A nonlinear least squares inversion to the single-scattering three-dimensional radiative transfer equation for satellite-based tomography

A major uncertainty in weather and climate models lies in the effect of clouds and aerosols. The IPCC's Fifth Assessment Report states that while models include more cloud and aerosol processes, there is still low confidence in the representation and quantification of the processes in climate models. One way to increase our understanding and representation of cloud and aerosol properties is by improving retrievals of these properties from our radiation measurements. The algorithms used for remote sensing from passive satellite instruments are outdated. For example, all state-of-the-art retrieval algorithms that are in operational use for retrieving cloud and aerosol properties assume the radiative transfer to be one-dimensional (1-D). It has become clear that using 1-D radiative transfer is unsuitable for atmospheres with horizontal heterogeneity, and that algorithms need to be developed to invert satellite measured radiance with 3-D radiative transfer for improved retrievals of cloud and aerosol properties. This thesis will cover the development of a single-scattering 3-D radiative transfer inversion algorithm. By producing a single-scattering inversion algorithm, we can learn many lessons and gain insight on how the eventual development for a multiple scattering inversion model should be approached. The inversion algorithm required the development of a single-scattering 3-D radiative transfer forward model. Results from the forward model are verified with the I3RC Monte Carlo model. The inverse problem was approached by way of tomography, a method that is commonly found in medical imaging. A non-linear least squares inversion model was developed to reconstruct the distribution of scattering properties in a heterogeneous domain. The inverse model produced excellent reconstructions in many cases, demonstrating the correctness of our formulation. However, due to the ill-posedness of the inverse problem, the inverse model did not always produce correct results. Future work should focus on further development of the inverse model by adding constraints.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Atmospheric Sciences

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• King, Megan R

• nonlinear least squares
• single-scattering
• three-dimensional radiative transfer equation
• tomography
• atmospheric remote sensing
• inverse problems

Statement dc:rights

• Copyright 2015 Megan R. King

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/78557

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/78557