I am a geoscientist meaning I strive to understand how the world works. Narrowing down further, I am a hydrologist, so I want to know how water moves and is stored on or below the Earth’s surface. Focussing even more, I work in flood risk, so I use this understanding to predict where risk might occur and help people plan and prepare. All of these things are immensely complex, with processes occurring that are both smaller and larger than a human’s normal frame of reference. To make sense of this complexity we use models.
Could a video game help us design new ways to plan for flooding? (screenshot from Cities:Skylines)
Models are simplifications of real-world phenomena based on rules determined through observation and experiments. They can take many forms, from conceptual diagrams, to physical scaled-replicas, to full numerical digital twins. By approximating the real-world, they do not perfectly recreate it but tell us information about it in a format and timeframe that are useful to us.
Gaming is another arena that requires simplified representations of the real-world. Like modelling in geoscience, including all of the complexity we currently understand would make that representation equally complex. The majority of this complexity is entirely superfluous to the gaming experience and would actually detract from the enjoyment of the players, either through introducing complicated rules or by requiring massive computational resource.
Simulation is the realm where models and games meet. It is a realm where the differences between the two are determined by the goal (what is useful) rather than the means. What is the difference between a group of live action roleplayers embarking on a quest and a group of emergency planners wargaming a disaster response? The physics of projectile motion can become both a targeting tool for an artillery operator and Angry Birds.
Like Angry Birds, models can become games. They lend themselves to it so much that scientists using models to run ‘what if’ scenarios are often accused of ‘game playing’. However, to make a model a game requires an act of worlding – the use of narrative and art to create a setting and a purpose for a player. To paraphrase Frank Lantz, these are the operas out of which you build your bridge.
Whereas geoscience models must still be rooted in the real-world enough to remain useful – as Dr Erica Thompson puts it, at some point we need to escape from Model Land – game-based models do not have such restrictions. The usefulness of game-based models is the pleasure of the player (or more cynically, the profits of the developers), so escaping from the game-based Model Land is undesirable. This lack of any restrictions means games developers have a freedom to explore different model approaches to real-world issues, using model types that offer potential but have yet to reach a development stage to be applied yet. Could such games become our models of tomorrow?
Screenshot of vehicles interacting with a flooded road junction from the game Cities:Skylines.
Cities:Skylines is a city-builder simulation game where the players act as the all powerful Mayor managing every aspect of their growing city. At their finger tips is a plethora of spatial information on anything they might want to check – traffic flows, energy supply, crime rates, and much more. They can watch the entire city grow and also follow individually named citizens. This in itself demonstrates the potential of smart cities where networks would provide similar spatial information to city planners. Under the hood of the game there are a lot of models running together to simulate an entire city. The models that interest me are the hydraulic model that simulates water flows and the agent-based models (ABMs) controlling citizens and traffic.
I forced the game to simulate a river flood by increasing the input for the hydraulic model. As the model caused areas of the city to become inundated by flood water, the ABMs responded, rerouting people and traffic. No longer served by transport links, flooded areas entered decline, starved of goods, services, workers, and customers. As the waters receded, the areas recovered. Although these models created something that looked and felt real in the world the game has created, they are clear nonsense when subjected to any critical analysis against the real-world. But what it does do is demonstrate the potential useful information this combination of models, essentially a digital twin of a smart city approach, could provide if designed for flood risk purposes.
Fundamentally, there is so much cross-over between models and games, particularly numerical models and video games, that they arguably could be described as the same thing, or at least on the same spectrum. The realm where they meet is rich with potential and I believe there is much to be gained by delving deeper into this realm. That’s where you’ll find me.
Author bio: Dr Chris Skinner is the Director of FloodSkinner – Creative Flood Education, visiting researcher at the Energy and Environment Institute, University of Hull, and senior hydrologist at the Environment Agency. He is lead convenor of the popular Games for Geoscience session at the European Geoscience Union General Assembly. His work includes game-based exhibits, immersive storytelling, and hosting the FloodSkinner YouTube channel.
YouTube: https://www.youtube.com/@floodskinnersmodellife
Twitter: @FloodSkinner
MultiPlay 