Digital twins for cities

Let’s start by “filing” this under a few ideas so that readers might better understand why this article on city administrations creating digital twins of their municipalities deserves their time. It’s kind of fascinating, it’s partially a new vision of smart cities, it’s quite aspirational so far (some might say it’s mostly hype), and it’s an intriguing topic where urbanism, architecture, challenges like the climate crisis, and gaming intersect.

Gaming? Not only because these digital twins might look like SimCity, but also because in many projects the technology behind the visuals is the same as the one used for games like Fortnite, Unreal Tournament, or Among Us.

So what is a digital twin?

[D]igital twins allow cities not just to create virtual models, but to run simulations of new policies or infrastructure projects and preview their potential impacts before making a decision in the real world.

This kind of concept is being created and invested in around the world in cities like Las Vegas, Los Angeles, New York, Phoenix, Singapore, Helsinki and Dubai. The potential is seen as enormous and the breadth of uses is quite wide, but which will pan out and which will prove too hard or unrealistic? Some of the numbers advanced are staggering… and likely over-optimistic.

The technology could help officials cut operating costs and carbon emissions of new construction, and avoid costly modifications after a project is completed. Amid an ever-looming climate crisis facing urban areas, it could enable cities to test the effectiveness of various measures against rising sea levels and urban heat. By one estimate, digital twins could save cities some $280 billion by 2030.

We also mentioned smart cities, these digital twins require “huge datasets” and lots of monitoring, something very similar to smart city visions of recent years.

Virtual Singapore is already a live test case for what the tech can do. The island-nation’s model comprises more than 3 million images captured at street level and 160,000 images taken from the air, along with billions of data points plotted in 3-D, amounting to more than 100 terabytes of raw data. The foundation of the model will rely on 14 core datasets on everything from land use to tree cover to underground utilities.

At this point, you might be thinking “this already exists in Google Maps,” or “I’ve seen better in games.” And you’d be half right. Visually, the current examples are similar or even inferior to existing services and games. Where it gets more valuable is that these are not just images, a true digital city model “knows” what things are.

What makes the model smart, Khoo says, is that it distinguishes buildings from trees and roadways from sidewalks, and so on, making it easier to test how individual elements react in various simulations. Windows, rooftops and the facades of a building are also treated as individual assets.

So why should we be looking at this kind of application on a website about Fab Cities? Well, for one thing, if they do fulfil their promise, such applications might help to determine “how new development might impact the environment. When developers propose a new building, the model is used to analyze how it might affect wind flow, shadows and the already insufferable urban heat-island effect.” All things make a city more liveable, and all things that help us better understand the space we inhabit.

Second, although these models are largely presented for local impacts, by being digital they could also easily exchange data, share conclusions, and share experiments being run on them. Local impact, global action, and consideration for the lived experience are all very Fab City.

More: Not a digital twin, but a fun example of how applications that “understand” what is shown can build on that. JveuxDuSoleil uses a 3D map of the world to project shadows according to the time of day and help you decide which terrasse you should spend your afternoon on 😎.

Image: A 3-D “reality mesh” of Singapore’s national garden, Gardens by the Bay, enables the mapping team to capture the shape and other attributes of vegetation. Courtesy of Singapore Land Authority.