From this initiative a two year project emerged with three
workgroups: (1) Benchmarking of forest fire prevention and combat
systems, (2) Support to forest fire prevention and combat information
systems, and (3) Forest surveillance, detection and alert on forest
fires.
workgroups: (1) Benchmarking of forest fire prevention and combat
systems, (2) Support to forest fire prevention and combat information
systems, and (3) Forest surveillance, detection and alert on forest
fires.
With a direct participation in this
initiative, trough the coordination of the second work group, Critical
Software implemented the Integrated Forest Fire Risk Map, originally
developed in the scope of Premfire project, a European Space Agency project, awarded to Critical Software in 2001.
initiative, trough the coordination of the second work group, Critical
Software implemented the Integrated Forest Fire Risk Map, originally
developed in the scope of Premfire project, a European Space Agency project, awarded to Critical Software in 2001.
Based on information collected through
satellite imagery (land cover, vegetation moisture and meteorological
data), integrated forest fire risk results in 232 meters spatial
resolution map with two versions (risk for the current day and a
forecast for the next day) produced on a daily basis.
satellite imagery (land cover, vegetation moisture and meteorological
data), integrated forest fire risk results in 232 meters spatial
resolution map with two versions (risk for the current day and a
forecast for the next day) produced on a daily basis.
Since land cover could easily become
outdated due to the large forest fires occurring every day, a
methodology was developed to automatically detect forest burnt areas.
This methodology, based on MODIS imagery, was implemented in order to
update land cover, thus eliminating false alerts occurring in these
areas. First results on this automatic burnt areas detection
methodology point to 18 % omission errors and 20 % commission errors,
although possible improvements were already identified and better
results are to be expected in the near future.
outdated due to the large forest fires occurring every day, a
methodology was developed to automatically detect forest burnt areas.
This methodology, based on MODIS imagery, was implemented in order to
update land cover, thus eliminating false alerts occurring in these
areas. First results on this automatic burnt areas detection
methodology point to 18 % omission errors and 20 % commission errors,
although possible improvements were already identified and better
results are to be expected in the near future.
With the end of forest fire season, risk
index effectiveness was calculated with the daily evaluation of burnt
areas incidence in the different risk classes.
index effectiveness was calculated with the daily evaluation of burnt
areas incidence in the different risk classes.
As it could be shown, in 2005, nearly 60%
of burnt areas larger than 50 ha occurred in the two higher risk
classes (Extreme and Very High), which corresponded to 39 % of the
Portuguese territory. However, when the same study is made with burnt
areas between 50 ha and 500 ha, the percentage of burnt area in the two
higher risk classes rise to 72%. These results mean, whenever a fires
run out of human control and reaches huge proportions, even low risk
areas are affected.
of burnt areas larger than 50 ha occurred in the two higher risk
classes (Extreme and Very High), which corresponded to 39 % of the
Portuguese territory. However, when the same study is made with burnt
areas between 50 ha and 500 ha, the percentage of burnt area in the two
higher risk classes rise to 72%. These results mean, whenever a fires
run out of human control and reaches huge proportions, even low risk
areas are affected.
|
Risk class
|
Area
of each class in the total territory /em> |
/div> | /em> |
Extreme
|
17.3
|
70.8
|
32.4
|
57.6
|
|
Very High
|
21.5
|
21.9
|
27.1
|
38.6
|
||||
|
High
|
24.5
|
7.3
|
25.3
|
27.8
|
||||
|
Medium
|
18.5
|
0.0
|
14.2
|
19.6
|
||||
|
Low
|
6.6
|
0.0
|
1.0
|
2.1
|
These forest fire risk maps became crucial tools to allocate firefighters on the field. Automatic burnt areas detection allowed not only to update fire risk maps but also provided an invaluable tool for rapid damage assesmet.
(Credits Critical Software S.A.)