Systematic simulations of aerosol optical property retrieval uncertainty for scanning polarimeters
Abstract
Scanning polarimeters, which utilize multi-angle, multispectral polarimetric observations from aircraft and orbit, represent the next generation of instruments capable of determining aerosol and cloud properties remotely. Retrieval of these properties from observations, however, are not straightforward. Iterative minimization techniques are often used to match radiative transfer simulations to the observations, where the aerosol and cloud parameters of the optimal model match are considered the best estimate of what is physically present in the scene. If the radiative transfer model perturbation sensitivity, expressed as a Jacobian matrix, can be assessed at the solution, then the observation uncertainty can be projected into the domain of the retrieved parameters. These parameter uncertainties provide are an extremely useful means to assess retrieval success. Another aspect of our iterative minimization techniques is the need for a reasonable initial estimate of optical properties. This estimate is provided by matching observations to a Lookup Table (LUT) of pre-computed radiative transfer scenes. This LUT spans a wide range of aerosol and cloud optical properties, and also includes numerical estimates of the Jacobian matrix at each element in the LUT. Using the Jacobians, we can estimate the retrieval uncertainty for all elements of the LUT, and therefore build a table of expected uncertainty. This paper presents how this approach is used in a systematic manner, and how it can be used to test retrieval capability for various combinations of polarized, multi-angle and multispectral observations.
Keywords
Aerosol; uncertainty; polarimetry; remote sensing
Full Text:
PDFDOI: https://doi.org/10.1478/C1V89S1P050
Copyright (c) 2015 AAPP | Physical, Mathematical, and Natural Sciences