 |
| Titel |
Comparing Methods Dedicated to the Retrieval of Atmospheric CO2 from Space Borne Observations of Backscattered Near-Infrared Sunlight |
| VerfasserIn |
A. Butz, H. Boesch, O. P. Hasekamp, A. Cogan |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250029975
|
|
|
|
|
|
| Zusammenfassung |
| The Orbiting Carbon Observatory (OCO) and the Greenhouse Gases Observing SATellite
(GOSAT) target at inferring atmospheric CO2 abundances with high accuracy and precision
and global coverage. The observational strategy relies on measuring sunlight backscattered
by the Earth’s surface and atmosphere of two near-infrared CO2 absorption bands and the O2
A band. Several methods are proposed to retrieve the column average dry air mixing ratio
XCO2 from such measurements.
Commonly, these retrieval methods differ by the implementation of the forward model
that simulates the satellite measurements and by the inverse method that infers the target
quantity XCO2 given simulated and measured radiance spectra. The forward model itself
consists of a radiative transfer model that requires input from several sub-modules providing
information on spectroscopic line parameters, surface properties, atmospheric state variables
and aerosol properties of the sampled air masses. Differences in the implementations of the
forward model and the inverse method can result in significant biases in XCO2 retrieved from
the different algorithms.
Here, we aim at comparing different retrieval approaches and identifying and characterizing
(and potentially removing) differences between them. Our bottom-up-approach starts with
comparing radiance spectra generated by the forward models for simple scenarios such as
purely Rayleigh scattering atmospheres so that the radiative transfer model and the treatment
of molecular absorption and atmospheric state variables can be verified. Then more complex
scenes such as aerosol loaded atmospheres where polarization of radiation is important are
considered. To this end we simulate an ensemble of radiance spectra by either of the
considered retrieval methods and cross-retrieve XCO2 among all approaches. These retrieval
comparisons will then allow us to draw conclusions on differences in XCO2 introduced by
the inherent assumptions of the forward models and inverse methods, which has to be
taken into account when interpreting the OCO and GOSAT XCO2 data products. |
|
|
|
|
|
|