Publication Abstracts
Martin 2014
, 2014: Advancements for Three-Dimensional Remote Sensing of the Atmosphere. Ph.D. thesis. Columbia University, doi:10.7916/D8WM1BZC.
Climate modeling efforts depend on remote sensing observations of clouds and aerosols in the atmosphere. This dissertation presents a foundation for using three-dimensional (3D) remote sensing techniques to retrieve cloud and aerosol properties in complex cloud fields.
The initial research was aimed at establishing a set of single-scattering properties that could be used in subsequent 3D remote sensing applications. We looked carefully at the sensitivity of single-scattering measurements to aerosol properties. The thesis presents an analysis of several hypothetical sets of single-scattering measurements at three visible wavelengths: 405, 532, and 780nm. Full data includes all angularly dependent scattering-matrix elements and both the volume-scattering and volume-absorption coefficients. Information content and microphysical parameter uncertainty is computed for full data and for several incomplete data subsets. Results show that even the phase function alone offers reasonable constraints on the size distribution for larger particles, yet all single-scattering measurements fail to constrain the size distribution for very-small particles. This leads to the inherent instability in estimates of number concentration. However, particle concentration is well-constrained when represented by volume or area density. Measurements of polarization improve estimates of size distribution and complex refractive index, provided that the particles are large enough to exhibit Lorenz-Mie scattering behavior. For smaller particles, approaching the Rayleigh limit, measurements of the volume absorption coefficient become increasingly important. These results are derived for laboratory and in situ instruments that measure extremely rich sets of single-scattering data, and provide useful intuition for satellite and aircraft based remote sensing of aerosols in the atmosphere.
Subsequent research focused improving satellite observations of aerosols near clouds, which are needed to understand the aerosol indirect effect. There is mounting evidence that we need to model 3D effects to retrieve cloud and aerosol properties together in certain key regions, including broken cloud fields and the regions near cloud edges. However, there remains some hesitancy in how to make progress without a clear plan for how to (efficiently) use a 3D vector radiative transfer solver in a multi-pixel retrieval framework. In particular, there is a problem of scale. Remote sensing of the atmosphere and surface in 3D regions will require thousands of unknown parameters and millions of radiometric measurements. The inverse problem is much larger than those based on 1D radiative transfer and the independent pixel approximation. Our work showed how to use adjoint methods to organize forward solver operations (i.e. calls to a 3D radiative transfer code) as efficiently as possible. The first solve computes the residual misfit with measurements, and the second solve computes the gradient of the misfit function with respect to all unknowns. In this way, the adjoint method allows us to adjust cloud and aerosol properties with only two radiative transfer calculations per wavelength — a number that is independent of the number of retrieval parameters, measurement view angles and pixels. This scalability property makes adjoint methods ideally suited to multi-pixel retrievals with many aggregated columns and to radiometric measurements with high spatial resolution, multiple angles, and polarimetric sensitivity.
Taken together, these two bodies of work contribute to the long-term goal of retrieving cloud and aerosol properties in the regions near cloud edges — retrievals which would greatly improve the modeling of cloud-aerosol interactions and the prediction of how these interactions impact weather and climate.
Export citation: [ BibTeX ] [ RIS ]
BibTeX Citation
@phdthesis{ma04010a,
author={Martin, W. G. K.},
title={Advancements for Three-Dimensional Remote Sensing of the Atmosphere},
year={2014},
school={Columbia University},
address={New York},
doi={10.7916/D8WM1BZC},
}
[ Close ]
RIS Citation
TY - THES ID - ma04010a AU - Martin, W. G. K. PY - 2014 BT - Advancements for Three-Dimensional Remote Sensing of the Atmosphere DO - 10.7916/D8WM1BZC PB - Columbia University CY - New York ER -
[ Close ]
