"It would take us four, maybe five hours to hike to the crater," Barry Voight remembers. "Then to get into position which meant going up and around and down the other side.
"There were places we had to rappel.
"It's not impossible to go up and down the volcano, it's just a question of—
"The weather conditions are such that you don't have the whole day to work with.
"You have a couple of hours. It's not a big deal."
Barry Voight is a geologist, a landslide specialist, a volcanologist teaching at Penn State and on call with the United States Geological Survey (USGS), which sent him to glacier-capped Volcan Ruiz in January 1986 in response to urgent request from Colombian Government to U.S. Government . . . as ground deformation patterns strongly suggested imminent massive slope failure.
"They'd promised me a helicopter. They went to the effort of trying to contract one from the Philippines and have it shipped in all the way to Colombia. I couldn't believe they were going to do that. But I wasn't surprised they didn't have one operating—the volcano's over 17,000 feet high. Helicopters, once they get over 14,000 feet, are really straining.
"So we scrambled around. If the weather wasn't—"
Voight has a gruff voice, a voice with gravel and meltwater in it; it trickles and dries up; he runs his hands through ash-gray hair; then the idea rumbles in his head and his voice comes rushing out, tumbling, thick with extra consonants and repeated words, almost out of control until he laughs, a short, sarcastic huff, and the dam holds, the words stay in their channel.
"You have the possibility of getting locked in there, that was part of the problem, because it snowed, because the weather could come in and you could get a white-out condition where you couldn't move, so you were a little bit afraid of that because you had no—"
Huff.
"It's one thing to be hanging from a rope down a cliff in a quarry where you know what's above you.
"There was a seismic alert.
"And so you were a little bit afraid that thing would erupt and if you had a little hot water eruption and a little bit of snowmelt and you were right in the spigot—you were dead."
When Voight smiles, it is mostly in his eyes, sky-blue—if the crevices time and weather have scored in his wide-open face shift at all, it is imperceptible.
"So, anyway, it took us about a week to do it, going up every day, or trying to—some days we made it, some days we didn't. But ultimately we got the instruments in place and checked out."
The November before, after almost a year of fuming and trembling, the volcano cradled in Nevado del Ruiz had belched. The eruption was small, melting less than 5 percent of the 25,000 square meters of ice capping the mountain, but this wall of meltwater scoured down through the precipitous river gorges and burst out on the town of Armero in a 120-foot-high river of mud and debris. Of Armero's 29,000 people, 20,000 were entombed and killed, another 5,000 injured; 3,000 more were killed in nearby river valleys. After the disaster, the Colombian government increased its surveillance of the volcano. When cracks began widening in the snowpack on the eastern rim of the crater, a sign that the whole northeast flank of the mountain might break off and the magma blast out—as if, Voight says, the top had been popped off a shaken Coke bottle—the Colombian government called on the United States Geological Survey. Even with Armero erased, the river valleys below Nevado del Ruiz still supported 40,000 people.
The USGS sent Voight because he was the geologist who had predicted, one month before the great May 1980 eruption, that when and if Mount St. Helens blew, it would blast out sideways. Voight had even foretold how far the avalanche and the mud lahars would flow.
Mount St. Helens was Voight's first volcano. While he'd scrambled on Mount Kilimanjaro for pleasure and had mapped the lava plateaus of Iceland, for 23 years, ever since double-majoring in geology and civil engineering at Notre Dame, Voight's chosen puzzle had been landslides—from little collapses of clay on the shores of Lake Michigan to glacier falls and avalanches in Italy, New England, Alaska, and Norway—and he had edited "the standard reference work on mass movements" (according to its publisher), Rockslides and Avalanches. Volume I came out in 1978 and volume II in 1980, just before, thanks to the USGS, Voight found himself pitching his tent on a ridge by Maratta Creek, 10 miles west of Mount St. Helen's summit, in a pretty campsite that a month later would be a flat field of gray ash. "The mountain was spouting ash then, and steam," he told a reporter in 1983. "At dusk, it sent up these marvelous orange columns."
As Voight flew into the shadow of Volcan Ruiz in January 1986, he saw below, he later wrote, "a pale liquid crescent onlapping a gently undulating and fertile plain. The ghosts of razed foundations peer in rows from the crescent's luminous surface. A prow of land is speckled with trees and tin-roofed adobe and concrete dwellings, marking higher ground—adobe and gravestones above a steaming sea of lethal mud—Armero."
Voight would spend the next two years on two projects: devising a new method for predicting when a volcano will erupt, and tracking the letters and telegrams and communiques that would explain why Armero's 20,000 weren't evacuated. The manuscript version of his "Method for Prediction of Volcanic Eruptions" begins with two epigraphs (which were deleted by Nature when the journal printed the report as its March 10, 1988, cover story): "Ach goddam . . . Rabaul him-he buggerup finish for all time," from Mytinger's Headhunting in the Solomon Islands (Rabaul, of course, is a volcano), and "Thou, Nature, art my goddess; to thy law my services are bound," from Shakespeare's King Lear.
"Countdown to Catastrophe," Voight's published analysis of why volcanic hazard management—in spite of international cooperation and the latest technological advances—had failed Armero, is also prefaced by an epigraph:
We are all of us fellow passengers on the same planet and we are all of us equally responsible for the happiness and well-being of the world in which we happen to live.—Van Loon's Geography (1932).
And the report is organized around quotations from Albert Camus' The Plague, such as this: "People in town are getting nervous, that's a fact . . . and of course all sorts of wild rumors are going around. The Prefect said to me, 'Take prompt action if you like, but don't attract attention.' He personally is convinced that it's a false alarm." And this: "In this respect our townsfolk were like everybody else, wrapped up in themselves. . . . They disbelieved in catastrophes. A catastrophe isn't a thing made to man's measure; therefore we tell ourselves that it is a mere bogy of the mind, a bad dream that will pass away. But it doesn't always pass away and, from one bad dream to another, it is men who pass away . . . because they haven't taken their precautions."
Dangling from mountaineer's cord below the crater of Volcan Ruiz in February, three months after Armero's catastrophe, sick and dizzy from the altitude, with no tent, no cache of food, no first aid kit, "what I was doing," Voight says, back in his office at Penn State, remembering, "was setting up a monitoring system for the front"—a series of 3-inch reflectors, strategically placed, that could be located by a laser beam from as far as 10 kilometers away. The changes in their relative positions would tell him what to expect—should the people below be evacuated?—and when.
"There were some very big cracks opening there.
"Even people who had flown over could recognize them.
"One reflector, which had already been placed toward the front of the crater, was showing that very large movements were occurring."
Voight rummages through the rubble of books and papers and files and folders and rocks and shells and antelope skulls and photographs and maps and stubs of chalk that give relief to the institutional surfaces in his office, and picks out a squat black-and-chrome cylinder like a car headlight, and a handful of boxes of color slides.
"It's a problem to establish one of these reflectors on rock," he says, tossing down the slides. "I ended up using very hard nails that I could hammer in, nails with a screw on the end, and the surface of the rock was a little bit weathered so you could do that, barely. It wasn't really great."
In one slide, he and Oscar Ospina Herrera (an "Andenista," the South American version of an "alpinist") from the Observatorio Volcanologico de Colombia, perch on a rubbly shelf, hands bare, in stocking caps and sunglasses, their thighs criss-crossed by rappelling slings, picks and hammers cluttering a table-like rock, Ospina fiddling with the mirror, Voight mugging for the camera.
"It wasn't a terrific installation. You gotta figure out what type of technique will work and you just don't have a suitcase full of 57 varieties of machines, you have limitations on what's available, on what you can carry up to that altitude, we had to get up there to see what was going to work—"
Huff.
"And we couldn't just go up and slap in the instruments. You damn well want them in the right place."
Ruiz is now the most heavily monitored volcano in South America, Voight believes, even though the reflectors he placed there are now covered with dust. "I tried to put them in placeswhere you wouldn't get snow—I put them on a very steep face, a cliff—but nonetheless there're meltwater streams—you're right below a glacier—and you can get rain up there, and you get snow, and you get snowmelt, and the rivers come down choked full of sediment and they're turbulent and they're tossing up dirt into the atmosphere, and you have updrafts—"
Huff.
"It wasn't all fun.
"But there aren't many people who are experienced in knowing just what you want to measure and where you've gotta do it. You've gotta be familiar with what this data means, in terms of what the stability of the entire mass is, so you have to make a field inspection—look at what the fracture systems are, whether there's any evidence of faulting. You're not just sticking instruments in at random, you're making a geologic inspection of the place."
In another slide, taken low on the mountain, cactus in the foreground, the white hump of Ruiz's lethal ice-cap blotting out the sky, Voight stands next to a tripodded black box, the Ranger V-A laser distance meter that can spot a reflector's creeping from 10 kilometers away when the distance covered is mere centimeters. He is pointing a pair of binoculars at the rough face of the volcano, picking out the exact ledge to which he clung while planting a reflector. After the shutter clicked, Voight had trained the laser on the ledge until the reflector glowed back with a ruby eye. Then he had marked its precise position on another photograph, one to be left with the Comite de Estudios Volcanologicos in Manizales, a city of 350,000 in the next drainage over from Nevado del Ruiz, so that its survey crews could monitor the mountain.
"So we were sticking in instruments and doing some mapping of fractures around the edge and trying to evaluate the risks for that particular sector of the cone—
"So the next step is, what does it mean? Will the crater fall apart? If it falls apart, it's like Mount St. Helens. You have the possibility of getting a large amount of material moving out—there's magma down below there. And you'd get melting, you'd get mudflows in all sectors again. There's that river valley down below, 40,000 people—"
Huff.
"They moved out for a while, after Armero, but then they moved back."
He collects up the slides and shoves the pile under a Spanish language textbook he's been studying in preparation for the next time a Latin American volcano rumbles.
"It was a question of evacuation," he explains. "There had been some evacuations since Armero. They were costly. And there were false alarms. Some were false in the sense that nothing happened. During one of them, there was a small eruption that produced a small mudflow but that was still a false alarm"—huff—"false in the sense that the mudflow didn't kill anybody. It didn't go far enough."
With his mirrors set and monitored for a little over a month by his ruby laser, Voight could conclude in early March that the crack opening up on Ruiz's flank was merely the creeping of the glacier's ice, not a fissure in the volcanic rock below: The top of Ruiz's Coke bottle wa not coming off. "These results," he wrote, "provided a measure of relief to those responsible for the hazard response at Ruiz." The U.S. Embassy in Bogota sent Voight a nice letter. The study should remind us, Voight wrote, that "effective observational monitoring depends more on the questions to be answered than on the number of instruments used."
A year before, in March of 1985, a geologist for the United Nations' Office of Disaster Relief, investigating reports of tremors and fumaroles at Nevado del Ruiz, had suggested that INGEOMINAS, Colombia's geology and mines bureau, prepare a hazard map and risk report for the volcano. Work on the map began in earnest in September. On the same day that the final map was officially released—November 10, 1986—Volcan Ruiz began three days of continuous tremor.
On November 11, INGEOMINAS announced that Armero had a lead time of two hours and could be effectively evacuated. (The newspaper El Tiempo published the report on November 14.)
On November 12, a day before the catastrophic eruption, a team of scientists hiked to the summit for geochemical samples. "They took a look at the surface, sampled some little fumaroles, and got off," said Voight. "They didn't see anything really peculiar."
At 3:00 in the afternoon of the 13th, a phreatic eruption—a steam and ash eruption—brought on a sulfurous, dirty rain that oxidized metal roofs in Armero. At 5:00, when the Technical Emergency Committee for the Colombian province of Tolima met for a previously scheduled meeting in the nearby town of Ibague, they learned of the ashfall. They alerted all police and civil defense stations to prepare for floods and mudflows. The Red Cross radioed a warning to its field staff. The Emergency Committee then discussed the agenda for a forthcoming meeting to fine-tune all regional emergency response plans.
By 7:30, when the meeting adjourned, the ashfall had ended; still, several committee members went directly to the Red Cross communications center to request that Armero be "prepared to evacuate." Another of Voight's sources reported that an "order to evacuate" was made—which, Voight wrote, "is not at all the same thing." Local officials in Armero were "almost certainly aware of the alert; but no decision to evacuate had been made, and it is uncertain whether any 'preparations' were being carried out in earnest." Radio Armero was still broadcasting "reassuring, calming messages."
At 9:00, a strong earthquake occurred within Ruiz. "Torrents of meltwater cascaded from the ice cap into the river channels flanking the cone. . . . Mudflows were mobilized in the channels of the major rivers. Campesinos of the upper valleys of the Lagunillas and Guali reported hearing the lahars." Meanwhile, the volcano was spouting hot rocks and ash. "The pumice scorched many roofs, but caused no fatalities. Meanwhile, the lahars were racing down-valley, growing more voluminous by scraping a meter or so of rain-soaked colluvium from valley walls, and entraining several meters of bogs, sediment, and pore-water from the valley bottom."
Between 9:45 and 10:00, officials in Ibague attempted to order the evacuation of Armero, but the torrential dirty rain of a few hours earlier and the ash of the current eruption had fouled the communications network. Up river from Armero, the noise of the passing mud lahar was so loud that people living near the river channels had to shout to be heard. Vibrations could be felt kilometers away.
Between 10:00 and 10:45, the towns of Libano, Murillo, and Ambalema radioed Armero to evacuate. At 10:45, "the lahar disemboweled the landslide dam at Sirpe and released a cool-water flood wave that raced ahead of the hot lahar. Time had run out for Armero . . .
"Many survivors took flight only after hearing commotion in the streets, as the first flood waves struck the village. Electric power failed and confusion reigned in the darkness. Though many attempted to escape on foot, over 20,000 died. Thousands of the injured managed to reach high ground, but by noon next day, only 65 of the one to two thousand residents still trapped alive in the Armero mud had been rescued.
"There seems to be no evidence that the village population was given specific evacuation orders by officials. . . . Following the late afternoon ashfall, Radio Armero and the church public address system had advised calm, and even by 11:00, the mayor of Armero, in ham radio contact with Ibague, was sufficiently unconcerned by the threat that he remained in the village with his family."
A catastrophe isn't a thing made to man's measure; therefore we tell ourselves that it is a mere bogy of the mind.
"Volcanic emergency management is faced with a seldom-win situation," wrote Voight, in the section of "Countdown to Catastrophe" subtitled "Retrospection." "What does one cry out when there may be a wolf?
"The volcanologist's problem is to record accurately the probabalistic chance for error, and yet maintain the specificity and credibility necessary to encourage the appropriate government and management action, and public acceptance and response to the message."
There are 1,300 active volcanoes on the planet. In a 1989 grant proposal to the National Science Foundation's division of Natural and Manmade Hazards, Voight writes: "Among the most awesome and feared of natural phenomena, volcanic eruptions have claimed 250,000 lives since 1700. In 1902 alone, three cataclysmic eruptions in the Western Hemisphere took over 36,000 lives, at La Soufriere on St. Vincent, Mont Pelée on Martinique, and Santa Maria in Guatemala. These events led to the establishment of volcano observatories in order to systematically study the behavior of volcanoes before, during, and after eruptions. Impressive investigations have now been carried out at some locations, including stratigraphic and chronologic studies, mapping, and volcano monitoring.
"But despite such efforts, successful eruption forecasts have been rare. . . . Population growth has made the problem particularly urgent—for example at Mayon Volcano in the Philippines, where 1,200 died by surge and mudflow in 1812, the current population in the zone of high risk approaches 800,000."
In the News and Views section of the March 10, 1988, Nature, USGS geologist Robert I. Tilling assesses Voight's method of predicting volcanic eruptions as "an intriguing new way of looking at this long-standing, vexing problem." (Tilling Voight identifies as "one of the guys who's guided volcanological projects through the years.")
"Voight's prediction method," Tilling continues, "represents a significant refinement in the interpretation of monitoring data, emphasizing the evaluation of critical rate change-time relationships by mathematical analysis rather than by visual inspection."
Ach goddam . . . Rabaul him-he buggerup finish . . .
"The more successful of empirical eruption-prediction models," reads the manuscript copy of Voight's Nature paper, "likely have a common foundation in first principles, suggesting that practical solutions to problems such a Rabaul may be sought in natural laws." If the ancient Greeks considered the volcano Etna the god Hephaestus' smithy, in which he forged thunderbolts for Zeus, and the medieval Danes saw Iceland's Hekla as the fiery mouth of Hell, to Voight a volcano is an illustration of "rock mechanics" in action.
"The notion that rock failure accompanies eruptions is obvious," notes Tilling, "but Voight extends this idea by arguing that the eruption processes themselves can be described in terms of the fundamental law that governs the failure of materials, ·_-_ ¨_ - A = 0."
In 1986, Voight recognized that this relation, first put forward by the Japanese engineer T. Fukuzono to codify his experiments on landslides, applied generally to failures of ice, soil, metals, polymers, concrete, or almost any substance under strain, including the shell of a volcano. The equation, Voight wrote, was a proposition of "striking if incompletely understood generality" that could supply "a mathematical analogy" between a material failure and a volcanic eruption.
To predict when a volcano will erupt, Voight assigns to _ (omega) several different kinds of data, such as measurements of distance change (what he set up at Ruiz), tilt, fault slip, and seismicity. (Gas emission rates and chemical ratios may also be possible omegas.)
For each variety of omega, Voight takes the derivative with respect to time (·_), which gives him the rate—how fast the distance or other variable is changing. He then takes the second derivative of omega (¨_), which gives him acceleration—how fast the rate of change itself is changing. A and _ (alpha) are empirical constants. Voight can estimate these numbers by plotting the rate (·_) versus the acceleration (¨_) and calculating the slope and intercept of the line on the graph. Or he can plug in the various rates of change that had been observed up to the present time and, by the procedure of "numerical least-squares curve-fitting," find the empirical contants A and _ .
This integrated equation can also be shuffled around to solve for the time-to-failure, relative to an arbitrary day—that is, how many days from today, when Voight is busy calculating away in a volcanological outpost low on the volcano's flank, will the crater blow its top? The latest possible day of reckoning can be pegged by assuming that ·_f, the rate of change at the time of failure, is infinite. It's a logical assumption: If ·_ is standing for something like how fast the ground is tilting, the speed at which this is happening at the instant the volcano erupts (and spits the instrument into the sky) is so fast you may as well call it infinitely fast.
A quicker way than calculating for the time-to-failure is to graph the inverse of ·_, the rates of change, plotting them against time. "Experience suggests," notes Voight, "that for volcanoes, _ is frequently near 2." In which case, the inverse rate curve will be nearly a straight line. Says Voight, "graphical extrapolation is straightforward": The day of reckoning is the day the projected curve crosses the time axis.
"In experienced hands," Voight writes, "the method provides a way to improve the art of prediction at particular volcanoes, and may allow extrapolation of predictive patterns to volcanoes elsewhere."
But the method has an Achilles' heel: Its accuracy ultimately depends on the precision and frequency of the observations, which means, Voight says, "You gotta go monitor the volcano.
"And preferably, you have to get there—you want to get there, you'd like to get there—well before it erupts.
"And certainly if some eruptive activity starts, even if you don't have a monitoring program beforehand, if something's going on, you're getting a little bit of steaming, you think, 'We're going to have an eruption soon!'—you'd like to get there very fast, because the problem isn't only to predict an eruption, but to predict an eruption of consequence. "Not all eruptions are dangerous."
In April of 1989, Voight was called back to Colombia by the United Nations' Disaster Relief Organization.
Galeras, a volcano overlooking the northernmost Incan ruins, was belching and shaking, while on its slopes, in Pasto, a city of 400,000 and capital of the coffee-rich province of Nariño, "People were selling out. There were runs on the banks. They remembered Armero."
Galeras, some 300 miles southwest of Ruiz, is, at 14,700 feet, a much more accessible volcano—a dirt track goes right to the rim of the crater, the state radio antenna perches on its lip—but it still posed its obstacles to volcanology—not glacial crevasses or extreme altitude, but land mines laid to secure the communications yard from guerrillas. "You stepped where the guy ahead of you stepped," says Voight, "when you wanted to take a leak, when you were setting up the Geodimeter—you didn't just shift over a foot without worrying about it." Voight and Dick Janda, a USGS geologist and Penn State Alumni Fellow who has worked often with Voight since they were at field camp together in 1959, studied the volcano for a month. By June, they had defined the "zones of lethality" for Galeras and had prepared a hazard map for the area. Pasto, they decided, was not in the "high-hazard sector," although some populated areas nearby "appeared to be at risk."
"Some tricky geography around the volcano makes it unlikely the city will be hit," says Voight. "But we also know enough to know that we can get fooled."
Galeras erupted while Voight was there—"From Pasto it looked very bad, a big ash cloud," although it did no damage.
But Voight had not predicted it. "It was a phreatic eruption, a steam and ash eruption. We measured the deformation of the crater mound the day before and the day after, got zero. The deformation didn't accelerate. No warning. For a phreatic eruption, the deformation method is maybe useless. We shot down from the upper caldera rim to the base of the inner cone—we would've seen any deformation if there had been any.
"We're not surprised. Worldwide, there's not a whole lot of data on whether deformation accelerates or not during a phreatic eruption, so now we've proved the point—at least for this instance—there's no accelerating deformation.
"But the monitoring network is in the right place for a prediction. We made sure we got the baseline data. It's in the bag of tricks for the next time, let's put it that way." On his two-hours' leave in Pasto, Voight bought a watercolor of the city and its volcano.
The Indonesian island of Java has been called "one of the most densely populated agricultural areas on the globe," and the volcano Merapi, a squat, notched 9,500-foot pyramid looming over the island's rice paddies, "the most regularly lethal volcano in the world."
In the past 200 years, Merapi has erupted 30 times, and each time the end of its slumber has been marked by a seeping of fresh magma and the slow growth of a dome, culminating in a violent, explosive eruption with flashing, fulgurous clouds and devastating mud lahars. From 1967 through 1984 Merapi was awake and active; a cat-nap followed until September 1986, when the volcano again began building a dome. The last big eruption, in the 1930s, killed about 1,400; today, a million people live on the volcano's slopes.
The volume of the current dome of lava building inside the crater is greater than 5 million cubic meters. "This represents a current source of concern in regard to volcanic hazards," Voight wrote to the National Science Foundation's Division of Natural and Manmade Hazard Mitigation. "Experience indicates that hazardous eruptions at Merapi are associated with dome volumes greater than 3 million cubic meters."
In 1988, Voight traveled to the Merapi Volcano Observatory in Yogyakarta, the city at the foot of the volcano; he was sent by the USGS's Volcano Hazards Coordination Project and the US-AID's Office of Foreign Disaster Assistance to evaluate the hazard of the volcano's steep south flank, alongside of which the dome was sitting, tilted rakishly.
As Voight reported to the USGS on his return, "Merapi is believed by the Volcanological Survey of Indonesia to be Indonesia's most dangerous volcano"—of the country's 120 known active ones—"yet no deformation monitoring program was found to exist." Voight had established one, hiring porters to carry concrete up to the crater so that he could cement benchmarks for the portable reflectors into position; fixed reflectors, he noted, would have been stolen. "Surveying on the summit of Merapi at dawn is marvelous," he later wrote. "An array of dark cones—Sumbing (9,000 feet), Sundoro (7,500), Dieng (7,500), and Slamet (9,000)—pierce the sea of clouds, marking the spine of Java. A golden band rims the horizon. From small villages far below rise faint cascades of silvery notes in richly harmonic chords—gamelan orchestras greeting the dawn.
"Also, up there," he adds, mocking his own sentiment, "the equatorial climate is endurable."
Merapi, Voight is convinced, is a place where his volcano prediction method will work—if only he can get back there in time to site more instruments, get good baseline data, and arrange for frequent, perhaps automated, monitoring before the lid flies off.
"I've always tried to work in places that are beautiful," he says. He is seated, still for a moment; he places his hands square on his knees. "Volcanoes have a 'terrible beauty,' they say, and of all the volcanoes on earth—"
Huff.
"If you wanted to run out of time by this particular method," he says, "Merapi is a good place to go." He leans to the floor and pulls a heap of papers out from under his feet.
In 1983, a reporter asked Voight about his campsite at Mount St. Helens: If you had been there—then—would you have died? Voight had answered, softly, Yeah. The reporter: Why were you so close? Voight: There were lots of people in close. . . . You had this volcano, and it was close to populated areas. . . . You're willing to take certain risks for certain purposes.
"I don't have a long residence time at these volcanoes," he says now, tossing the stacked papers a few inches to one side, "what's the term?" He springs up and begins sorting through a filing drawer. "Here—the 'encounter probability.'
"My 'encounter probability' is pretty low.
"Volcanologists in general don't fare too badly—at Mount St. Helens, they live with the volcano, they know when to keep away.
"No, that's not right, is it." He smiles and looks away. "Volcanologists never know when to keep away. The only reason they're not there is they can't afford to get there."
In July 1989, the National Science Foundation's Division of Natural and Manmade Hazard Mitigation funded Voight's quarter-of-a-million dollar proposal to test his method for prediction of volcanic eruptions on Merapi.