That technology has allowed us to achieve amazing feats like building a system that can land a rover on Mars using artificial intelligence is amazing. So why are we so often unprepared for disasters, like the recent Texas storms or hurricanes that batter the Caribbean and United States? And why are responses often so inadequate?
Indeed, were it not for social media, the human costs of these disasters could be much higher, suggests Dhiraj Murthy, an associate professor of journalism and sociology at the University of Texas at Austin.
Living in Texas, he’s had a front-row seat as the challenges following recent storms unfolded. And he’s found that social media is filling a critical role that emergency services or public utilities can’t fill when getting vital information to people in crisis.
For example, Murthy pointed out, it’s not the announcement of, say, the location and status of power outages that provide the most useful information for victims, but the public conversation that ensues.
“Within the comment threads [was] the substance where people were posting information about where they had suffered greater outages, or where there was issues of burst pipes or other things. So that the comment thread within the Twitter post actually had a tremendous amount of interaction between people and actually formed a sort of community,” he told Spark host Nora Young.
“For example, I saw a lot of users on Twitter asking questions such as, ‘Is it safe to run my gas stove with pots of boiling water? So I get the steam and the heat?'”
And members of the community would jump in and help people find solutions, he said.
“There’s a perception that if they post on social media, people are listening. And I don’t think it’s an unfounded perception. It may not be that emergency services are listening. But it’s definitely the case that community members are listening.”
He pointed to another app, Nextdoor, that is even more community focused and a great tool troubleshooting and problem solving.
“Nextdoor is a really popular neighborhood-based platform, and you have to have a verified address within that neighborhood. So you have a certain certitude that the people who have joined a community with a Nextdoor [group] are parts of that. So in there, I’ve noticed that people will perhaps be less anonymous about their posting,” Murthy said.
In a broader sense, officials can also use the information exchanged on social media to get a better sense of how the disaster–and its response–is unfolding.
“The machine learning algorithms could go through and aggregate that, for example, say this user is in this part of New York in the case of [Hurricane] Sandy, and they report that they don’t have power, they report that they have a flooded street right in front of their apartment, and then aggregate that at scale. So when they have enough inputs, they’ve got real time data that’s coming in.”
Wealthy countries still ‘get taken by surprise, even by very predicted disasters’
Despite the usefulness of social media, a bigger issue might be that authorities, even in wealthy countries, still don’t do enough to prepare for disasters even when they’re eminently predictable, said Malka Older, a disaster sociologist and humanitarian aid worker (and also a Hugo-nominated science-fiction writer).
“Even in Japan, which I think is one of the countries that puts the most emphasis on preparedness, people still complain to me that there was not enough budget for what they wanted to do, and that the national level of the National Organization for mitigating disasters did not have enough people or budget to really take care of things. And that’s, that’s really the sort of best case,” she told Young.
“If you look at a lot of other countries, whether we’re talking about the US or the UK, or France, or really any of the countries that we think of as developed and safe, then you see that they still seem to get taken by surprise, even by very predicted disasters.”
Older said that it’s not the disasters per se that are the problem; it’s that governments haven’t adequately built infrastructure to manage the risks.
“It’s very easy, particularly with our sort of short political cycles and our short attention spans, to push it off and not do the preparation work and not put in the preparedness to make ourselves more resilient. But it’s a really false economy, because putting money into preparedness and putting time into preparedness is much, much cheaper and more effective than going in afterwards and trying to rebuild all the things that were destroyed. And obviously, of course, the people who are lost, we cannot we can’t rebuild those.”
She suggested it’s an attitude that these disasters are just natural phenomena that can’t necessarily be managed, and that’s not the case, she said.
“I think a big part of it has to do with hubris. And that a lot of the wealthy countries really think that they are kind of beyond the impacts of nature, really. And that they’ve got hurricanes covered. They build strong, permanent buildings. They have educated people and they don’t have to worry about this and it’s very much not true, you know, we’re facing natural hazards that can absolutely impact everything that we are able to build. And because we don’t spend the resources—the money, yes—but also really the time and the information resources in preparing for and thinking about these disasters, we keep getting taken by surprise, and we keep finding that they were really not ready.”
Predicting the impact, not just that a disaster will happen
To get emergency teams in place quickly following an extreme weather event, it helps to know in advance what the impacts on communities will be.
And while we can’t predict when exactly a natural hazard will occur, there are ways to determine the potential impact of these events and prepare for them, Seth Guikema, an associate professor of engineering at the University of Michigan, told Young.
He runs a research group that uses historic data from past weather events, along with machine learning, to predict the occurrence of future hazards and the impact they will have on infrastructure like the electrical grid and drinking water systems.
“So there we use a lot of historic data about what happened in past events to develop predictive models. So when you see a forecast event, [whether] it’d be a derecho, hurricane or winter storm, we try to predict, well, how many outages are there going to be? And where are they going to be?” he said.
And diversity in past data sets is key, he added, so that they represent the various weather conditions that may develop in the future.
In the case of Texas, where there’s a very high degree of interdependence between different types of infrastructure, communities can’t suddenly scramble to reinvent their systems right before an extreme event, said Guikema. And while power outages aren’t always preventable, what we can do is be more prepared to respond to them quickly and begin restoration, he added.
“I think we will always be in the situation of having impacts from these disasters; we can’t prevent them. And as long as people want to live in areas where natural hazards happen, we’re going to have to deal with them. What we can try to do is help people understand what those risks are upfront, so that they can take more appropriate risk management actions, to limit the impacts to respond faster, and in some cases, to move out of the area.”
Written by Adam Killick and Samraweet Yohannes. Produced by Nora Young and Samraweet Yohannes.
— to www.cbc.ca