Remote Sensing of Terrestrial Photosynthesis
517
many ecosystem features (Ustin et al. 1993). The green vegetation fraction
is highly correlated with NDVI (Gamon et al. 1993), and may provide
estimates of FPAR, along with additional information.
The question of which vegetation indices are related to which vegetation
parameters and why has several facets. Some of these are biophysical, some
are physiological, and some are ecological. Choudhury (1987), Kumar and
Monteith (1981), Sellers (1985), and Sellers (1987) have explored the biophysical basis of the relationships and more or less agree on the conclusion
that FPAR should be more nearly linear with SR than with NDVI. FPAR
gradually saturates as LAI increases above 1 ,implying that the ability of the
vegetation indices to predict LAI should decrease at high LAI. Because
canopy photosynthesis also saturates with increasing LAI, the two curvilinear functions tend to cancel each other, making photosynthesis a linear
or nearly linear function of FP AR. This is the conceptual basis for the
numerical results of Choudhury (1987), Sellers (1985), and Sellers (1987),
who used models that assumed constant photosynthetic capacity per unit of
leaf area, from the top to the bottom of the canopy. With a gradient in
photosynthetic capacity tuned to parallel the gradient in average illumination, the linear SR/photosynthesis relationship is preserved, and the slope
of the relationship becomes much less sensitive to the spatial organization of
the canopy within the view area (Sellers et al. 1992a,b).
From the ecological perspective, parallel saturation in FP AR and canopy
photosynthesis with increasing LAI are not entirely sufficient to account for
a strong relationship between a vegetation index and yield or net primary
production. This kind of relationship also implies that photosynthesis and
growth vary in parallel with light absorption. The explanation for this
phenomenon has trivial and deep components, with relative importances
that remain to be quantified. The trivial component is related to the fact that
growth drives light absorption. Light absorption, therefore, varies in parallel
with size in anything that grows, including animals and inanimate objects.
The deep component is related to the observation that the efficienty of light
utilization in plant growth (the ratio of annual plant yield to absorbed
radiation) is surprisingly constant across a range of ecosystems (Russell et
al. 1989) and over a range of environmental stresses (Fig. 24.3). This
observation implies that plants generally limit investments in light harvesting
(canopy development) to levels consistent with the availability of the factor
that limits yield or net primary production, independent of whether that
factor is a resource (e.g., water or nutrients) or an inherent constraint of
growth potential (Field 1991).
The balancing of canopy development to the availability of the resource
or resources that limit growth is likely to be regulated on different time
scales and with different fidelities in different ecosystems. In herbaceous
ecosystems, FP AR, growth, and photosynthetic capacity potentially track
growth-limiting resources on a time scale of days to weeks (Figs. 24.2 and
24.3). In evergreen ecosystems, persistence of a canopy through periods
517
many ecosystem features (Ustin et al. 1993). The green vegetation fraction
is highly correlated with NDVI (Gamon et al. 1993), and may provide
estimates of FPAR, along with additional information.
The question of which vegetation indices are related to which vegetation
parameters and why has several facets. Some of these are biophysical, some
are physiological, and some are ecological. Choudhury (1987), Kumar and
Monteith (1981), Sellers (1985), and Sellers (1987) have explored the biophysical basis of the relationships and more or less agree on the conclusion
that FPAR should be more nearly linear with SR than with NDVI. FPAR
gradually saturates as LAI increases above 1 ,implying that the ability of the
vegetation indices to predict LAI should decrease at high LAI. Because
canopy photosynthesis also saturates with increasing LAI, the two curvilinear functions tend to cancel each other, making photosynthesis a linear
or nearly linear function of FP AR. This is the conceptual basis for the
numerical results of Choudhury (1987), Sellers (1985), and Sellers (1987),
who used models that assumed constant photosynthetic capacity per unit of
leaf area, from the top to the bottom of the canopy. With a gradient in
photosynthetic capacity tuned to parallel the gradient in average illumination, the linear SR/photosynthesis relationship is preserved, and the slope
of the relationship becomes much less sensitive to the spatial organization of
the canopy within the view area (Sellers et al. 1992a,b).
From the ecological perspective, parallel saturation in FP AR and canopy
photosynthesis with increasing LAI are not entirely sufficient to account for
a strong relationship between a vegetation index and yield or net primary
production. This kind of relationship also implies that photosynthesis and
growth vary in parallel with light absorption. The explanation for this
phenomenon has trivial and deep components, with relative importances
that remain to be quantified. The trivial component is related to the fact that
growth drives light absorption. Light absorption, therefore, varies in parallel
with size in anything that grows, including animals and inanimate objects.
The deep component is related to the observation that the efficienty of light
utilization in plant growth (the ratio of annual plant yield to absorbed
radiation) is surprisingly constant across a range of ecosystems (Russell et
al. 1989) and over a range of environmental stresses (Fig. 24.3). This
observation implies that plants generally limit investments in light harvesting
(canopy development) to levels consistent with the availability of the factor
that limits yield or net primary production, independent of whether that
factor is a resource (e.g., water or nutrients) or an inherent constraint of
growth potential (Field 1991).
The balancing of canopy development to the availability of the resource
or resources that limit growth is likely to be regulated on different time
scales and with different fidelities in different ecosystems. In herbaceous
ecosystems, FP AR, growth, and photosynthetic capacity potentially track
growth-limiting resources on a time scale of days to weeks (Figs. 24.2 and
24.3). In evergreen ecosystems, persistence of a canopy through periods