Skip to main content
ARS Home » Northeast Area » University Park, Pennsylvania » Pasture Systems & Watershed Management Research » Research » Publications at this Location » Publication #220951
Title: Using Remote Sensing and Modeling to Estimate Gross Primary Productivity in Humid-Temperate Pastures

Author
item Skinner, Robert
item GILMANOV, TAGIR - SOUTH DAKOTA STATE UNIV
item WYLIE, BRUCE - USGS

Submitted to: Agronomy Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 4/7/2008
Publication Date: 7/13/2008
Citation: Skinner, R.H., Gilmanov, T.G., Wylie, B.K. 2008. Using remote sensing and modeling to estimate gross primary productivity in humid-temperate pastures [abstract]. In Plants & Soils: Montreal 2008 Program and Abstracts. p. 71-72.

Interpretive Summary: An interpretive summary is not required.

Technical Abstract: A complete understanding of the terrestrial carbon budget requires local and regional information on the carbon dynamics of individual biomes. Although northeastern USA forests have been extensively studied, less information is available for grasslands within the region. Techniques are also needed to scale fluxes from an individual pasture to whole-farm, watershed, or regional scales. The purpose of this research was to combine information from micrometeorology, modeling, and satellite-based Normalized Difference Vegetation Index (NDVI) data to provide a basis for estimating gross primary productivity (GPP) across grazed pastures in the northeastern USA. Eddy covariance systems have continuously collected carbon dioxide flux data on two pastures in central Pennsylvania since January 2003. Light curve analysis and relationships between nighttime temperature and flux were used to partition 2003 data into GPP and ecosystem respiration (Re). Bi-weekly Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI estimates were obtained for each pasture in 2003. In addition, a pasture simulation model was run for each pasture to obtain estimates of GPP, forage yield, and soil moisture stress. Results were combined across pastures and model- and NDVI-estimated GPP were compared with bi-weekly averaged GPP from eddy covariance measurements. Both the model- and NDVI-based estimates were able to reasonably track seasonal changes in measured GPP with some exceptions. The model tended to overestimate GPP during May and September but underestimated GPP in July and August. NDVI estimates tended to be higher than observed GPP during spring green-up and again in December but lower in May. Estimates of annual GPP were 4918 g CO2 m-2 for eddy covariance, 5050 g CO2 m-2 for the model and 5168 g CO2 m-2 for NDVI. Modeling and remote sensing successfully predicted annual GPP of humid-temperate pastures to within 3 to 5% of observed values and were able to accurately capture seasonal flux dynamics. Used in combination, remote sensing and modeling have the potential to provide a powerful tool for extrapolating eddy covariance measurements from individual pastures to whole-farm and larger scales.