CanadaFireSat: Towards high-resolution wildfire forecasting with multiple modalities
Bibliographic record
Abstract
Canada experienced in 2023 one of the most severe wildfire seasons in recent history, causing damage across ecosystems, destroying communities, and emitting large quantities of CO 2 . This extreme wildfire season is symptomatic of a climate-change-induced increase in length and severity of fire seasons affecting the boreal ecosystem. Therefore, it is critical to empower wildfire management in boreal communities with better monitoring solutions. Wildfire probability maps are an important tool for understanding the likelihood of wildfire occurrence and the potential severity of future wildfires. Fire forecasting tools based on Earth observation data exist, but they are limited both by the lack of label information and by their reliance on coarse-resolution environmental drivers and satellite products, which leads to wildfire occurrence prediction of reduced resolution, typically around ∼ 0 . 1 °. To tackle these two limitations, this paper presents a benchmark dataset, CanadaFireSat available on the Hugging Face Hub , and baseline methods for high-resolution wildfire forecasting. We model wildfire forecasting as a binary patch classification task at 100 m, where each patch is labeled as fire if at least one pixel within the patch has burned at the native label resolution (10 m). CanadaFireSat leverages multi-modal data from high-resolution multi-spectral satellite images (Sentinel-2), mid-resolution satellite products (MODIS), and environmental factors (ERA5). We experiment with convolutional (CNN) and transformer (ViT) architectures. We observe that using multi-modal temporal inputs outperforms single-modal temporal inputs across all metrics, achieving a peak performance of 60.3% in F1 score for the 2023 wildfire season, a season never seen during model training. This demonstrates the potential of multi-modal deep learning for wildfire forecasting at high-resolution and continental scale. The code is available on GitHub for the data generation and the model benchmarking .
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".