Data Sources
Comprehensive data collection for accurate wildfire prediction and monitoring
Satellite Imagery
- MODIS: Moderate Resolution Imaging Spectroradiometer
- VIIRS: Visible Infrared Imaging Radiometer Suite
- GOES: Geostationary Operational Environmental Satellites
- Landsat: High-resolution earth observation data
Provides fire locations, radiative power, and temporal data for real-time tracking.
Ground Sensors
- ALERTWildfire: Network of high-definition cameras
- FireWatch: Early detection sensor network
- FireScout: AI-powered camera systems
Delivers video feeds and IoT sensor data for ground-level detection and verification.
Weather Data
- NIFC: National Interagency Fire Center data
- NOAA: National Oceanic and Atmospheric Administration
- CAL FIRE: California Department of Forestry data
Provides temperature, humidity, wind speed, and fuel moisture measurements.
Terrain Data
- USGS: Topographical information
- Vegetation Maps: Density and fuel type analysis
Enables terrain-aware fire spread prediction and risk assessment.
Government Datasets (Data.Gov)
Specialized Datasets
1. Satellite-Based Wildfire Detection
NASA FIRMS
Global fire detection from MODIS and VIIRS satellites
Format: CSV, GeoTIFF, KML, JSON
Sentinel-2 Fire Dataset (ESA)
High-resolution imagery with burned area segmentation
Format: GeoTIFF
Landsat Burned Area Data
Burned area mapping from Landsat 8
Format: GeoTIFF
California Fire Perimeters
Polygon shapefiles of historical fire perimeters
Format: Shapefile, KML, GeoJSON
2. Ground-Based & Aerial Fire Imagery
ALERTWildfire Public Dataset
High-resolution live fire detection camera feeds
Use: Train ML models for early detection
FireNet Dataset (Kaggle)
40,000+ annotated wildfire images
Format: PNG/JPEG
Global Wildfire Smoke Detection Dataset
High-resolution smoke images for ML training
3. Fire Weather & Sensor Data
NOAA Historical Fire Weather Data
Wind, humidity, and temperature data related to historical fires
Use: Train ML models to predict fire ignition likelihood
RAWS Weather Stations Data
Live and historical weather sensor data from wildfire-prone regions
Data: Wind speed, humidity, temperature, fuel moisture
UCI Wildfire Prediction Dataset
Focused on predicting wildfire ignition events
Data: Sensor and climate data, labeled by fire occurrences
4. Fire Spread Simulation & Modeling
FARSITE Fire Behavior Prediction Data
Fire spread model data used in the FARSITE simulator
Use: ML-based fire propagation simulation
GEOMAC Wildfire Perimeter Data
Fire perimeters tracked over time using geospatial data
Use: Training fire spread models with past real-world fire progression
FireCast (Global Fire Risk Prediction)
AI-powered fire probability maps updated in near real-time
Data: Fire likelihood based on climate, topography, and satellite
Our Data Mission
Pyre AI combines these diverse data sources through advanced machine learning algorithms to create a comprehensive wildfire prediction and monitoring system. By integrating satellite imagery, ground sensors, weather patterns, and historical data, we provide the most accurate and timely information to protect communities and natural resources.