The Santa Clara Valley Water District board members continued their examination of the Bay Delta Conservation Plan with a workshop session on November 14 that focused on Delta levees, as well as the BDCP’s effects on the ecosystem.
During the workshop, Martin McCann and Curt Schmutte gave presentations on the Delta levees and seismic risk. They were followed by Chuck Bonham, Director of the Department of Fish and Wildlife, and David Zippin and Jennifier Pierre from ICF International. This post will cover Martin McCann’s presentation; Curt Schmutte’s will follow later this week. The remainder will be covered in the upcoming weeks.
Dr. Martin McCann is President of Jack R. Benjamin & Associates, Inc. and a consulting professor of Civil and Environmental Engineering at Stanford University. He is technical director of the Delta Risk Management Project (DRMS), an effort by DWR to assess the risks associated with levee failures in the Delta and Suisun Marsh.
Martin McCann began by saying that his talk would have a narrow focus on the seismic risks to the Delta, noting that there were a lot of other aspects the other speakers would be addressing.
“Needless to say, what goes on in the Delta – in it, about it – is very complicated,” he said. “Technically complex, socially complex, and politically complex – by just about any metric, it’s not an easy problem. But it is possible, I think, to talk about the seismic risk problem in it’s very fundamental aspects so that you can appreciate the ingredients, appreciate the vulnerability, and appreciate the scale of the problem in terms of what levee failures in the Delta could mean.”
The issue is really quite straightforward if we stick to those fundamentals, he said. “From my perspective, looking at things in a risk framework puts an order to things. My experience in the Delta and elsewhere suggests that a non-risk framework often leaves loose ends technically from a communication perspective, and a risk framework puts a theoretical as well as a pragmatic framework to the events that could take place and how likely they are.”
A risk analysis answers certain basic questions like what can happen, how likely are these things to occur, and how much confidence is there in the assessments, Mr. McCann explained. “Fortunately, in a lot of ways, talking about events like this in a risk context is inherently familiar to people. They understand risks, whether it’s from a betting perspective, or whether it’s from the perspective of your doctor who’s explaining the options to deal with a certain health problem … in all of these cases, there are benefits, there are costs or downsides, and there are intangibles that have to be dealt with. A risk conversation puts things on a platform that’s really understandable to a lot of people.”
That being said, with more complex problems and low probability/high consequence events, things get to be a little more difficult to understand, he acknowledged. “Things aren’t always as intuitive as one would like. There’s been a tremendous amount of study that makes it clear that people often have a hard time just simply cognitively understanding and dealing with uncertainty so again, a risk analysis provides that framework.”
So what are the elements of a risk analysis for the Delta? “The first is that we have some kind of an exposure to some hazard or some challenge to the system, in this case, earthquakes,” he said. “Second, we have a vulnerability. We have something that might not perform as well as we would like. If things fail, we have consequences and certainly in the Delta we have a far ranging set of consequences, both within the Delta, outside the Delta, and throughout the state. All of those things have some chance or likelihood that they would occur, and that they would collectively occur or conspire, if you will, to lead to very adverse consequences. The Delta has all of these.”
In California, we tend to think mainly about the large faults – the San Andreas Fault, the Hayward Fault, and the Calaveras Fault, and those are clearly highly active with the possibility to generate large magnitude earthquakes, he said. “But if you sit back and look at seismic hazards in general, literally every place in the US is exposed to the potential of earthquake occurrences and strong ground shaking,” he said. “Now that’s an important point because one of the ideas we often hear about in the Delta is that we’ve never had an earthquake in the Delta of significance and we’ve never had a levee failure. That’s because you didn’t have the earthquakes to cause the levee failure, but when you look at the basic science and the data that we have in the Delta specifically and around the country in general, what we find is that there is geologic seismic evidence for these events to occur. The issue that it comes down to is how likely. That frequency does change rather dramatically.”
Unfortunately, there are two problems with the Delta: the potential for earthquakes in the immediate vicinity of the Delta and the occurrence of earthquakes on the periphery, such as the highly active faults in the Bay Area, he said.
He then presented a map from the USGS that was a graphical depiction of the seismic risk for the nation. “It is simply a graphical depiction of the fact that literally everywhere in the country is exposed to the potential of earthquake hazards. This map is for a particular frequency of occurrence of earthquakes, so it tells us the likelihood, and the colors tell us the magnitude. Obviously California and a few other places have the darker redder colors and the potential for higher ground shaking at this probability of occurrence, but everywhere has a potential for some probability of occurrence of certain levels of ground shaking.”
He pointed out that in the eastern U.S. where the frequency of occurrence of earthquakes are low, they are still considered the dominant contributor to the likelihood of a core damage accident in nuclear power plants. “Whether you’re talking about Pennsylvania or New Madrid, it really doesn’t make any difference,” he said. “Earthquakes are a dominant contributor and that’s because their frequency of occurrence is high enough that it is a potential hazard.”
He then presented a slide that had two maps; one showing the location of active faults and their historic seismicity in the greater Delta and Bay Area, and the other map focusing on the Delta and showing the earthquakes from the mid-60s to present. “By and large, the earthquakes that have occurred since the 60s have been small, less than magnitude 5, but from a geologic and seismologic perspective, the occurrence of these earthquakes along with other data is part of the evidence that’s built up to identify the sources of earthquake occurrences in the future, their potential magnitude, and more importantly, their rate of occurrence,” said Mr. McCann. “So while the Delta clearly has a lower rate of earthquake occurrences within the Delta, or in the immediate vicinity of the Delta, earthquakes do occur, and it’s only a question of what their rate is, when we talk about earthquakes of engineering significance, greater than magnitude 5.”
“So earthquakes do occur and there is a presence of a seismic hazard,” he continued. “It almost seems self-evident but it is a rhetoric that we often here with regard to the Delta and seismic risk that they haven’t occurred. Well, they are occurring and the consensus is really unanimous that we do see earthquakes and we can see earthquakes of substantial magnitude that would be of engineering significance to the Delta levees and all other structures in that area.”
The next part of the seismic risk problem is the vulnerability, he said. “In the Delta, there is roughly 1100 miles of levees on approximately 60 islands, depending upon how you count, and in a very simple but a very real way, these levees are not engineered and certainly not engineered even if they’ve been upgraded from a seismic perspective, by and large.”
Delta levees face two levels of vulnerability. “One is the vulnerability that they might actually fail immediately during the strong shaking; the shaking is simply strong enough to literally knock the levee down,” he said. “Far more problematic is the fact that if the levee survives that level of loading, the chance is that it’s going to be damaged. And that damage is far more problematic because it’s more likely that it’s going to be more far reaching in terms of the extent.”
“Unlike a dam where you can lower the pool after the earthquake, meaning remove the load, we don’t have that opportunity in the Delta,” he said. “You can’t remove the water on the water side the way we can for a dam, so there’s going to be a constant load resistance problem that the levee has to deal with and the likelihood that they survive that all of the time is not very high.”
One can make the argument that embankments like the levees do perform quite well in earthquakes, Mr. McCann said, noting that there are hundreds if not thousands of small dams around the country that have experienced strong ground shaking that performed well, even though they had no seismic design when they were constructed. 
“But there’s one exception,” he said. “That exception is if the embankment or foundation is vulnerable to liquefaction. If it is, than by and large, most professionals will say all bets are off. The devil will be in the details of the character of those soils, and so on, but it’s almost like an on/off switch. If liquefaction is a possibility, the level of vulnerability drops tremendously and the expectation of moderate to good performance is now off the table.”
He then displayed a slide that displayed the makeup of soils in a Delta levee. “We have liquefiable soils in the embankment itself, we have liquefiable soils in the foundation, and either one of those can lead to a failure or a level of damage during the earthquake that makes the levee vulnerable if not outright breach it during the period of strong ground shaking.”
He then showed a picture of a damaged levee in the Imperial Valley as the result of the 1940 El Centro earthquake. “Two points to take away from this picture … The first point is the amount of settlement that’s taken place – it measured 7 feet. The second thing to note is that as far as you can see in the picture, you have damage,” said Mr. McCann. “Now in this particular case, not unlike the Delta, the water has not overtopped the levee as a result of this damage, but left unattended, the levee is highly vulnerable, because of new leakage paths that will be created by the cracking that has occurred and the instability that would result as a result of this damage.”
He then showed a slide with two pictures from Kobe, Japan, noting that the picture on the left was the levee prior to the earthquake, and the picture on the right is the levee after the earthquake, the damage a result of liquefaction. “The levee was dry during the earthquake and you can see the extent of damage that has taken place; it’s there as far as the eye can see,” said Mr. McCann. “So in the context of the Delta, where we’ve got 1100 miles of vulnerability, on any given island where we may have tens of thousands to over 100,000 feet of levee, if we think of the levee as a chain with links in the chain and we’ve got 1000s and 1000s of feet of these links, the potential opportunity for damage and then post-earthquake breaching as a result of damage of this type is really quite substantial.”
Delta levees aren’t really levees as much as they are dams, he said, presenting a slide showing damages to dams during earthquakes. “Levees are typically dry but our Delta levees have water against them all the time, and so they behave like dams,” said Mr. McCann. “We can certainly have the circumstance shown here on the right hand side which is not dissimilar to the Delta situation, which is a dam that is hydraulically filled during its construction, meaning the soil is placed with hydraulic means, and it is highly susceptible to liquefaction during an earthquake, and so the Sheffield Dam in fact slid and breached during the 1925 Santa Barbara earthquake.”
The potential for damage if the levee doesn’t breach is the potential for damage, he said, pointing to the picture on the bottom right of the lower San Fernando Dam. “That dam experienced tremendous slumping and failure of the embankment during that earthquake,” he said, noting that it was fortunate that the reservoir pool was down and that an intake tower survived that provided the opportunity to lower the pool even further soon enough to prevent a dam failure.”
Because the Delta islands have subsided and are below sea level, when a levee fails, there is an influx of salinity onto the island and into the Delta as a whole, said Mr. McCann: “Depending upon a number of factors, the islands that breached, the volume of those islands, the time of year, where those islands are located in the Delta, and then ultimately the timing and the order in which repairs are carried out after the breaches, all effect the impacts to the Delta water quality and ultimately to water exports.”
The severity of salinity intrusion is a function of a number of factors, he said. “In this particular example, roughly 20 islands were flooded, and within a period of a month, literally the entire Delta, the islands themselves, the channels and sloughs, now have become saline and so at that point, we’re not going to be able to pump water to Central and Southern California.”
He then displayed a slide with four graphs, depicting both three island and twenty island failures in both normal and varied hydrology; the graphs plot export deficits in terms of acre-feet against the time until the Delta has recovered. “These figures simply make the point that the impact in terms of just about any metric you want to talk about, duration of exports being disrupted, how many days, months … All of those metrics are a function of those variables that I mentioned, and the charts on this page point out simply two of them. One is how many islands and the examples shown here are sort of two extremes. One is three islands only, the other one is 20; and exactly which three make a difference, and exactly which 20 make a difference, and of course we don’t know what that will be.”
The other factor is simply timing, he said. “Does it occur at the end of the summer? Does it occur during the winter? Has it been a wet year? Will it be a wet year next year? Was it a wet year prior? Was it a dry year?” said Mr. McCann. “The variation in hydrology changes things. And it turns out those two factors alone have the equivalent impact to salinity intrusion and ultimately to exports.”
Mr. McCann then displayed some graphs to give a sense of scale of the impact of levee failures in the Delta. He noted that the results are presented as a function of the number of flooded islands and the volume of flooded islands. The chart on the upper left shows the time to restart exports as a function of flooded islands, and the chart on the bottom right shows the time until the Delta is recovered and is fresh again. “The data points are simply the result of the various random factors that I pointed out earlier that affect how severe these results could be, and it’s pretty significant again looking at the 20 flooded island case, we go from effectively 0 to 600 days. … The variation is really quite significant, depending upon these various factors. Which flooded islands have actually been breached by the earthquake. Where are they in the Delta. What was the hydrology at the time.”
He then presented another chart that plotted export deficits against the number of flooded islands. “The deficits vary rather significantly from effectively 0 to over 6 MAF,” he noted.
He presented another set of charts. “These charts give you a sense of how much of that spread depends upon the time of the year as measured in terms of the position of X2,” said Mr. McCann. “The lime green is where X2 is furthest east into the Delta, meaning the Delta is most vulnerable during that period of time, and the blue and the orange line show you other periods during the year. You can see there’s quite a divergence, and that explains part of that scatter that you’re looking at.”
The chair of the board asks Mr. McCann to explain what X2 is so that people in audience can understand. Curt Schmutte, waiting in the wings as the next panelist, explains: “The X in X2 is the distance from the Golden Gate Bridge, and it’s measured in kilometers, so today, X2 is at 81 kilometers. The 2 stands for 2 parts per 1000 salinity, or about 6% seawater. So today, the 6% sea water line is located at 81 kilometers, and it’s an indicator of basically the dynamic nature of the food web in the eastern or central part of the Suisun Marsh. It’s a salinity that is desirable as a certain biological foodweb production.”
“We can see that as X2 is greater than 84.5 kilometers,” Mr. McCann continued. “It’s now considerably eastward from the Golden Gate Bridge, and therefore there is more salinity already present in the Delta, and with the failure of the levees, there’s that much less that has to happen and therefore more salinity that’s being in effect drawn into the Delta as a result of the levee failures.”
He then presented a similar slide that displayed the average time to restart as a function of the location of X2, noting that the upper left is the average time to restart exports, and the bottom is the same, except with one difference. “One of the factors that determines how severe the event will be is the strategy that’s used in terms of repairing the levees,” said Mr. McCann. “There have been two fundamental strategies that have been looked at, and one of those is referred to as the Middle River corridor, which says you order your repairs to create a corridor on the eastern part of the Delta to in effect, as fast as possible, create a freshwater channel so we can get freshwater down to the pumps.”
That strategy does make sense, but not in all circumstances, he noted. “Below the million acre-foot mark, there isn’t an awful lot to be gained by that strategy,” he said. “On the other hand, once we get to a million acre-feet, we begin to see an advantage, so taking the same amount of repair resources, if we reorder things in terms of we undertake those repairs, if we’ve got more than one million-acre-feet of flooding that has taken place, then we’re gaining an advantage and the difference here is about 100 days, so the times to restart exports improves by 100 days if we invoke the Middle River strategy.”
There is another concern with levee failures that isn’t really talked about much in California and with the Delta, and that is the public health risk, he said. “We learned from Katrina all too clearly that when levees fail, the public is at risk and people do die in these events. And as a result of that, there was a movement to more formally look at the risks the public are exposed to, and tolerable risk criteria have been established,” noting the Corps has now adopted the dam standards to the levees.
The standards determine what the frequency or chance that people might lose their life in an event; there are levels that society simply won’t tolerate, he said. “We all recognize there are risks, but there are levels at which we won’t tolerate it and we’ll either stop the activity, remove a dam, upgrade the dam, etc., and bring it down to a tolerable level. This chart has probability on the vertical axis and loss of life on the horizontal axis. And if you plot the results for the Delta, what you find is the risks to the public are substantially above the tolerable risk criteria that are being used by these agencies.”
“So you can sit back and say that the risk studies that have been done, the numbers are too high or I don’t like the details of how they got there, but at the end of the day, you have to recognize that the details of the numbers are no longer really are the point,” said Mr. McCann. “The point is that there’s considerable exposure to the public, not to mention to the state’s water supply, because that line is one or two orders of magnitude than what would be considered the minimum acceptable tolerable risk criteria that federal agencies and others are looking at.”
“We are inherently accustomed to manage risks in our personal lives, professional lives, and it’s even become the language of choice when dealing with a lot of these types of problems,” he said. “It’s is difficult, but it is easily recognized. We all have homeowner’s insurance that’s going to protect us against a fire at our house. I’ve never had a home fire, I’ve never actually seen one in my neighborhood, but as a risk management approach to things, that’s a really good idea because the cost is relatively low, your consequences extremely high, both financially and the impact to your family and so on. The point being even though things haven’t happened in our direct, immediate experience, the idea of recognizing risks and managing them is rather inherent.”
“It gets more difficult as we begin to talk about low probability, high consequence events, because by definition, low probability events we don’t see very often, and so they are not within our immediate experience,” he said. “We know homes burn down because we have a fire department in town and they do go out and they deal with house fires, but we don’t have them necessarily personally. And that’s very much the case in earthquakes, even here in the Bay Area where we have active faults. Many of us have not experienced strong shaking,” he said, noting that even though he’s been here since the mid 70s, he’s been out of town for every single earthquake. “It’s been very distressing; this is my business and I missed Loma Prieta,” he joked. “So the point is that local probability events are hard to get our heads around.”
“In summary, as a plumbing system that we would like to have high reliability for, the Delta levees simply don’t meet the fundamentals of a reliable system if they are to be the core of our plumbing system,” said Mr. McCann. “They don’t meet reasonable engineering standards and therefore aren’t reliable. And looking at the public safety side, we tend to look at levees for providing flood protection. Dry most of the time, there when we need them when the water rises and protects everybody behind the levees. Delta levees aren’t quite that, they are more like dams, but at the end of the day, as devices for protecting the public, current standards would argue that the risks they present to those behind the levees would not be tolerable.”
For more information:
- To view the agenda and meeting materials for this workshop, click here.
