A12B-01 10:20h
Estimates of Black Carbon Emissions from Open Biomass Burning
Emissions from biomass burning may have significant climate effects through their direct and indirect climate forcing. We have used the TOMs AI product together with an inverse model to estimate black carbon emissions from open biomass burning for the year 2000. We used the University of Michigan version of the LLNL IMPACT global transport model together with the GSFC Data Assimilation Office meteorological fields for the year 2000, the assumed aerosol size distributions and the satellite viewing angle to estimate the modeled AI and then we minimized the cost function between measured and modeled AI to estimate emissions from 9 different regions. A priori emissions based on both Ito and Penner [2004] and Arellano et al. [2004] were used. Emissions estimates from Africa are well constrained by the inversion method, but those from S. America, Indonesia, S. Asia, Australia, temperate boreal forests, and the rest of the world depend significantly on the a priori emissions as well as the error estimates that are used in the analysis. Our best-fit a posteriori emissions are 4.72 Tg/year, almost a factor of 3 larger than the bottom up estimates of Ito and Penner [2004]. The importance of these large estimates for climate forcing is discussed. Arellano, A. F., Jr., P. S. Kasibhatla, L. Giglio, G. R. van der Werf, and J. T. Randerson (2004), Top-down estimates of global CO sources using MOPITT measurements, Geophys. Res. Lett., 31, L01104, doi:10.1029/2003GL018609. Ito, A., and J. E. Penner (2004), Global estimates of biomass burning emissions based on satellite imagery for the year 2000, J. Geophys. Res, 109, D14S05, doi:10.1029/2003JD004423.
A12B-02 10:35h
Residence Time Maps for Inverse Modeling in Complex Terrain
We will use inverse modeling to estimate the distribution and magnitude of European emissions of halogenated green house gases. Simulated annealing will be employed to combine measurement data with modeled residence time maps of Europe. The latter are obtained from a backward Lagrangian Particle Dispersion Model (LPDM). The measurements are taken at Jungfraujoch, a remote site in the northern part of the Swiss alps at 3580m asl. From a measurement point of view, the advantage to sample at Jungfraujoch lies in the possibility to sample both background air and air polluted by European emissions. Due to its remote and elevated position Jungfraujoch is not influenced by emissions nearby the station. From a modeling point of view, Jungfraujoch is demanding, since the measurement station is situated in a very rugged terrain. The quality of the modeled residence time maps is however, of central importance as they are an essential input to the inversion technique. In this context, several aspects have to be checked. One point is the quality of the used meteo fields. Despite the 7km x 7km grid spacing of the MeteoSwiss alpine model (aLMo) we use, the topography is still only approximately captured. The residence time maps also depend on the steering parameters of the LPDM, in particular the starting height, the wind field resolution, the model time step, the run time, and the number of released particles. We examine the sensitivity of the residence time maps on these input parameters. We link the resulting residence time maps with measurements to choose an optimized set of steering parameters.
A12B-03 10:50h
Simmulations and Inverse Modeling of Global Methyl Chloride
Methyl chloride (CH$_{3}$Cl) is one of the most abundant chlorine-containing gases in the atmosphere and thus, it is a major contributor to stratospheric chlorine. In spite of its important role in atmospheric chemistry, the known sources and sinks of CH$_{3}$Cl are unbalanced. Global simulations of atmospheric CH$_{3}$Cl are conducted using the GEOS-CHEM model. In addition to the known sources (1.5 Tg yr$^{-1}$) from ocean, biomass burning, incineration/industry, salt marshes, and wetlands, a hypothetical aseasonal biogenic source of 2.9 Tg yr$^{-1}$ is added in order to match needed emissions. Observations from 7 surface sites and 8 aircraft field experiments are used to evaluate the model simulations. The model results with a priori emissions and sinks reproduce CH$_{3}$Cl observations at northern mid and high latitudes reasonably well. However, the seasonal variation of CH$_{3}$Cl at southern mid and high latitudes is severely overestimated. Simulated vertical profiles show disagreements in the vicinities of major sources, principally reflecting the uncertainties in the estimated distributions of our added pseudo-biogenic and the biomass burning sources. Inverse modeling is applied to obtain optimal source distributions of CH$_{3}$Cl on the basis of surface and aircraft observations and model results, especially for the biogenic and biomass burning sources. The inversion of the pseudo biogenic CH$_{3}$Cl source is modeled for 4 seasons and 6 geographical regions. The inversion of the biomass burning source is modeled for 4 seasons and two hemispheres. The inversions of the other sources and oceanic and soil uptake are also performed. Model simulations using a posteriori emissions are in better agreement with the observations particularly at the southern high latitudes. The a posteriori source of biomass burning CH$_{3}$Cl is lower by about 27% than the a priori estimate, although there is little change for the biogenic source. The inverse modeling results suggest a clear seasonal pattern of the biogenic source, peaking in spring and fall.
A12B-04 11:05h
Constraining Global Isoprene Emissions With GOME Formaldehyde Column Measurements
Biogenic isoprene plays an important role in tropospheric chemistry. Current isoprene emission estimates are highly uncertain due to lack of observations. Formaldehyde (HCHO) is a high-yield product of isoprene oxidation. The short photochemical lifetime of HCHO allows observations of this trace gas to help constrain isoprene emissions. We use HCHO column observations from the Global Ozone Monitoring Experiment (GOME) satellite instrument. These global data are particularly useful for studying the tropics where most of the global budget of isoprene is emitted, and where in situ observations are notably sparse. We present results from an inverse model study that uses GOME data from September 1996 to August 1997 to fit modeled sources of HCHO from the GEOS-CHEM chemistry transport model. Isoprene emissions are divided into 10 vegetation types. Column contributions to HCHO from these 10 biogenic sources, in addition to biomass burning and industrial sources are considered. We fit these 12 sources of HCHO to the observed column data for 8 geographical regions (North America, Europe, East Asia, India, South Asia, South America, Africa, and Australia). The a priori simulation highly underestimates global HCHO columns over the 8 geographical regions (bias: -14 - -46%; R:0.52 - 0.84). The a posteriori solution shows generally higher isoprene and biomass burning emissions. The a posteriori simulation with a posteriori isoprene emissions improved the model bias for all regions (bias: -5.9 - -26%; R = 0.57 - 0.85). The a posteriori estimate of the annual global isoprene emissions of 565 Tg C yr$^{-1}$ is about 50 % larger than the a priori estimate. This increase of global isoprene emissions significantly affects tropospheric chemistry, decreasing the annual global mean OH concentration by 10.8% to 0.95 x 10$^{6}$ (molecules cm$^{-3}$ ) and increasing the annual global tropospheric O$_{3}$ burden by 1.5% to 333 Tg. The atmospheric lifetime of CH$_{3}$CCl$_{3}$ increases from 5.2 to 5.7 years.
A12B-05 11:20h
Estimating global CO$_{2}$ surface fluxes using CO$_{2}$ and $\delta^{13}$C data from the NOAA/CMDL network.
In the last several years there has been considerable attention devoted to using global time-space patterns of CO$_{2}$ in the atmosphere to infer surface fluxes. Here, we present flux results derived from both CO$_{2}$ and $\delta^{13}$C, using data from the NOAA/CMDL sampling network in a three-dimensional inversion framework. $\delta^{13}$C data has been used previously but either in a two-dimensional transport framework, or using data from a small number of sampling sites. While it is clear that atmospheric $\delta^{13}$C contains unique information about surface flux patterns, it remains unclear exactly where and when $\delta^{13}$C is useful, given uncertainties in the $^{13}$C budget. We will use a Bayesian inversion setup in which CO$_{2}$ and $\delta^{13}$C data will be used to optimize not only surface fluxes but also patterns of isotopic fractionation and disequilibrium. Within this framework, we will be able to formally assess how much information $\delta^{13}$C data add compared to CO$_{2}$ alone, given uncertainties in CO$_{2}$ data, $\delta^{13}$C data, fractionation, disequilibrium, and first guesses of fluxes. This analysis will allow us to answer the question of what advances need to occur so that the atmospheric $\delta^{13}$C signal can be used most effectively. Finally, using our best uncertainty estimates, we will compare surface fluxes derived from CO$_{2}$ and $\delta^{13}$C data.
A12B-06 11:35h
On the Robustness of Air-Sea Flux Estimates of Carbon Dioxide from Ocean Inversions
Inverse methods analogous to those used for atmospheric inversions have been adapted to estimate regional air-sea fluxes of carbon dioxide using ocean interior observations of dissolved inorganic carbon and related tracers and an Ocean General Circulation Model (OGCM). We estimate seperately the preindustrial component and the component due to the anthropogenic perturbation of atmospheric carbon dioxide. Previous sensitivity studies have shown that model circulation is one of the most important sources of error in the ocean inversion. We present estimates of preindustrial and anthropogenic air-sea carbon dioxide exchange using a suite of nine different OGCM's in order to quantify the robustness of our results and explore the role of different representations of ocean circulation in the inversion. Most of the large scale features of the inverse estimates are robust across all models. The preindustrial inverse estimates generally follow the expected pattern of uptake at high latitudes and out gassing in the tropics; however, all of the models call for out gassing in the Southern Ocean between 44S and 58 S. The greatest anthropogenic carbon uptake occurs at mid- to high- latitudes, with a large anthropogenic carbon sink in the Southern Ocean, while the bulk of the anthropogenic carbon storage occurs at mid-latitudes. Preliminary results also suggest interesting, robust differences between these inverse estimates and estimates from forward model simulations. Both the preindustrial and anthropogenic carbon dioxide flux estimates are most robust at mid and high northern latitudes, except for the high latitude North Atlantic. The carbon dioxide flux estimates are most uncertain in the Southern Ocean, where the inverse estimates are strongly dependent on the rates of deep water ventilation in the OGCM. The preindustrial inverse estimates for the Indian Ocean are also sensitive to the choice of OGCM, and the anthropogenic estimates have significant uncertainties in the tropical Pacific. Over large spatial scales, inverse estimates based on different OGCM's are in better agreement than estimates based on forward simulations of the same models, but this is not necessarily true for smaller model regions.
http://quercus.igpp.ucla.edu/OceanInversion
A12B-07 11:50h
Inversion of CO2 sources and sinks using satellite data: application to TOVS
The current observation surface network for CO2 poorly informs about space and time variations of CO2 fluxes over land at a global scale. Additional information is sought to compensate this lack of knowledge: the surface observations gets denser and efforts are being made to measure atmospheric CO2 from space. The HIRS instrument has been operated since 1979 on-board the polar obiting NOAA satellites. Eventhough it was designed for meteorological purposes, CO2 variations in the upper troposphere noticeably impact the radiances that it measures. CO2 retrievals in the tropics (20N-20S) have been produced only recently and first validations have shown that such data agree with aircraft observations reasonably well. Properly handling them to constrain the inversion of CO2 sources and sinks is an important challenge given the length of the instrument record. In this paper we describe how CO2 retrievals from HIRS can contribute to the monitoring of the CO2 surface fluxes. A variational inverse scheme has been designed, that uses the LMDZT transport model. Error characteristics of the CO2 retrievals are estimated using both model computations and in situ observations. Results are shown and prospects for other CO2-sensitive satellite instruments, like AIRS, SCIAMACHY and the CO2-dedicated OCO instrument are discussed.
A12B-08 12:05h
Constraining Global and Regional Budgets of CO From MOPITT Retrievals
The MOPITT instrument onboard the NASA TERRA satellite provides the first multi-year CO retrievals that have been recently used for development and evaluation of tracer assimilation schemes and inverse modeling of surface emissions. In this paper we review our recent results for the CO data assimilation and inversion of CO surface fluxes using the CMDL surface data and mid-troposphere MOPITT CO retrievals in the MOZART CTM. We will discuss applications of various assimilation/inversion schemes, including the specification of the CTM errors, evaluation of potential retrieval and forecast biases, and assignement of prior emission errors. The estimated 2000-2004 year-to-year variations of the monthly CO emissions on the global and regional scales will be highlighted.