|
|||||
|
|
Assessing FPAR source and parameter optimization scheme in application of a diagnostic carbon flux model |
|
| Authors: | David P. Turner William D. Ritts Christoph Thomas T. Andrew Black |
| Affiliation: | a Forest Ecosystems and Society, Oregon State University, Corvallis OR, USA b Atmospheric Sciences, University of California, Davis CA, USA c Atmospheric Sciences, Oregon State University, Corvallis OR, USA d Ecology and Evolutionary Biology, University of Colorado, Boulder CO, USA e Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, CO, USA f Biometeorology and Soil Physics Group, University of British Columbia, Vancouver BC, Canada g Department of Land, Air, and Water Resources, University of California Davis, Davis CA, USA |
| Abstract: | The combination of satellite remote sensing and carbon cycle models provides an opportunity for regional to global scale monitoring of terrestrial gross primary production, ecosystem respiration, and net ecosystem production. FPAR (the fraction of photosynthetically active radiation absorbed by the plant canopy) is a critical input to diagnostic models, however little is known about the relative effectiveness of FPAR products from different satellite sensors nor about the sensitivity of flux estimates to different parameterization approaches. In this study, we used multiyear observations of carbon flux at four eddy covariance flux tower sites within the conifer biome to evaluate these factors. FPAR products from the MODIS and SeaWiFS sensors, and the effects of single site vs. cross-site parameter optimization were tested with the CFLUX model. The SeaWiFs FPAR product showed greater dynamic range across sites and resulted in slightly reduced flux estimation errors relative to the MODIS product when using cross-site optimization. With site-specific parameter optimization, the flux model was effective in capturing seasonal and interannual variation in the carbon fluxes at these sites. The cross-site prediction errors were lower when using parameters from a cross-site optimization compared to parameter sets from optimization at single sites. These results support the practice of multisite optimization within a biome or ecoregion for parameterization of diagnostic carbon flux models. |
| Keywords: | Carbon flux Diagnostic model FPAR Parameter optimization CFLUX Gross primary production Ecosystem respiration Net ecosystem exchange |
| 本文献已被 ScienceDirect 等数据库收录！ | |