Incendi attivi

Risksv00.60 credits/per_km2

Responsible:Niccolò Calandri

Active Fire — VIIRS Fire Detection

Overview

The Active Fire layer uses data from NASA's SNPP and NOAA-20 satellites equipped with the VIIRS (Visible Infrared Imaging Radiometer Suite) sensor to detect thermal anomalies indicative of active fires in near-real-time. The native spatial resolution is 375 m (I-Band), which is refined to 30 m through bicubic interpolation for enhanced spatial detail.

The detection algorithm exploits the strong mid-infrared emission (~3.7 µm) produced by combustion processes. Each detected hotspot is assigned a confidence level ranging from 0 % to 100 %, reflecting the likelihood that the thermal anomaly represents an actual fire event. Multi-sensor aggregation (SNPP + NOAA-20) increases temporal revisit frequency, while cloud and water masking ensures that false detections from reflective or cold surfaces are minimized.

Active fire monitoring is essential for environmental risk assessment, forest fire management, air quality forecasting, and carbon emission estimation. The near-real-time availability of VIIRS data (typically within 3 hours of acquisition) makes it a critical tool for emergency response.

Calculation Methodology

Detection algorithm:

The VIIRS Active Fire product uses the contextual fire detection algorithm applied to the I-Band 4 channel (~3.74 µm). The algorithm:

Identifies potential fire pixels based on brightness temperature thresholds in the mid-infrared
Applies contextual tests comparing each pixel to its neighboring background pixels
Rejects false alarms from sun glint, desert hot surfaces, and industrial heat sources
Assigns confidence values (0–100 %) based on the strength of the thermal anomaly relative to background

Multi-sensor processing:

Data from both SNPP (launched 2011) and NOAA-20 (launched 2017) are aggregated
Combined revisit: approximately every 6 hours at mid-latitudes
Cloud masking via VIIRS Cloud Mask (VCM)
Water body masking via static water mask

Post-processing:

Bicubic interpolation from 375 m to 30 m output resolution
Spatial smoothing preserves fire perimeter shape while enhancing visual detail

Confidence classes:

Class	Threshold	Description
No Fire	0 %	No thermal anomaly detected
Low	1–30 %	Weak signal, possible false positive
Medium	30–60 %	Moderate signal, likely fire
High	60–90 %	Strong signal, confirmed fire
Very High	> 90 %	Very strong signal, large or intense fire

Data Sources

Code	Name	Provider	Type	Coverage	Resolution	Availability
`NASA/LANCE/SNPP_VIIRS/C2`	VIIRS Active Fire SNPP C2	NASA LANCE	Thermal detection	Global	375 m	2012 — present
`NASA/LANCE/NOAA20_VIIRS/C2`	VIIRS Active Fire NOAA-20 C2	NASA LANCE	Thermal detection	Global	375 m	2018 — present

Documentation: NASA FIRMS

Quality Range and Levels

Indicator	Unit	Range	Inverted
`active_fire`	% confidence	[0, 10, 30, 60, 90, 100]	Yes

Inverted = Yes: a higher confidence value indicates a worse condition (active fire present).

Level	Thresholds	Interpretation
A (No Fire)	0 %	No active fire detected
B (Low)	1–30 %	Low confidence detection, possible false alarm
C (Medium)	30–60 %	Medium confidence, likely active fire
D (High)	60–90 %	High confidence active fire
E (Very High)	> 90 %	Very high confidence, intense fire

Technical Notes

The 375 m native resolution represents a significant improvement over the older MODIS 1 km fire product
Bicubic interpolation to 30 m does not add new fire information but improves spatial representation
Night-time detections tend to have higher confidence due to reduced solar contamination
Small or low-intensity fires (e.g., agricultural burning) may fall below detection thresholds
Urban heat islands and industrial facilities can occasionally produce false positives at low confidence levels

References

Schroeder, W., Oliva, P., Giglio, L., Csiszar, I.A. (2014). "The New VIIRS 375 m active fire detection data product: Algorithm description and initial assessment". Remote Sensing of Environment, 143, pp. 85–96.
Giglio, L., Schroeder, W., Justice, C.O. (2016). "The collection 6 MODIS active fire detection algorithm and fire products". Remote Sensing of Environment, 178, pp. 31–41.
Li, F., Zhang, X., Kondragunta, S., Csiszar, I. (2018). "Comparison of fire radiative power estimates from VIIRS and MODIS observations". Journal of Geophysical Research: Atmospheres, 123, pp. 4545–4563.

Data Sources & References

Sources

WRD_VRSNP_991d

WRD_VRN20_991d

Methodology

La mappa degli incendi attivi viene generata combinando i dati di due sensori VIIRS: Suomi NPP (NASA LANCE SNPP) e NOAA-20 (NASA LANCE NOAA20), entrambi con risoluzione nativa di 375 metri. I dati delle due collezioni vengono uniti per aumentare la copertura temporale. Per migliorare la risoluzione apparente, viene applicato un ricampionamento bicubico a 30 metri. Viene selezionato il valore massimo di confidenza nel periodo temporale analizzato. I corpi idrici vengono mascherati utilizzando ESA WorldCover (classe 80). Il livello di confidenza viene classificato in 5 categorie: 0 (nessun incendio), 1 (bassa confidenza, 1-30%), 2 (media confidenza, 30-60%), 3 (alta confidenza, 60-80%), 4 (confidenza molto alta, 80-100%). Viene inoltre eseguita un'analisi dei cluster per identificare raggruppamenti di pixel di fuoco, calcolando numero totale di incendi, area per livello di confidenza, e distanza dal centro dell'area di interesse.