safety efforts, officials measure a community’s distance from a potential risk—the landfall of a hurricane, the path of a tornado, the reach of a rising river—and then evacuate the area accordingly. Before ordering an evacuation from Merapi, Indonesian police ask vulcanologists at the National Agency for Disaster Management (where Sarwidi is an adviser) to declare a danger zone measured from the mountain’s peak. Everyone within the perimeter must leave. A perimeter is two-dimensional, however, and a mountain is not. A volcano’s topography, the wind, even the construction materials of nearby houses means risk from a volcano can be greater 10 miles from its spout than two miles.

In 2010, many of those who died were among the farthest from Merapi’s peak. The emergency-response teams had evacuated the wrong people—or not enough people, at least—from areas they hadn’t realized were under threat. Those areas are where Sarwidi wants to place Rulindas.

He also wants to be smarter about identifying risks. Currently, the early warning systems for many of the world’s volcanic areas are based on visual observations. On Merapi, that means a ring of lookout towers manned 24 hours a day during eruptions, plus instruments that measure tremors and changes in the mountain’s pressure.

In 2010, a few weeks into the eruption, many of the instruments were knocked out. The final explosion came at night, when it was too dark for the lookouts to judge the cloud’s size and shape. The surge of superheated ash landed on top of people without warning, since “no one expected it to go to seven miles” from the peak, Sarwidi says.

He hauls out a laptop and loads images taken by a German satellite during the 2010 eruption. The satellite recorded changes in the mountain. Sarwidi insists that the imagery could have been used to map the likely path Merapi’s debris cloud would take in the weeks that followed. “But we only use local data,” he complains.

Had officials used topographic analysis from satellite imagery to target evacuations better, and built cheap shelters for stragglers, they could have saved hundreds in 2010, Sarwidi claims. He adds that his shelters could also keep someone caught in the wrong place at the wrong time, like rescue personnel, alive for a few vital minutes—long enough for the cloud of suffocating ash to pass.

BUT HIDING from the world’s deadliest volcano in what looks like a public bathroom requires a gigantic leap of faith. None of the three Rulinda prototypes has ever been properly tested.

All three were occupied by Kaliurang residents during the 2010 Merapi eruption, but the cloud missed the town. The proof will be in the survival rate after the next direct hit from a volcanic cloud.

Making people fully “safe” on Merapi isn’t realistic, says Sarwidi. “Safer,” though, could be. Cas, the Australian vulcanologist, agrees. “Absolutely, it’s better to have a shelter,” he says, even an untested one, since the odds of survival without one are zero.

Sarwidi thinks he can boost those odds with his little brick boxes, and do it inexpensively enough to build them beside active volcanoes anywhere in the world. So far, though, he has only convinced his hometown. Above it, Merapi continues to smolder.