In July 2018, the Vicksburg Engineer District announced an emergency closure at Columbia Lock on the Ouachita River in the northeast corner of Louisiana. District personnel investigating seepage and sand boils at the structure had discovered voids under a lock wall.
Two and a half years prior, a sand boil had developed inside the lock chamber, with water upstream of the lock forcing its way under the miter gate, through the aquifer beneath the lock and up through the lock floor. Unlike many locks that have a heavy, continuous concrete floor, the lock floor at Columbia is made up of a sand and gravel base layer that acts as a filter, with heavy concrete blocks on top to hold that filter in place. The porous floor was intended to allow seepage between the upper pool and lock chamber.
Columbia Lock provides a maximum 18-foot lift between the lower pool to the upper pool.
“In late 2015, there was the sand boil inside the lock chamber floor, so we knew those blocks and the filter system had become compromised and had displaced some of those blocks,” said Lanny Barfield, dam safety officer for the Vicksburg District. “We now had water running under the lock and boiling up into the lock chamber.”
When Corps officials discovered that initial sand boil in November 2015, they addressed the issue by raising the water level in the chamber to that of the upper pool, much like how the Corps will sandbag around a sand boil on the land side of a levee.
“We raised the pressure and stopped it and then did an emergency repair with our in-house crew,” Barfield said. “We were just trying to get the lock back open as fast as we could.”
That worked until June 2018, when a sand boil formed below the downstream miter gate.
“You can’t do anything about that,” Barfield said. “You can’t raise the water level. That’s just the river level. So now we really had a problem because it was both in the lock chamber and below the lock chamber.”
Columbia’s lock walls have preformed holes filled with sand, likely intended to be a relief well. Barfield’s geotechnical branch team drilled through those in search of voids.
“When we broke through the bottom, our drilling fluid was going into voids, and it was showing up downstream of the lock,” Barfield said. “That was very concerning because it told us that we had a continuous void. We were drilling on the upstream of the lock, which is 600 feet long and 84 feet wide, and within minutes our drill fluid was showing up coming out of a sand boil downstream.”