Resources
N/A
Related Research Areas
Carbon Cycle & Ecosystems

Data collected during the "Balanco Atmosferico Regional de Carbono na Amazonia" (BARCA) experiment, carried out in November 2008 and May 2009, will be used in a multi-tiered modeling effort to generate observationally constrained monthly and annual budgets of CO2, CH4 and CO for the Amazon Basin. The BARCA data set is unique both in its temporal coverage, capturing diurnal variability during both seasonal transitions (dry to wet, wet to dry), and its spatial coverage, with flights spanning over 15 degrees latitude and ~25 degrees longitude and 160 vertical profiles. We have 1 Hz data for the target gases and O3, plus flask samples, aerosol data, etc. Local sources and sinks of the carbon species are evident, attributable to a variety of processes such as wetland emissions, photosynthesis, respiration, and biomass burning.

The final product will be a time-resolved carbon budget for Amazonia, developed in a multi-stage model-data synthesis. The first stage involves creating an ensemble of a priori models to generate predicted CO2 values at measurement points. The second stage is optimization of the underlying emissions models via inverse analysis, and stage 3 is production of an ensemble of a posteriori models that can be driven by remote sensing and meteorological inputs. The a posteriori modes will then be run for the full year, to generate time resolved budgets for sources and sinks of CO2. The final stage is a rigorous uncertainty analysis of the ensemble results. The a priori CO2 models will couple both Lagrangian Particle Dispersion Models (LPDMs) and Eulerian models with surface flux models. The Stochastic Time Inverted Lagrangian Transport (STILT) and FLEXPART are LPDMs that link CO2 fluxes at the surface to concentrations in the atmosphere. Both will be driven by a variety of meteorological products (BRAMS - Brazilian developments on the Regional Atmospheric Modeling System, GFS, WRF, MM5, ECMWF, etc.). Surface flux models are ED-2 (Ecosystem Demography) and VPRM (Vegetation Photosynthesis and Respiration Model). VPRM is a remote-sensing driven surface flux model which ingests and assimilates comprehensive environmental, satellite and tower flux data in a very simple mathematical structure, providing accurate simulation of CO2 fluxes at the landscape scale. ED-2 predicts surface CO2 fluxes at the tree scale, and then rigorously aggregates via probability distribution functions to ecosystem scale; it is fully coupled to BRAMS, resulting in a prognostic biosphere and atmosphere model that can provide surface input to the a priori CO2 model. The coupled ecosystem-transport model produces a priori CO2 values for direct comparison with observations.

Optimization of the a priori CO2 model will use two inverse approaches, the classical Bayesian top-down approach, and a novel geostatistical approach, to parameterize VPRM and ED-BRAMS and to create an a posteriori model that matches observed CO2 data from BARCA. Using the optimized a posteriori model and environmental drivers from remote sensing and assimilated meteorology, a year long model run will generate a carbon budget for the Amazon Basin in 2008-9. The results will be tested using continuous surface site data and periodic aircraft flask sampling, helping to constrain the error bounds on the model output and the associated regional carbon budget.

Project PI: Steven Wofsy/Harvard University

Room 110D Pierce Hall 29 Oxford Street Cambridge, MA 02138

Phone: (617) 495-4566

Fax: (617) 495-4551

Email: swofsy@seas.harvard.edu

http://harvardforest.fas.harvard.edu/profiles/wofsy.html

Tags

Everyone's Tags

Popular Resources

Nothing to see here at the moment. Check back later.

Discussions

Log in to start a discussion.

Project Highlight

1 members

Started: Sep 03, 2010

Last Activity: Feb 09, 2011

Admin:

What can I do on this project?
  • Only approved users can join
  • Anybody can view this project
  • Any registered users can leave comments
  • Anybody can view comments

New Member

Need help?

Visit our help center