User Tools

Site Tools


Report a bug, broken link, or incorrect content

Fire Anomalies

by Megan Kanaga Creutzburg

Other Names:

Thermal anomalies
MODIS Thermal anomalies/fire product


Thermal remote sensing can be a useful method for tracking wildfires and assessing fire damage. MODIS satellites have been collecting thermal/fire anomaly data since 2000, and the data have been used successfully by fire ecologists to track and monitor fire activity across the globe. MODIS data can be used to monitor many aspects of fire patterns and consequences, including monitoring burn scars, vegetation composition and condition, smoke emissions, and water vapor / clouds. Fire is detected from the MODIS images using a contextual algorithm that uses the strong emission of mid-infrared radiation as a signature of fire activity (described in detail in Giglio et al. 2003). This produces an initial categorization of pixels into six categories: missing data, cloud, water, non-fire, fire, or unknown. The initial categorization often classifies nonfire pixels as fire pixels due differences in temperature and reflectance values across different locations. All pixels classified as potential fire pixels are then compared to neighboring pixels in a series of contextual threshold tests to determine the degree of temperature differentiation of fire pixels from the background of non-fire pixels. The resulting map identifies pixels that are likely to contain active fires, and classifies the active fire pixels based on confidence (i.e. high, nominal or low confidence that the pixel represents an active fire in reality). A database of active fire products is available with information on fire occurrence and location, the rate of thermal energy emission, and an estimate of the smoldering/flame ratio of the fire. All MODIS fire anomaly images have a spatial resolution of 1 kilometer and each satellite (Terra and Aqua) views the entire surface of the earth every one to two days.

Three MODIS fire products currently exist: MOD14 is the most basic fire product by MODIS that identifies active fires and volcanic activity. It is a level 2 product that is used to generate the higher level (MOD14A1 andMOD14A2) products. MOD14A1 is a level 3 (more highly processed than level 2) image of thermal activity that is composited every 24 hours and packaged into 8-day products. Images are distributed as tiles, which cover a large square geographic region, and pixel size within each tile is 1 kilometer. MOD14A2 is a level 3 image of thermal activity that is composited over an 8-day period as a summary product. Images are distributed as tiles and pixel size within each tile is 1 kilometer.

Similar Methods

  • Surface temperature - the MODIS fire products are created from remotely-sensed thermal imagery. There are other satellites/sensors with thermal bands, but without the high temporal frequency of MODIS, they are generally not as useful for mapping active fire locations.
  • Normalized Burn Index is a remote sensing technique for measuring the extent and severity of a burn after the fire has gone out.


MODIS produces global maps of fire occurrence through the MOD14, MOD14A1, MOD14A2 MODIS products. Burned area data that map the spatial extent of recent fires are also available (product MCD45A1)

Successful Rangeland Uses

  • Takahata et al. (2009) used remotely sensed MODIS fire data to understand fire concentration and temporal patterns of burns in sensitive habitats.
  • Urbanski et al. (2009) used MODIS fire data to map areas burned by wildfire across the western U.S.

Application References

  • Tahakata, C., R. Amin, P. Sarma, C. Banerjee, W. Oliver, and J.E. Fa. 2009. Remotely-Sensed Active Fire Data for Protected Area Management: Eight-Year Patterns in the Manas National Park, India. Environmental Management. DOI 10.1007/s00267-009-9411-8.
  • Urbanski, S.P., J.M. Salmon, B.L. Nordgren, and W.M. Hao. 2009. A MODIS direct broadcast algorithm for mapping wildfire burned area in the western United States. Remote Sensing of Environment 113: 2511–2526.

Technical References

  • Giglio, L., Descloitres, J., Justice, C.O., Kaufman, Y. 2003. An enhanced contextual fire detection algorithm for MODIS. Remote Sensing of Environment 87:273-282.
  • Justice, C., L. Giglio, L. Boschetti, D. Roy, I. Csiszar, J. Morisette, and Y. Kaufman. 2006. MODIS Fire Products Algorithm Technical Background Document. Version 2.3, EOS ID# 2741.
  • Morisette, J.T., L. Giglio, I. Csiszar, A. Setzer, and W. Schroeder. 2005. Validation of MODIS active fire detection products derived from two algorithms. Earth Interactions 9: 1-25.


Small or relatively cool-burning fires may not be detected via MODIS imagery. Some images have a large number of pixels that are classified as unknown, but more recent algorithms are able to classify pixels with greater confidence. Additionally, while the Terra and Aqua satellites have very high temporal frequency for a satellite imager, fires burning in arid environments (especially those dominated by fine fuels like annual grasses) may change significantly between successive satellite revisits.

Data Inputs

Remotely sensed images from two MODIS satellites (Terra and Aqua) are processed into level 2 and 3 MODIS thermal anomaly products, which are available for free to the public.

Software/Hardware Requirements

Software and hardware requirements will depend on the use of fire anomaly products. A GIS or remote sensing program will be required to visualize maps, and other programs may be required for additional image processing.

Sample Graphic

MODIS image of thermal anomalies in a tile of northern Australia. Gray pixels represent non-fire land areas, blue represents water, white represents active fire, and yellow pixels are unknown. This image is cloudless, but clouds would be shown in purple if cloud cover existed.

High-resolution Landsat ETM+ images showing an unburned landscape in Tanzania (upper left) burn scars from a 2000 fire (upper right) and the fire pixels detected by MODIS satellites overlaid on the burn scars (lower right). (source: Case study 5.2: Remote sensing of fire disturbance in the Rungwa Ruaha landscape, Tanzania, CBD Technical Series Number 32, Convention on Biological Diversity. Available online at

Additional Information

Existing datasets

Web Search Results



You must have an account and be logged in to post or reply to the discussion topics below. Click here to login or register for the site.

remote_sensing_methods/fire_anomalies.txt · Last modified: 2013/02/07 12:58 by gtucker