update
tion on
its, the
t larger
mation
ction of
IE/EEA
change
iposium
ROM is
| for the
f the 4^
| for the
uropean
echnical
odology
Project,
'ciences,
asibility
essment.
Ireland),
al guide
gramme,
Photo-
ared for
ver data
e. Phare
ng — ett
REMOTE SENSING TOOLS FOR
FOREST FIRE PREVENTION AND ASSESSMENT
Emilio Chuvieco, Javier Salas and David Riafio
Department of Geography
University of Alcalá de Henares
Colegios, 2 - 28801 Alcalá de Henares (Spain)
KEY WORDS: Forest Fires, Fuel Moisture Content, Fire Detection, Burned area mapping.
ABSTRACT:
Forest fires are a major environmental concern in Tropical and Mediterranean countries, since they are the
main factor of vegetation degradation. Satellite remote sensing data could play a crucial role in improving
current tools of fire risk estimation and fire effects assessment, since satellite images provide frequent
coverage at proper resolution to monitor vegetation status. Real-time monitoring of fuel moisture content
and burned land maps are examples of products closely connected to fire management planning. Future
systems designed for operational fire detection might greatly improve current alarm systems. This paper
explores present applications and future requirements of remote sensing systems for forest fire studies, with
special emphasis in the needs of European Mediterranean countries.
1. INTRODUCTION
Remote sensing from space offers several
opportunities for forest fire research. The wide
area and frequent coverage provided by satellite
sensors make them a very valuable tool for
prevention, detection and mapping of wildland
fires (Chuvieco and Martin, 1994). Remote
Sensing contributions may be classified into four
types of applications, related to the temporal scale
that is being emphasised (Chuvieco and Cocero,
1996): short-term fire danger estimation, long-
term fire risk assessment (both are related to pre-
fire management), fire detection (during the fire),
and fire effects assessment (post-fire). The most
recent research issues related to these three
phases of fire management are reviewed in this
paper.
2. THE ESTIMATION OF FUEL
MOISTURE CONTENT
Current methods of fire danger estimation rely
upon the estimation of fuel moisture content
(FMC) from meteorological variables (Viegas et
al., 1994). Most meteorological danger indices are
assumed related to the moisture content of dead
fuels, but this relationship has not been
extensively proven by field data, since only the
Canadian system (Van Wagner, 1987) is based on
actual FMC measurements. On the other hand,
litlle attention has been paid to the estimation of
FMC of live fuels, which are also quite critical in
fire's rate of spread. Furthermore, even accepting
a good agreement between meteo-indices and
FMC status, meteorological variables are only
available for specific observation points, which
may not be properly located for fire applications.
Most commonly, weather stations are located
close to urban areas, sometimes far from forested
areas and, therefore, interpolation-extrapolation
algorithms need to be applied to infer danger
conditions of fire-prone areas. The performance
of those algorithms may be poor, because of
topographic variability or the influence of local
conditions (Hubbard, 1994).
Satellite data might be a complementary source of
information to improve current fire danger
indices, avoiding some limitations of current
meteo-indices. Satellite-borne sensors offer a
systematic spatial sampling and are directly
related to vegetation status. However, the
relations between satellite detected radiances and
FMC need to be proven. Considering that dead
fuels are on the floor, they will not be directly
sensed by satellites if located in a densely
vegetated area. Therefore, satellite observation is
mostly reduced to live fuels located at the canopy
layer. On the other hand, moisture content is a
particularly difficult parameter to estimate from.
reflectance measurements, as it accounts for little
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 639
