HR: 17:15h
AN: B22E-06    [PDF]
TI: A Model-based Approach to Scaling GPP and NPP in Support of MODIS Land Product Validation
AU: * Turner, D P
EM: david.turner@oregonstate.edu
AF: Forest Science Department, Oregon State University, Corvallis, OR 97331 United States
AU: Cohen, W B
EM: warren.cohen@oregonstate.edu
AF: USDA PNW Research Station, 3200 SW Jefferson St, Corvallis, OR 97331 United States
AU: Gower, S T
EM: stgower@facstaff.wisc.edu
AF: Department of Forest Ecology and Management, University of Wisconsin, Madison, WI 53706 United States
AU: Ritts, W D
EM: david.ritts@oregonstate.edu
AF: Forest Science Department, Oregon State University, Corvallis, OR 97331 United States
AB: Global products from the Earth-orbiting MODIS sensor include land cover, leaf area index (LAI), FPAR, 8-day gross primary production (GPP), and annual net primary production (NPP) at the 1 km spatial resolution. The BigFoot Project was designed specifically to validate MODIS land products, and has initiated ground measurements at 9 sites representing a wide array of vegetation types. An ecosystem process model (Biome-BGC) is used to generate estimates of GPP and NPP for each 5 km x 5 km BigFoot site. Model inputs include land cover and LAI (from Landsat ETM+), daily meteorological data (from a centrally located eddy covariance flux tower), and soil characteristics. Model derived outputs are validated against field-measured NPP and flux tower-derived GPP. The resulting GPP and NPP estimates are then aggregated to the 1 km resolution for direct spatial comparison with corresponding MODIS products. At the high latitude sites (tundra and boreal forest), the MODIS GPP phenology closely tracks the BigFoot GPP, but there is a high bias in the MODIS GPP. In the temperate zone sites, problems with the timing and magnitude of the MODIS FPAR introduce differences in MODIS GPP compared to the validation data at some sites. However, the MODIS LAI/FPAR data are currently being reprocessed (=Collection 4) and new comparisons will be made for 2002. The BigFoot scaling approach permits precise overlap in spatial and temporal resolution between the MODIS products and BigFoot products, and thus permits the evaluation of specific components of the MODIS NPP algorithm. These components include meteorological inputs from the NASA Data Assimilation Office, LAI and FPAR from other MODIS algorithms, and biome-specific parameters for base respiration rate and light use efficiency.
UR: http://www.fsl.orst.edu/larse/bigfoot/index.html
DE: 0400 Biogeosciences
DE: 1615 Biogeochemical processes (4805)
DE: 1640 Remote sensing
SC: Biogeosciences [B]
MN: 2003 Fall Meeting