Dallas' most complex and aggressive public works project ever is mired by the U.S. Army Corps of Engineers' concerns about something in the Trinity River bottoms as old and common as dirt.
Sand, to be exact.
It began gumming up the city's Trinity River Corridor Project just when supporters thought the project was finally unstoppable - on the day in November 2007 that voters rejected a referendum that would have kept the project's tollway from between the river's levees.
Now worries about the floodplain sand will delay parkway construction by nearly two years and add tens of millions of dollars to the project's price tag. Before the corps approves the Trinity Project's toll road, signature bridges, lakes and raised-deck park, it must first understand the extent of underground sand and the potential risks it poses to Dallas' flood protection.
Yet this isn't a story of a sudden, alarming discovery. Engineers have measured and studied sand in the city's floodplain for decades. Rather, internal documents show a slow-moving collision between Dallas' prime directive to "keep the dirt flying" and a fundamental rethinking by the corps of its approach to flood protection.
At issue are basic questions of engineering: How much risk is acceptable and how should it be measured? If, as one Trinity veteran says, rivers are among the world's most complex geological systems, how much data is needed to change a floodplain safely?
The failure of the New Orleans levees after Hurricane Katrina in 2005 prompted Congress to order an examination of levees nationwide. The federal agency launched a top-to-bottom review of flood-control standards and practices - an overhaul that is still under way.
Top corps officials have repeatedly reassured Dallas leaders that the agency is still an enthusiastic project partner. But corps engineers evaluating the project day-to-day have grown increasingly cautious, frustrating local counterparts eager to finish the Trinity project as quickly as possible.
"They're just very, very risk-averse now," said Trinity River Corridor Project director Rebecca Dugger. "They're hesitant to say yes to us. I think they're also hesitant to say no."
The friction became public when the corps announced this spring that the Dallas floodway might not hold back a major flood. The corps explained that not enough is known about floodplain sand - or how it might interact with bridges and other construction in the river bottom to weaken the city's flood-control armor.
Their worry: Water could seep through the sand and undermine the levees. Construction could also create gaps in the hard clay covering the flood works, and a big-enough flood could force enough water through the sand and erode the levees from within.
That's never happened in 70-plus years that the levees have stood between the river and the city, and city officials believe there's no imminent public safety threat.
Yet sand has been an engineering challenge since the city began building the levees in the 1930s. Between 1953 and 2006, engineers found sand in about half of 400 boring samples taken along the river and levees. Since January this year, they have hit sand in an even higher percentage of borings in the floodplain. Most of the sand that's been found lies beneath layers of clay.
For the last half-century, the corps has analyzed the Trinity's sand. At one point in the 1950s, the corps searched for flaws in the levees by forcing water into underground sand deposits - in the very same area that worries corps officials today. They wanted to see whether the levees would leak or blow out, but neither happened.
Then, and for the decades that followed, the corps concluded that sand wasn't a serious or pervasive problem.
Now, however, they're worried about what they still don't know.
So the city recently extended the Trinity project's timetable by 20 months to analyze the soil. That includes 1,500 test borings, double the number done in the floodplain over the last half-century. The $29 million price tag equals what consultants said it would cost the city to raise and refurbish the entire levee system as recently as August 2003.
The corps has imposed stiff permit requirements for construction that their engineers had already blessed. Texas Department of Transportation officials say new permitting will cost at least $1 million each for the Margaret Hunt Hill Bridge and the Interstate 30 signature bridge. Resulting design changes could cost more.
The city, the corps, TxDOT and other agencies say they're still a team and still committed to the ambitious project. They've also tried to keep their differences behind closed doors. The corps recently told the city not to release years of federal comments and guidance on the project. North Texas Tollway Authority officials have withheld their consultants' analyses.
But e-mails and other documents released by the city, the tollway authority and TxDOT reveal their frustrations about how the corps keeps changing rules of the game.
Some officials say they've felt "blindsided" by endless corps what-ifs and policy shifts.
"I think one of the fundamental challenges for us and all the local [flood-control] partners across the country is that the corps is in flux," said Jill Jordan, Dallas assistant city manager.
Hurricane Katrina "was really pretty devastating for the corps, in that they got blamed for things that were not their fault. They've been really trying to make sure that that never happens to them again. They are fundamentally questioning all of their guidelines, all their design criteria."
"We have a major levee system," Jordan said. "We don't want to wait three years before we figure out what [the] study methodology is going to be."
Kevin Craig, the corps point man for the Trinity project, said the devastation in New Orleans "made the corps realize that piecemealing projects together is not as safe a way to do things as looking at things holistically. We've identified the need for national consistency in the way that we are approaching, inspecting, approving and analyzing projects. ...
"We're looking at things much more stringently."
As floodwaters still swamped New Orleans, a corps official reassured North Texans that Dallas' flood protection worked. "We're not worried about it failing," senior corps engineer Gene Rice told The Dallas Morning News in September 2005.
But corps officials quietly began scrutinizing projects once considered routine.
Among the first along the Trinity was the new bridge at Hampton and Inwood roads. The corps surprised TxDOT officials with demands for design changes in January 2006, more than a year after receiving design plans and months after TxDOT hired a contractor to build it.
By mid-2006, the corps was questioning everything from bridge heights to parkway penetrations of the levee and excavations planned in the floodplain. That prompted meetings in Washington and large meetings of engineers in Dallas.
City and consulting engineers feared the corps was becoming increasingly inflexible, and that could complicate or even kill key project plans, e-mail records show. Edginess was apparent in minutes of workgroup meetings over the summer of 2007.
In one, tollway engineers offered a database of about 400 soil borings from 26 studies conducted along the river since the 1950s. Rice countered that more sophisticated data would be needed to properly evaluate parkway plans.
A tollway engineer asked: Couldn't the corps give a little and offer even a preliminary OK, so that designers could move forward with plans to excavate floodway lakes and use the dirt for the toll road and levees? A second corps engineer responded crisply: Corps headquarters was calling the shots, and headquarters wasn't saying anything "at this time."
Rice finally said that the corps' concerns about sand and water seepage could be allayed if the tollway authority built diaphragm walls in areas where the parkway penetrated the levees. Tollway officials estimated that fix alone would cost $45 million.
Construction began that summer on the Margaret Hunt Hill Bridge, and parkway supporters won a referendum in November 2007 to keep the parkway between the levees. Mayor Tom Leppert soon declared the parkway would be built by 2013.
Though no one would realize it for months, the mayor's new slogan - "Beat 2014" - was doomed on the day of the Trinity vote.
A TxDOT inspector's construction diary told what happened. As bridge construction crews drilled into the floodplain to build pier 6-D, their hole began crumbling 12 feet down.
It was a common bridge-construction problem, fixable by using temporary steel casing to keep the hole open. The required casing didn't come before nightfall, and the contractor decided to backfill the hole temporarily and move on to other piers.
On Dec. 5, after rain delays and holidays, pier 6-D was finished with only one more complication: Two cranes pulling together couldn't budge 14 feet of casing from the pier hole. So the expensive steel was now a permanent part of the bridge, at contractor's expense.
Four months later, corps engineers watched drilling of bridge piers into the west levee. The observers grew alarmed when they saw drillers hit sand about 42 feet down - 12 feet below the levee structure, according to Craig, the corps' Trinity chief. Craig said the engineers asked questions and heard anecdotes that made them think there were serious sand issues with other bridge piers. Construction and TxDOT officials were later unable to confirm those anecdotes.
One corps engineer quickly arranged for 11 new test-borings near the bridge. Her late May e-mail to two city staffers explained why: The discovery of large amounts of sand beneath the levee during bridge drilling "is a significant levee safety issue," so the corps needed to figure out how much sand was there and "design an appropriate fix."
It sounded routine to city staffers - yet another round of borings and corps design tweaks. Dugger, the project director, later said she didn't hear of those borings until the following year. "It just didn't raise eyebrows," she said.
By summer's end, data from the 11 borings suggested there was far more sand near the bridge than corps engineers had previously realized. It wasn't clear where the sand ended, either, and that deepened the corps' concerns, but the federal agency wouldn't alert the city or TxDOT for months.
"We wouldn't have had a whole lot of additional information to give," Craig said.
In mid-November 2008, representatives from the five local, state and federal agencies in the project gathered for what was described in minutes as part pep rally and part wonk session.
Corps Brig. Gen. Ken Cox gave a gung-ho welcome, pointing out the group's coffee mug bearing a project model and the agencies' names. "The corps is committed to this project," he declared. "It's a major project for us."
As the group hashed out a new partnering agreement, someone said agency leaders needed to make clear to staff members: "This is a new day."
"Not saying, 'No,' " their agreement declared, "but saying, 'How do we get to yes?' "
That lasted about three weeks. Then a corps e-mail hit City Hall.
Rice, the senior corps engineer, wrote on Dec. 3 that he and his colleagues had just met to review construction related to the Margaret Hunt Hill Bridge. The upshot: Sand caused "significant problems" in drilling a bridge piling and posed "significant risk" to the floodway. Without big fixes, these were "show stoppers."
Incredulous, Dugger e-mailed her boss. This was the first she'd heard of such problems, Dugger wrote. Worse, the corps had met without including anyone else. "Calling anything 'a showstopper' is totally against our partnering agreement isn't it???" she wrote.
She pressed Rice. He responded that sand pockets were so big, wet and unstable that one cave-in threatened bridge-drilling equipment.
A TxDOT senior regional manager e-mailed Dugger to say that Rice's information was news to his agency. TxDOT and construction company officials would later say that no one recalled any out-of-the-ordinary construction issues involving bridge piers and sand.
"We've even gone back and looked at the photographs," said Tracey Friggle, TxDOT's chief regional engineer. "Looking from a distance, you can't see a big sinkhole or anything."
The news got worse. Officials wrangled for weeks before getting the corps' OK for 300 new floodplain soil borings needed for parkway design. The chief tollway engineer told a City Council committee he was hanging by fingernails to meet the project's aggressive schedule.
And on Feb. 6 - a late Friday night - the corps' deputy district engineer e-mailed City Manager Mary Suhm that a report from the corps' most recent levee inspection was nearly done.
Two years in the works, the corps would declare Dallas' levees "unacceptable." The e-mail said "most critical" problems included sand-related seepage at the Margaret Hunt Hill Bridge and the threat of similar problems around other bridges, utility towers and even the county jail.
City officials scrambled. Most perplexing, city officials later said, was the fact that the corps' last levee inspection, in 2006, had rated the Dallas levees "excellent." Corps officials say the new inspection was more comprehensive and was revised midway through to meet strict post-Katrina inspection standards.
Within a week, Suhm and Leppert were in Washington. In a meeting arranged by Sen. Kay Bailey Hutchison, R-Texas, the corps promised a full-time czar for the Trinity project. Craig, who is the second-ranking civilian in the corps' regional office, would be agency point man. He would have an office in Dallas and brief the mayor and congressional delegation weekly.
The tough inspection report was released April 1, and the corps imposed new restrictions, including a ban on construction within 50 feet of the levees. Despite previous authorizations, the corps told TxDOT that it needed extensive new soil testing and more stringent permits for the Margaret Hunt Hill Bridge.
In early June, the mayor announced the 20-month delay in the overall project and $29 million study to identify risks and floodway fixes.
Though Trinity project supporters are "wondering what's the next shoe that's going to drop," Dugger said, the city is focusing on moving forward. She and other city officials said that their communication with the corps has improved.
"The fact that there is sand, and it is a fact that there is sand out there, doesn't necessarily equate to there being a problem," she said. "We still feel strongly that the Trinity Parkway can strengthen those levees."
She and other city officials hope resolving the corps' questions will take less time and money than expected. But it's unclear how or when the corps will give the Trinity project a final "yes" - much less how expensive that answer might be.
"I wish it could be done tomorrow, but there are rules and processes we have to go through," Craig said. "There is always a desire to lean forward, but never at the expense of safety."