Gross primary production (GPP) represents the total rate of production by photosynthesis, including the organic matter used up by respiration. Net primary production (NPP) is the net rate of production in excess of respiration (what's left after respiration has been accounted for). The rate and amount of both GPP and NPP are determined by available resources (including nutrients and water) and by the physical environment (including soils and climate).

NPP provides an indication of the amount of carbon dioxide removed from the atmosphere that ends up "fixed" in vegetation. Accurate global maps of NPP can help in the study of uncertainties and perturbations in the carbon cycle, and resolve discrepancies in the location and magnitude of terrestrial carbon sources and sinks. Maps of NPP over large areas are also relevant to issues of land use change, food security and vegetation feedback on climate.

A number of approaches to NPP estimation have been developed; a promising technique utilizes satellite remote sensing data to infer vegetation light capture and light use efficiency. The radiation used by plants in photosynthesis is discussed below in detail (Section 6.1). Modeling of light interaction with vegetation began with simple models that described light penetration through plant canopies as an exponential reduction of light with canopy depth. These models assumed that canopies behave like big leaves. Because canopies are not spatially uniform, this assumption is not robust. Plant canopies typically have gaps, clumps and a variety of discontinuous spatial patterns.

More sophisticated models have since been developed that trace light ray paths through plant canopies, and some models mathematically characterize the canopies using fractal geometry. Such models require a great deal of computing power and information about plant canopies not usually available over large areas. Thus these models are not practical for use on regional or global areas unless some simplifying principles can be applied to generalize the process (canopy light absorption) described by the models.

Vegetation light capture and absorption is generally not obtained directly by remote sensing instruments. Instead, spectral absorption is inferred from remotely sensed spectral reflectance. Simplified, the proportion of light absorbed is 1 minus the proportion of light reflected (measured by remote sensors). The more light reflected in a given wavelength band means that less light is absorbed in that same band.

6.1 Photosynthetically Active Radiation (PAR)

Photosynthetically active radiation (PAR) is actually restricted to just a portion of sunlight's spectrum -- from 400 to 700 nanometers (nm) or 0.4 to 0.7 micrometers (µm) which is comparable to the range of light the human eye can see. Absorption of light by chlorophyll takes place largely within narrow bands that peak at 680 nm and 700 nm. Even though green light (~550 nm) is within the PAR region, a greater proportion of it is reflected compared to other PAR wavelengths, which is why plants appear green to humans. Light absorbed outside the PAR region may be disadvantageous to plants because it can raise leaf temperatures and require heat dissipation (through transpiration).

A number of terms involving PAR need to be distinguished and explained.

6.1.1 Incident PAR -- The amount of PAR incident at the top of the atmosphere is known, modeled using Earth-Sun geometry and known solar radiation. It varies seasonally with changes in Earth's orbit and tilt. The amount of PAR incident at the top of a vegetation canopy varies with latitude and local topography, and throughout the day with changing sun angle, variations in cloud cover, and atmospheric properties (such as moisture and particulates). Thus any given place may have very different amounts of incident PAR. Simple maps of surface incident PAR have been generated using top-of-atmosphere values adjusted by satellite observations of cloud cover to account for the reduction of PAR caused by clouds.

6.1.2 Intercepted PAR (IPAR), absorbed PAR (APAR), and fractional PAR (fPAR) -- Intercepted PAR (IPAR) is the amount of PAR caught by various canopy layers as the PAR incident at the canopy top travels down through canopy layers to the ground. Absorbed PAR (APAR) is the amount of PAR actually absorbed by canopy layers. APAR takes into account PAR reflected off the canopy top back into the atmosphere and PAR reflected off the soil or background material back into the canopy.

The difference between IPAR (intercepted PAR) and APAR (absorbed PAR) depends on canopy closure, coverage over the background materials, canopy composition, density, and reflectance. If a canopy has complete, dense coverage and consists of green leaves, then IPAR may be a good approximate of APAR since healthy green leaves do not reflect much PAR. For open canopies or canopies that consist of both green and brown materials, IPAR and APAR may differ. In sparse or open canopies, less PAR is intercepted and more PAR is transmitted to (and potentially reflected back by) the ground surface (Figure 4.06).

Fractional PAR simply computes what fraction of the incident PAR (reaching the canopy top) is either intercepted (fIPAR) or absorbed (fAPAR).

fIPAR (the fraction of intercepted PAR) = IPAR/PAR incident at the canopy top

fAPAR (the fraction of absorbed PAR) = APAR/PAR incident at the canopy top

This terminology is technically the most specific and its use is recommended, however much literature uses the term fPAR (fractional PAR) interchangeably with fAPAR (the fraction of absorbed PAR). The values of fIPAR and fAPAR have a potential range from 0 (no interception or no absorption) to 1 (total interception or total absorption).

6.1.3 Instantaneous fAPAR , average daily fAPAR, and integrated APAR -- Plant canopy absorption of PAR is an instantaneous process that varies throughout the day and seasonally throughout the year. To know exactly how much light is absorbed by a plant canopy through time, it would seem necessary to monitor instantaneous fAPAR continuously. However, many current satellite sensors do not collect data continuously over the same spot, rather they only get measurements from the same areas every few (or more) days.

By establishing the systematic variation of instantaneous fAPAR throughout the day for a particular canopy (this can be measured in the field or by remote sensing efforts that are intensified for limited durations, or it can be modeled), the average daily fAPAR value can be approximated from an instantaneous measurement. Figure 4.07 shows how the variation of fAPAR relates to time of day and LAI in a spatially uniform canopy. Spectral vegetation indices (SVIs) have been related to instantaneous fAPAR for a variety of canopies.

A measurement of instantaneous fAPAR can be combined with a daily integration (sum) of PAR incident at the top of the canopy to derive an estimate of the amount of PAR absorbed (APAR) by the canopy on a daily basis. The amount of daily APAR can also be integrated over all days of the growing season (using remotely sensed observations of phenology) to obtain a value for the amount of energy available for plant growth on an annual basis. These integrated quantities are measures of the energy available for photosynthesis on a daily or annual basis that are important to models of plant primary production and studies that compare the light use efficiency of different plants.

6.2 Light Use Efficiency (LUE)

The ratio of net primary production (NPP) to absorbed photosynthetically active radiation (APAR) is called light use efficiency (LUE). If LUE values are similar for all plant types and biomes, then estimating NPP using remote sensing data would be simplified because NPP would be directly proportional to APAR. If LUE values vary widely, however, then representative values have to be determined for each distinct vegetation type or biome. LUE variability within and between ecosystems also must be assessed.

There is evidence that gross primary production (GPP) is determined largely by the long-term response of vegetation to environmental conditions that determine the amount of foliage that develops, and thus the amount of light that is captured and absorbed. Figure 4.08 shows that gross primary production (GPP, or annual assimilation) per unit APAR for aspen and spruce (two boreal tree species) is similar. Yet a plot of APAR vs. NPP for the two species (Figure 4.09) shows two separate relationships, indicating different light use efficiency (LUE) values for the different species. The broadleaf deciduous stands (aspen) produce about 1.0 gram of biomass per megajoule (MJ) of APAR (energy available for photosynthesis), whereas the needleleaf evergreen (spruce) stands produce about half that amount (0.5 g) of biomass with the same amount of energy. The difference in LUE in this case is largely due to differences in respiration (which is higher for the spruce).

Production efficiency models (PEMs) are statistical models of primary production using remotely sensed observations and light use efficiency (LUE). Recent sophisticated PEMs incorporate a suite of algorithms (formulas) to derive factors that affect LUE, including absorbed photosynthetically active radiation (APAR), surface and air temperature, vapor pressure deficit (VPD), surface soil moisture, and biomass.