The boat ride into camisea starts at a low-slung provisions town called Ivochote in the rainforests of southeastern Peru. It is the last place to use a phone, sleep in a decent bed and buy supplies. Open-air cargo boats muscle against the Urubamba River's rain-swollen current. Machiguenga families with pigs, onions, rice and beer wait for ferries that will take them adentro, deep into one of the world's most isolated and ecologically rich rainforests.
Many locals talk about la empresa, shorthand for the $1.6 billion Camisea Gas Project. Backed by a Dallas firm hoping to get Peru's natural gas to U.S. markets by 2010, the Camisea project includes gas wells deep in the jungle, a distribution system in Lima and Callao and two pipelines—one carrying gas and the other natural gas liquids. The project is run by Transportadora de Gas del Perú (TGP), an incestuous consortium partly owned by Techint, the same Argentine company TGP hired to build the pipeline and the same company that made the pipes used in the project.
The project is an engineering feat by any measure. The two pipelines traverse 340 miles of some of the world's toughest jungle terrain, top nearly 16,000-foot Andean peaks and snake down to the Peruvian coast. The natural-gas pipeline turns north to supply Lima's market, while the liquids pipeline turns south to Callao, where it will feed a new $2.4 billion facility that will process the liquids before shipment. Peru LNG—another consortium, headed by the Texas company Hunt Oil (holding a 50% stake) and including SK Corporation, from South Korea, and the Spanish Repsol YFP—built the facility. This "export phase" of the project, known as Camisea II, will bring Peru's gas to international markets.
Hunt Oil, led by Ray Hunt, a Halliburton board member and major Republican Party donor, has requested a $400 million loan from the Inter-American Development Bank (IDB), along with the bank's help in raising an additional $400 million more in private investment. But serious questions about the pipeline phase of the project may be putting Camisea II at risk: Last October the IDB said it would wait for a technical audit of the pipeline before deciding on the new loans. The bank's decision is expected in mid-2007.
As Camisea neared completion in August 2004, its backers spoke glowingly of the project. The IDB had already thrown its support behind the project in the form of a $75 million direct loan and a syndicated loan of $60 million. Bank officials promised to keep Camisea's environmental and social impacts to a minimum, but critics have long said the bank has failed to fulfill that commitment.
On December 22, 2004, fewer than four months after Camisea went on line, a 14-inch section of the pipe ruptured, dumping about 48,343 gallons of natural gas liquids. Eight months later, in August 2005, another rupture occurred but without a substantial spill. Two more breaks occurred the following September and November.
Then, in March 2006, near Echarati, a village in the department of Cuzco, the liquids pipeline broke again, spilling enough gas onto the rainforest floor to fill the tanks of 30,000 cars, and leading to an explosion that severely burned a mother and her child. By then, international environmentalists, Urubamba communities and politicians, development experts and journalists wanted to know why a brand-new billion-dollar pipeline had ruptured so often in such a short a time.
Just days before the explosion, far from Peru's mahogany rainforests, Bill Powers, a California engineer, had addressed a meeting at IDB headquarters in Washington, presenting a detailed engineering report on Camisea. Powers and his team at E-Tech, a California-based nonprofit engineering company, had written the report relying heavily on information provided by Carlos Salazar, a Peruvian welding inspector turned whistleblower who had worked on Camisea and said he saw evidence of Techint's negligence.
The report dropped like a bombshell: Techint had rushed the job, the report said, in order to avoid a $90 million contractual late fee; 40% of the pipes used were leftovers from other projects; welders lacked certification; and pipes were corroding after having been left exposed during the rainy season.
While activist groups, such as Amazon Alliance and Amazon Watch, publicized the E-Tech story, Techint denied everything, at one point producing receipts it said proved its pipes were new. The company also threatened Powers with a lawsuit, and Salazar received threats back in Peru. But Techint's lawsuit dissolved when the March 2006 break occurred at a precise spot E-Tech had told the IDB was vulnerable. "They were ready to come after us, but the break came and showed we were right," Powers said.
In late october, powers introduced me to "Andrés," an engineer who agreed to talk on condition of anonymity, saying he did not want to endanger his family. Andrés worked on Camisea as an inspector. He provided me with his ID issued by Peru's College of Engineers and offered technically detailed accounts of problems along the high-pressure pipeline. He also provided a copy of an inspection report issued by Gulf Interstate Engineering, the Houston-based quality-control company hired by TGP to oversee Techint's work. The report, written when construction was in its final stages, noted that one of Techint's welding teams was unfit—contradicting the company's public insistence that all welders had been thoroughly qualified.
Andrés emphasized that Techint had been unprepared for the jungle terrain, since it had failed to carry out a geotechnical study of the pipeline route. He said the company used a route proposed by Shell Oil when it was contending for rights to Camisea. "TGP took Shell's plan and modified it without even inspecting the route from the ground. They only flew over in a helicopter," he said. Héctor Gallegos, president of the Peruvian College of Engineering, agreed that Techint had failed to gather geotechnical data. "In engineering there is a basic probability equation for the design and construction of a work, which is 'Danger times vulnerability equals risk,'" Gallegos said. "In this case, the danger—the geology and soil—was not known, so it was impossible for them to fix the vulnerability."
E-Tech's report was full of claims, backed by field inspectors contracted by Techint, that the job was sloppy and rushed. The inspectors reported irregular practices at pipeline job sites and provided Powers with a "pipe book," or daily log of construction activities, in which "every piece of pipe, every weld, every welder, and every X-ray of every weld is methodically recorded," Powers said. The pipe book, which pertained to a three-mile section in the mountains, confirmed what Powers had suspected: that "Techint played it fast and loose with required procedures, especially in the homestretch."
"In one case," Powers explained, "Techint didn't bother to study the terrain before choosing a path for the pipeline, only to discover a deep gulch filled with loose soil as the company prepared to lower the now-welded pipeline into the ground." Powers said Techint improvised by hurriedly cutting pipe and mixing and matching 40-foot sections, which were custom-bent to fit the original trench route. "In the process, critical information—who welded what, welds needing repair and the location of the welds needing repair—was lost. TGP was facing major fines for every day of delay beyond the mid-August 2004 completion deadline, and the problem was buried," Powers said.
Techint denied that it had rushed the pipeline job and that it had used faulty materials. The company said the breaks were caused by landslides, which are frequent during the Urubamba river valley's tumultuous rainy season. Robert Montgomery, an IDB representative whom I interviewed in spring 2006, agreed with Techint's explanation.
(Andrés also told me he saw an unpublished transcript of a closed-door Peruvian congressional hearing in which a Hunt Oil representative admitted that Techint had failed to conduct key studies. I contacted Hunt Oil about that claim but did not hear back by press time.)
Back in Peru, the March 2006 explosion, together with the E-Tech report, intensified public attention on the Camisea project. Soon after the explosion, the Peruvian government announced it would investigate. A Peruvian newspaper, La Republica, reported that Peruvian Prime Minister Pedro Pablo Kuczynski, who had adamantly defended Camisea and suggested the fifth break was sabotage, had once worked both as an IDB consultant and as financial assessor to Ray Hunt. The IDB, for its part, fell under criticism for not properly supervising the project and for moving to conduct a long-delayed technical and environmental audit only after the fifth break. Meanwhile, as Peru approached a presidential runoff election, leftist candidate Ollanta Humala pledged to root out Camisea's problems and renegotiate contracts (he lost to former Peruvian President Alan García, who fully backs the project).
By the summer, Carlos Armas Vela, a Peruvian lawmaker, was spearheading a congressional investigation. He and fellow lawmakers complained that the proposed audit was too limited; that it was controlled by the same bureaucrats at Peru's Ministry of Energy and Mines (MEM) who had failed to manage the project in the first place; and that the audit should be a job for the incoming government. Peru's College of Engineers, which had a representative on the selection committee, resigned in protest. Around that time, Fernando Trigos, an engineer and one of Armas' political aides, told me the congressional committee wanted a thorough audit of the pipeline. "They want transparency from start to finish," he said.
The transparent technical audit upon which so much was riding would be performed by a company selected by the government. But which government agency would select it? Camisea critics I spoke to said the obvious choice would have been the Energy Investment Supervisory Organization (or OSINERG, its Spanish acronym), a semi-independent state regulatory agency that had charged Techint several times for various violations. But the task was given to the MEM, which has overseen the spill-prone project from the start.
MEM officials convened a selection committee, generated what critics say were ridiculously limited audit specifications and opened bidding to 11 companies in early summer 2006. This first round of bidding was canceled because the MEM received only one bid, from the Mexican subsidiary of the German company Germanischer Lloyd (GL). The same company, GL-Mexico, won the bidding several weeks later, beating out three other companies.
Not long after the first round was canceled, I went to a Buenos Aires café to interview a Texas engineer who knows Camisea well. With 30 years' experience in jungle pipeline inspections, he had represented one of the 11 companies that tried to win the first inspection contract. He asked for anonymity, fearing professional reprisal, but provided me with the materials issued by the MEM's solicitation committee. He pointed out evidence he says shows the ministry tried to steer the audit to friendly companies. "The fix," he said, "was definitely in on this one."
The documents show Peruvian officials gave foreign companies only 19 business days to translate the 101-page bid specifications, write a proposal and translate it into Spanish for submission. On June 12, the MEM sent the companies formal invitations to visit Camisea on June 20, leaving engineers only eight days to get vaccines and make arrangements. At the last minute, the Texas engineer says, he and another representative of a U.S. company were denied travel permission because their yellow fever vaccines were less than 10 days old. Peruvian officials also gave companies only one day to examine data culled from the June 20 inspection, formulate questions and have them translated. "For foreign companies that was impossible," the Texan said.
What's more, the MEM commission only gave itself three days to evaluate the detailed proposals from the 11 companies, with each proposal more than 100 pages long. "In cases as complex as Camisea, the industry standard is two or three weeks," said the Texan, who suspects Peruvian officials gave themselves only three days because they figured the impossible deadlines would knock out most companies.
Another oddity in the bid was a "point system" based on experience and technical merit, which the MEM commission designed to rank potential companies. But in four of the 11 categories, officials required only that companies had worked on jungle construction projects—there was no requirement that the auditor have any specific experience with pipelines. "And why were no Brazilian companies asked to bid, since they have the most experience building jungle pipelines?" the Texan asked.
Only GL-Mexico managed to get its materials in on time. Any questions about how the company managed to do so only deepened in October when the company won the contract with a bargain-basement $1.9 million. Though it is normal for officials to choose the lowest bid, GL-Mexico's puny offering astounded Camisea watchers. The next lowest bid was nearly $6 million. The Texan told me his team had budgeted 30,000 labor-hours alone to conduct the audit, and expressed serious doubts any company could do the job for $1.9 million. "I have talked to others involved in the bid, and we agree that the job cannot be done for that amount," he said.
The Texas engineer also said he wouldn't be surprised if Techint paid GL-Mexico to bid low, get the job and gloss over any problems. "If Techint gave them $10 million," he said, "it would still be cheap compared to what it would cost if all their dirty laundry is aired. They have already spent more than that on remedial work." There is another theory: that Techint may have promised GL-Mexico work on Techint construction projects under way in Mexico (two new liquid natural gas receiving terminals, for example, together worth $1.2 billion, which Techint has been hired to build).
GL-Mexico said in October it had no business dealings with Techint—an official prerequisite for winning the contract. However, I contacted Olaf Mager, a spokesperson at Germanischer Lloyd's global headquarters in Germany, who said the company "had been working for a subcontractor [of Techint] in the area of verification and certification of various installations." He added that, due to confidentiality agreements, he could say no more. "Something is up," said Héctor Gallegos of the College of Engineers about GL-Mexico. On October 10, his group published an open letter asking the MEM to suspend the audit.
It is unclear what weight the audit will have on the IDB's looming $400 million loan decision. The bank could be putting the decision off because it is looking into irregularities in the auditor selection process. Indeed, the Texas engineer showed me an e-mail in which an IDB investigator was told about the selection committee's impossible deadlines.
The IDB said in October it would await the Peruvian government's technical audit of the pipeline before making a decision on the new loan, a decision the banks says it expects to make in mid-2007. Meanwhile, the IDB is conducting its own audit. But in September a dozen environmental groups, including U.S.-based Amazon Watch, sent a joint letter to IDB President Luis Alberto Moreno saying the process needs transparency. The group complained that the bank's audit is being carried out by the IDB's Private Sector Department—the same unit that the activists say has failed to oversee the project. Angry that the bank is already proceeding with due diligence on the Camisea II loan, they want independent groups like the Netherlands Commission for Environmental Impact Assessment brought into the mix.
Perhaps the most troubling problem is revealed in one of the MEM's solicitation documents from its bidding process. It shows that Techint knew, or should have known, of the gas pipe's vulnerabilities before it began rupturing. An annex to one of the documents lists seven failures during hydrostatic pressure tests, which are standard procedure on all sections of a pipeline before it is put in operation. All five widely reported ruptures, including the one that caused an explosion that maimed the two villagers, occurred in the first 130 miles of the liquids pipeline—the same section of pipe where the seven test failures took place, according to the solicitation documents.
Other concerns center on the thickness of the pipe Techint used, which some engineers say was too thin for the steep, unstable jungle terrain where soil movements can place tremendous pressure on the pipe wall. "I talk to inspection company people, who see a lot of pipelines, and they can't believe that 0.219 pipe [one-fifth of an inch] was used in this location," said my Texas source. Not surprisingly, he notes that Techint's pipe-making company produces pipes of that size.
Also noteworthy is that bid documents do not mention special internal pipe-integrity tests conducted on the pipelines by Tuboscope, a global testing company hired by Techint in summer 2006. That oversight feeds other concerns. The March 2006 break was in relatively flat terrain, not a place where shifting soil would have likely caused problems. That has led some to wonder if Techint is running the pressure too high.
"Abrupt changes in elevation mean big differences in internal pressure inside a liquids pipeline," said Powers. "Were the pressures Techint subjected the pipelines to during the pre-commissioning phase high enough to reveal all the weaknesses, which is the fundamental point of doing these tests, or are the pressures being reached now with a 'roll of the dice' in uncharted and untested pressure territory?"
The MEM did not respond to interview requests. A Hunt Oil representative told me to talk to TGP, as did Techint's headquarters in Buenos Aires. I finally spoke to Eva Izquierdo, the assistant to TGP President Ricardo Markous in Lima. At her request, I submitted a list of questions. I asked TGP how much of the pipe was manufactured by companies controlled by Techint and how much engineers' decisions to use certain grades and weight of pipe were influenced by generating profits for the Techint pipe mill. I asked if TGP had a geotechnical analysis of the pipeline route that was produced before construction started in 2002. I asked the company about the five breaks, and if there would have been fewer if they had used different materials and better quality control or engineering methods. I also asked whether the company would change any part of the design or do anything differently if it could.
Two weeks later, after several e-mails, phone calls and promises that I would get my questions answered, Izquierdo politely e-mailed me saying her boss had read the questions and that all my answers could be found on the company Web site.
What do Camisea's critics want? Is it to end gas extraction, to pull modernity out of the remote rainforests altogether? The people I talked to, those who had risked much to come forward, said they simply want responsible engineering, transparency and oversight. Powers is a tireless, witty engineer with a degree from Duke University and a master's degree in environmental science. Over nine months of contact with him, I watched him become the de facto face of the Camisea opposition. He says he wants to bring professional engineering practices to ecologically and socially sensitive rainforests, where extraction operations are ruled by good-old-boy ways, unchecked by weak and corruptible governments. The same goes for Andrés, who told me, "The time has come to tell the truth about Camisea." In October, as this story went to press, I sent an e-mail to Carlos Salazar, the whistle-blower who arguably put most on the line to raise the curtain on Camisea. He told me the threats had stopped now. He said he has no regrets.