A Probabilistic Wildfire Risk Model for Canada: Insights for Data, Science and Policy
Bibliographic record
Abstract
Wildfires are increasing in intensity globally, causing death, displacement, elevated health risks due to smoke inhalation, and billions of dollars in damages. In Canada, wildfires are the costliest disaster by event occurrence and recent fires have had international impacts, including on air quality in the United States. These direct consequences of intense wildfires, in addition to the impacts on financial and insurance markets, necessitates a comprehensive and science-based understanding of wildfire risk. This presentation will provide an overview of the wildfire risk in Canada and discuss the innovative scientific approach to risk modelling. In Canada, single wildfire events have caused billions of dollars of losses to residents, governments, and insurers along with considerable social and health impacts, including community displacement and smoke inhalation. The record-breaking 2023 fire season demonstrated that all of Canada can be affected by wildfires, and the intensity of recent wildfires internationally, including in the United States, illustrate the urgent need to better understand intense wildfires. In Canada, Public Safety Canada’s (PS’s) mandate is to “keep Canada safe from a range of risks such as natural disasters”. The Data Science and Engineering Team at Public Safety connects data, analysis, and engineering to policy development, to work towards meeting this ambitious mandate and to support effective adaptation. The team is imbedded in a policy directorate and provides data analysis and technical policy input for programs including a federally-backed Insurance Program, Disaster Financial Assistance modernization, and creating and sharing Canada-wide risk ratings for natural hazards, including wildfires. Canada has significant wildfire risk across much of the country. Many small and medium settlement areas are directly exposed to wildfire risk, and the past few years have seen unprecedented destruction throughout the country. Canada’s extensive forests as well as our dispersed and often remote settlement patterns create a unique and problematic landscape of wildfire risk that will be discussed in the presentation. Several recent wildfire events have led to insurer losses well beyond the historical maximum of losses for any year over the last 50 years. In Canada and across the world, insurers have begun to reduce their risk exposure for high risk properties by refusing policy renewals, reducing capacity to write policies, and raising premiums. PS is responsible for building a holistic understanding of natural hazards across Canada. In this presentation, we will share our team’s and scientific collaborator’s novel efforts across three federal departments to fill knowledge gaps and perform a ‘first of its kind’ wildfire impact assessment and quantification for residential structures in Canada. This includes progress developing and validating novel approaches for characterizing and communicating probabilistic wildfire hazard, developing a Canada-wide wildfire risk assessment, quantifying house loss to structures, and providing wildfire risk information for Canada-wide financial risk analysis. We will highlight our efforts towards data-informed, aligned and effective adaptation policy in addition to how we are bridging the gap between data, science, and policy to keep Canadians safe.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.009 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.009 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".