HR: 14:40h
AN: A13G-04    [Abstracts]
TI: A Grid-Based Inverse Modeling Approach For Constraining the CO Budget
AU: Stavrakou, T
EM: jenny@aeronomie.be
AF: Belgian Institute for Space Aeronomy, Avenue Circulaire 3, Brussels, 1180 Belgium
AU: * Muller, J
EM: jfm@aeronomie.be
AF: Belgian Institute for Space Aeronomy, Avenue Circulaire 3, Brussels, 1180 Belgium
AB: The adjoint modeling technique has been recently used to invert for the emissions of ozone precursors (CO, biogenic VOCs, and NOx) in the IMAGES global chemical transport model [1]. The advantages of this framework compared to other inversion methods are that (1) non-linearities in the response of the calculated concentrations to changes in the emissions are taken into account (an important feature given the strong chemical feedbacks existing between ozone precursors), and (2) it can cope with an arbitrarily large number of control parameters. This latter feature is explored in the present study. The observations used in this work are monthly averaged total CO columns from MOPITT (May 2000 - April 2001) binned on the model grid. Two distinct approaches are compared for inverting CO emissions: in the first inversion setup, the annual fluxes of CO and NMVOC from anthropogenic, biogenic and biomass burning sources over large regions are optimized; in the second one, the parameters to be determined are the monthly emissions from every model grid cell and from each of three broad categories (anthropogenic, biogenic, biomass burning). The total number of unknowns is equal to 18 in the first, "big-region" approach, while it is on the order of 30,000 in the second, "grid-based" scheme. The inversion is overdetermined in the first case and underdetermined in the second one: about 6,000 observational elements are used, but most of them are redundant due to the large chemical lifetime of CO and strong horizontal mixing in the troposphere. To reduce the underdetermination, additional constraints are introduced, in the form of correlations between the a priori errors on the control parameters. Spatial correlations are estimated based on geographical distance, country boundaries, and ecosystems. Temporal correlations are also considered. The results from both the big-region and grid-based approaches are compared and evaluated against independent observations from aircraft campaigns and ground-based measurements. The resulting emissions are discussed and compared with previous inverse modeling studies. [1] J.-F. Muller and T. Stavrakou, Atmos. Chem. Phys., 5, 1157-1186, 2005.
DE: 0300 ATMOSPHERIC COMPOSITION AND STRUCTURE
DE: 0322 Constituent sources and sinks
DE: 0365 Troposphere: composition and chemistry
DE: 0368 Troposphere: constituent transport and chemistry
DE: 1615 Biogeochemical cycles, processes, and modeling (0412, 0414, 0793, 4805, 4912)
SC: Atmospheric Sciences [A]
MN: Fall Meeting 2005