1. Stand Structure in Terrestrial Ecosystems
Scatterometry has also been used to study changes
in backscatter associated with changes in vegetation moisture status and soil moisture content (e.g.,
Way et aI., 1991). The main purpose of such studies has been to develop and validate models for
microwave scattering and attenuation by vegetation, and to develop a sound basis for interpreting
SAR images.
Profiling radars are nonimaging, airborne scatterometers that illuminate the surface with a nearvertical beam. These systems have direct application for forest inventory (e.g., Hyyppa and
Hallikainen 1996). Over the past decade several
systems have been deployed and tested in Europe.
For example, the Helsinki University of Technology scatterometer (HUTSCAT) is a multipolarization C- and X-band helicopter-borne scatterometer with vertical resolution of 0.65 m
(Hallikainen et aI. 1993). Profiling radars return
strong signals from the canopy top and from the
ground, permitting direct calculation of tree height.
Profiling radars also have demonstrated potential
to discriminate structurally different vegetation
types in diverse forests.
Synthetic Aperture Radar
Of available radar data, SAR data from aircraft
and orbiting satellites provide the most useful
information on stand structure over large areas
(reviewed by Waring et aI. [1995] and Kasischke
et aI. [1997]). SAR systems acquire multiple observations of a targeted area made from different
locations in order to obtain backscatter properties
at relatively high spatial resolution. Most spaceborne SAR systems have a fixed look angle between 20 and 50° with images covering a few degrees from near edge to far edge. An exception is
RADARSAT, whose angle can be varied from
from 20 to 59°. Today's spaceborne SAR systems
also have relatively fine spatial resolution of 20 to
40 m and sampling frequencies of 20 to 40 days.
SAR imagery has been used to map landcover
classes, inundation patterns, phytomass, canopy
cover, and the freeze/thaw status of trees (Hess et
al. 1990; Ranson and Sun 1997; Way et aI. 1997).
Despite their utility in cloudy regions and for deriving vegetation structural properties, SAR data
have had little application in ecosystem ecology.
Impediments to wider use include the complex na17
ture of polarized backscatter data (especially in
areas with significant vertical relief), poor data
coverage, image speckle, and the need for specialized image processing software for calibration,
rectification, and feature extraction (Kasischke et
al. 1997). However, in recent years several operational systems have been placed in orbit, calibrated data are more widely available, and software has become available for the nonspecialist.
Differences in the types of information that can
be obtained from radar and optical systems, such
as LANDSAT TM, can be illustrated by comparing images collected over the same region (Fig.
1.4). Spaceborne imaging radar systems that are
now in operation include: ERS-l, JERS-l, SIR-C/
X-SAR, RADARSAT, and ERS-2. The ERS-2
follow-on, known as advanced SAR or ASAR,
will be deployed on the planned European
ENVISAT and consists of a dual-polarized C-band
SAR. It will have both copolarized channels (HH
and VV), but not crosspolarized (HV). The
planned "wide swath" mode has a 400-km swath
width and 100-m resolution.
Radar Interferometry
Interferometric measures of vegetation structure
have the potential of providing canopy height,
crown depth, canopy attenuation (extinction coefficient), and the density of canopy elements (Hagberg et aI., 1995; Treuhaft et al., 1996), as well as
more standard measures of ground surface elevation (Zebker and Goldstein, 1986). For interferometric radar, the phase of each returning microwave pulse is measured along with backscatter
power and polarization. If two pulses travel
slightly different distances, they can interfere constructively, increasing the amplitude of the return
pulse, or destructively, creating a decrease. Different path lengths can be generated by shifting the
origin of the pulse, either by using two microwave
antennae separated by a known distance, known as
the baseline, or by synthesizing a baseline by analysis of two SAR data sets recorded on different
orbits or flights. Single-pass systems that use two
antennae separated by a known baseline may be
best for vegetation study, because vegetation
movement between orbits causes image decorrelation. However, the decorrelation itself may yield
information about vegetation structure and permit
Scatterometry has also been used to study changes
in backscatter associated with changes in vegetation moisture status and soil moisture content (e.g.,
Way et aI., 1991). The main purpose of such studies has been to develop and validate models for
microwave scattering and attenuation by vegetation, and to develop a sound basis for interpreting
SAR images.
Profiling radars are nonimaging, airborne scatterometers that illuminate the surface with a nearvertical beam. These systems have direct application for forest inventory (e.g., Hyyppa and
Hallikainen 1996). Over the past decade several
systems have been deployed and tested in Europe.
For example, the Helsinki University of Technology scatterometer (HUTSCAT) is a multipolarization C- and X-band helicopter-borne scatterometer with vertical resolution of 0.65 m
(Hallikainen et aI. 1993). Profiling radars return
strong signals from the canopy top and from the
ground, permitting direct calculation of tree height.
Profiling radars also have demonstrated potential
to discriminate structurally different vegetation
types in diverse forests.
Synthetic Aperture Radar
Of available radar data, SAR data from aircraft
and orbiting satellites provide the most useful
information on stand structure over large areas
(reviewed by Waring et aI. [1995] and Kasischke
et aI. [1997]). SAR systems acquire multiple observations of a targeted area made from different
locations in order to obtain backscatter properties
at relatively high spatial resolution. Most spaceborne SAR systems have a fixed look angle between 20 and 50° with images covering a few degrees from near edge to far edge. An exception is
RADARSAT, whose angle can be varied from
from 20 to 59°. Today's spaceborne SAR systems
also have relatively fine spatial resolution of 20 to
40 m and sampling frequencies of 20 to 40 days.
SAR imagery has been used to map landcover
classes, inundation patterns, phytomass, canopy
cover, and the freeze/thaw status of trees (Hess et
al. 1990; Ranson and Sun 1997; Way et aI. 1997).
Despite their utility in cloudy regions and for deriving vegetation structural properties, SAR data
have had little application in ecosystem ecology.
Impediments to wider use include the complex na17
ture of polarized backscatter data (especially in
areas with significant vertical relief), poor data
coverage, image speckle, and the need for specialized image processing software for calibration,
rectification, and feature extraction (Kasischke et
al. 1997). However, in recent years several operational systems have been placed in orbit, calibrated data are more widely available, and software has become available for the nonspecialist.
Differences in the types of information that can
be obtained from radar and optical systems, such
as LANDSAT TM, can be illustrated by comparing images collected over the same region (Fig.
1.4). Spaceborne imaging radar systems that are
now in operation include: ERS-l, JERS-l, SIR-C/
X-SAR, RADARSAT, and ERS-2. The ERS-2
follow-on, known as advanced SAR or ASAR,
will be deployed on the planned European
ENVISAT and consists of a dual-polarized C-band
SAR. It will have both copolarized channels (HH
and VV), but not crosspolarized (HV). The
planned "wide swath" mode has a 400-km swath
width and 100-m resolution.
Radar Interferometry
Interferometric measures of vegetation structure
have the potential of providing canopy height,
crown depth, canopy attenuation (extinction coefficient), and the density of canopy elements (Hagberg et aI., 1995; Treuhaft et al., 1996), as well as
more standard measures of ground surface elevation (Zebker and Goldstein, 1986). For interferometric radar, the phase of each returning microwave pulse is measured along with backscatter
power and polarization. If two pulses travel
slightly different distances, they can interfere constructively, increasing the amplitude of the return
pulse, or destructively, creating a decrease. Different path lengths can be generated by shifting the
origin of the pulse, either by using two microwave
antennae separated by a known distance, known as
the baseline, or by synthesizing a baseline by analysis of two SAR data sets recorded on different
orbits or flights. Single-pass systems that use two
antennae separated by a known baseline may be
best for vegetation study, because vegetation
movement between orbits causes image decorrelation. However, the decorrelation itself may yield
information about vegetation structure and permit