Bay Bridge

Caltrans reports show engineers misrepresented or ignored test data

A recently released California Department of Transportation investigation, which concluded that testing errors and falsifications of data did not affect the safety of state bridges, revealed for the first time that improprieties extended beyond one rogue technician.

The report shows that four state engineers misrepresented or ignored data or other "consequential" information on four freeway structures, including the Benicia-Martinez bridge. Two still work for Caltrans.

Caltrans conducts tests to help verify the safety and stability of bridges, retaining walls and other freeway structures.

Caltrans declined to make any official available for an interview or to respond to most written questions. Caltrans executives previously said that other records for structures with suspect testing had been examined, and deemed all to be safe.

In addition, in a written statement, Caltrans Director Malcolm Dougherty said that "Going forward, Caltrans has changed its rules and practices to ensure the integrity of the department's work into the future."

The Caltrans report was produced in response to Bee stories about problems in integrity testing involving the new eastern span of the San Francisco-Oakland Bay Bridge and other structures. Duane Wiles, a technician fired by Caltrans in 2011, admitted falsifying foundation testing data.

The report, released last month, examined years of data from tests conducted by the Caltrans Foundation Testing Branch. The Caltrans experts who wrote the report concluded that engineers who analyzed and certified data collected by Wiles and other technicians improperly added, changed or omitted data or evidence of equipment failures in their analysis or descriptions. The Caltrans report did not state that the engineers falsified data, nor did it name them. But it exonerated technicians from wrongdoing in those cases, and flagged improprieties in the engineers' analyses and reporting.

The Bee obtained signed test records for each of the projects cited in the Caltrans investigation. Those separate test reports show the identities of the engineers who prepared them: Toua Vang and Tejinderjit Singh, who were unlicensed at the time of the tests. Both are still employed by Caltrans and Singh is now licensed, according to state records. They worked with two licensed engineers, Michael K. Harris, who has since retired, and Constantin I. Mercea, who now works for a different state agency.

Mercea referred questions to Caltrans. Vang and Harris declined to comment. Singh would not comment without Caltrans approval, which was not provided. Caltrans did not respond to requests to confirm their identities.

In his written statement, Dougherty called the report "the most exhaustive study in the Department's history," and that it "identified every possible issue, no matter how significant or small." A federal peer-review panel praised the investigation as thorough and well conducted.

The Caltrans report authors conceded that the loss of thousands of essential data files due to poor archiving practices, unresolved technical questions and time constraints means the full story will never be learned about the number of testing lapses associated with the Caltrans Foundation Testing Branch, and their significance.

State Sen. Mark DeSaulnier, D-Concord, chairman of the Transportation and Housing Committee, said Caltrans faces a larger problem of credibility in such self-examinations.

"Until there are external reviews of Caltrans you will continue to rely on an agency investigating itself," he said in a written statement. DeSaulnier last month introduced Senate Bill 486, which would move the agency's audits and investigations unit into the California Transportation Commission, and require oversight reporting to the governor, the Legislature and the public.

"This will introduce transparency and accountability without increasing costs," he said.

For the recent investigation, Caltrans experts reviewed thousands of data files and reports involving tests of reinforced concrete piles that form the foundations of freeway structures.

Technicians test piles by lowering radiation-emitting probes into tubes cast near the outer edges of a pile. The report noted that in a few cases, instead of testing each tube – essential to accuracy – Wiles retested a single tube and submitted the data as if it had come from a second tube.

The authors of the report wrote that they suspected numerous such "reruns," among other problems, but limits imposed by Caltrans on the scope of their work sometimes made reaching firm conclusions impossible.

Another problem described in the report involved failures by test instruments to record data accurately. Normally, the probes measure concrete density 10 times per foot for each pile. Ten files from five structures showed a significant number of "null values," in which no data were recorded – possibly the result of a technician error, according to the report. The gaps in data represented untested sections of a pile.

Null-value readings were first described in a Bee examination of data files published in August. The Caltrans report, which looked at The Bee's findings and similar examples, found three "consequential" cases. They included a retaining wall and freeway undercrossing, both in Southern California, and a freeway connector ramp near Oroville. The Caltrans report found that null values were either ignored or omitted, or false data were swapped in to replace the missing readings during data analysis and test-report writing – tasks conducted by engineers.

"The omission or replacement of null values in these eight files is considered consequential and may have significantly impacted the engineering interpretations presented in the (test) Report," the Caltrans investigation noted. The original engineering reports contained no mention of the null values or their possible significance.

In a written statement, Caltrans spokesman Will Shuck dismissed that conclusion by the agency's experts.

"A null value is not a finding, it is a computer glitch that provides no useful information," he wrote. "We have not found a single case where an engineer replaced a null value with an actual, usable number. Where no data was gathered, no data was reported. There were no misrepresentations regarding null values."

Data files from the tests in question show some large gaps due to null values – in one case, most of a three-foot section of a 42-foot foundation pile. None was noted by the engineers, leaving the impression that the data were normal. According to the Caltrans investigation and the agency's testing rules, engineers should have reported the data gaps as potential flaws that required re-examination, and if needed, repair.

In another case, involving the Benicia Bridge, engineers copied a portion of data for one testing tube and pasted it into the data file for a different tube. They did so to replace incorrect data apparently caused by a malfunctioning test probe, suggesting falsely that no malfunction occurred, according to the Caltrans report.

The report did not include interviews with Wiles, his former supervisor Brian Liebich, or any of the engineers. Wiles and Liebich have been suspended by the Federal Highway Administration from working on federally supported jobs, and that agency has begun proceedings that could ban them from such work for up to three years.

Doug Hecox, a spokesman for the federal agency, said in an email that sanctions for the four engineers, if any, would be up to the FHWA's Office of Inspector General or Caltrans.

Caltrans officials did not respond to questions about whether the engineers had been interviewed or disciplined.

Willful data falsification is punishable by fines and imprisonment, according to the FHWA. But Wiles was not prosecuted and Caltrans allowed him to retire with full benefits after he appealed his firing. Liebich, who formerly led the testing branch, also was fired. His appeal is pending before the State Personnel Board.