Close Window

Upscaling Carbon Fluxes and Uncertainty Across the Northern Forest Ecoregion Using a Hierarchical Bayesian Approach

Kusum J Naithani, Penn State, kjn18@psu.edu (Presenter)
Kenneth James Davis, Penn State, kjd10@psu.edu
Klaus Keller, Penn State, klaus@psu.edu
Jingfeng Xiao, University of New Hampshire, j.xiao@unh.edu
Nathan Urban, Princeton University, nurban@princeton.edu
Paul Bolstad, University of Minnesota, pbolstad@umn.edu
Dong Hua, University of Minnesota, dhua@umn.edu

Quantifying uncertainty in carbon flux upscaling is critical to improving our understanding of the terrestrial carbon cycle. Here we use a hierarchical Bayesian framework to develop a simple diagnostic model and assimilate eddy covariance flux observations. The hierarchical Bayesian approach incorporates parameter uncertainty associated with tower location, stand age and plant functional type in addition to the inherent natural variability in flux measurements. We use this framework to (1) develop and test a hierarchical Bayesian model using a combination of in situ flux data and remotely sensed data, (2) evaluate multiple sources of uncertainty and scale-up model parameters and uncertainty, (3) evaluate the trade-off between spatial (number of towers across a region) and temporal (length of data record) extent of eddy covariance towers and reduction in parameter and prediction uncertainty in CO2 upscaling from towers to a region, and (4) test the model outside of the region where it was developed– northern Wisconsin and Upper Peninsula of Michigan. The study area is located in the U.S. Northern forest ecoregion, and utilizes the observations collected by the Chequamegon Ecosystem-Atmosphere Study (ChEAS). We show how parameter uncertainty, selection and grouping of eddy covariance flux sites, the choice of land cover products, and spatial and temporal extent of flux tower data have substantial impacts on regional carbon budgets.

Presentation Type:  Poster

Session:  Other   (Tue 11:30 AM)

Associated Project(s): 

  • Davis, Ken: Probabilistic Carbon Flux Upscaling Across a Northern Forest Ecoregion ...details

Poster Location ID: 252

 


Close Window