Understanding the dynamics of ecosystem carbon cycling requires an accurate determination of the spatial and temporal distribution of photosynthetic CO₂ uptake by vegetation. Optical sampling using spectral reflectance can provide the observations required for this type of analysis without physical contact with the vegetation, to supply inputs to a light use efficiency (LUE) model. LUE models have been used to estimate productivity for a number of ecosystems, and take this form:

    G = ε fAPAR Qin = ε  APAR,

where: G is gross ecosystem production (GEP); the uptake of carbon through photosynthesis; Qin is the incoming photosynthetically active radiation (PAR); fAPAR is the fraction of PAR absorbed by green vegetation; and ε is the LUE. 

This study examined the use of spectral reflectance to determine inputs to the LUE model, utilizing narrow band spectral indices to calculate fAPAR and ε, and examined the remotely determined ecosystem GEP over diurnal periods and throughout the growing season.  fAPAR was calculated using the Normalized Difference Vegetation Index (NDVI, using red and near-infrared wavelengths) and ε was estimated using the Photosynthetic Reflectance Index (PRI).  PRI detects changes in Xanthophyll cycle pigments using reflectance at 531 nm compared to a reference band at 570 nm.

Data were collected at the Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) fields (39.03°N, 76.85°W) at USDA Beltsville Agricultural Research Center. Agricultural Research Service researchers grew corn (Zea mays L., 'Pioneer 33A14') and measured CO₂ fluxes using eddy covariance methods throughout the 2007, 2008, and 2009 growing seasons. Over daylight periods on multiple days, which spanned the growing seasons, hyperspectral reflectance measurements were made using an ASD FieldSpec spectroradiometer along a fixed transect in the cornfield. A sensor mounted on the flux tower measured incoming PAR. The net CO₂ fluxes were partitioned into gross ecosystem production (GEP) and ecosystem respiration.
Results/Conclusions

The LUE model driven by optical measurements provided a good description of GEP, both for diurnal trends within days and for trends over weeks and months, and when sampled over multiple growing seasons.  GEP calculated from spectral information was strongly related to GEP from tower measurements, with R²=0.76 for 130 observations and a standard error of 0.29 mg m^-2 s^-1.  This approach and our results highlight the importance of narrow spectral band data for detecting physiological changes in vegetation.