Forecasting and tracking wildfires and flooding is extremely complicated as multiple variables are involved. Weather, topography and groundwater flows are just some of the elements that need to be considered when preparing or responding to natural disasters.
Artificial intelligence (AI) and machine learning (ML) technologies are enabling these variables to be analyzed and interpreted quickly and more accurately than ever before allowing decision makers and first responders to make informed decisions that can save lives and safeguard infrastructure.
University of Waterloo researchers from across engineering fields, from chemical and civil to environment engineering, are developing solutions to support communities and first responders in building their preparedness and resiliency.
Dr. Joshua Pulsipher, a Chemical Engineering professor, is developing modelling that use machine learning to incorporate data to more accurately uncover meaningful patterns and predictions.
"Fires are very complicated, especially wildfires," Pulsipher says. "Because fires change dynamically over time, we need to address these complexities with models that are able to accurately and reliably forecast what a fire will do."
Pulsipher's team aims to have these models support decision-making tools for predicting and combatting fires.
Waterloo researchers are also using innovative modelling methods to predict flooding. Velocity start-up,Heron Hydrologic Ltd., uses a new approach for floodplain mapping that can support communities in flood mitigation and response. Most flood forecasting in Canada focuses on streamflow forecasting, and real-time inundation mapping is often not available due to technological limitations.
While completing his PhD in the Faculty of Engineering, Heron's co-founder Robert Chlumsky (BASc '15, MASc '17 and PhD '25) worked alongside Engineering professors, Dr. James Craig and Dr. Bryan Tolson to develop a software called Blackbird. The software maps and models floodplains to support landuse planning and risk assessments.
"The resulting flood maps can be both real-time and two-dimensional, facilitating more effective emergency response to flood events," Chlumsky explains. "Flooding is a growing problem and that's why it is imperative that we have the right tools to inform our risk around flooding and provide real-time decision support to ensure we protect people when the floods do come."
Since launching, Heron continues to grow, hiring several coop students and two additional full-time employees. The company was supported by Velocity including through UpStart and the Velocity Pitch Competition.
As the number and severity of wildfires and flooding increase, communities across Canada and around the world are investigating ways to lower their risks and augment their resiliency.
AI and ML models are widely used in manufacturing and industry to optimize operations and offer the same benefits when used to support disaster preparedness and mitigation through supporting policies and action plans.
"We're looking to be more proactive in our behaviour, so instead of simply reacting, we are ready," Pulsipher says.
Pointing to extreme weather events like the one that shut down the Texas power grid or the fires in Los Angeles, Pulsipher says optimization modelling can get the right personnel and resources in the right spots before they are needed.
"We've certainly heard interest from stakeholders, but with any new technology there is skepticism," Pulsipher says. "It is imperative that we continue to collaborate with stakeholders to further develop these models to ensure we have a technology that can be trusted and works."
Meanwhile for Heron Hydrologic, they continue to pilot Blackbird in a number of projects including with Natural Resources Canada, the Government of Newfoundland and Labrador and an upcoming project with the Alberta Environment and Protected Areas. Chlumsky says they are working towards a long-term scalable business model through developing additional flood data products that support clients in the insurance and forecasting spaces.
"As the amount of data grows, so do the opportunities for integrating that data into new innovative technologies," Chlumsky says. "These new technologies can really make a difference in ensuring that our communities are safer and more resilient."
