The Light Environment of Plant Canopies
where NDVImi, is the NDVI with no vegetation and NDVI,, is the NDVI
with dense vegetation. Carlson, et al. (1995) set NDVI* equal to the fraction of vegetative cover, and this is a reasonable approximation, especially
when solar zenith angles are small. Clearly NDVI* varies from zero to
one over the range of vegetation cover and accounts for the observation
that NDVI increases more rapidly than the fraction of vegetation cover
as vegetation density increases.
Remote sensing from satellites has the possibility of sampling the
entire land surface of the earth daily at a 1 km spatial resolution on
the ground and a spatial resolution of 10 m, or less with less frequent
temporal sampling. Because of this phenomenal spatial sampling, much
effort has been expended to determine what biophysical quantities are
most closely related to the remote sensing observations. An examination
of Eq. (1 5.28) provides some useful insights here. Remember that optical
remote sensing from satellites is possible only under relatively clear-sky
conditions when atmospheric transparency is high, because satellites need
to view the surface with minimal contamination from the atmosphere.
From Eq. (15.28), when L, is small, reflection from the soil dominates
(third term on the right of Eq. (15.28)). As L, increases, the dominant
term in Eq. (15.28) becomes the scattering of intercepted near-infrared
beam radiation (first term on the right of Eq. (1 5.28)), which also happens
to be closely related to the intercepted PAR radiation. Table 15.2 contains
values of the three terms in Eq. (15.28), NDVI, NDVI*, and IPAR and
the fraction of canopy cover (f,) assuming
fc = ex~(-Kbe(O)Lr).
(15.33)
Clearly NDVI* is most closely related to IPAR and fraction vegetative
cover (fc) when @ is small (30"). The relation between NDVI* and IPAR
is likely to be better at other solar zenith angles because both NDVI*
and IPAR change but fc is fixed with @. The close relation between
NDVI* and IPAR occurs because intercepted solar radiation dominates
both variables; interception in the visible portion of the solar spectrum
dominates IPAR and interception in the NIR portion of the solar spectrum
dominates NDVI*.
The effects of leaf angle and sun zenith angle can be seen from Table
15.3. Clearly NDVI* is a reasonable predictor of fraction of IPAR for a
modest range of conditions. Since IPAR is closely related to vegetation
productivity potential (Eq. (14.13) with St replaced by IPAR and conversion efficiency e adjusted accordingly [e is about doubled]), remote
sensing has something significant to contribute to global vegetation studies. The robustness of the relation between NDVI* and fraction of IPAR
is further established by studies that have shown NDVI* to be related to
the fraction of IPAR associated with the green vegetation in canopies that
have both green and dead foliage.
Example 15.3. Compare the nadir, near-infrared BRF (BRFN) for a
canopy with a spherical leaf angle distribution (x = 1) with the hemi-
where NDVImi, is the NDVI with no vegetation and NDVI,, is the NDVI
with dense vegetation. Carlson, et al. (1995) set NDVI* equal to the fraction of vegetative cover, and this is a reasonable approximation, especially
when solar zenith angles are small. Clearly NDVI* varies from zero to
one over the range of vegetation cover and accounts for the observation
that NDVI increases more rapidly than the fraction of vegetation cover
as vegetation density increases.
Remote sensing from satellites has the possibility of sampling the
entire land surface of the earth daily at a 1 km spatial resolution on
the ground and a spatial resolution of 10 m, or less with less frequent
temporal sampling. Because of this phenomenal spatial sampling, much
effort has been expended to determine what biophysical quantities are
most closely related to the remote sensing observations. An examination
of Eq. (1 5.28) provides some useful insights here. Remember that optical
remote sensing from satellites is possible only under relatively clear-sky
conditions when atmospheric transparency is high, because satellites need
to view the surface with minimal contamination from the atmosphere.
From Eq. (15.28), when L, is small, reflection from the soil dominates
(third term on the right of Eq. (15.28)). As L, increases, the dominant
term in Eq. (15.28) becomes the scattering of intercepted near-infrared
beam radiation (first term on the right of Eq. (1 5.28)), which also happens
to be closely related to the intercepted PAR radiation. Table 15.2 contains
values of the three terms in Eq. (15.28), NDVI, NDVI*, and IPAR and
the fraction of canopy cover (f,) assuming
fc = ex~(-Kbe(O)Lr).
(15.33)
Clearly NDVI* is most closely related to IPAR and fraction vegetative
cover (fc) when @ is small (30"). The relation between NDVI* and IPAR
is likely to be better at other solar zenith angles because both NDVI*
and IPAR change but fc is fixed with @. The close relation between
NDVI* and IPAR occurs because intercepted solar radiation dominates
both variables; interception in the visible portion of the solar spectrum
dominates IPAR and interception in the NIR portion of the solar spectrum
dominates NDVI*.
The effects of leaf angle and sun zenith angle can be seen from Table
15.3. Clearly NDVI* is a reasonable predictor of fraction of IPAR for a
modest range of conditions. Since IPAR is closely related to vegetation
productivity potential (Eq. (14.13) with St replaced by IPAR and conversion efficiency e adjusted accordingly [e is about doubled]), remote
sensing has something significant to contribute to global vegetation studies. The robustness of the relation between NDVI* and fraction of IPAR
is further established by studies that have shown NDVI* to be related to
the fraction of IPAR associated with the green vegetation in canopies that
have both green and dead foliage.
Example 15.3. Compare the nadir, near-infrared BRF (BRFN) for a
canopy with a spherical leaf angle distribution (x = 1) with the hemi-