How Earth Observation images are used to detect and map forest fires?
Forest fires are one of the main causes of destruction of natural resources and this is observed on a global scale. Many countries worldwide are affected by this calamity, which is induced by nature and human activities. Using satellite data it is revealed that in last decade nearly 350 million hectares of land was affected by vegetation fires worldwide.
Remote sensing and GIS technologies are useful disciplines in studying land features. It assists in monitoring and knowing the causes of forest fires and understanding how to reduce its affects and find solution to numerous issues attached to it. Many researchers, for obtaining appropriate solutions have been working for more than last two decades. Jia et al. (2006) and Arroyo et al. (2008) used remote sensing techniques for fuel mapping; Fraser et al. (2003) and Maingi (2005) applied remote sensing for mapping burned land surfaces, whereas Flannigan and VonderHarr (1986) used the techniques for monitoring forest fires. In the literature it can be found that many workers studied on post-fire vegetation regeneration and fire management. Among them, Solan-Vila and Barbosa (2010); Diaz-Delgado and Pons (2001); Duncan et al. (2009); and Keramitsoglou et al. (2008) have cited good examples on post-fire studies using remote sensing approaches. Spectral signatures of Earth’s features have unique characteristics for identifying feature objects. Healthy living vegetation reflects radiation in near infrared (NIR) and absorbs red light in the visible region of the spectrum. Whereas, burnt areas reflect more radiation in the visible and short-wave infra red (SWIR) region and absorb more radiation in the NIR. Thus, features like forest burnt areas can be depicted easily using spectral signature technology and using specific algorithms that finds indices. Using indices has more advantage over using original bands of satellite data for interpretation because it helps in reducing anisotropic as well as atmospheric effect (Chuvieco et al., 2002).
Thermal Mapping of Forest Fires (TMFF) subsystem
This system is connected to the USGS website (http://e4ft101.cr.usgs.gov) to download MODIS (MOD14A1 and MYD14A1) fire products. MOD14A1 and MYD14A2 correspond respectively to the Terra and the Aqua satellite platforms of the MODIS sensors. These products are produced twice a day (one from Terra and one from Aqua) and are in raster hdf format, tiled and projected on a conical projection. The consisting pixels refer to the possibility for each area to be fire. The TMFF mosaics, resamples and reprojects these data. A user has only to give the spatial extend (location) and dates; the TMFF will download the corresponding MODIS products. For TMFF system a number of algorithms are written, in this work, to make the system automatic. The algorithms are written in Interactive Data Language (IDL), which traces and maps forest fire areas in shape files. The pixels that have the highest possibility of being of fires events are selected and vectorized. These vectors are imported to a GIS and the Regions of Interests (ROIs) are extracted.
Read more about detecting fires and mapping burned areas here.
