Wildfire Prediction Model

AltaML’s advanced Wildfire Prediction model predicts the likelihood of fire outbreaks with same-day and next-day notice.

Powered by Microsoft’s state-of-the-art AzureML service, the model predicts fire risk in designated zones using historical data along with influencing factors, such as ecological zones, weather forecasts, global CO2 levels, Canadian Forest Fire Weather Indices, and human factors such as holidays and weekends. By consolidating all of these data sources into a Machine Learning model, wildfire management branches can use the fire risk predictions to make crucial decisions about deployment of their heavy equipment, aircraft, and fire fighting personnel, ensuring a balance that is both cost-efficient and responsible in terms of risk management. Annual operational cost savings can be substantial, on the scale of millions of dollars.

Customization and Configuration

This offering includes a modelling pipeline that produces models, inference pipeline for making same-day and next-day predictions, and two standard PowerBI dashboards (customized for fire specialist vs. duty officer).

For accurate fire predictions in your area, some additional services will be required to retrain the model on your region’s historical data and ecological zones, calculate Fire Weather Indices, as well as any configuration required to meet your specific prediction requirements and/or dashboard updates to ensure the end result fits seamlessly into your workflow. Please contact AltaML for additional information.

Key Outcomes

Situational Awareness

• Allocate limited presuppression resources where they are most needed.

Efficiency Improvement

• Reduce time required to respond to wildfires, leading to a faster response time.

Strategic Planning

• Advanced planning for resource sharing agreements.

• Manage personnel time off flexibly.