6. REMOTE SENSING OF FOREST FIRES
49
limited radiometric resolution that may lead to confusions with other covers
(Kennedy et al, 1994). Several prototypes (such as the Fuego and the Focus
programs) are being studied to overcome the deficiencies of AVHRR data,
namely to increase the temporal coverage and radiometric sensitivity. However,
even with dedicated optical systems, fire detection from space can still be
unsuccessful, because of cloud contamination or topographic shadows, which
may hide the fire underneath. Little experience is available in using microwave
systems have not been tested for fire detection.
2.3
Fire assessment
Burned land inventorying and mapping are very important to assess effects
of fire on vegetation and soil. Spectral discrimination of burned areas is critical
for fire effects assessment (Pereira et al., 1997). However, this method faces
several difficulties, such as the confusion with non-vegetated surfaces or
shaded areas. Several single-date and multi-temporal techniques have been
proposed to improve burned land mapping. The most successful approaches
have relied on the use of analytical techniques (vegetation indices, principal
component analysis or spectral mixture analysis), and simple multi-temporal
methods (subtraction or ratios). When working at a global scale, AVHRR data
need to be corrected from atmospheric and cloud effects, which introduce
severe distortions in the original data.
High-resolution data, such as Landsat or SPOT images, are more adequate
then the global low resolution sensors for the detail assessment of fire-affected
areas. In this case, it has been possible to discriminate burning intensities, by
comparing radiances from before and after the fire (Chuvieco and Congalton,
1988). This information may be critical to reduce soil erosion after fire, since
rapidly growing species may be planted in those sectors more severely affected
by the fire (Isaacson et al., 1982). Monitoring recovery after fire also provides
critical information to better understand fire effects on vegetation species.
Finally, satellite images may be used to analyze changes in landscape pattern as
a result of fire.
3.
THE NEED OF USER-ORIENTED DATA
In all three aspects previously reviewed, the use of satellite data greatly
enhances forest fire management. However, limitations of current systems
notably reduce the operational application of remote sensing technologies to
this field and, to some extent, to other environmental hazards.
These shortcomings are partially due to the lack of specific orientations
of remote sensing missions. Most satellite platforms were designed for
49
limited radiometric resolution that may lead to confusions with other covers
(Kennedy et al, 1994). Several prototypes (such as the Fuego and the Focus
programs) are being studied to overcome the deficiencies of AVHRR data,
namely to increase the temporal coverage and radiometric sensitivity. However,
even with dedicated optical systems, fire detection from space can still be
unsuccessful, because of cloud contamination or topographic shadows, which
may hide the fire underneath. Little experience is available in using microwave
systems have not been tested for fire detection.
2.3
Fire assessment
Burned land inventorying and mapping are very important to assess effects
of fire on vegetation and soil. Spectral discrimination of burned areas is critical
for fire effects assessment (Pereira et al., 1997). However, this method faces
several difficulties, such as the confusion with non-vegetated surfaces or
shaded areas. Several single-date and multi-temporal techniques have been
proposed to improve burned land mapping. The most successful approaches
have relied on the use of analytical techniques (vegetation indices, principal
component analysis or spectral mixture analysis), and simple multi-temporal
methods (subtraction or ratios). When working at a global scale, AVHRR data
need to be corrected from atmospheric and cloud effects, which introduce
severe distortions in the original data.
High-resolution data, such as Landsat or SPOT images, are more adequate
then the global low resolution sensors for the detail assessment of fire-affected
areas. In this case, it has been possible to discriminate burning intensities, by
comparing radiances from before and after the fire (Chuvieco and Congalton,
1988). This information may be critical to reduce soil erosion after fire, since
rapidly growing species may be planted in those sectors more severely affected
by the fire (Isaacson et al., 1982). Monitoring recovery after fire also provides
critical information to better understand fire effects on vegetation species.
Finally, satellite images may be used to analyze changes in landscape pattern as
a result of fire.
3.
THE NEED OF USER-ORIENTED DATA
In all three aspects previously reviewed, the use of satellite data greatly
enhances forest fire management. However, limitations of current systems
notably reduce the operational application of remote sensing technologies to
this field and, to some extent, to other environmental hazards.
These shortcomings are partially due to the lack of specific orientations
of remote sensing missions. Most satellite platforms were designed for