The "atmospheric river" that struck California in December dumped more than five inches of rain on the city of San Francisco in a single day, submerging cars, clogging Highway 101 and sending residents to their kayaks.
As climate change brings more catastrophic storms, along with rising seas, what investments should San Francisco and other coastal cities make to lessen the blow of the next disaster? Should they rebuild drainage systems, lift roadways, relocate flooded homes to drier ground, all of the above?
Artificial intelligence is helping mayors and city players figure out how to make their cities more resilient to the ravages of climate change. A new kind of simulation, called a "digital twin," incorporates extensive amounts of data on each particular city, including its infrastructure, populations and climate threats. Built with a type of AI called machine learning, digital twins are helping planners run cost-benefit analysis on climate projects and prioritize climate programs.
Becoming a flood-resilient city can take some serious urban engineering, ranging from creating new parks and planting tens of thousands of trees, to constructing levees or digging new water channels. Given that such mega-projects can cost hundreds of millions of dollars and may or may not solve the problem, many cities have been hesitant to jump in. "There are solutions, but cities don't know where to start," says Isabel Belliard, an executive at French urban-planning services company Siradel, which is among a handful of companies, government agencies and research groups developing digital twins.
The adoption of digital twins is still in the early days. The French cities of Paris and Angers, as well as the principality of Monaco, have started using an early version of a digital twin made by Siradel. In January, the company released its "Urban Climate Resilience Toolkit" that enables city planners and leaders to get a simulated 3-D look at how flooding and extreme heat brought on by climate change is likely to impact their populations—and to virtually experiment at little cost with possible fixes to see what will work best.
Another company advancing the digital-twin approach is Bentley Systems, a global engineering company headquartered in Exton, Pennsylvania. Bentley is using its digital-twin software to help cities manage the impact of climate change on infrastructure, from bridges, dams and roads to the electric grid and water supply. "We have to update the way we measure and calculate the risk around cities' decisions, so we can better represent what really happens in the world," says Rodrigo Fernandes, who heads sustainable development at Bentley. "A digital twin can do that in a more dynamic and accurate way."
Thanks to climate change, floods are becoming a bigger and bigger problem in much of the world and in the U.S. in particular. According to the U.S. Environmental Protection Agency, flooding from rising sea levels in cities near coastal areas of the U.S. is already five times as common as it was in the 1950s, and the trend is accelerating.
Living away from the coasts won't necessarily help. A study by the U.S. National Center for Atmospheric Research and other research groups, published in April of last year in Nature Portfolio's Communications Earth & Environment journal, predicted that the middle of the U.S. is becoming increasingly prone to flash flooding.
Meanwhile, most of America is likely to experience crushing heat waves in the coming decades. Summer temperatures reaching 125 degrees will become the rule for more than 100 million people in the U.S. by 2053, according to a peer-reviewed study published by nonprofit research group First Street Foundation. Cities will suffer the most, because concrete, asphalt and other construction materials efficiently absorb the heat and radiate it back to surroundings, including people. Scientists call it the "urban heat island" effect.
The simplest and one of the most effective ways for a city to reduce its vulnerability to flood and heat damage is to plant trees and create green spaces. Exposed soil and vegetation can soak up vast quantities of water that would otherwise careen down streets and into buildings and homes. Trees, vegetation and soil also don't get as hot as concrete, and they disperse most of the heat they do grab down deeper into the ground.
But urban planners are hardly in a position to tear up miles of roadway or pull down hundreds of buildings and homes in order to replace them with vast swaths of trees, parks and gardens. Given the limited space that can be devoted to greenery, the question becomes one of where limited numbers of green plots can be placed to do the most good.
A digital twin is designed to consider such scenarios, but its effectiveness depends on the richness of the data it has about the particular city, and on accurate mathematical models about weather impacts. In its years-long effort to build its Urban Climate Resilience Toolkit, Siradel worked with climate and weather scientists to incorporate their formulas into its own software, and with construction, materials and agronomic engineers to ensure the software could accurately model how buildings, roads and parks interact with sunlight, heat, wind and water.
Loading the resulting models with detailed data about specific cities was an even bigger challenge. To get data on 3,000 major cities around the world, Siradel first grabbed what publicly available information it could on city layouts, building sizes and greenery. To flesh those basic pictures out, it pulled in what 3-D digital models were available for individual buildings—models that architects now routinely develop for their projects, and typically leave on file with cities. "Some of these models go down to the level of individual bricks," says Belliard.
That still left plenty of gaps in most cities. To close them, Siradel sent people driving around cities with car-roof-mounted video-mapping devices aimed at buildings. And since cities in Europe, Asia and some other regions are webbed with tiny alleys too small for cars to get through, the company also sent people with backpack-mounted versions of the video-mapping equipment walking through the alleys.
Finally, Siradel layered demographic data into its urban databases. That way the models could be informed not only by how many people live in a given building or neighborhood, but also about whether they are especially vulnerable to weather-related pressures. Heat waves, for example, are particularly dangerous for low-income residents who are less likely to have air conditioning and for those who are elderly, sick, pregnant or very young.
The resulting digital twins offer urban planners a way to garner insights into what sort of mitigations would offer the most protection against what sort of weather, and how many people would benefit. Assessing specific proposals can require complex tradeoffs, says Belliard. Putting in a solar-panel array on the roof of a large building might sound like an easy win for sustainability, he notes, but a rooftop garden might do more good as a heat and rainwater defense. In communities at high risk of flooding but with little space for greenery, a planner can determine the costs and benefits of replacing sidewalks, plazas and roads with newer types of concrete and asphalt that can absorb water and drain it into the soil below.
Bentley used its digital-twin technology to help the city of Rotterdam in the Netherlands modify its massive system of water pumps to better conserve water and reduce the pumps' energy usage and carbon emissions. The project involved installing sensors that monitored the pumps' operations in real time and fed that data into a digital twin that included the city's water system. Machine-learning software flagged when and where pumps were pumping too much or too little water. By letting the software make adjustments to the pumps on the fly, the city reduced the water system's energy-usage and emissions by 30 percent and sharply cut the amount of water it wasted.
The company also has been working with Singapore on a more complex digital twin that includes all streets and buildings and will eventually include maps of the city's massive underground utilities infrastructure. The model is accurate to inches and includes 25,000 gigabytes of data. Among the goals of the project is to find the ideal locations for solar panels throughout the city and to manage flooding. Yet such projects are just scratching at the surface of what digital twins might be able to do, claims Fernandes. "A single digital twin that simulated everything in the city would be a more massive undertaking," he says. "But that's our vision."
Digital twins can already help with a range of complex questions such as where to place a large city park to provide the most heat-and-water protection to the most at-risk people, or how to manage building development across a city section so as to create natural ventilation from winds. For cities that are prepared to tackle massive flood-control projects such as canals or levees, being able to model the impacts of the proposed changes could save tens or even hundreds of millions of dollars and maximize the number of people who benefit. For example, Bilbao, Spain, dug a canal across the foot of the city in 2017, severing the city from the mainland, to reduce its growing problems with river flooding. The project has so far lowered the river level by a foot, but it may be decades before the city knows for sure if the canal was the most effective way to solve the problem.
Besides Paris, Angers and Monaco, Belliard says Siradel is in discussion with other cities around the world, including several in the U.S. Orlando, Florida, the home of Disney World, is also considering using a digital twin, according to the mayor's office. In addition to Singapore and Rotterdam, Bentley has worked with the Argentine city of Mendoza, and Brazil's Joinville.
In the end, extreme weather from climate change will have to be fought with action, not maps on a computer screen. But if a map can point the way to the right action, it's a good first step.
About the writer
To read how Newsweek uses AI as a newsroom tool, Click here.
