Floods are one of the most common types of natural disaster around the world. The damage they cause can be physically, emotionally and financially devastating.
But what if we knew where floods would occur in advance? That’s what Danielle Rainville, a master of applied science student in the civil engineering program, is tackling: a model to predict floods in Ottawa and mitigate damage.
A faster method of flood prediction
Rainville is developing new solutions to sustainability and environmental issues. She credits her supervisor, Professor Hossein Bonakdari, as a big inspiration. “While this topic was new to both of us, we used this opportunity to learn together,” says Rainville.
Rainville’s goal was to use statistics to determine the probability of floods, even in situations where limited environmental data is available. This is especially useful in regions with extreme climates, isolated regions or developing regions where it’s more difficult to record data. “If my predictions are more flexible than current measures, this could be an improvement,” says Rainville. “Future floods would be able to be predicted possibly even weeks ahead.”
A data-driven approach to public safety
Rainville’s research on the likelihood of floods is based on a probabilistic approach known as a Bayesian network. It involves creating graphs to show how different environmental factors affect the chance of flooding in Ottawa.
The first step in building this model is collecting the right data. “I collected the public data for different parameters (e.g., air temperature, flow rate, precipitation) and extracted their probabilities,” she says. “These are then compiled into software that tries to make sense of the data you give it. The hardest part is to make sure the results make sense. This is done by re-arranging the parameters in the model and with multiple sensitivity analyses.”
Once the model was completed, Rainville was able to confirm the accuracy of her results by comparing it to data on previous floods in the Ottawa region. Thus, she was able to accurately predict the flood risk of the Ottawa River.
Using machine learning to predict climate scenarios
Testing the accuracy of her predictive model is the most crucial step. With an accurate model, Rainville can begin developing methods to go from testing historical data to using current data to predict future floods. Rainville is looking to develop machine learning models to predict floods in different climate scenarios. She’d like her research to help prevent severe damage to local infrastructure, wildlife and communities.
Rainville hopes for a long career in civil engineering. Her research on natural disasters is just the beginning.
Sustainable, resilient infrastructure: A uOttawa Engineering priority
One of the five primary research areas at uOttawa’s Faculty of Engineering is sustainable and resilient infrastructure. The University is dedicated to developing infrastructure that can withstand harsh conditions and hazards.
Rainville finished in first place in the sustainable and resilient infrastructure category at the 2026 Engineering and Computer Science Poster Competition, held during Engineering Research Celebration Day. She also finished first place in the Women in Science and Engineering (WISE) category. Her poster, titled Integration of Hydrological Parameters Through a Bayesian Framework to Determine Flood Likelihood for the Region of Ottawa, won over judges with her method of protecting the local community.
Discover other winning engineering poster competition projects.