Title: Fire Weather and Risk Dataset for Hérault, France (2020–2025)

Version: 1.0

Authors:
- Nathalie Neptune

---

Description:
This dataset provides daily meteorological and fire risk data for the department of Hérault (INSEE code 34) in southern France, covering the years 2020 to 2025. It integrates:

- Meteorological variables (from SAFRAN-SMI)
- Fire danger index (IFM) and derived fire risk indicators
- Soil moisture and water balance variables (e.g., SWI, ETP, etc.)
- Commune-level metadata (INSEE code, name, coordinates)
- Historical fire occurrences from the BDIFF database

The data is intended to support modeling and evaluation of daily fire danger risk and can serve as a benchmark for machine learning models.

---

Files Included:

1. **herault_ifm_weather_dataset.csv**
   - Daily meteorological and fire risk data for SAFRAN grid points in Hérault.
   - Includes commune metadata and IFM values.

2. **herault_fire_incidents_communes.csv**
   - Fire occurrence data derived from the BDIFF database. 
   - This file contains official fire incident records for the department of Hérault, derived from the BDIFF database operated by the French Ministry of Agriculture.
   - It includes one row per fire event, with surface burned, location (INSEE commune code), and optional details on affected land cover and damage. 
   - The file is filtered to include only significant fire events (≥5000 m² of surface burned).
   - Column names are in French and reflect the original structure of the BDIFF export. They have not been translated to preserve traceability.
   - This file is provided as **secondary data** for model validation and benchmarking purposes. It is not required to use the main dataset.
   - Key columns (original French names):
		- `Date de première alerte`: Date of the fire (first alert)
		- `Code INSEE`: Official commune code
		- `Nom de la commune`: Commune name
		- `Surface parcourue (m2)`: Total area burned
		- `Surface forêt (m2)`: Forest area burned
		- `Nombre de décès`, `Nombre de bâtiments totalement détruits`: Severity info (if reported)



---

Column Highlights (main CSV):

- `DATE`: Observation date (YYYY-MM-DD)
- `LAMBX`, `LAMBY`: SAFRAN grid coordinates (Lambert 2 étendu)
- `LAT_DG`, `LON_DG`: Latitude/Longitude (WGS84)
- `T_Q`, `HU_Q`, `FF_Q`, `PRELIQ_Q`, `ETP_Q`: Core weather variables
- `SWI_Q`: Soil Wetness Index
- `ifm`: Fire danger index (IFM)
- `code_insee`, `nom_commune`: Commune identification
- Derived indicators: rolling means, lagged features, fire memory indicators

---

Temporal Coverage:
2020-01-01 to 2025-12-31 (daily)

Spatial Coverage:
- Region: Hérault, France
- Department Code: 34
- CRS: EPSG:4326 (WGS84)

---

Data Sources:
- SAFRAN-SMI and SAFRAN-IFM (Météo-France)
- BDIFF Fire Incident Data (ONF)
- Commune boundaries: France GeoJSON (https://france-geojson.gregoiredavid.fr)

---

License:
CC-BY 4.0

---

Licensing Notes:
- The meteorological data in this dataset is derived from the SAFRAN model (Météo-France), originally published under the Etalab Open License 2.0. 
- Commune boundaries used in this dataset were obtained from the France GeoJSON project by Grégoire David (https://france-geojson.gregoiredavid.fr), published under the Etalab Open License 2.0.
- The file `herault_fire_incidents_communes.csv` is derived from the BDIFF database operated by the French Ministry of Agriculture. Redistribution is permitted for non-commercial, research, and educational use, provided proper attribution is maintained and no modifications are made to the original content. Source: https://bdiff.agriculture.gouv.fr  © Ministère de l’Agriculture et de la Souveraineté Alimentaire

---

Suggested Citation:
Nathalie Neptune (2025). Fire Weather and Risk Dataset for Hérault, France (2020–2025) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.17069738
