HR: 08:30h
AN: B51E-03    [Abstracts]
TI: Constraining Regional Fossil Fuel CO2 Emissions and Choosing an Optimal Measurement Network Using 14C Observations, Modeling, and Inversions.
AU: * Bergmann, D
EM: dbergmann@llnl.gov
AF: Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94551-0808, United States
AU: Cameron-Smith, P
EM: pjc@llnl.gov
AF: Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94551-0808, United States
AU: Guilderson, T
EM: tguilderson@llnl.gov
AF: Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94551-0808, United States
AU: Grant, K
EM: keg@llnl.gov
AF: Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94551-0808, United States
AU: Graven, H
EM: hgraven@ucsd.edu
AF: Scripps Institute of Oceanography, 9500 Gilman Dr. UCSD MC 0244, La Jolla, CA 92093- 0244, United States
AB: We are developing a capability to separately retrieve regional fossil fuel and biospheric CO₂ sources, using a combination of high resolution global modeling of CO₂ tracers, highly accurate ¹⁴C measurements, and TransCom style direct inversions. Our primary data-set is a suite of 7 Scripps clean air sites that construct a trans Pacific latitudinal profile, and existing published data. In these initial experiments we have retrieved the fossil fuel emissions from 3-5 regions, and are selecting additional measurement locations to optimize our retrievals for the 22 TransCom regions. In the future we will build on this work to separately retrieve regional fossil-fuel and biospheric CO₂ emissions for each region using C-12,13,14 measurements from all potential sites. We use the LLNL global atmospheric chemical transport model, IMPACT, at 2x2.5 degree resolution and GEOS4 meteorological data to perform a TransCom 3 style forward simulation using 22 separate tracers (11 land regions and 11 ocean regions) to form 22 basis functions. Using this forward simulation and a limited number of both conventional ¹⁴CO₂ measurements and more recent AMS ¹⁴C measurements, we estimate the fraction of the CO₂ which is from fossil fuel emissions. Model results and measurements are then used in an inversion algorithm to retrieve regional fossil fuel emissions from each geographical region. We are using a linear inversion algorithm based on single value decomposition. This formulation of linear inversion allows us to investigate retrievable degrees of freedom, the information retrievable from measurements for particular regions, and the marginal contributions from particular measurement sites. With initial data only available at 7 sites, we have combined the 22 regions into 3 or 5 latitude bands. Then using the model results and the inversion we can select the site for an additional measurement which will maximally reduce the retrieval error. In this way we can design an optimal observation network by adding one site at a time.
DE: 0322 Constituent sources and sinks
DE: 0368 Troposphere: constituent transport and chemistry
DE: 0414 Biogeochemical cycles, processes, and modeling (0412, 0793, 1615, 4805, 4912)
DE: 0428 Carbon cycling (4806)
DE: 0454 Isotopic composition and chemistry (1041, 4870)
SC: Biogeosciences [B]
MN: 2007 Fall Meeting