Accounting for Aerosol Scattering in the Remote Sensing of Greenhouse Gas

Citation

Zhang, Qiong (2017) Accounting for Aerosol Scattering in the Remote Sensing of Greenhouse Gas. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z9XW4GRQ. https://resolver.caltech.edu/CaltechTHESIS:06152016-152036651

Abstract

This thesis includes three different projects related to the remote sensing of Earth's atmosphere. The first part, comprising Chapter 2 and Chapter 3, focuses on the retrieval of Level 1 product, particularly the effect of aerosol scattering in the remote sensing of greenhouse gases. In Chapter 2, we study the aerosol induced bias in the retrieval of column averaged CO₂ mixing ratios (X_CO2). Ground based remote sensing data from the California Laboratory for Atmospheric Remote Sensing Fourier Transform Spectrometer (CLARS-FTS) are used. We employ a numerical radiative transfer model to simulate the impacts of neglecting aerosol scattering on the CO₂ and O₂ slant column densities (SCDs) operationally retrieved from CLARS-FTS measurements. These simulations show that the CLARS-FTS operational retrieval algorithm likely underestimates CO₂ and O₂ abundances over the LA basin in scenes with moderate aerosol loading. The bias in the CO₂ and O₂ abundances due to neglecting aerosol scattering cannot be canceled by ratioing each other in the derivation of the operational product of X_CO2. We propose a method for approximately correcting the aerosol-induced bias. Results for CLARS X_CO2 are compared to the direct-sun X_CO2 retrievals from a nearby Total Carbon Column Observing Network (TCCON) station.

In Chapter 3, we explain why large X_CO2 retrieval errors are found over deserts in the space borne Orbiting Carbon Observatory-2 (OCO-2) data. We argue that these errors are caused by the surface albedo being close to a critical surface albedo (α_c). Over a surface with albedo close to α_c, increasing the aerosol optical depth (AOD) does not change the continuum radiance. The spectral signature caused by changing the AOD is identical to that caused by changing the absorbing gas column. The degeneracy in the retrievals of AOD and X_CO2 results in a loss of degrees of freedom (DOF) and information content (H). We employ a radiative transfer model to study the physical mechanism of X_CO2 retrieval error over a surface with albedo close to αc. Based on retrieval tests over surfaces with different albedos, we conclude that over a surface with albedo close to α_c, the X_CO2 retrieval suffers from a significant loss of accuracy.

In the Appendix, we put in a Chapter based on my work with Prof. Andrew Thompson on ocean The second part, mainly in Chapter 4, focuses on the application of Level 2 product. In this Chapter, we examine the uncertainties in middle atmospheric HO_x chemistry by comparing the Aura Microwave Limb Sound (MLS) OH and HO₂ measurements with the simulations of the Caltech-JPL KINETICS photochemical model. The model using the standard chemical kinetics underestimates OH and HO₂ concentrations in the mesosphere. To resolve the discrepancies, we use MLS OH and HO₂ measurements as benchmark to adjust the involved chemical rate coefficients within reasonable uncertainty ranges with an optimal estimation algorithm. The results show that four key reaction rate constants and the O₂ cross section at Lyman-α (121.6 nm) are the most sensitive parameters for determining the HO_x profiles. We conclude that the rate coefficient of H + O₂ + M → HO₂ + M requires a very large adjustment beyond the uncertainty limits recommended in the NASA Data Evaluation, which suggests the need for future laboratory measurements. An alternative explanation is that radiative association plays a significant role in this process, i.e. H + O₂ → HO₂ + hv, which has never been measured or computed.

In the Appendix, we put in a Chapter based on my work with Prof. Andrew Thompson on ocean submesoscale turbulence.

Thesis Files