The Economist's Ukraine War-Fire Model

The model detects war-related fires in Ukraine during the ongoing armed conflict with Russia to approximate patterns of military activity. It combines satellite imagery from NASA's Fire Information for Resource Management System (FIRMS) with synthetic-aperture radar (SAR) data. FIRMS, originally developed to monitor forest fires, uses infrared sensors and provides information about fires from multiple satellites that pass over each location at least twice per day. Because it is sensitive to cloud cover, it often fails to record fires when skies are overcast, especially in winter. SAR, which works by sending out pulses of energy and recording the reflections, captures changes in the shape of objects and is unaffected by weather conditions, although it may miss activity in sparsely populated areas. To distinguish war-related fires from other types, The Economist developed a machine-learning algorithm built on 100 separate models trained to predict fire activity in non-war years. When at least 95 of these models confirm that fire activity in a specific location and time is abnormally high compared with pre-war numbers, the incident is coded as war-related. The model has several limitations. First, satellites detect fires by sensing the heat they produce, but this process may be obstructed by weather conditions (such as cloud cover) or if the fires have stopped and the area cooled down before a satellite passed overhead. Second, the statistical method used by The Economist to detect war-related fires is probabilistic and relies on strict thresholds, which can result in real war-related incidents being excluded because they do not meet the criteria of abnormality. Finally, the model generally cannot distinguish war-related fires from non-war-related ones during periods of extremely high temperatures. The model's results are published and regularly updated in The Economist, while the source code and underlying data are available on The Economist's GitHub page, which also hosts plots and interactive visualizations. The folder "output-data" contains downloadable .csv files on territorial control (based on the Institute for the Study of War data), cloud cover across Ukraine, dates of successfully acquired fire data, FIRMS updates, fires by oblast, strikes by location and day, war-related fires, and several other indicators.

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2025-09-18
DOI
https://doi.org/10.5281/zenodo.17151723
Primary Topic
Geochemistry and Geologic Mapping
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

The Economist's Ukraine War-Fire Model

Zenodo (CERN European Organization for Nuclear Research)
Geochemistry and Geologic Mapping
article

The Economist's Ukraine War-Fire Model

article en

Abstract

The model detects war-related fires in Ukraine during the ongoing armed conflict with Russia to approximate patterns of military activity. It combines satellite imagery from NASA's Fire Information for Resource Management System (FIRMS) with synthetic-aperture radar (SAR) data. FIRMS, originally developed to monitor forest fires, uses infrared sensors and provides information about fires from multiple satellites that pass over each location at least twice per day. Because it is sensitive to cloud cover, it often fails to record fires when skies are overcast, especially in winter. SAR, which works by sending out pulses of energy and recording the reflections, captures changes in the shape of objects and is unaffected by weather conditions, although it may miss activity in sparsely populated areas. To distinguish war-related fires from other types, The Economist developed a machine-learning algorithm built on 100 separate models trained to predict fire activity in non-war years. When at least 95 of these models confirm that fire activity in a specific location and time is abnormally high compared with pre-war numbers, the incident is coded as war-related. The model has several limitations. First, satellites detect fires by sensing the heat they produce, but this process may be obstructed by weather conditions (such as cloud cover) or if the fires have stopped and the area cooled down before a satellite passed overhead. Second, the statistical method used by The Economist to detect war-related fires is probabilistic and relies on strict thresholds, which can result in real war-related incidents being excluded because they do not meet the criteria of abnormality. Finally, the model generally cannot distinguish war-related fires from non-war-related ones during periods of extremely high temperatures. The model's results are published and regularly updated in The Economist, while the source code and underlying data are available on The Economist's GitHub page, which also hosts plots and interactive visualizations. The folder "output-data" contains downloadable .csv files on territorial control (based on the Institute for the Study of War data), cloud cover across Ukraine, dates of successfully acquired fire data, FIRMS updates, fires by oblast, strikes by location and day, war-related fires, and several other indicators.

Zenodo (CERN European Organization for Nuclear Research)
Uppsala University (SE)
Openalex Percentile: Top 35%
Geochemistry and Geologic Mapping
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.