AI for the Resilient City

Many local and regional governments have begun or ramped up the transformational change to take climate action and plan for future resilience but among the barriers to getting started or making informed decisions are the foresight, tools and applications staff need to understand how the impacts of climate change will affect specific geographic regions and how climate solutions may lessen, mitigate or adapt to those future impacts, stresses, hazards and events. Municipalities are eager to create, develop, or access the climate tools needed to understand which communities are most at risk from the impacts of climate change like extreme heat or flooding, and how capital investment can be best utilized to get the largest return on investment and benefit local communities, especially those that are the most vulnerable.

The City of Calgary utilized the tool to better understand the parts of communities that are consistently exposed to high temperatures and to inform climate adaptation planning decisions, prioritize City resources, and inform decisions made by community planners, social services, and emergency management groups to direct cooling interventions and foster climate resilience to heat-induced impacts.

In its current form, the UHI tool provides three main modes for better resiliency planning and decision-making. Story Mode allows users to see data insights (like heat temperature, Infrastructure, and demographic data) as easily digestible stories. Explore Mode provides a granular view of the data identifying UHI hotspots, identifying different building types and clusters of buildings, and understanding high-level breakdowns of population age demographics and how that may relate to vulnerability and infrastructure stress. Compare Mode allows users to compare correlating variables over different time frames, like changes in extrema heat in comparison to building age, vegetation cover, or pervious and impervious surfaces.

In the newest Mode (currently only available to the Region of Peel), Evergreen has created a machine-learning Scenario Mode that allows users to input custom variables (such as building count and % of vegetation healthiness/greenness) and understand a timeline of how these changes to variables will alter the land surface temperature and extreme heat of specific measured areas.

Through the UHI maps produced by the tool, the city can also provide evidence that supports and informs the following actions:

1. Utilizing the tool to inform five planning projects (currently in draft, non-approved form) and approximately 10 public infrastructure projects. It helped to inform a project that developed a value for the heat reduction effects of natural assets in the City of Calgary.
2. Providing quantitative data that demonstrated the importance of natural assets for temperature reduction and the value of increasing and enhancing the natural assets from a climate risk perspective.
3. Demonstrating the effect of roads and paved spaces on heat, with the highest temperatures being around paved parking lots and rooftops. This information can improve the design of these spaces to be better equipped for climate change and become more climate resilient for the future.
4. Demonstrating the cooling effect of parks and natural assets, particularly of water bodies.
5. Indicating that heat amplifies the impact of paved spaces and reduces the impacts of greenspaces extending beyond the geographic boundaries of the respective feature affecting parts of the community nearby.
6. The Climate Adaptation team and the Climate Governance and Strategic Planning Team at the City of Calgary used the tool to inform and develop “Community Climate Risk Profiles”. The Community Climate Risk Profiles are created for each community in Calgary and assist Community Planners in making evidence-based decisions related to climate risk reduction as they complete multi-community planning projects.

There were many lessons learned from the pilot program with the City of Calgary’s application but the ability to build in a predictive function was the most important one. This could help to predict, for example, how a climate solution like planting street trees would affect UHI and extreme heat years and decades down the road. As Evergreen built the tool and, in the process, consulted with scientists from NASA and other data science experts, the current scope and data inputs that existed in the current tool seemed inadequate to create a predictability function at the resolution of street level. With additional climate models, a wider range and greater depth of data and a wider mode resolution (neighbourhood/ community level vs street level), the ability to create the predictive function is possible.

As the program and tool have moved through the pilot stages and working with our technical partners at Gramener Inc, we have learned that the tool could be used to better inform GHG emissions sources based on buildings/infrastructure in compilation with heat data, and a number of other factors. In the future, Evergreen is hoping to integrate the machine learning abilities of the tool to identify UHI and extreme heat hotspots and provide an understanding of local area cooling and emissions reductions through the implementation of retrofits such as green or white rooves.

The program has since expanded through several phases (as mentioned in brief above) with partnerships and ambitions to the Region of Peel and Toronto & Region Conservation Authority. This partnership is important as the tool geographically expands to incorporate a regional government that within its boundaries has three cities: Mississauga, Brampton, and Caledon. This brings on new opportunities and challenges which Evergreen is running headfirst at by innovating and growing the program to provide governments and cities with the resources to take climate action. As for the next steps, Evergreen has recently completed the creation of a new Scenario Mode which allows users to create custom scenarios from the existing tool to understand how planning, climate or infrastructure projects could change the community’s ability to mitigate UHI or extreme heat. While this Mode is new, it is expected to be a strategic next step for the program, and partners in planning climate action solutions and better arming government staff and planners with a tool to make more informed decisions.

Next steps for the tool moving forward:

• The direct integration of socio-economic datasets to better understand a risk and vulnerability profile layer at the community level and how it correlates to UHI and extreme heat, directing users on where investments would have the greatest impact on the most vulnerable to climate change (in the context of warming temperatures)
• The machine learning integration of determining building or community wide GHG emissions reductions through the correlation of weather and albedo, understanding the direct implications of white and green rooves as a starting point.
• Scaling of the tool and program to two new municipalities by the end of 2024.

Link to Full Case Study

Additional Resources:

• AI for the Resilient City Toolkit
• Climate Projections for City of Calgary
• City of Calgary Community Climate Risk Index Summary Report