Methodology, Provenance & Scientific Validation
Himalayan Radiometric Telemetry & Topographic Integration
FIREINTEL integrates satellite radiometric thermal observations from NASA FIRMS (MODIS Terra/Aqua and VIIRS Suomi-NPP), high-resolution 30-meter elevation and relief models from the Copernicus GLO-30 DEM, and official geospatial boundaries to produce an empirical, peer-verifiable wildland fire observatory for the Western Himalayas (Uttarakhand, India).
In satellite remote sensing, each recorded data point represents a radiometric pixel hotspot detection identified by automated thermal anomaly algorithms (MOD14/MYD14 or VNP14IMG). A single large, actively spreading wildfire incident can trigger multiple contiguous pixel detections across a single overpass and additional detections over consecutive satellite overpasses. Conversely, small sub-canopy smoldering fires beneath dense sal or oak canopies may remain radiometrically obscured until crown ignition occurs. All statistics in FIREINTEL explicitly reflect validated satellite observations, avoiding artificial incident aggregation assumptions.
Satellite Remote Sensing Data Acquisition
Thermal anomaly observations are sourced directly from NASA's Fire Information for Resource Management System (FIRMS), spanning January 1, 2025 to December 31, 2025. Data acquisition is partitioned across two operational polar-orbiting satellite constellations:
Operates the 375m I-Band (I4: 3.55–3.93 µm). 6,623 confirmed boundary-validated observations (87.83% of total).
Primary High-Res ContributorOperates 1km Channels 21/22 (3.9 µm) and Channel 31 (11.0 µm). 918 validated detections (12.17% of total).
Long-Term Radiometric BaselineRigorous inspection confirms NOAA-20 archives are absent from the active 2025 dataset.
Verified Absent from BaselineSwath Geometry & The Optical Bow-Tie Effect
Understanding the difference between nominal nadir spatial resolution and actual ground-projected pixel footprint is essential for interpreting fire detections in steep mountainous terrain.
MODIS utilizes a 40-sided rotating mirror scanning ±55° across a 2,330 km swath. As scan angle increases off-nadir, optical distortion causes substantial pixel footprint growth:
- Nadir resolution: 1.0 km × 1.0 km
- Edge of scan footprint: ~2.0 km (track) × 4.8 km (scan)
- Area expansion: ~9.6× larger at scan edge
VIIRS incorporates specialized onboard sample aggregation (3:1 aggregation at nadir transitioning to 2:1 and 1:1 near scan edge) over a 3,060 km swath:
- Nadir resolution: 375 m × 375 m
- Edge of scan footprint: ~800 m × 800 m
- Area expansion: Constrained to ~4.5× (vs ~10× for MODIS)
Uttarakhand Boundary Validation & Spatial Filtering
Raw NASA FIRMS CSVs encompass wide regional swaths that overlap neighbouring territories (Himachal Pradesh, Uttar Pradesh, Nepal, and Tibet). To achieve uncompromising territorial precision, FIREINTEL executes a two-stage geospatial verification:
Filters raw FIRMS downloads to coordinates within latitude 28.5°N–31.5°N and longitude 77.5°E–81.2°E.
Point-in-polygon ray-casting test against official Survey of India boundary vectors in EPSG:4326, validating exactly 7,541 internal observations.
Fire Radiative Power (FRP) & Confidence Semantics
Fire Radiative Power (FRP) in Megawatts (MW)
Fire Radiative Power quantifies the instantaneous rate of radiant thermal energy emitted by the flaming/smoldering combustion front, expressed in Megawatts (MW):
Because thermal emission in the MWIR band (~3.9 µm) scales approximately linearly with combustion temperature, FRP correlates directly with fuel consumption rate (~0.368 kg/MJ for forest biomass), providing a direct energetic metric far superior to burned-area estimates.
Decoupled Confidence Semantics
NASA FIRMS employs fundamentally different confidence metrics across instruments, which must not be conflated into a single naive filter:
Continuous statistical deviation test against background brightness temperature and viewing angle.
Discrete qualitative flags derived from 375m I4 brightness temperature, glint tests, and cloud adjacency flags.
Digital Elevation Model & Topographic Derivatives
FIREINTEL integrates the European Space Agency's Copernicus GLO-30 DEM at 30 Meters horizontal spatial resolution across the EGM2008 vertical geoid datum (129 m to 7,804 m in Uttarakhand).
Environmental Data Integration Status
While Copernicus DEM 30m terrain is fully ingested, ECMWF ERA5-Land meteorological reanalysis remains pending in remote CDS queues:
ECMWF Climate Data Store API client architecture is deployed, pending download completion. FIREINTEL does not synthesize or extrapolate unverified atmospheric grids.
Scientific Limitations & Remote Sensing Caveats
- Dense Canopy Attenuation: Sub-canopy smoldering surface fires under dense Himalayan broadleaf canopies (e.g. moist Sal, dense Oak) may emit insufficient MWIR radiant flux to trigger spaceborne detection algorithms until crown ignition occurs.
- Cloud & Monsoon Masking: Persistent monsoonal cloud decks during July–August completely obscure ground thermal emissions, explaining the 0.07% minimum detection rate during peak monsoon.
- Topographic Occlusion & Radar Shadows: In precipitous mountain gorges, steep opposing ridges can partially shield off-nadir line-of-sight sensor paths.
Data Provenance & Acquisition Lifecycle Matrix
Full transparency regarding downloaded, processed, and pending external datasets.
| Dataset / Instrument | Source & Provider | Nominal Resolution | Status | Observation Count / Scope |
|---|---|---|---|---|
| VIIRS (Suomi-NPP) | NASA FIRMS (VNP14IMGTDL) | 375 m (nadir) | Downloaded & Validated | 6,623 Hotspots (87.8%) |
| MODIS (Aqua & Terra) | NASA FIRMS (MCD14DL) | 1,000 m (nadir) | Downloaded & Validated | 918 Hotspots (12.2%) |
| VIIRS (NOAA-20 / JPSS-1) | NASA FIRMS (VJ114IMGTDL) | 375 m (nadir) | Verified Absent | 0 Hotspots (100% VIIRS is Suomi-NPP) |
| Copernicus GLO-30 DEM | European Space Agency (Copernicus) | 30 m (1.0 arcsec) | Downloaded & Extruded | Complete Uttarakhand DEM (129m–7,804m) |
| State Boundary (Uttarakhand) | Survey of India / Official Boundary | Vector Polygon (EPSG:4326) | Boundary Validated | Strict point-in-polygon spatial filter |
| Master Analytical Grid | FIREINTEL Geospatial Framework | 0.05° (~5.5 km × 5.5 km) | Synthesized Grid | 2,174 Discrete Analytical Cells |
| ERA5-Land Reanalysis | ECMWF Copernicus Climate Data Store | 0.1° (~9 km native) | Pending External CDS Queue | CDS API architecture ready; awaiting remote fetch |
Reproducibility & Audit Trail Standard
Orbital Visualization: Orbital tracks rendered on the planetary globe represent nominal Keplerian orbital geometries (inclination ~98.2°, altitude ~705–824 km, Sun-synchronous). They are designed for Earth-observation mission education and do not constitute real-time NORAD/Two-Line Element (TLE) ephemeris tracking.Notice: Illustrative orbital geometry — not live spacecraft tracking.
Reproducibility Guarantee: All 7,541 fire records are stored in boundary-validated CSV (fireIntel/data/processed/boundary_validated/uttarakhand_fires_boundary.csv) with full audit columns: latitude, longitude, brightness, scan, track, acq_date, acq_time, satellite, instrument, confidence, version, bright_t31, frp, daynight, month.