Geoscience Reference
In-Depth Information
Profi ling LiDAR instruments like SLICER (Ni-Meister et al., 2001; Lefsky et al.,
2002; Kotchenova et al., 2004) or ICESAT-GLAS (Zwally et al., 2002; Harding
and Carabajal, 2005; Lefsky et al., 2005; Lefsky et al., 2006) do not produce images
but have the advantage of recording full waveforms of signal returns, which can be
used to estimate forest canopy height and more generally the 3D target structure.
Environmental Applications
These basic techniques of Earth Observation are now being applied to address a
vast array of practical concerns ranging from environmental monitoring to mapping
the three-dimensional pattern of major cities for fl ight simulator software, to inter-
planetary observation and monitoring compliance with international conventions
on nuclear non-proliferation. However, most of the research done by geographers
and environmental scientists has focused on applying Earth Observation methods
and data to understand various processes of change in land cover and the earth
surface.
Fires
Biomass burning is a major contributor to global carbon dioxide fl uxes from the
land to the atmosphere (Patra et al., 2005). Wildfi res also dramatically decrease
the surface albedo (Eugster et al., 2000), with potential feedbacks to the climate
system. With climate change, the frequency and intensity of fi re regimes in some
parts of the world are increasing. Because of their economic impacts on the forestry
sector, as well as on air quality and ecosystems, wildfi res are often classifi ed as
disasters.
Remote sensing is used to routinely monitor the occurrence of wildfi res. Three
basic approaches can be distinguished: (i) the detection of 'hot spots' (active fi re
detections) using thermal infrared data (Kant et al., 2000; Stroppiana et al., 2000;
Barducci et al., 2002; Schultz, 2002; Lasaponara et al., 2003; Li et al., 2003; Soja
et al., 2004; San-Miguel-Ayanz et al., 2005; Csiszar et al., 2006); (ii) 'burned
area' mapping, typically using near-infrared or middle infrared wavelengths
(BourgeauChavez et al., 1997; Fraser et al., 2000; Diaz-Delgado and Pons, 2001;
Brivio et al., 2003; Sukhinin et al., 2004; Balzter et al., 2005; George et al., 2006;
Gerard et al., 2003); and (iii) the measurement of fi re radiative power, to infer fi re
radiative energy which can be related to carbon release (Wooster et al., 2003).
The method of active fi re detection can serve as a near-real time early warning
system, but it has the drawback that it can fail to detect fi res under dense clouds
or between satellite overpasses. Depending on the sensor characteristics, false detec-
tions caused by hot surfaces or sun glint can also be a problem. Active fi re data are
generally not an accurate estimate of burned area. Examples of global active fi re
datasets are the World Fire Atlas at the European Space Agency (http://dup.esrin.
esa.int/ionia/wfa/), the Fire Information for Resource Management System (FIRMS,
http://maps.geog.umd.edu/fi rms/) or the World Fire Web at the European Commis-
sion's Joint Research Centre.
Burned area mapping approaches are based on various types of change detection
methods. Some authors have used the changes in Normalised Difference Vegetation
Index (NDVI) before and after a fi re to map the burned area (Fraser et al., 2000).
Others have used the Short-wave Infrared Index (NDSWIR) to map burned areas.
