Flood Risk

Protecting Dam Breach Inundation Zones

Dam failures are a ‘low probability but high consequence’ event best addressed by preparation and maintenance.
By: | February 28, 2017 • 7 min read

After a five-year drought, the rains finally returned to California this winter. Lake Oroville, which was formed in 1967 at the foot of the Sierra Mountains by the nation’s tallest dam, began to refill.

An atmospheric river, colloquially known as a “Pineapple Express,” continued dropping water at such a pace that it replenished the reservoir and then some. The swollen lake forced the excess water onto an emergency spillway alongside the Oroville Dam for the first time in half a century.

The spillway cement crumbled and sent a cascade of water down the mountainside. Engineers feared the erosion compromised not only the spillway but also the 770-foot-tall earthen dam.

An emergency evacuation was hastily ordered and nearly 190,000 residents were forced to flee their homes. Traffic clogged roads. Fortunately, no life was lost and water levels eventually receded.

“There has to be a more robust conversation around flood.” — John Dickson president, NFS Edge Insurance Agency

For many, this crisis is a wake-up call for renewed assessment of the aging infrastructure of the U.S. dam system and the emergency response plans drawn to prevent loss of life in the event of a failure.

Dams Exceeding Effective Dates

Nearly all of the 700 dams managed by U.S. Army Corps of Engineers are more than 30 years old. More than half have reached or exceeded the 50-year service lives for which they were designed. Oroville Dam turns 50 next year.

Tim McCarty, risk control manager, Trident Public Risk Solutions

Tim McCarty is a risk control manager at Trident Public Risk Solutions, which insures a wide range of municipalities including those in close proximity to a dam. His statistics on the aging dam system in the U.S. are even more alarming.

“The average age of our dams nationwide is 56 years,” he says. “And, the average age of a failed dam is 62 years.

“So we’re kind of reaching that point where we’re starting to have some very old infrastructure and if it’s not properly maintained we may see repeats of this type of an incidence,” McCarty said.

The Federal Emergency Management Agency says dam failures are a “low probability but high consequence” event.

Since they rarely occur, many people who live downstream in “dam breach inundation zones” are completely unaware of the potential hazard. But they are, in fact, at the mercy of the dam’s ongoing health.

“The only time the concept is front and center is when water is rushing over the dam,” said John Dickson, president of NFS Edge Insurance Agency.

“I worry constantly that that’s a total disservice to the American people. We need to have the conversation when the sun is shining.”

Advertisement




Dams are a vital part of the U.S. infrastructure. They provide flood protection, water supply, hydropower, irrigation and recreation. But all it takes is one busy muskrat to compromise the integrity and cause a breach.

The operators of the Oroville Dam were advised in 2005 to shore up the spillway with concrete. For a dozen years, the expensive recommendation was tabled. The reasoning was the emergency spillway was never used because the lake water never rose high enough.

“With a dam breach, when it fails, it fails quickly, the water comes out quickly and there’s a limited amount of time.” — Dr. Louis Gritzo, vice president and manager of research, FM Global.

But when it finally did, the emergency response — mass evacuations and helicopters conducting air drops of dirt and boulders to fortify the dam —averted a catastrophe.  But it also was an expense way beyond the cost of maintenance.

Perhaps most alarming, not all people received the emergency notification when the evacuation was called, according to the Associated Press.

Dams Fail Quickly

Every four years, the American Society of Civil Engineers issues a “Report Card for America’s Infrastructure.” Bridges, roads, tunnel systems are evaluated on capacity, condition, funding, future need, operation and maintenance, public safety, resilience and innovation.

Dams received a “D” grade in the 2013 report. The next report is due March 9, and there is no expectation for a major shift in that grade. The Army Corps of Engineers estimated it would currently take $24 billion to fix all dams that need repairs.

State agencies regulate most of the nation’s 84,000 dams. Unlike bridges and roads, the U.S. government may inspect the dams but they don’t maintain most of them. More than 65 percent are privately owned and those owners may lack the money needed for adequate maintenance.

Louis Gritzo, vice president and manager of research, FM Global

Experts say dam owners need to know how their structures are aging and prepare for the repairs that need to be done. Develop a timeline for a replacement and how to respond if the dam fails.

“We’re chasing an aging population,” said Louis Gritzo, vice president and manager of research with FM Global.

It is most important to conduct regular inspections and deal with what seems to be minor issues immediately.

“Many people just think you build a dam and it’s just there and fades into the background,” McCarty said.

“But it really is an active system that has to be maintained just like a road or a bridge.”

Gritzo said he discusses with clients opportunities to retrofit facilities and fill in the weak spots. More often than not, an engineer looks at an aging structure and weighs when it is best for a new build. The cost of that is often too challenging, Gritzo said.

Unlike a river flooding over its banks, a dam breach may occur with little to no warning. For that reason, an emergency response plan is needed to lay out how much time is required to get mitigation efforts in place and who is responsible for completing each task.

Conduct Hazard Assessments

“With a dam breach, when it fails, it fails quickly, the water comes out quickly and there’s a limited amount of time,” said Gritzo.

Risk managers need to conduct a very good hazard assessment in the event one of these dams has a catastrophic failure. FM Global uses tools such as complex computer models to calculate different breach scenarios and determines where the water might go and how much flooding might occur when it gets there.

As of 2015, a quarter of dams designated “high hazard” don’t have an EAP.

The insurer works with clients to decide if that amount of water is something they can protect a facility against, or if it’s too large.

“If it’s a meter of water or less you can protect against it,” Gritzo said. “That’s an easy cut-off point.”

More than a meter of water and there’s much fewer protection measures available.

The risk manager that is associated with the facility needs to know the risk exposure, how to react and who to contact in an emergency.

As of 2012, there were 13,991 dams classified as “high hazard,” up 3,000 from a decade earlier.

When a dam is designated a “high hazard,” it means there’s a potential for loss of life if it fails. All high hazard dams require an emergency action plan (EAP). Dams with a “low hazard” or “significant hazard” may have a low expectation for loss of life but still carry potential for damage to surrounding terrain, roads or buildings.

As of 2015, a quarter of dams designated “high hazard” don’t have an EAP. Sometimes, the owners don’t even know they are labeled high hazard, McCarty said.

Risk managers should contact the authority or municipality that creates the EAP to start an ongoing communication and organize emergency drills.

“The time to be swapping business cards and introducing yourself is not when there’s an emergency,” Gritzo said.

John Dickson president, NFS Edge Insurance Agency

McCarty periodically reviews the state assessments of dams in his clients’ communities to try to ascertain the condition of the dams and what kind of ongoing maintenance has been required. He also checks to make sure each community has an EAP in place should something occur.

“Our clients are doing a great job with it, but we know that there are a lot of dams that aren’t getting the attention they need,” McCarty said.

“There has to be a more robust conversation around flood,” said Dickson of NFS Edge.

Advertisement




“The way we do that is not responding in the face of imminent disaster but having the conversation when the levee is not about to be breached or the spillway is not being activated because the dam is at historic highs.”

Municipal leaders should prepare to address the ongoing maintenance that is needed, the EAP they have in place and the responsibilities that go along with having a dam, McCarty said.

“We’ve had an incident here that has heightened our awareness. Those things tend to tail off as time passes,” McCarty said.

“We really need to keep this in our collective memories otherwise we’ll see more and more of these incidents occur.”

Juliann Walsh is a staff writer at Risk & Insurance. She can be reached at [email protected]

More from Risk & Insurance

More from Risk & Insurance

Black Swan: Cloud Attack

Breaking Clouds

A combination of physical and cyber attacks on multiple data centers for cloud service providers causes economic havoc. Even the most well-prepared companies are thrown into paralyzing coverage confusion.
By: | July 27, 2017 • 10 min read

Scenario

By month 16 of the new presidential administration, the Sunshine Brigade is more than ready to act.

Stoked by their anger over rampant economic inequality, the mostly college-educated group of what might best be called upper-middle-class anarchists — many of them from California, Oregon and Washington State — put in motion the gears of a plan more than two years in the making.

Their logic, to them at least, is unimpeachable. Continued consolidation of economic power into the hands of fewer and fewer corporations is creating a world where the rich increasingly exploit and shut out the poor.

Advertisement




The rise of the techno giants is accelerating this trend, according to the Sunshine Brigade’s de facto leader Emily Brookes, an All-American rugby player and a graduate of Reed College in Oregon.

With a new presidential administration seemingly bent on increasing the economic advantages of the rich with no end in sight, nothing to do then but break things up; and in so doing break the hold of this technology oligarchy.

As Emily Brookes so forcefully put in her instant messages to the other members of the brigade: Break the Cloud.

With more than 500 members, many of them with ample financial and technical resources, the Sunshine Brigade is very capable of delivering on its plan for a two-pronged attack.

It is also radicalized enough to justify the loss of some human life, even its own countrymen, to “save” — in its collective logic — the tens of millions of global citizens that are living as virtual slaves in this callous, exploitative global economy.

With websites and digitally connected services large and small down for days, irritation turns to fear.

The first wave in the attack is an attempt to infect and shut down the data centers for the top three cloud service providers. It takes months to set up this offensive.

Rather than rely on a phishing scam from outside the firewalls of the service providers, The Sunshine Brigade uses its social and business connections to place three members on each of the cloud provider’s payrolls. An infected link from someone you know, someone in the cubicle right next to you, seems like an unstoppable play.

It only partially works. Only one of the cloud service providers is harmed when an unsuspecting employee clicks on a link from their traitorous co-worker. The released malware manages to cripple a major cloud service provider for 12 hours.

With millions of users affected, the act creates substantial disruption and garners global headlines. Insured losses are around $1.5 billion. But this is just the beginning.

The morning after, the Sunshine Brigade unleashes a far more devastating and far more ruthless Round Two.

Using self-driving trucks, the Sunshine Brigade smashes into five data centers; three on the West Coast, and two in the Midwest. Fourteen employees of those cloud servers are killed and another 23 injured; some of them critically.

This time the Brigade gets what it wanted. The physical damage to the data centers is substantial enough that it significantly affects three of the top four cloud service providers for five days.

With websites and digitally connected services large and small down for days, irritation turns to fear.

Small and mid-sized banks, which host their applications on clouds, are shut down. Small business owners and consumer banking customers immediately feel the brunt. Retailers that depend on clouds to host their inventory and transaction information are also hit hard.

But really, the blow falls everywhere.

In the U.S., transportation, financial, health, government and other crucial services grind to a halt in many cases.

Not everyone is disrupted. Some of the larger corporations are sophisticated enough in their risk management, those that used back-up clouds and had steadfast business resiliency plans suffer minimal disruption.

Many small to mid-size companies, though, cannot operate. Their employees can’t get to work and when they can, they sit idly in front of blank computer screens connected to useless servers.

For the man on the street, this is hell.

Advertisement




Long lines blossom at the likes of gas stations, banks and grocery stores. A population already on edge from a steady diet of social media provocation becomes even more inflamed.

By nightfall of Day Five, the three major cloud service providers are recovered, and digital “normalcy” begins to creep back. But for many small and medium-sized businesses, the recovery comes way too late.

Economic losses promise to register in the tens of billions. It’s not being too imaginative to think that losses could hit the $100 billion mark.

Two multinational insurers based in the U.S., three Lloyd’s syndicates and a Bermuda insurer signal to regulators that their aggregate cyber-related losses are so great that they will most likely become insolvent.

Emily Brookes and her cohorts were willing to kill more than a dozen people to promote their worldview. In their youthful naiveté, they could not know just how much suffering they would cause.

Observations

For some commercial insurance carriers, the aggregated losses from a prolonged disruption of cloud computing services could be catastrophic, or close to it.

“It’s on a par with any earthquake or hurricane or tornado,” said Scott Stransky, an associate vice president and principal scientist with the modeling firm AIR Worldwide.

AIR modeled the insured losses for the Fortune 1,000 were Amazon’s cloud service to go down for one day. They came up with a figure of $3 billion.

Now consider that most businesses in this country are small businesses, with not nearly the risk management sophistication of the Fortune 1000. Then consider a cloud interruption of five days or more.

Mark Greisiger, president, NetDiligence

“Almost any company you talk about today would rely to some extent on the cloud, either to host their website, to do invoicing, inventory, you name it — the cloud is being used across the board,” Stransky said.

“It’s a significant issue for insurers and one we think about a lot,” said Nick Economidis, an underwriter with specialty carrier Beazley.

“Should a cloud service provider go down, everybody who is working with that cloud service provider is impacted by that,” he said.

“Now, pretty much every software maker is on the cloud,” said Mark Greisiger, president of NetDiligence.

“In the old days, someone would come in and install software on your servers and come in annually for maintenance. That’s all gone bye-bye. Everybody who makes software is forcing you onto their private cloud,” Greisiger said.

The aggregation risk for carriers is complicated by the degree of transparency they have into which insured’s applications are hosted on which cloud provider.

Now here’s the even trickier part. Clouds outsource to other clouds.

“It’s almost becoming a spider’s web of interdependencies on who has access to what in terms of upstream and downstream providers,” Greisiger said.

Determining which of their insureds is hosted on which cloud, and in turn, where that cloud is outsourcing to other clouds can be very difficult for carriers to determine.

Even if a company is careful to diversify the risks they’re taking, they might not realize that a high percentage of insureds are even with the same cloud provider. They could be hit with devastating losses across their entire portfolio of business, said an executive with BDO consulting.

AIR’s Stransky said his company launched a product in April, ARC, which stands for Analytics of Risk from Cyber, which is designed to help carriers gain that much needed transparency.

Among insureds, surviving an event of this magnitude will depend not only on the sophistication of their risk management department, but on the company’s overall ability to negotiate contracts with vendors and suppliers that will indemnify the company in the case of a cloud outage of this duration.

It will also depend on organization’s understanding that there is no off-the-shelf solution that will prevent an event like this or make a company whole after it.

Shiraz Saeed, national practice leader, cyber, Starr Companies

Experts say contracts with cloud service providers, customers and suppliers must be structured so that a company is defended should it lose cloud access for as much as five days or more.

Best practices also include modeling just what your losses would look like in this area, and vetting your full portfolio of insurance policies to understand how each would respond.

One broker said buyers can’t be blamed if the complexities of the coverage issues at stake here are initially hard to grasp.

“It’s becoming a spider’s web of interdependencies on who has access to what.” —Mark Greisiger, president, NetDiligence

“I think it’s the broker’s job to inform the client of this exposure,” said Doug Friel, a vice president with JKJ Commercial Insurance, based in Newtown, Pa.

“You may have business interruption coverage for direct physical damage to your building. But have you ever thought about your business income if your IT structure goes down?” Friel said.

He said many buyers might not realize there is a difference.

Advertisement




Large businesses should have the resources to demand from their cloud service providers that they be indemnified for the entirety of a cloud failure event. There will be a fee for that, but it will be well worth paying, Friel said.

“You have to push,” Friel said. “They are going to say, ‘Here is our standard contract, sign it.’ ”

Don’t settle for that, he said, although many do in ignorance, he added.

“Where possible, we would look for clients to negotiate their contracts. These business relationships should be mutually beneficial, even if one of these events occur,” said Shiraz Saeed, national practice leader, cyber, for the Starr Companies.

It’s a partnership, he said.

“It shouldn’t be a zero sum game on either side. I think there should be an understanding of what the potential loss might be and then designing a contract around that,” he said.

While cloud service providers are known for having high grade security systems, most average organizations don’t have the means for that. But no matter what a company’s resources, the first step is modeling where your digital assets are, and what you and your customers stand to lose if you lose access to them.

“Most insureds don’t seem to understand the amount of individual loss that you could be subject to,” said Jim Evans, leader of insurance advisory services at BDO Consulting. “Usually this stuff is measured in hours,” he said. “But what if a cloud provider is out for three or four days?” he said.

“Trying to quantify what you did lose in an event is hard enough. Trying to do a modeling exercise about what you could lose? It’s something that just doesn’t get done enough,” he said.

Once you have an understanding of what you own and what you stand to lose, the next step is prioritizing the protection of the assets you have. That means drilling into your contract with your cloud service providers to get the maximum indemnification.

It also means spreading your risk so that if at all possible, not all of your assets or your customers’ assets are housed by one cloud service provider. Cloud platforms can be public, private, or a hybrid of the two.

Understanding where your assets are in that architecture is crucial. Spending the money to insure that they are protected behind a diverse menu of firewalls is highly advisable.

Navigating the different iterations of business interruption coverage in property, cyber and kidnap and ransom policies is also important.

Make sure your broker can provide clarity on the different types of coverages and tailor them to your needs, experts said.

The concept of design thinking is really what’s in play here. Organizations have to work with vendors in every aspect of their operations to design a risk management system that can sustain this kind of hit.

“Build a better mousetrap to protect yourself,” said JKJ’s Friel.

“Depending on your service, you need to have the best and the brightest designing this stuff. Spread the risk.”

“Don’t be afraid to ask for more,” he said.

Postscript

In engineering an attack on the cloud, Emily Brookes and her cohorts accomplished the opposite of what they set out to do.

Advertisement




Only the largest corporations with the most sophisticated risk management programs were able to survive the attempt to break the cloud with manageable losses.

Small businesses, the true backbone of the U.S. economy, suffered terribly. Entrepreneurs who put their life’s work into their business lost it in many cases.

Those on the lowest part of the economic scale, the working poor, lost their jobs and their ability to cover their rent and grocery bills. They joined the ranks of those subsidized by the government by the millions.  The attempt to break the cloud resulted in an even more polarized society. &

Dan Reynolds is editor-in-chief of Risk & Insurance. He can be reached at [email protected]