War of Nerves

by Jonathan Tucker

  The Sarin used in the M55 warheads had been manufactured at Rocky Mountain Arsenal between 1953 and 1957 in a series of 431 lots. The first 241 batches had gone through two distillation steps, yielding a product that was 92 percent pure and had a low level of acidity. For the remaining 190 lots, however, the chief engineer of the Chemical Corps had decided to expedite production and save money by dropping the second distillation step. These so-called round-out lots of Sarin had a purity specification of only 88 percent. The M55 Action Team determined that heavy-metal impurities and high concentrations of acid in the round-out lots had caused the nerve agent to corrode the rocket’s thin aluminum shell, resulting in numerous pinhole leaks.

  The M55 investigation also found a second explanation for the leaking rockets. In 1965, the Chemical Corps had begun to dispose of the obsolete M34 Sarin cluster bomb, of which some 10,000 units had been manufactured in the late 1950s. Because the Sarin in the cluster bomblets was about 90 percent pure, it was recovered and reused in some of the early lots of M55 rockets. During its residence in the M34 bomblets, however, the nerve agent had picked up copper impurities from the welds. Two or three years after the M34 Sarin was recycled into the M55 rockets, the copper impurities reacted with the aluminum shell in a process known as bimetallic corrosion, causing tiny pits that eventually turned into pinhole leaks.

  By 1968, the defective M55 rockets had begun to leak like sieves, leaving puddles of Sarin on the floor of the storage igloos. Because of the extreme safety hazard, inspectors had to enter the affected bunkers wearing full-body protective suits, even in the midsummer heat. The final report of the M55 Action Team, submitted in March 1968, recommended the prompt disposal of the leaking rockets. About 50,000 of the defective rounds were secretly destroyed by open-pit burning at Dugway Proving Ground. Chemical ordnance specialists dug twelve large trenches and placed the M55 rockets nosedown, covered them with dunnage and gasoline, and set them on fire. The results were disastrous: many of the rocket motors ignited, causing the weapons to fly erratically through the air, spewing their deadly contents and contaminating the ground.

  After this misadventure, the Army decided to dispose of its leaking M55 rockets under Operation CHASE (an acronym for “cut holes and sink ’em”), an existing program for the ocean dumping of obsolete ordnance. Initially limited to conventional weapons, CHASE had begun accepting chemical munitions in 1967. Thirty leaking M55 rockets were stacked inside massive steel containers called “coffins,” which were then filled with concrete and welded shut. Each filled coffin weighed 6.4 tons. In a 1968 mission called CHASE 5, several ton containers of mustard agent and hundreds of M55 coffins were loaded onto an aging Liberty ship, the S.S. Corporal Eric Gibson. The Navy towed this vessel from Colt’s Neck Naval Pier in New Jersey to a spot 200 miles east of Atlantic City and scuttled it in 7,200 feet of water.

  Open-air burning pits at Dugway Proving Ground in Utah were used in 1968 to destroy about 50,000 defective M55 rockets that were leaking Sarin.

  THROUGHOUT THE 1960S, as recommended by the Project 112 report, the Deseret Test Center conducted an extensive series of field trials involving nerve agents and chemical simulants, which were held at sea in the Pacific Ocean and on land in Alaska, Hawaii, and the Panama Canal Zone. Of 134 planned trials, 50 were actually carried out (19 at sea and 31 on land), some of them jointly with Britain and Canada under the Tripartite Agreement. Roughly five thousand U.S. soldiers took part in the ship-based trials and about five hundred in the land-based ones.

  A subset of the Project 112 operation, called the Shipboard Hazard and Defense (SHAD) program, involved six sea trials with chemical agents in the Pacific Ocean from 1964 to 1968, three with live nerve agents and the others with simulant chemicals. The Deseret Test Center’s naval fleet consisted of five light tugboats and two converted Liberty ships, the U.S.S. George Eastman and the U.S.S. Granville S. Hall. These vessels were “citadel” ships, meaning that the portholes and hatches could be sealed and an advanced wash-down system eliminated all traces of nerve agent after a test.

  In addition to assessing the vulnerability of Navy warships to nerve agent attack, the SHAD trials tested procedures for detection and warning, crew protection, and ship decontamination while maintaining a warfighting posture. The Navy feared that an enemy attack with a few VX-filled bombs or missile warheads might put an entire warship out of action, including an aircraft carrier operated by approximately 5,500 servicemen. Seawater could not be used to decontaminate the carrier deck because it would damage the insides of the aircraft irreparably.

  The first SHAD test series, code-named “Flower Drum,” extended over several months in 1964 off the coast of Hawaii. A gas turbine mounted in the bow of the George Eastman generated a cloud of vaporized Sarin that wafted over the ship, while crew members wearing gas masks took air samples from various parts of the vessel to measure contamination levels. The second phase of Flower Drum involved spraying a towed barge with a solution of VX mixed with a fluorescent dye to assess the effectiveness of a water wash-down system for protection and decontamination. In another SHAD trial called “Fearless Johnny,” which took place southwest of Hawaii in August–September 1965, a Navy A-4B aircraft took off from an airfield on the island of Kauai and sprayed the George Eastman with VX (or a simulant chemical) mixed with fluorescent dye. The purpose was to measure the extent of exterior and interior contamination caused by aerial spraying of VX under various levels of shipboard readiness, to assess the impact of VX contamination on military operations, and to determine the effectiveness of the ship decontamination system.

  The Deseret Test Center also conducted numerous open-air releases of Sarin and VX on land to study the dispersion and persistence of nerve agents under various climatic conditions. A test series in April–May 1967, code-named “Red Oak,” involved the detonation of Sarin-filled artillery shells and M55 rockets in Upper Waiakea Forest Reserve, a dense rain forest on the island of Hawaii. In addition, from 1964 to 1968, the Army Tropic Test Center, near Fort Clayton in the Panama Canal Zone, conducted “environmental” tests of nerve agent munitions to determine the effects of tropical climate on long-term storage. Over a two-year period, mines, rockets, and artillery shells filled with VX, and rockets filled with Sarin, were placed on pallets outdoors under ventilated covers and periodically tested for leaks, pressure, corrosion, and agent purity.

  The Army performed live-agent trials under arctic conditions at the 1,200-square-mile Fort Greely Military Reservation, about a hundred miles southeast of Fairbanks, Alaska. This vast military reservation encompassed mountains, glaciers, forests, tundra, rivers, and lakes, and its remoteness made it ideal for arctic warfare exercises and cold-weather trials. In winter, temperatures plummeted to minus 50 degrees for days on end. Between 1962 and 1967, the Deseret Test Center conducted hundreds of open-air trials at the Gerstle River Test Site, about thirty miles south of Fort Greely, to measure the effects and persistence of nerve agents or chemical simulants under subzero weather conditions. These tests were code-named “Elk Hunt,” “Whistle Down,” “Night Train,” “Sun Down,” “Devil Hole,” “Swamp Oak,” “West Side,” and “Dew Point.” In some cases, the Army fired artillery shells and rockets containing Sarin or VX into spruce or aspen forests or open terrain with snow cover to study the behavior of the agent cloud. Other trials involved the dispersal of nerve agents from mines, bombs, rockets, torpedoes, and spray tanks.

  During Elk Hunt, Phase I, Sarin-filled munitions were detonated in place by remote control while cameras filmed the effects of the vaporized agent on sheep or goats tethered nearby. Elk Hunt, Phase II, held in 1965, involved a series of trials to determine how much VX would be picked up by military vehicles and troops passing over contaminated terrain. Army vehicles ran over and triggered VX-filled mines, after which soldiers wearing gas masks and protective suits washed down the vehicles and themselves with decontaminating solution. The vehicles and protective suits were then tested for residual traces of VX.
>
  Although the nerve agent trials in Alaska were shrouded in secrecy, an accident at Gerstle River provided a rare glimpse into the testing program. In February 1966, Army personnel left 200 artillery shells filled with VX and three M55 rockets with Sarin warheads on the frozen surface of Blueberry Lake, a body of water about a thousand feet across in a remote portion of the testing ground. The munitions had been prepared for demolition and then apparently forgotten. When the ice melted in the spring thaw, the shells and rockets sank to the bottom of the lake. In August 1968, more than two years later, the incoming head of the special projects division at the Arctic Test Center heard about the missing weapons and ordered Blueberry Lake pumped dry in the spring of 1969. Workers recovered the munitions from the lake bottom and destroyed them by chemical neutralization.

  In January 1971, the Gerstle River incident became public thanks to investigative reporting by a Fairbanks journalist named Richard A. Fineberg, whose articles sparked concern in Congress over the Army’s cavalier handling of deadly nerve agents. Fineberg determined that the 203 shells and rockets left on the frozen lake had been part of a large stockpile of surplus chemical arms that the Arctic Test Center had begun to destroy in 1964. In response to probing questions from lawmakers, Army officials explained that the missing weapons had been “aggregate leftovers from a number of tests in the past” and had not been destroyed immediately because of “the priority of test operations over disposal operations.” This explanation implied that the lost chemical munitions were a tiny fraction of the total number tested in Alaska. The massive release of nerve agents into the pristine arctic wilderness may have caused serious and lasting environmental damage. In July 1972, animal protection experts at Fort Greely reported the mysterious deaths of fifty-three caribou in the vicinity of Blueberry Lake, where the sunken chemical munitions had been recovered three years earlier.

  In the spring of 1968, the Army was implicated in another accident involving nerve agents, this one involving thousands of deaths. Although the victims were sheep and not people, the incident was to have profound implications for the future of the U.S. chemical weapons program.

  CHAPTER ELEVEN

  INCIDENT AT SKULL VALLEY

  ON MARCH 13, 1968, Chemical Corps technicians wearing gas masks and protective suits were preparing for an open-air release of VX nerve agent at Dugway Proving Ground. The Army’s chemical and biological testing site covered 1,315 square miles of northwest Utah, an area one quarter larger than the state of Rhode Island. It encompassed a varied landscape of barren salt flats, sand dunes, and rugged cliffs, surrounded on three sides by mountain ranges. The base also included an airfield with a 13,000-foot runway that could accommodate aircraft of any size.

  Security at Dugway was tight, and armed guards and military aircraft patrolled the 210-mile perimeter. The main entrance gate was on the eastern edge of the proving ground, an eighty-mile drive from Salt Lake City on U.S. Highway 40 through twisting mountain passes and across the flat desert floor. Visitors encountered a military checkpoint with a sign clearly designed to intimidate: WARNING: DANGEROUS INSTRUMENTALITIES OF WAR ARE BEING TESTED ON THIS POST. CAUTION: DO NOT HANDLE ANY UNIDENTIFIED OBJECTS. REPORT THEIR LOCATION TO SECURITY.

  About one thousand civilians and five hundred servicemen and their families lived in the residential section of the post, near the southern tip of the snowcapped Cedar Mountains. The officers and civilian scientists had detached houses like those in a modest suburban subdivision, while the troops were housed in barracks. Despite the arid climate, the lawns were watered and wild horses often grazed on the lush grass. The scientists at Dugway were mainly college graduates who had majored in chemistry or biology, along with a few Ph.D.s. Most had been attracted by the generous salaries, which compensated for the isolation and harsh weather of the Utah desert. The post had a PX and a movie theater, but the closest town was Tooele, thirty miles away. On weekends, staff members drove their RVs into the hills to explore abandoned gold mines or follow the tracks of the old wagon trains. During the week, they focused on the technical aspects of their work and spent little time ruminating about its broader political or moral significance.

  Roughly fifteen miles west of the quiet residential streets of Dugway, the proving ground began on the salt flats and continued up into the rugged hills. Marked out on the desert floor were circular grids ranging in size from 1,800 square feet to 150 square miles. Tower Grid, for example, had a 75-foot artillery tower and a 300-foot rocket tower on which chemical munitions could be mounted and detonated, releasing cone-shaped plumes of toxic agent that floated downwind over the two-mile-long range. The thirty acres in the center of the grid were instrumented with 3,200 battery-powered air samplers mounted on poles, which measured the concentration of chemical agent at multiple points in space and time. These data could be converted into density maps of the toxic plume and later into application tables and mathematical models of chemical weapons effects, including downwind transport, dilution, and deposition.

  The principle underlying the safety of nerve agent testing at Dugway was that the vast majority of droplets in an agent plume would settle to the ground within a mile of the point of release, while the rest would be diluted to harmless levels by the time they reached the border of the proving ground. Under unusual atmospheric conditions, however, toxic clouds were known to travel long distances. Every month or so, Dugway officials placed a confidential call to the sheriff of Tooele County asking him to patrol Highway 40, about thirty-five miles north of the proving ground, and tell people who had stopped by the side of the road to get back into their cars and keep driving. The sheriff never asked the reason for this request, and Dugway officials never volunteered any information.

  THE LIVE-AGENT TRIAL planned for March 13, 1968, involved the spraying of VX from a high-performance aircraft. Over the previous fifteen years, Dugway had conducted some 1,200 tests in which roughly a million pounds of nerve agents had been released. This particular trial was the third in a series of three to evaluate the TMU-28B spray tank being developed for the Air Force. Exhaustive testing of the full configuration—two spray tanks loaded with VX and mounted under the wings of a fighter aircraft—was necessary so the developers could assess the weapon system’s strengths and limitations under various meteorological conditions.

  Although the morning dawned clear and sunny, in late afternoon the skies over the proving ground clouded over, the temperature dropped, the wind began to gust, and the distant flash of a thunderstorm could be seen on the horizon. The weather conditions were far from ideal for an open-air release of nerve agent, but Colonel James H. Watts, the commanding officer of Dugway, gave the go-ahead.

  At 5:30 p.m., an Air Force jet fighter flew over the proving ground. Mounted under its wings were two pressurized spray tanks filled with 320 gallons of VX solution, weighing a total of 2,600 pounds. The nerve agent had been mixed with a dark red dye to make it easier to observe the agent cloud and measure the droplet sizes. With a crackling roar, the fighter leveled off at an altitude of 150 feet above the desert floor, the height of a fifteen-story building. A series of burning smoke pots marked the flight line, a half-mile upwind of the target grid. According to the test plan, the pilot would release a linear cloud of VX, jettison the empty spray tanks, and then climb to a higher altitude.

  After conducting two practice runs to make sure he was flying in the correct pattern, the pilot received the order to start the release. He opened the valves on the pressurized spray tanks and the nozzles began to discharge two parallel sprays of VX-dye mixture, forming pinkish contrails behind the fighter.

  Five seconds later, at the end of the run, the ejection equipment malfunctioned and one of the tanks failed to drop. The pilot had no way of stopping the pressurized flow of agent, and as the plane climbed rapidly, the roughly twenty pounds of VX left inside the tank continued to spray out up to an altitude of about 1,400 feet. Although the Dugway technicians were aware of this problem, they assumed that the preva
iling winds would carry the toxic cloud along the west side of the Cedar Mountains, allowing it to dissipate harmlessly over the barren flats of the Great Salt Desert. Their work done for the day, the team members packed up their equipment and headed home to dinner.

  Shortly after the cloud of VX droplets had been released, some unusual weather conditions developed. A weak cold front passing over Dugway generated cumulus clouds with strong updrafts that sucked up and retained the oily droplets. The clouds then merged into a broad crescent formation that was carried by the wind beyond the edge of the proving ground. About an hour and a half after the VX test, the wind made a 180-degree shift in direction and began to blow from the west, gusting at up to thirty-five miles per hour. Because of this change in wind direction, the VX-tainted clouds did not remain on the western side of the Cedar Mountains but were carried over the high ridge and into Skull Valley, a desert rangeland twenty-seven miles northeast of the test grid.

  Although the federal government managed most of the land in Skull Valley, some of it was owned by private ranchers. Thousands of horses and cattle pastured in the valley throughout the year and flocks of sheep grazed there from November to May, eating native plants such as cheatgrass and bud sage. As night fell, intermittent showers of snow and rain developed and continued until morning. The precipitation washed the oily droplets of VX out of the air and deposited them on several flocks of sheep grazing in Skull Valley and on the slopes of the Stansbury and Onaqui Mountains that formed its eastern edge. The wind also carried VX droplets through a pass in the Onaqui Mountains and into Rush Valley beyond, some forty-five miles from the test site, where they were rained out onto another flock of sheep. Exposure to moisture degraded the VX into a chemical derivative that, while still highly toxic, was not absorbed as readily through the skin. Nevertheless, the sheep ingested the poisonous substance by eating contaminated vegetation and licking snow, their primary source of moisture.