Seahawks ‘fan quakes’ and Microsoft help scientists track real earthquakes

Jose Miguel Parrella, a Microsoft senior product manager, was watching the Seahawks-Panthers playoff game at home in January, when he knew Seattle had made a big play — before it was even shown on TV.

Parrella’s tipoff? A Seahawks fan quake.

The play had caused fans at Seattle’s CenturyLink Field to jump in the stands, making the ground vibrate. Data of their seismic activity streamed onto Parrella’s phone three seconds later, seven seconds faster than the TV broadcast of the play. It was Seahawks safety Kam Chancellor’s stunning, 90-yard interception return touchdown in the fourth quarter.

So-called “fan quakes” have become a Seattle phenomenon, referring to ground tremors triggered by fervent, cheering Seahawks fans at the Seattle stadium. The quakes became famous when elated fans shook the ground during the now-legendary, 67-yard touchdown run by Marshawn “Beast Mode” Lynch in the 2011 playoffs.

A seismometer near the stadium had detected the shaking, equivalent to a magnitude-1 or -2 earthquake. Quickly dubbed “Beast Quake,” it prompted scientists to view Seahawks football games as a fertile testing ground for developing an earthquake warning system.

For the last two seasons of playoff games, seismologists have placed sensors inside CenturyLink and live-streamed vibration data of jumping, stomping fans. What they want to know: How quickly can data reach thousands of people far away? Where does it hang up? Would the system crash in a real emergency?

“It’s somewhat the problem we’re facing with early earthquake warning. How do we characterize earthquakes and push warnings out to large numbers of people very rapidly, without being clobbered by requests for data?” says Bill Steele, spokesman for the Pacific Northwest Seismic Network (PNSN), of the fan quake studies. The federally funded group tracks the region’s earthquakes and volcanoes.

The group foresees a day when machines can detect an earthquake and blast a warning within seconds onto people’s phones, TVs, computers and other devices — before a quake’s most damaging waves arrive. The alert would automatically shut off gas mains, reduce train speeds, and stop and open elevators, hopefully saving lives and reducing property damage.

“We want to be able to serve information, without the demand for information bringing us to our knees,” Steele says.

Enter the cloud.

On the night before the Seahawks-Panthers game in January, PNSN software engineer Jon Connolly realized his live data feed was going to fail. The national media was writing about the fan quakes, and the growing traffic to his “QuickShake” site made it fumble. It began throwing errors. The data flow got “gappy.”

“It just got crushed,” says Connolly, who was using a Web-hosting service at the time. In a panic, he called Nik Garkusha, then a senior product manager with Microsoft Azure. Garkusha and two other Azure open-source advocates, Parrella and Cory Fowler, had helped build last year’s fan quake platform.

The trio quickly agreed to help again. They started moving code to Azure’s cloud platform and had a system ready for load testing in about three hours. They worked throughout the night and into the next day. “We were developing down to the wire, until kickoff,” says Connolly.

During the game, Azure’s open, flexible system was able to stream data in real time, while thousands of people watched simultaneously. At one point, more than 3,000 people were on the site, watching seismographic wiggles seconds after the earth shook. QuickShake was a hit.

“The benefit of using Azure is we were able to serve as many clients as possible,” says Connolly, whose servers couldn’t have handled the traffic. “It allowed us to scale out and serve more and more people.”

With two different technology stacks at play in Azure, the small volunteer team from Microsoft was able to quickly develop and load-test the platform. Azure’s built-in components were also essential. Visual Studio Online helped the team live-edit code, while Excel captured load-testing data to improve performance for the next game.

“We have a very complete ecosystem,” says Parrella, a program manager in Microsoft’s open source strategy team. “It was amazing to see how it worked – seismic activity was transmitted to the cloud and to my phone faster than the TV image could travel through satellite and cable.”

For the Microsoft employees who worked on the project, contributing to a study that could one day help scientists develop an earthquake alert system was incredibly rewarding.

“That just skyrocketed the amount of geekiness around the whole thing, and it was really exciting to be a part of it,” says Fowler, a program manager in the Azure App Service group.

The timing couldn’t have been better. This year, PNSN began working with Microsoft and other groups to distribute a beta warning system for the earthquake-prone region. Scientists estimate an 84 percent chance of a Northwest earthquake in the next 50 years, similar to the 6.8-magnitude Nisqually quake of 2001.

The groups will help improve software that displays regional alerts of a coming earthquake. They will also look at how alerts can automatically trigger public safety actions to save lives and reduce damage. Developed by the University of California in Berkeley, the so-called ShakeAlert software is the crux of California’s earthquake warning system, now rolling out as a test operation.

A West Coast warning system that includes the Northwest could be available in two to five years, depending on funding and adoption by the U.S. Geological Survey, scientists say.

In a race against an earthquake, warnings must beat the most destructive seismic waves, which travel more than 2 miles per second to the surface. Alerts need to hit many targets, from individual phones to critical infrastructure.

To do that, scientists will need help from technology leaders, says PNSN’s Steele.

“Our expertise is assessing earthquakes rapidly,” he says. “But Microsoft and other companies have been focused for decades on how to push information to consumers and provide people the information they need rapidly on their electronic devices.

“We’re not experts in that field, and we would really like the private sector help us carry that ball.”