The spire of the new San Francisco-Oakland Bay Bridge eastern span majestically climbs hundreds of feet above the bay, an emerging icon of California's engineering and aesthetic prowess.
Scheduled for completion and public use in 2013 at a projected cost of $6.3 billion, the bridge is the largest public works project in state history. Its designers placed one quality above all others: the strength to withstand the strongest anticipated earthquake.
Yet a Bee investigation has found that the state Department of Transportation technician who conducted key testing to ensure structural integrity of the span's foundation was later disciplined for fabricating test results on other projects. The technician, Duane Wiles, also failed to verify that his testing gauge was operating properly, as required by Caltrans to ensure the gauge's accuracy, before he examined parts of the Bay Bridge tower foundation.
When Caltrans officials became aware of the problems with Wiles they did not thoroughly investigate his earlier work - despite public safety concerns raised by other test employees and an anonymous whistleblower.
Never miss a local story.
Until contacted by The Bee for comment, Caltrans had not assessed Wiles' work on the Bay Bridge tower.
Although Caltrans says the bridge is safe, The Bee's findings raise questions about its structural integrity that are not easy to answer. Outside experts say the bridge tower foundation probably is reliable but suggest further review.
Questions about Caltrans' testing extend to other projects. A Bee examination of nearly 50,000 test documents regarding foundations for bridges, overpasses and other freeway features showed that structures across the state were approved after questionable work by Wiles. In three confirmed cases, Caltrans documents show that he fabricated results.
Wiles also routinely discarded raw data files that provide the best information to detect fabrications.
Federal Highway Administration investigators have launched a detailed investigation of past tests, Caltrans chief engineer Robert Pieplow acknowledged in a recent interview. He said the U.S. Department of Transportation had completed a separate investigation of fraud, waste and abuse in the Caltrans Foundation Testing Branch, where Wiles worked, and that Caltrans was conducting a similar probe. Caltrans would not release any findings.
In a written statement, Pieplow said that for legal reasons he "cannot confirm or deny the identities of employees," but that Caltrans "has identified the full extent of this technician's actions (and) taken appropriate remedial measures."
On Nov. 8, three years after Wiles' fabrications were discovered and three weeks after The Bee contacted Caltrans about them, Wiles and his supervisor were placed on administrative leave, according to a Caltrans spokeswoman.
Wiles declined to comment.
In an interview, Pieplow called the type of tests conducted by Wiles one aspect of the agency's quality assurance program.
"As for the Bay Bridge," Pieplow said in a written response to questions, "the (tower foundations) are safe and it would be highly misleading and irresponsible to suggest otherwise."
Nor, so far, had federal investigators found falsified test data for the Bay Bridge in their ongoing investigation, he said. The Federal Highway Administration declined to discuss its investigation or findings with The Bee.
But Caltrans data and engineering diagrams cast doubt on the adequacy of testing of the foundation for the span's signature feature, the tower, and raise questions about its structural integrity.
In 2006 and 2007, Wiles tested seven of the 13 concrete and steel shafts, or "piles," buried deep in the bedrock beneath the bay to support the new Bay Bridge tower.
Wiles' tests, meant to confirm pile strength, showed results roughly the opposite of those captured by other technicians. In all but one case, Wiles' data showed no significant problems, while his colleagues detected many areas of questionable concrete density that required further scrutiny or repair.
Two internationally known bridge-foundation experts who examined hundreds of Caltrans data files, engineering diagrams and reports at The Bee's request, have questioned whether any of the 13 piles were tested adequately before their approval and the tower's construction atop them. Part of the problem, they said, was that the pile design made the structures extraordinarily difficult to cast and test.
The experts agreed that the bridge tower piles probably are reliable even with undetected weaknesses, because they were "overbuilt" to withstand a quake stronger than scientists believe will occur on the nearby San Andreas or Hayward faults. They disagreed on the degree of uncertainty introduced by testing irregularities and design concerns.
"The pile foundations for that structure would be categorized as highly redundant," said Dan Brown, who runs a small but influential foundation design and testing firm and has received numerous national and international honors for his deep-foundation work. "A structural defect in one or two of them would not really be a game changer."
But Bernard Hertlein, a principal scientist at Aecom Technology Corp., a global engineering and construction firm, and co-author of a reference text on foundation testing, said the adequacy of Caltrans tests and of the pile design raise significant questions with no clear answers.
Defects under the main tower would be nearly impossible to detect now, Brown and Hertlein agreed.
"Fixing the foundation in any significant way is pretty much impossible," said Hertlein, who has tested thousands of foundation piles. "Replacing the piles is not a viable alternative. Trying to beef up those existing foundations, which are already so massive," he added, "is probably not economically viable under any realistic circumstances."
Asked how he would address uncertainties about the bridge tower, Hertlein suggested a formal review of Caltrans engineering and test procedures by top bridge-building experts, "to come up with new procedures that at the very least would prevent this level of doubt on any future project."
He allowed that such a plan, combined with Caltrans' assumption that the completed bridge will be reliable, might not soothe wary drivers who would make about 100 million trips over the span annually.
"That's where the problem gets thorny," Hertlein said. "The only way to reassure the public is to do a complete review, with top-notch bridge designers. To look at the design, to assume you do have as much as a 40 percent-flawed foundation, and try and make an educated guess about how the structure would behave in a worst-case scenario."
The 1989 Loma Prieta earthquake killed 63 people as it pancaked double-deck roadways in the Bay Area, flattened buildings and interrupted the World Series. It collapsed a section of the upper deck of the Bay Bridge, killing one motorist.
Lessons from the disaster transformed road and bridge building. Billions of dollars were spent on seismic upgrades.
State officials determined that the Bay Bridge eastern span, which resembles a giant Erector Set toy, must be replaced. It would be impractical, they concluded, to retrofit the structure to withstand a far more powerful temblor expected to hit eventually.
The centerpiece of the chosen design, a "self-anchored" suspension bridge, connects Yerba Buena Island with the elevated roadway leading to Oakland. Self-anchored refers to a single tower that supports roadways with cables tied to the roadways themselves. It would be the largest such span ever built.
Massive piles up to more than 8 feet in diameter, cast in concrete reinforced with rebar, underpin the main tower. Each one reaches 196 feet below the waterline into the bedrock below. Their upper portions, surrounded by mud and water, are jacketed in steel and welded to the bridge footing.
To build the piles, boreholes were drilled and steel jackets set in place. Rebar reinforcing cages were then lowered into the holes, along with thin steel and plastic pipes set near the edge of each pile. Then the concrete was poured into the boreholes, casting the rebar and pipes in place.
The pipes allow tests of concrete density using two methods - gamma-gamma logging and cross-hole sonic logging.
In the first, Caltrans technicians, including Wiles, passed a radiation source down the length of the hollow tubes. The process can find changes in density that reflect structural defects within a few inches of a tube. Typically, about one in five piles shows anomalies that require further examination and possible repair, according to Caltrans.
Sonic tests, conducted in this case by a contractor, transmit compression waves "cross-hole," between steel pipes, to check average concrete density in the interior of a pile. But for most of the piles tested by Wiles, apparently no sonic tests were conducted.
Builders repair major defects to ensure that piles can withstand extreme stress.
James LoCoco, sales manager for the manufacturer of the gamma-gamma loggers, Denver-based Mount Sopris Instrument Co., called the gauges precision devices if used properly and verified for accuracy. But they go wrong periodically. Caltrans has returned devices to Mount Sopris for repair half a dozen times in the last decade, LoCoco said.
Records show that Wiles was responsible for verifying the accuracy of devices he used. That simple exercise involves examining a concrete block of known density. Expected results mean the device is fine, while unanticipated results suggest a malfunction. Wiles failed to verify the devices, according to Caltrans logs and memos.
For the Bay Bridge tower, six of the seven piles Wiles tested were found to be fine; five of the other six piles tested by different technicians showed significant anomalies.
Likewise, Wiles neglected to verify his devices during nearly all of 2006 and 2007. For some of the structures, Wiles recorded the required verifications, according to a Caltrans memo, but the results showed obvious accuracy problems that he ignored.
A Bee review of Wiles' work records shows that he failed to verify the testing devices for work on at least 25 structures, including several portions of the Bay Bridge, as well as bridges, freeway ramps, pedestrian walkways and large road signs in San Diego, San Jose, Anaheim, Oroville, Placerville and Los Angeles.
In his written comments, Pieplow, Caltrans' chief engineer, downplayed the importance of verifying the equipment. "If it were not functioning, the engineer conducting subsequent data analysis would readily notice" readings that fall outside the expected range, he said.
Brown, the geotechnical expert, viewed the failure to verify as far less troubling than the basic design of the Bay Bridge tower piles. It includes an extraordinary volume of rebar, ostensibly to increase pile strength, but he said that causes unintended effects.
"They put so much steel in it that there is no way in hell for all that concrete to get though" without trapping voids or weak pockets of contaminated concrete, particularly at the edge of the pile, Brown said, calling one section "ridiculous" in its congestion. "It's like trying to squeeze concrete through a screen door."
Other elements of the design made detecting such defects difficult. Each pile alternated steel and plastic pipes, because plastic is necessary for gamma tests and steel offers greater accuracy for cross-hole sonic logging.
Although the tower piles were cast after Caltrans issued rules in 2005 to ensure adequate testing of structural integrity, they were designed earlier.
That design violated key elements of the 2005 rules. For example, gaps between gamma-test pipes should not exceed 33 inches. In these piles, the pipes were cast 37 inches apart, so less of the pile was tested than required. The rules also call for at least 3 inches of clearance between the test pipe and adjacent rebar, which can alter readings. In this congested design, average clearance was about 2 inches; some testing pipes were surrounded by rebar supports as little as a half-inch away.
Pieplow said that the pile design did not contradict current test requirements, noting that Caltrans expects the engineer who evaluates test data to consider the unique character of each pile.
Hertlein, the other testing expert, identified another design problem: The pipes were about 8 inches from the edge of the pile, but the maximum range of gamma testing is about 4 inches.
The point of gamma-gamma tests is to detect casting problems that often occur at the periphery, particularly where rebar congestion occurs. The periphery of a pile performs a crucial role in load bearing and lateral stabilization during a major earthquake. If such flaws were present, they would likely have been missed. (In other structures, such as a large freeway interchange in Southern California, test pipes were placed up to 16 inches from the pile edge.)
More important, Caltrans said it had no records of sonic logging to supplement gamma tests for six shafts examined by Wiles. Therefore, apparently just 5 percent of the volume of those piles was examined for structural soundness - and that small portion was impeded by rebar congestion and examined by Wiles with a device that was not verified for accuracy.
Again, the independent experts' conclusions varied. Brown said that the foundation's redundancy - deploying more piles than absolutely required - would protect against undetected defects, which probably are minor.
Hertlein was less sanguine.
"They slipped up quite badly in their quality control," he said.
Caltrans memos show that Wiles' falsification of test data first came to light in September 2008. It involved an examination of a freeway overpass in Riverside. Jason Wahleithner, an engineer, found the deception while evaluating the data for a report, according to an internal Caltrans email obtained by The Bee. Foundation Testing Branch engineers were troubled by the fabrication and how the branch chief, Brian Liebich, responded.
The following month, Michael Morgan, a Caltrans foundation test engineer at the time, expressed concern to Liebich, in a memo obtained by The Bee, about a meeting soon after the fabrication came to light, in which Liebich told engineers to distribute test assignments evenly among the group's technicians. This included Wiles. Morgan urged Liebich to bar Wiles from testing duties, at least until after other falsifications had been ruled out.
"We are putting the reputation and integrity of the (testing branch) at stake..." Morgan wrote. "We are a public agency and the data and conclusions contained in our reports can end up costing contractors thousands to millions of dollars. Our work also can be linked to the safety of the traveling public." He said that he refused to "participate in a charade" about Wiles' work.
About a month later, he emailed Liebich and Liebich's boss Mark Willian to say that he had taken a cursory look at thousands of data files created over two years to check for obvious falsifications. Morgan wrote that he found suspicious data - "duplicate time or out of time sequence files" that imply someone changed or created files after tests ostensibly were conducted. He suggested a methodical examination to see if the data were falsified.
"The work that has been done so far is inconclusive and barely scratches the surface of what could reasonably be called a thorough or comprehensive search for falsified data," Morgan wrote, noting that he lacked the time to complete the forensic task.
Yet, in a January 2009 memo, Liebich told Willian, who oversaw various geotechnical services for Caltrans, that Morgan's review proved Wiles' had falsified just once, on the Riverside freeway overpass. Caltrans retested before completion of the structure.
Liebich described Morgan's work as a definitive validation of not just Wiles' other work, but "all of the Gamma-Gamma Logging data collected by the Branch."
He also cited an "intensive analysis" by Tejinderjit Singh, another engineer whom Liebich assigned to examine 18 months of Wiles' prior work. Singh found no evidence of other fabrications, Liebich wrote, apparently validating Wiles' claim that he falsified data only once.
Based on the review cited by Liebich, in April 2009 Willian issued a written reprimand to Wiles, obtained by The Bee, calling the fabrication "a critical and inexcusable breech of ethics." He directed Wiles to behave in an ethical manner in the future.
Pieplow said "the individual" who falsified data refused to say why he did so.
Caltrans memos show that less than two months after Wiles was reprimanded, Wahleithner, apparently on his own initiative, found two other falsifications on a freeway sign in Oakland and a heavily traveled I-405 freeway bridge over Braddock Drive on the west side of Los Angeles.
Liebich told Willian in a memo that the additional falsifications had been missed in the first review because they fell outside the time frame examined. But the incidents occurred in April 2007 and March 2008, during the period scrutinized.
In both cases, Liebich wrote, the "pile in question needs to be evaluated for potential safety concerns to the traveling public."
Liebich did not return calls seeking comment, and Willian referred questions to Caltrans public affairs officials. Last week, Liebich was placed on administrative leave and Willian was reassigned to head drilling services, one of the 11 units he previously supervised.
Brown, the foundation expert, said that the significance of false data would depend on considerations unique to each structure, but only severe defects could create life-threatening dangers.
"Systematic flaws ... that's where you would worry," he said. "That could make the thing fall down."
As of last year, the Federal Highway Administration had deemed more than 7,000 California bridges "structurally deficient." Few fall down, even in earthquakes, said Bengt H. Fellenius, a foundation consultant and former engineering professor at the University of Ottawa in Ottawa, Canada, although most were built before today's tests were required. Fellenius and other experts said that designers build in a margin for error that usually compensates for construction or testing lapses.
But uncertainty rises when tests are deliberately fabricated, which he called "a rare thing," adding: "It's like murder."
James Davis, Caltrans acting director for engineering services, said in an interview that the agency assessed the Oakland sign, assuming serious defects in the portion of the pile for which data was falsified. Even so, it would hold during a maximum anticipated earthquake on the nearby Hayward fault, the agency concluded. So no repairs were made.
Davis said falsifications for the I-405 bridge pile were detected before the bridge was completed, and the pile was replaced.
Caltrans memos, logs and test reports about the structure, obtained by The Bee, tell a different story - that long after the structure was built, testing branch chief Liebich warned that it was a safety wild card.
In response to written questions about Davis' account, Pieplow said the I-405 pile "was found safe during multiple analyses over several years. ... Upon learning of the falsified data, another analysis was conducted, and the pile was yet again found to be safe."
Caltrans declined to release documents showing its analyses.
Pieplow defended the review ordered by Liebich shortly after the first case of falsified data was found as "thorough," and said it confirmed "the safety of all structures."
Morgan, the engineer who conducted much of that review, disagreed.
About a year after Wiles was reprimanded, Morgan contacted Dolores Valls, then deputy division chief for geotechnical services and Willian's boss. In an email obtained by The Bee, he wrote that no genuine analysis of Wiles' work had been performed.
Morgan branded Liebich's exoneration of the Foundation Testing Branch - based largely on Morgan's own cursory analysis - "a well-crafted misrepresentation" with conclusions "unsupported by the facts."
Morgan added: "There has never been a comprehensive organized effort to analyze past gamma-gamma data and discover falsification."
Part of the problem involved missing data.
Wahleithner, the engineer who detected the fabrications, asked the manufacturer how to examine suspicious data, and he learned that the key was a special "RD" file, he told a colleague in a June 2009 email obtained by The Bee.
RD refers to unalterable raw data files generated by the test device. If data were in doubt, an RD file could be used to generate complete and accurate results for comparison against dubious data. In other words, if a technician falsified data, deleting RD files would be advisable to avoid getting caught.
Wiles routinely kept only "LAS" files - simple text files that anyone can change or edit - to display the results. He nearly always discarded the RD files.
Caltrans records show that other technicians retained RD files irregularly at best.
The manufacturer "pretty much read me the riot act when I told them we don't save the RD files," Wahleithner continued in his email.
Pieplow said that in 2008, after the first falsification was detected, Caltrans began to require that all RD files be retained - and that all test employees complied. But Caltrans archives indicate that only a small fraction of RD files were retained until July 2009. Records also show that in at least 30 cases, tests took place but no files - RD or LAS - were saved. In 85 other cases, no files were retained, but it's unclear whether testing took place.
Pieplow said that federal investigators delivered a progress report that their probe has not uncovered other falsified data so far.
Wiles kept on the job
In January 2009, an anonymous whistle-blower sent a detailed explanation of Wiles' fabrication problem, along with hundreds of pages of background files from testing logs and other Caltrans files, to the State Bureau of Audits. The Bee obtained a copy of the materials recently.
After the first falsification became known, the whistle-blower noted, other testing branch employees recalled that Wiles earlier had spoken openly "about his ability to change the data or make up for missing data by opening up and modifying the data files" created by the gamma-gamma software.
The whistle-blower added: "Wiles' past statement, combined with the speed and dexterity with which he was able to falsify the data on a recent job where he was caught, would lead a reasonable person to believe that he probably has falsified/modified (gamma) data in the past."
The whistle-blower also documented Wiles' failures to verify his devices to ensure accuracy and urged a full investigation to safeguard "the integrity of bridge foundations traveled over by the public."
A spokeswoman from the State Bureau of Audits said the agency does not comment on unpublished work.
"There generally seems to be no interest in finding out about problems with testing," the whistle-blower wrote to Will Kempton, then Caltrans director, a few months later. The author reiterated comments made to state auditors and warned that if no formal investigation ensued, concerns would be directed to the Association of Drilled Shaft Contractors trade group and to the Federal Highway Administration, because federal funds support many California projects.
Several members of the trade group declined to comment.
Asked whether fabricating bridge data and misrepresenting the extent of the fabrications is a firing offense at Caltrans, Pieplow, the chief engineer, called it "a human resources matter," handled case by case.
Even after falsifications came to light, Wiles kept his job testing piles. Later he was moved into collecting data on rock and soil conditions, and recently was moved out of technician work pending a personnel evaluation. Wiles was placed on administrative leave only after The Bee raised questions about his work.
Although Liebich no longer manages foundation testing, the unit's slogan remains the same: "One test is worth a thousand expert opinions."