IgniWise: An Intelligent Fire Prevention System Protecting Spanish Forests with Machine Learning

IgniWise is a machine learning-based planned burn window prediction system in Spain. By analyzing historical fire data, real-time meteorological information, and geographic data, it automatically assesses safe burn timing to help prevent forest fires. It uses the Random Forest algorithm as its core, integrates multi-source data (such as Copernicus Program data and official Spanish fire records), and provides tools and datasets in an open-source manner to offer scientific support for fire prevention decisions.

Forest fires are a major global environmental challenge. Mediterranean climate regions like Spain are prone to fires in hot and dry summers, and climate change has intensified the frequency and intensity of extreme fires. Planned burning is a preventive measure, but it faces issues such as narrow time windows (only 20-30 days per year), complex assessments (requiring integration of meteorology, terrain, vegetation, etc.), high risks (easy to get out of control), and time-consuming and error-prone manual assessments. IgniWise was developed to address these problems.

IgniWise uses the Random Forest algorithm as its core model for reasons including strong interpretability, training on over 10,000 historical fire events, high computational efficiency, and good robustness. Data fusion includes: historical fire data (Spain's MITECO records from 2001 to 2024), real-time meteorology (OpenWeatherMap API), terrain (Copernicus DEM GLO-30 elevation data), vegetation (NDVI index from Copernicus Sentinel-2 satellite images), and land cover (CORINE 2018 edition).

1. National coverage: Covers 48 provinces of the Spanish Peninsula; 2. Automated prediction: Results are automatically updated every 6 hours; 3. Visual design: Color coding with green (safe), yellow (cautious), and red (dangerous); 4. Open-source and free: Open-sourced under the MIT License, datasets shared under CC BY 4.0.

• Fire prevention professionals: Narrow the assessment time range, provide objective references, and reduce the risk of subjective errors; - Policy makers: Grasp the national situation from a macro perspective, optimize resource allocation and strategy formulation; - Research community: Open-source code and public datasets support verification and reproduction, improvement and expansion, and cross-regional comparisons.

1. Tool nature: Only a decision-making aid, cannot replace professional judgment and official authorization; 2. Data risk: Meteorological data relies on third-party APIs, which may have service availability issues; 3. Model limitations: Trained on historical data, may not fully adapt to new risks from climate change; 4. Regional limitations: Optimized only for the Spanish Peninsula; other regions require retraining. It is recommended to refer to the official warnings from Spain's National Meteorological Agency (AEMET).

IgniWise chooses the practical Random Forest algorithm, focuses on data integration, engineering implementation, and user experience, and promotes technology inclusion through open-source and open methods. In the context of climate change, such AI applications for public interests are of great value, and the open-source nature also provides a foundation for community collaboration and improvement.