The destruction of Oroville Dam’s main spillway in February likely occurred because it was built on highly erodible rock, according to several experts interviewed by Water Deeply. If confirmed by a forensic investigation now underway, rebuilding the spillway will require a much more expensive and time-consuming effort.
The Oroville spillway was ripped apart in February as California Department of Water Resources (DWR) released water from the dam to make room for heavy storm runoff into the reservoir.
It’s an important reminder that no matter how carefully built and maintained a dam might be, it will always remain vulnerable to unknowns.
“You go up where that spillway was and the upper 40 to 80 feet of soil is garbage,” said J. David Rogers, an expert on dam safety and a professor of geological engineering at Missouri University of Science and Technology. “It’s a special kind of metamorphic mess.”
The water releases in February, however, were not out of the ordinary. In fact, the spillway saw much greater flow during a heavy runoff event in 1997 – without any serious damage to the spillway.
The difference this time, Rogers said, is simply the effect of 20 years of additional aging of the spillway, which may have expanded cracks and seams in the concrete.
Rapidly changing moisture conditions also may have contributed: Weeks of heavy rain, after five years of drought, caused the hillside on which the spillway rests to become rapidly saturated. Photos show drains built into the spillway gushing water, both before and after it broke apart.
Rogers pointed to aerial photography of the disaster’s aftermath that reveals dark red or orange earth beneath the section of the spillway that was ripped apart. This, he said, is a type of rock known as saprolite, which literally means “rotten rock” in Greek.
It’s likely, Rogers said, that water flowing down the spillway entered cracks in the concrete, creating a vortex of high-velocity water that eroded the rock from underneath. This is a failure mechanism known as “stagnation pressure.”
It’s also possible the spillway was weakened from below by water flowing through the drain system under the spillway that eroded the saprolite rock, creating voids that weakened the overlying concrete.
A four-member board of consultants hired by DWR also noted this possibility in a report to the Federal Energy Regulatory Commission.
“It seems likely that piping of foundation material beneath the chute slab may be responsible for the voids that have been found and repaired in the past,” the consultants wrote in the March 10 report.
The report noted that it appeared some of these voids had been repaired in the past by filling them with compacted clay, a material that would not resist subsequent erosion.
The dark red rock and soil that appears to run under and alongside the spillway, Rogers said, is an indication of saprolite: It has been colored by heavy weathering, just as rust turns steel a dark red color. This makes the rock crumbly and prone to erosion.
“That’s one of the oldest exposed erosional surfaces in the state of California,” Rogers said. “They’re going to have to go back in there and excavate all that red stuff out and replace it with roller-compacted concrete. They’re probably going to have to replace the whole spillway.”
Russell Shapiro, chair of the geological and environmental sciences department at Chico State University, agreed with most of Rogers’s analysis. Shapiro toured the damaged spillway on March 13.
“The critical issue is how deeply weathered were the rocks in the upper part of the spillway failure?” Shapiro said. “It is hard to say from only one field day. But the abundance of red [rock] suggests there could have been sufficient weathering.”
Lauren Bisnett, a spokeswoman for the DWR, declined to comment on whether geology contributed to the spillway failure.
“The factors possibly involved in the spillway failure are being investigated by a forensic analysis team,” Bisnett said. “We will not speculate on their work and look forward to their analysis.”
Why wasn’t the erodible saprolite excavated and replaced with concrete when the spillway was built in the 1960s? This would be the recommended method to address such a weakness in a spillway built today. It’s an obvious question with a surprising answer.
Rogers said the erosive risks of saprolite were not recognized 50 years ago when Oroville Dam was built. In fact, it wasn’t even recognized by geologists as a distinct rock type until the 1970s, when it was given the name saprolite.
Geologic maps prior to the 1970s identify the entire Oroville Dam region as a single rock type, a kind of metamorphic greenstone, and make no mention of saprolite. The dam itself is built on this greenstone, considered a very solid type of bedrock material, and some of it can also be seen in photos of the eroded spillway.
It was another dam disaster in 2005 that first raised red flags about saprolite. The Taum Sauk Dam in Missouri overflowed due to a combination of errors, eroding saprolite from underneath the concrete dam itself, causing the dam to collapse.
In February, engineers became concerned about a similar failure at Oroville Dam’s emergency spillway as the disaster unfolded.
The emergency spillway is simply a giant concrete curb, more than 40 feet high, that spills onto a wooded hillside when the reservoir fills to capacity. It is intended as a last-ditch safety valve, and has a minimal foundation of its own and no concrete chute to contain the water.
DWR had shut off the main spillway after it collapsed, causing the reservoir to rise to the emergency spillway as additional runoff poured in. The emergency spillway had never been used before.
As water began to flow over the emergency spillway, it immediately caused heavy erosion in the hillside, which soon began eating a path back uphill toward the spillway lip, raising concerns it would be undermined and topple over. This would have caused a rapid, uncontrolled release of water that could overwhelm downstream levees.
Nearly 200,000 people downstream were evacuated as a precaution in case the emergency spillway failed. And DWR was forced to reopen the main spillway, causing additional damage there.
The hillside below the emergency spillway exhibits the same sort of saprolite-like rock that was revealed under parts of the main spillway.
As a result, DWR’s board of consultants warned it is “absolutely critical” to avoid using the emergency spillway until a complete redesign of the structure can be planned and carried out.
As for the main spillway, an interim fix is possible in time for winter 2018, the consultants said. But a permanent fix will likely take an additional year.
They also said the upper section of the main spillway chute, which remains intact, probably should be completely replaced and rebuilt as well.
Depending on further geologic testing, this may require extensive excavation to remove the weak saprolite, and replace it with compacted concrete.
“It’s a hard lesson to learn,” Rogers said. “We don’t have a lot of experience with it, because you only get these kinds of flows once every 50 or 100 years. There’s not a big body of knowledge on how these older spillways are going to hold up.”
Never miss an update. Sign up here for our Water Deeply newsletter to receive weekly updates, special reports and featured insights on one of the most critical issues of our time.