Radiative Transfer Modeling for the Retrieval of CO₂ from Space

Author: Natraj, Vijay

Year: 2008

Degree: Dissertation (Ph.D.)

Advisor: Yung, Yuk L.

Committee Members: Yung, Yuk L.; Seinfeld, John H.; Wennberg, Paul O.; Flagan, Richard C.

Option: Chemical Engineering

DOI: 10.7907/XSPZ-5F22

Abstract

The Orbiting Carbon Observatory (OCO) mission was proposed to deliver the first temporally and spatially resolved global observations of CO2 to improve our understanding of the sources and sinks of CO₂. A retrieval algorithm was developed to obtain the column-averaged dry-air mixing ratio of CO₂ (XCO₂) from spectroscopic measurements of absorption in the 0.76 µm O2 A band and two near-infrared (NIR) bands of CO₂ centered at 1.61 µm and 2.06 µm. An aerosol optical-property database was developed to aid with the retrievals. Principal-component analysis was used to speed up radiative transfer (RT) computations. To test the algorithm, column O₂ was retrieved from measurements of absorption in the O2 A band over the sea surface. Using a single sounding, the column O₂ was retrieved with an error of around 1%. Polarization was shown to have a significant impact on the retrieval-error budget. A new model based on computing two orders of scattering (2OS) was developed to compute polarization in the OCO spectral regions. The multiple-scattering, scalar model Radiant was combined with the 2OS model to create the R-2OS OCO RT model. Tests with simulated backscatter measurements at the OCO validation sites showed that the R-2OS model reduced the biases in retrieved XCO₂ to much lower than 1 ppm in most scenarios, compared to errors as high as 10 ppm using the scalar model. Aerosol vertical distribution, thin cirrus and surface bidirectional reflection need further study.

Files