16
illumination and atmospheric turbidity. In this section, we focus on three specific radar systems that
can provide valuable structural data. These include
profiling radars, imaging synthetic aperture radar
(SAR), and interferometric SAR.
Radar sensors operate by transmitting microwave radiation of a specific wavelength and polarization and then measuring the power of backscatter from the illuminated surface. Radars may
transmit horizontally (H) or vertically (V) polarized
microwave energy and may receive either polarization, resulting in four linear polarizations: HH,
HV, VH, and VV. Backscatter is measured as radar
"cross-section" in square meters. This scattering
"area" is equivalent to actual area only for a perfectly reflecting isotropic target. As a result of forward scattering and absorption, most natural targets
have cross-sections smaller than their physical
areas, but, in some cases, the scattering geometry
focuses energy back toward the antenna, causing
very bright returns. The "normalized backscatter
coefficient" (0"0) equals the backscatter crosssection, 0", normalized by actual horizontal pixel
area. This coefficient is unity for a perfectly reflecting isotropic surface and ranges over many orders
of magnitude depending on radar wavelength, polarization, and local incidence angle, as well as on
the size, orientation, and dielectric properties of the
target. For this reason, backscatter is usually reported as 0"0 in decibels (that is, 0"0 = 10 * 10glO(0"1
Area).
Structure and water content are the major controls of vegetation on radar backscatter. Soil surface roughness and moisture may also have a large
effect on backscatter from low-biomass or discontinuous vegetation and at steep radar incidence
angles where microwaves penetrate the canopy.
Because of their high moisture content, plant structural components (leaves, branches, trunks, fruits)
have high dielectric constants and thus behave as
microwave scatterers and attenuators. Both microwave scattering and attenuation depend on the size
and orientation of the scattering elements in relation to the wavelength, polarization, and incidence
angle of the radar. Longer wavelengths (P-band
[30 to 100 cm] and L-band [15 to 30 cm]) penetrate deeply into a forest canopy, attenuated to
some degree by leaves and small branches. Thus,
backscatter originates mainly from large branches
and (in some forests) trunk-ground "doubleFrank W. Davis and Dar Roberts
bounce" reflections, in which microwaves are reflected from the ground surface, then off of standing stems, and finally back to the radar. Shorter
wavelengths (C-band [3.75 to 7.5 cm] and X-band
[2.4 to 3.75 cm]) are backscattered by small
branches and leaves; penetration is shallow, except
as it may occur through canopy gaps. Dry soils are
more absorptive than wet soils, and flooded, vegetated surfaces can generate very high backscatter
because the smooth reflective water surface enhances double-bounce reflection (Hess et al.
1990). P- and L-bands may find their best applications in forests (e.g., detection of flooding, estimation of biomass, monitoring forest regrowth)
while C- and X-bands may be best applied to lowbiomass vegetation (e.g., rangelands, wetlands, agriculture). Because the various bands and polarizations have been found highly complementary,
multiband, multipolarization radars are desirable
for vegetation study.
The round-trip travel time of the microwave signal from the radar antenna to the target provides
a precise measure of distance to the target. Different systems utilize the positional information in
different ways. Applied to vegetation canopies, a
ranging scatterometer records backscatter as a
function of distance into the canopy. A scatterometer or radar profiler traveling above the canopy
and pointed downward samples the canopy depth
profile. Radars such as SAR that are aimed
obliquely down to the side of the flight track construct two-dimensional images of the canopy, in
which each pixel represents backscatter integrated
through the canopy depth. Interferometric imaging
SARs can provide information on the depth dimension as well.
Scatterometers and Radar Profilers
Any calibrated radar can be a scatterometer, but
scatterometers used for study of vegetation are
most commonly nonimaging, ranging scatterometers, which record backscatter as a function of
range distance (based on signal travel time). Typical resolutions are on the order of 10 cm to several meters. Laboratory scatterometers and truckmounted field scatterometers have been used
extensively to examine scattering and attenuation
properties of plant structural components and canopies, forest included (e.g., Zoughi et al. 1986).
illumination and atmospheric turbidity. In this section, we focus on three specific radar systems that
can provide valuable structural data. These include
profiling radars, imaging synthetic aperture radar
(SAR), and interferometric SAR.
Radar sensors operate by transmitting microwave radiation of a specific wavelength and polarization and then measuring the power of backscatter from the illuminated surface. Radars may
transmit horizontally (H) or vertically (V) polarized
microwave energy and may receive either polarization, resulting in four linear polarizations: HH,
HV, VH, and VV. Backscatter is measured as radar
"cross-section" in square meters. This scattering
"area" is equivalent to actual area only for a perfectly reflecting isotropic target. As a result of forward scattering and absorption, most natural targets
have cross-sections smaller than their physical
areas, but, in some cases, the scattering geometry
focuses energy back toward the antenna, causing
very bright returns. The "normalized backscatter
coefficient" (0"0) equals the backscatter crosssection, 0", normalized by actual horizontal pixel
area. This coefficient is unity for a perfectly reflecting isotropic surface and ranges over many orders
of magnitude depending on radar wavelength, polarization, and local incidence angle, as well as on
the size, orientation, and dielectric properties of the
target. For this reason, backscatter is usually reported as 0"0 in decibels (that is, 0"0 = 10 * 10glO(0"1
Area).
Structure and water content are the major controls of vegetation on radar backscatter. Soil surface roughness and moisture may also have a large
effect on backscatter from low-biomass or discontinuous vegetation and at steep radar incidence
angles where microwaves penetrate the canopy.
Because of their high moisture content, plant structural components (leaves, branches, trunks, fruits)
have high dielectric constants and thus behave as
microwave scatterers and attenuators. Both microwave scattering and attenuation depend on the size
and orientation of the scattering elements in relation to the wavelength, polarization, and incidence
angle of the radar. Longer wavelengths (P-band
[30 to 100 cm] and L-band [15 to 30 cm]) penetrate deeply into a forest canopy, attenuated to
some degree by leaves and small branches. Thus,
backscatter originates mainly from large branches
and (in some forests) trunk-ground "doubleFrank W. Davis and Dar Roberts
bounce" reflections, in which microwaves are reflected from the ground surface, then off of standing stems, and finally back to the radar. Shorter
wavelengths (C-band [3.75 to 7.5 cm] and X-band
[2.4 to 3.75 cm]) are backscattered by small
branches and leaves; penetration is shallow, except
as it may occur through canopy gaps. Dry soils are
more absorptive than wet soils, and flooded, vegetated surfaces can generate very high backscatter
because the smooth reflective water surface enhances double-bounce reflection (Hess et al.
1990). P- and L-bands may find their best applications in forests (e.g., detection of flooding, estimation of biomass, monitoring forest regrowth)
while C- and X-bands may be best applied to lowbiomass vegetation (e.g., rangelands, wetlands, agriculture). Because the various bands and polarizations have been found highly complementary,
multiband, multipolarization radars are desirable
for vegetation study.
The round-trip travel time of the microwave signal from the radar antenna to the target provides
a precise measure of distance to the target. Different systems utilize the positional information in
different ways. Applied to vegetation canopies, a
ranging scatterometer records backscatter as a
function of distance into the canopy. A scatterometer or radar profiler traveling above the canopy
and pointed downward samples the canopy depth
profile. Radars such as SAR that are aimed
obliquely down to the side of the flight track construct two-dimensional images of the canopy, in
which each pixel represents backscatter integrated
through the canopy depth. Interferometric imaging
SARs can provide information on the depth dimension as well.
Scatterometers and Radar Profilers
Any calibrated radar can be a scatterometer, but
scatterometers used for study of vegetation are
most commonly nonimaging, ranging scatterometers, which record backscatter as a function of
range distance (based on signal travel time). Typical resolutions are on the order of 10 cm to several meters. Laboratory scatterometers and truckmounted field scatterometers have been used
extensively to examine scattering and attenuation
properties of plant structural components and canopies, forest included (e.g., Zoughi et al. 1986).