Every July, the opening ceremony of the San Fermin Festival signalled the imminent beginning of the bull's run in Pamplona, Spain, with over 5,000 people crammed into the city's central square. The crowd begins the morning dressed in white. By noon, many of their clothes were dyed pink by free-flowing sangria.
Participants at the event described people's raucous crashes to Dennis Bartolo, a physicist at Ecole Normal Superiure in Lyon, France, not stepping into the Plaza itself. “The density of people is so high that it's not just that you feel uncomfortable,” he said. “It's painful so you can feel pressure in your chest.”
For more than a few years, he filmed and studied events with the goal of helping prevent stamps that could possibly become fatal at a large public event one day. In a paper published Wednesday in Nature Journal, Dr. Bartolo and his colleagues can predict the spontaneous movement of large crowds in confined spaces when people's density exceeds a critical threshold. I say it might be.
Studying dense crowds is notoriously difficult. “We cannot invite 1,000 people to the experiment,” Dr. Bartolo said. He added, even if he “will not know how to ensure their safety.”
That's why the San Fermin Festival was so fascinating. Each year, thousands of people gather together predictably and relatively gently.
Dr. Bartolo and his colleagues mounted cameras on the upper balcony of two buildings on the opposite side of the plaza to photograph attendees gathering below. “When you watch the video, the dynamics actually look unstable, chaotic, and disturbed,” he said. However, he wondered whether he could tease the principles of organizing that govern the movement of the crowd.
Analyzing the footage presented a similar challenge to studying water flow. “Of course, we can't detect the location of every molecule in water. That's not possible,” Dr. Bartolo said. Still, there are mathematical techniques from the field of fluid dynamics to allow researchers to measure material flow by examining its direction and velocity. Dr. Bartolo applied these same methods to the San Fermin Festival.
The crowd turned out to be less confusing than they appeared. Instead, researchers detected circular vibrations in people's oceans. “We're talking about tracing the same circular trajectory in sync, if not thousands, all in sync,” Dr. Bartolo said.
Furthermore, orbital movements in which each person follows a coarse circle from each individual starting point within the crowd took 18 seconds to complete in this particular square. The timing was so reliable, Dr. Bartolo said that even if the movement appears random in the beginning, he can “set the clock” on this crowd dynamic.
The researchers then applied what they had learned to a deadly stampede. They looked into surveillance footage from the 2010 Love Parade in Duisburg, Germany. “And we detected the exact same vibration,” it appeared just before the fatal stampede, Dr. Bartolo said.
When researchers constructed a mathematical model of crowd mechanics, they discovered that these circulatory movements appear spontaneously, beyond the critical density of people. They do not rely on internal or external forces, such as those who are actively pushing each other.
As a safety measure, Dr. Bartolo proposes to monitor the crowds that are packed for movement of these orbits. When they are detected, they can issue a pre-warning warning about the occurrence of dangerous, uncontrolled movements. By catching vibrations when they are small, he ensures that event organizers will be dispersed and stationary in the crowd before they become larger in trajectory and people are crushed or trampled. I say I can ask.
“We're not there yet,” said Analisa Kuani, a computational mathematician at the University of Houston, about such real-world applications. Photographing a bright venue with a high quality camera is one thing. But, for example, grainy nighttime security footage may not reveal the circular movement of the Telltail.
Nevertheless, Dr. Kuani called this study an important contribution to understanding the collective behavior of large crowds.
“It's a huge effort,” she said. “And one day we can use it in a practical environment.”
