An analysis of remote sensing of the fraction of absorbed photosynthetically active radiation in forest canopies

Photosynthetically active radiation (PAR) absorbed by vegetation canopies is a key factor determining plant productivity and transpiration. The capacity of vegetation canopies to absorb PAR has been estimated using combinations of remotely sensed data from the visible and near infrared wavelengths using Spectral Vegetation indices (SVI). The ability to use satellite remote sensing to make measurements of the fraction of absorbed PAR over the entire Earth can provide important data required to understand global carbon and water budgets.; This study developed GeoSail, a canopy reflectance model for forests, by combining the SAIL (Scattering from Arbitrarily Inclined Leaves) radiative transfer model with a geometric model. Needle and leaf reflectance and transmittance, twig reflectance, and background reflectance data were acquired for ten forest types. The leaf to twig areas of branches and the height to width ratios for the trees were also determined. These data were used as inputs into the model. Model results favorably compared with available forest stand level data from the Oregon Transect Terrestrial Ecosystem Research (OTTER) and the Superior National Forest (SNF) studies.; The GeoSail model results were then used to study the relationships between SVI and the fraction of absorbed PAR {dollar}rm (fsb{lcub}APAR{rcub}){dollar} for the different forest types and a prairie. Four two-band SVIs were examined: Normalized Difference Vegetation Index (NDVI), Simple Ratio (SR), Perpendicular Vegetation Index (PVI), and Soil Adjusted Vegetation Index (SAVI). The model indicated that shadows in forests were an important factor in determining scene reflectance. The presence of twigs in forest canopies was shown to be a secondary factor. All indices produced curved relationships with {dollar}rm fsb{lcub}APAR{rcub}{dollar} for forests. The SR, PVI, and SAVI were insensitive to variations in {dollar}rm fsb{lcub}APAR{rcub}{dollar} when {dollar}rm fsb{lcub}APAR{rcub}{dollar} was less than 60 percent. Characteristics of these three indices which make them insensitive to variations in background soil brightness also make them insensitive to variations in {dollar}rm fsb{lcub}APAR{rcub}{dollar} in open forest canopies. For all of the indices using the SVI-{dollar}rm fsb{lcub}APAR{rcub}{dollar} relationship developed for prairie grasses on forest canopies resulted in an underestimation of {dollar}rm fsb{lcub}APAR{rcub}{dollar} in all cases.