Near-real time monitoring of burned area at global scale based on deep learning
Bibliographic record
Abstract
Recent advances in land surface reflectance modelling and machine learning techniques opened new opportunities for deriving burned area at a near-real time (NRT) basis and globally. Built from these recent advances, this paper describes a new and computationally efficient approach used by the Copernicus Land Monitoring Service (CLMS) to map burned area from Sentinel-3 OLCI&SLSTR reflectance data and VIIRS active fires at NRT within one day after each image acquisition, complemented by a non-time critical (NTC) product delivered several months after. A neural network is designed to predict fractional burned area on a per-pixel basis from time series of surface reflectance data. The neural network is used to generate time series of burned area detection maps, which are revised by active fire detections spatiotemporal densities to filter out areas likely to be related to land surface changes other than fires, such as agricultural practices or fast vegetation senescence. The algorithm was calibrated with data from 2019, and its quality was assessed with data from 2020, along with other global products, through an intercomparison analysis and a validation analysis using a stratified global random sampling of 30 m reference burned area data. The quality assessment included the NRT and NTC products presented here (CLMBA40nrt and CLMBA40ntc, respectively) and other global products available for 2020, the NTC product currently distributed by the CLMS (CLMBA31ntc), the MODIS-MCD64 Collection 6 (MCD64), the Sentinel-3 OLCI Climate Change Service C3SBA11 and the Sentinel-3 OLCI&SLSTR ESA’s Climate Change Initiative Fire Disturbance FIRECCIS311. The accuracies of CLMBA40ntc and CLMBA40nrt (Dice coefficient (DC) 65.0% and 56.5% respectively) are higher than CLMBA31ntc, MCD64 and C3SBA11 (DC ∼45%) and higher and similar, respectively, compared to FIRECCIS311 (DC 56.0%). The new algorithm described here allows for unprecedented accuracy and timeliness of global burned area estimates.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 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 | 0.000 |
| 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".