Summary:
Published in Remote Sensing of Environment 113: 2463-2475
Estimation of photosynthetic light use efficiency (ε) from satellite observations is an important component of climate change research. The photochemical reflectance index, a narrow waveband index based on the reflectance at 531 and 570 nm, allows sampling of the photosynthetic activity of leaves; upscaling of these measurements to landscape and global scales, however, remains challenging. Only a few studies have used spaceborne observations of PRI so far, and research has largely focused on the MODIS sensor. Its daily global coverage and the capacity to detect a narrow reflectance band at 531 nm make it the best available choice for sensing ε from space. Previous results however, have identified a number of key issues with MODIS-based observations of PRI. First, the differences between the footprint of eddy covariance (EC) measurements and the MODIS footprint, which is determined by the sensor's observation geometry make a direct comparison between both data sources challenging and second, the PRI reflectance bands are affected by atmospheric scattering effects confounding the existing physiological signal. In this study we introduce a new approach for upscaling EC based ε measurements to MODIS. First, EC-measured ε values were “translated” into a tower-level optical PRI signal using AMSPEC, an automated multi-angular, tower-based spectroradiometer instrument. AMSPEC enabled us to adjust tower-measured PRI values to the individual viewing geometry of each MODIS overpass. Second, MODIS data were atmospherically corrected using a Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm, which uses a time series approach and an image-based rather than pixel-based processing for simultaneous retrievals of atmospheric aerosol and surface bidirectional reflectance (BRDF). Using this approach, we found a strong relationship between tower-based and spaceborne reflectance measurements (r2=0.74, p<0.01) throughout the vegetation period of 2006. Swath (non-gridded) observations yielded stronger correlations than gridded data (r2=0.58, p<0.01) both of which included forward and backscatter observations. Spaceborne PRI values were strongly related to canopy shadow fractions and varied with different levels of ε. We conclude that MAIAC-corrected MODIS observations were able to track the site-level physiological changes from space throughout the observation period.
Sector(s):
Forests
Catégorie(s):
Scientific Article
Theme(s):
Ecosystems and Environment, Forest Ecology, Forestry Research, Forests
Departmental author(s):
Author(s):
HILKER, Thomas, Alexei LYAPUSTIN, Forrest G. HALL, Yujie WANG, Nicholas C. COOPS, Guillaume G. DROLET and T. Andrew BLACK
Year of publication:
2009
Format:
Paper
How to get the publication:
Keyword(s):
ecosystems and environment, photosynthesis, carbon cycling, remote sensing, atmospheric correction, multi-angular, forestry research scientific article