Wildland Fire Detection from Space: Theory and Application

Abstract

New satellite instruments are currently being designed specifically for fire detection, even though to date the detection of active fires from space has never been an integral part of the design of any in-orbit space mission.Rather, the space-based detection of fires during the last two decades has been exploiting measurements obtained for other objectives. The current fire products have proved to be of great benefit and interest, but their usefulness is not fully understood. Part of the confusion about the utility of these measurements stems from the lack of detailed knowledge about the data and its acquisition. The remote sensing research community has spent considerable time and effort trying to rationalize the usefulness of existing satellite imagery for active fire detection. Unfortunately, uncertainties about instrument capabilities pervades much of this research and the true limits of fire detection from space have not been fully evaluated and understood.To analyze the active fire detection capability of any instrument, the flow of energy from the source to the instrument and the instrument's response to that energy must be considered. For this reason, an approach has been developed that models the energy emitted from surface fires, allowing for the fact that fire is itself a variable phenomenon. The energy transmission is then modeled along its path through the atmosphere and through the instrument's optical system. A fundamental concern is in the estimation of the total surface area that emits the energy which defines a single pixel in the image. Unfortunately, most of the fire detection modeling done to date is based on a misconception about the pixel and its actual size. Rather than using the radiometric footprint size, the instantaneousfield-of-view (IFOV) is used to describe the 'resolution' of the instrument. In fact, the radiometric footprint is considerably larger than the IFOV and greatly affects the energy modeling used to estimate the fire detection thresholds of a particular instrument. Based on knowledge of the radiometric footprint, the fire detection capability of AVHRR, DMSP-OLS, and MODIS are reviewed. 2"the pixel, the elementary unit of analysis in remote sensing..., is a delusion which may become a snare for the unwary", P. Fisher (1997).

ENVISAT multisensor data for fire monitoring and impact assessment

“…Each AATSR hotspot group coincided with fire scars of Landsat ETM. However due to the coarse spatial resolution of 1 km the AATSR instrument may not detect small or low-intensity fires similarly to MODIS (Siegert et al 2004) and AVHRR (Cahoon et al 2000).…”

Temporal and spatial variability in biomass burned areas across the USA derived from the GOES fire product

“…The quality flag represents the confidence of fire detections with six different levels that indicate a fire pixel being processed (flag 0 -subpixel instantaneous estimation of fire size and temperature), saturated (flag 1 -saturated fire pixel), cloud-contaminated (flag 2 -cloud-contaminated fire pixel), high probability (flag 3 -high probability fire pixel), medium probability (flag 4 -medium probability fire pixel), and low probability (flag 5 -low probability fire pixel). To minimize false fires caused by cloud edges, extreme solar zenith angles, and sensor noise related to uncertainty in radiance detection, interchannel spatial misregistration, geo-location, and Point Response Function (Cahoon et al, 2000;Giglio & Kendall, 2001;Robinson, 1991), the WF_ABBA uses a temporal filter to exclude the fire pixels that are only detected once within the past 12 h (Schmidt & Prins, 2003). In this study, we collect the GOES WF_ABBA fire data between January 2000 and December 2006 across CONUS.…”

Fire Detection and Fire Characterization Over Africa Using Meteosat SEVIRI

“…The confidence values of these data have a peak correctly shifted toward higher confidence values when compared to that of the false detections, and thus, the confidence parameter is behaving sensibly in this regard. The majority of the false-alarm fire pixels still have a confidence value exceeding 0.5, and this is simply a result of the brightness temperature limits used to define a strong confidence and weak confidence fire pixels (10) and the fact that the detection algorithm is designed only to confirm potential fire pixels as true fire pixels when there is a reasonable level of confidence that this is correct (i.e., to minimize errors of commission as far as possible). A similar effect relating to relatively high confidence values for false-alarm fire pixels has been noted in the MODIS fire products [20].…”