Abid (2026) The extended Algerian forest fires dataset

Identification

• Journal: Mendeley Data
• Year: 2026
• Date: 2026-04-22
• Authors: Faroudja Abid
• DOI: 10.17632/8rkd9cdgs5.1

Research Groups

• Faroudja ABID (Contributor)

Short Summary

This work presents an expanded dataset of Algerian forest fires, increasing the number of instances from 244 to 1215 to enhance the study of meteorological influences and fire danger indices.

Objective

• To construct a comprehensive and extended dataset of forest fire occurrences in Algeria by incorporating a wider range of geographical regions and temporal periods to improve fire risk assessment and prediction.

Study Configuration

• Spatial Scale: Regional (9 regions across Algeria: Tizi Ouzou, Bejaia, Bouira, Khenchela, Tiaret, Setif, Blida, Sidi-Bel-Abbess, and Tipaza).
• Temporal Scale: Multi-year observations focusing on the fire seasons (June to September) of 2012, 2017, and 2021.

Methodology and Data

• Models used: Canadian Fire Weather Index (FWI) system.
• Data sources: Meteorological observations and historical fire occurrence records.

Main Results

• Dataset Expansion: The dataset was increased to 1215 instances, consisting of 633 'fire' and 582 'not fire' classes.
• Feature Set: The dataset includes 12 variables:
  • Meteorological parameters: Temperature (21–43 °C), Relative Humidity (11–96%), Wind Speed (1.39–13.89 m/s), and Rain (0–20.2 mm).
  • FWI Components: Fine Fuel Moisture Code (FFMC), Duff Moisture Code (DMC), Drought Code (DC), Initial Spread Index (ISI), Buildup Index (BUI), and Fire Weather Index (FWI).

Contributions

• Provides a significantly larger and more geographically diverse dataset for the Algerian territory compared to the initial 244-instance dataset, facilitating more robust training and validation of predictive models for forest fire danger.

Funding

• Not specified

Citation

@article{Abid2026extended,
  author = {Abid, Faroudja},
  title = {The extended Algerian forest fires dataset},
  journal = {Mendeley Data},
  year = {2026},
  doi = {10.17632/8rkd9cdgs5.1},
  url = {https://doi.org/10.17632/8rkd9cdgs5.1}
}