Coanda has been selected by the Canadian Space Agency (CSA) for a project to use artificial intelligence (AI) and machine learning (ML) to analyze large amounts of space data to create an improved wildfire prediction tool. The project is funded by the CSA through the Innovative Solutions Canada initiative.
Current wildfire prediction tools, such as the Canadian Wildland Fire Information System (CWFIS, along with a variety of other fire “indices”) primarily use climate and meteorological data and soil moisture measurements from earth-based weather stations. This data can be relatively sparse and costly to collect; weather stations are widely spaced, meaning that data must be interpolated between locations, and deploying resources to collect on-site measurements can be costly.
Coanda aims to employ deep learning architectures in combination with an ensemble of types of satellite data (spatial and potentially temporal), to provide improved wildfire predictions. The core part of the project is the development of the prediction model; however, the proposed solution will also include distillation of the model predictions into information and tools that address specific users’ needs.
“Annual expenditures on wildland fire management are $500 million to $1 billion – with improved prediction tools and data formats that more readily focus on users’ needs, better use of resources and money should be possible, as well as potential reductions in losses due to wildfires.” Dr. Neville Dubash, Senior Data Scientist.
Coanda is a leader in advanced mathematics and data science and offers a range of statistical and data analysis techniques to enhance the value of data for its clients.
Further information on the program can be found on the CSA website:
You can also download our overview document:
Download Overview DocumentWildfire Prediction Overview 403.05 KB