State of Wildfires 2024–25
Bibliographic record
Abstract
Abstract. Climate change is increasing the frequency and intensity of extreme wildfires globally, yet our understanding of these high-impact events remains uneven and shaped by media attention and regional research biases. The State of Wildfire Project systematically tracks and analyses global fire activity and this, its second annual report, covers the March 2024 to February 2025 fire season. During the 2024–25 fire season, fire-related carbon (C) emissions were totalled 2.2 Pg C, 9 % above average and the 6th highest on record since 2003, despite below-average global burned area (BA; 3.7 million km2). Extreme fire seasons in South America’s rainforests, dry forests and wetlands, and in Canada’s boreal forests pushed up the global C emissions total. Fire C emissions were over four times above average in Bolivia, three times above average in Canada, and ~50 % above average in Brazil and Venezuela. Wildfires in 2024–25 caused 100 fatalities in Nepal, 34 in South Africa, and 30 in Los Angeles, with additional fatalities reported in Canada, Côte d’Ivoire, Portugal, and Turkey. The Eaton and Palisades fires in Southern California caused 150,000 evacuations and US$140 billion in damages. Communities in Brazil, Bolivia, Southern California, and Northern India were exposed to fine particulate matter at concentrations 13–60 times WHO’s daily air quality standards. We evaluated the causes and predictability of four extreme wildfire episodes from the 2024–25 fire season, including in Northeast Amazonia (January–March 2024), the Pantanal-Chiquitano border regions of Brazil and Bolivia (July–September 2024), Southern California (January 2025), and the Congo Basin (July–August 2024). Anomalous weather created conditions for these regional extremes, while fuel availability and human ignitions shaped spatial patterns and temporal fire dynamics. In the three tropical regions, prolonged drought was the dominant fire enabler, whereas in California, extreme heat, wind, and antecedent fuel build-up were the dominant enablers. Our attribution analyses show that climate change made extreme fire weather in Northeast Amazonia 30–70 times more likely, increasing burned area roughly fourfold compared to a scenario without climate change. In the Pantanal–Chiquitano, fire weather was 4–5 times more likely, with up to 35-fold increases in burned area. In Southern California, climate change made larger burned area 89 % more likely, with burned area up to 25 times higher. The Congo Basin’s fire weather was 3–8 times more likely with climate change, with a 2.7-fold increase in burned area. Socioeconomic changes since the pre-industrial period, including land-use change, also likely increased burned area in Northeast Amazonia. Our models project that events on the scale of 2024–25 will become up to 57 %, 34 %, and 50 % more frequent than in the modern era in Northeast Amazonia, the Pantanal-Chiquitano, and the Congo Basin, respectively, under a middle-of-the-road scenario (SSP370). Climate action can limit the added risk, with frequency increases kept below 15 % in all three regions under a strong mitigation scenario (SSP126). In Southern California, the future trajectory of extreme fire likelihood remains highly uncertain due to poorly constrained climate-vegetation-fire interactions influencing fuel moisture, though our models suggest that risk may decline in future. This annual report from the State of Wildfires Project integrates and advances cutting-edge fire observations and modelling with regional expertise to track changing global wildfire hazard, guiding policy and practice towards improved preparedness, mitigation, adaptation, and societal benefit. Thirteen new datasets and model codebases presented in this work are available from the State of Wildfires Project’s Zenodo community (https://zenodo.org/communities/stateofwildfiresproject, last access: 11 August 2025).
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.008 | 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".