HR: 1340h
AN: A23C-1484    [Abstracts]
TI: Quantifying the impact of aggregation errors and model transport biases on top-down estimates of carbon monoxide emissions using satellites observations
AU: * Jiang, Z
EM: zjiang@atmosp.physics.utoronto.ca
AF: Department of Physics University of Toronto, 60 St. George Street, Toronto, ON M5S 1A7, Canada
AU: Jones, D B
EM: dbj@atmosp.physics.utoronto.ca
AF: Department of Physics University of Toronto, 60 St. George Street, Toronto, ON M5S 1A7, Canada
AU: Kopacz, M
EM: kopacz@fas.harvard.edu
AF: Division of Engineering and Applied Science Harvard University, 110J Pierce Hall, 29 Oxford Street, Cambridge, MA 02138, United States
AU: Liu, J
EM: jliu@atmosp.physics.utoronto.ca
AF: Department of Physics University of Toronto, 60 St. George Street, Toronto, ON M5S 1A7, Canada
AU: Henze, D K
EM: daven@its.caltech.edu
AF: Department of Chemical Engineering California Institute of Technology, 1200 E.California Blvd, Pasadena, CA 21041, United States
AB: Inverse modeling has become a widely used method for obtaining top-down estimates of surface emissions of atmospheric CO. These top-down estimates, however, are adversely influenced by systematic errors in the inverse model, such as biases in the transport fields and aggregation errors associated with choice of regional scales on which the emissions are aggregated for optimization (discretization of the state vector). We have conducted an inverse analysis of atmospheric CO, using the GEOS-Chem model and observations from the MOPITT satellite instrument, to quantify the potential contribution of model transport error and aggregation errors on top-down source estimates. We focus on quantifying CO emissions for September and October 2000, during the biomass burning season in the southern hemisphere. We employ a sub-optimal Kalman filter to assimilate MOPITT data to adjust the initial distribution of CO at the beginning of the inversion period, and then apply a 4- dimensional variational data assimilation scheme to optimize the CO emissions on the 2x2.5 grid of the model. The high-resolution, a posteriori source estimates are compared with estimates obtained from a coarse resolution, analytical Bayesian inversion to quantify the impact of aggregation errors in the coarse resolution inversion on the source estimates. We also carry out the coarse resolution analytical inversion using two different versions of the GEOS-Chem model, driven with different transport fields, to isolate the impact on the source estimates of systematic differences in transport (associated mainly with the different convection schemes) in the models.
DE: 0365 Troposphere: composition and chemistry
DE: 0368 Troposphere: constituent transport and chemistry
DE: 3315 Data assimilation
SC: Atmospheric Sciences [A]
MN: 2007 Fall Meeting