It’s a searing April’s day in 2016. A crush of 70 million pilgrims, tourists, policemen, ascetics and cows are milling about at the Kumbh Mela in Ujjain. The spiritually inclined have congregated here at one of the world’s largest religious gatherings for a dip in the Shipra river and a visit to multiple temples. But Ashish Verma is here on an entirely different mission. Armed with GoPro cameras, closed-circuit television cameras and remote-controls of drones, the associate professor of Transportation Engineering at Bengaluru’s Indian Institute of Science, and his 100-odd colleagues will do just one thing: watch how the crowd moves.
It’s early days but over the next year Verma hopes to be able to tease out patterns: do higher temperatures influence the speed at which crowds move? Are Indian pilgrims more resilient than foreign tourists? If panic spreads through a group squeezed to immobility, how soon can it be dissipated? “The answers to these could lead to better crowd management at religious congregations as well as prevent disasters such as the recent one in Mumbai,” he tells me over the phone.
When crowds collapse
Discussions on crowd management in urban spaces have surfaced since September 29, when a massive crowd of commuters, during the morning rush hour, were caught in a stampede on the staircase of a narrow foot overbridge at Mumbai’s Elphinstone station. Twenty-three people died and more than 30 were injured.
Human solutions for a human city
Urban planners have said that the tragedy was not due to complex, unexpected phenomena, but rather an absence of basic infrastructural measures such as multiple exits, barriers and segregated spaces that could have prevented particular spots from getting overcrowded. But crowd theorists, like Verma, want to find out if there are general rules that govern crowds across the world and whether these can be used to save lives.
While the Mumbai stampede may have zoomed focus on urban infrastructure, religious gatherings are — by a wide margin — the number one place for stampedes. A 2013 analysis in the International Journal of Disaster Risk Reduction found that 79% of stampedes in India have taken place at religious events, as opposed to miscellaneous or political events. (Worldwide, from 1980 to 2007, 215 stampedes have been reported with 7,069 deaths and more than 14,000 injuries.)
In 1954, a stampede at the Kumbh Mela killed nearly 800 because of an efflux that tore down crowd barriers. The annual Haj, which draws among the largest crowds in the world, took 2,236 lives in September 2015 due to a “progressive crowd collapse” as a review study describes it.
Academic literature abounds with theories that liken a large crowd to flowing liquid, with individuals bumping around like molecules. Some studies
describe masses in panic situations as being akin to “…mice escaping a water pool” or “dispersing ants”. There was even a “social force model,” proposed by the scientists, Helbing and Molnar, in 1995, where pedestrians in a crowd are imagined as “particles” that are “attracted” to their destinations and “repelled” by, say, someone in their path in the opposite direction. Then there are theories such as ‘cellular automata’ where rectangular cells subject to simple rules, flash on and off, and mathematicians study how groups of these cells evolve over time. Much like people.
Crowd mistook vendor’s ‘phool gir gaya’ cry as ‘pul gir gaya’: Elphinstone station stampede survivor
Verma, however, says that most of these models have limited applications and there is no substitute to capturing the movement of actual crowds live, and photographing and recording them from multiple angles, and having scientists from different disciplines — consumer research, advertising, Big Data analysts — study them and wait for patterns to emerge.
Amoeba, floating
One thing that jumps out is how groups move. Pilgrims, who aren’t local to the Kumbh Mela venue, seem to move in a specific pattern. “They will follow the crowd blindly, with heavy loads that retard their speed… as they move in groups, men form a soft boundary (often holding hands) with women and children in the middle,” Verma writes in a research paper. “The group acts like an amoeba in a floating medium.” Then there are cultural differences in how tolerant people are to being touched or maintaining a minimum distance from their neighbour.
In 2009, Partha Chakraborty, a professor of civil engineering at the Indian Institute of Technology, Kanpur, teamed up with researchers in Germany to study if Germans and Indians walked differently, especially when lined up or while walking in groups.
People, in separate experiments, were made to walk on a long corridor and timed at various intervals. When it wasn’t crowded, Indians and Germans walked at roughly the same speeds but as density increased, Indians generally moved faster. Intriguingly, Indians were also less averse to bumping into each other. “By visual inspection it appears that the Indian group of test subjects are less concerned about the personal space of other persons and thus the security distance (a minimum distance) is smaller compared to the German group,” say Chakraborty and colleagues in their paper. Importantly, the Indians studied were all groups of men and the Germans were a mixed crowd.
To Verma, this general disregard for “personal space” can be dangerous. In mass gatherings, Indians generally tend to be “aggressive” and this can lead to a “substantial transfer of body forces” that could pose a “security risk,” he notes. What if there were measures that, for instance, ensured that no more than two or three people collide in a crowd? Techniques to manage crowds haven’t been formalised yet, but Verma hopes he can soon give “general recommendations that are useful to the National Disaster Management Agency or other bodies.”