Combining Ecological and Atmospheric Observations to Understand Carbon Budgets Using Models with MODIS, LANDSAT, and GOSAT data.

Scott Denning, Colorado State University, denning@atmos.colostate.edu (Presenter)
Nick Parazoo, Colorado State University, nparazoo@atmos.colostate.edu
Katherine Corbin, Colorado State University, kdcorbin@atmos.colostate.edu
Baker Ian, Colorado State University, baker@atmos.colostate.edu
Rebecca McKeown, Colorado State University, beckym@nrel.colostate.edu
George James Collatz, NASA GSFC, jim.collatz@nasa.gov
Stephan Randolph Kawa, NASA GSFC, stephan.r.kawa@nasa.gov

We have developed and tested a method for estimating time-averaged exchange of carbon between terrestrial ecosystems and the atmosphere by combining observations form derived from MODIS, LANDSAT, and GOSAT sensors using models of terrestrial ecophysiology (a SiB variant) and atmospheric transport (PCTM). The analysis recognizes the heterogeneous distribution of vegetation at subgrid-scale and the need for intelligent interpolation between sparse observations. We have assimilated millions of estimates of leaf area from MODIS into a predictive phenology model, and use GEOS-5 weather analyses to compute photosynthesis and ecosystem respiration on an hourly basis. Carbon pools are initialized in equilibrium, and then allowed to evolve according to estimates of disturbance and recovery being derived from LANDSAT. Soil and biomass pools are then adjusted using ensemble data assimilation using GOSAT observations of column CO2 compared to predictions using the PCTM.

We have tested the ensemble assimilation system using synthetic data, and find that it has excellent skill at continental scales, even resolving some regional patterns that reflect CO2 and nitrogen fertilization, forest regrowth, and boreal warming.

Presentation: 2011_Oct03_PM_Denning_342.pptx (6641k)

Presentation Type:  Plenary Talk

Session:  Poster Speed Talks:

Presentation Time:  Mon 3:30 PM  (5 minutes)