Observed polarization and radiance images from a ground-based full-sky polarimeter are compared against a Successive
Order of Scattering (SOS) radiative transfer model for July 2011 cloud-free days in Bozeman, Montana, USA. The
imaging polarimeter measures radiance and polarization in 10-nm bands centered at 450, 490, 530, 675, and 780 nm.
The AERONET-retrieved aerosol optical depth, size distribution, and refractive index are used as inputs to the SOS
model. MISR BRDF retrievals are used for the surface reflectance. As shown in a previous work1 that considered 2009
data, model results generally agree with observation, but the simulated degree of polarization is typically higher than
observed data. Potential sources of this difference are considered, especially errors in the AERONET-retrieved aerosol
real refractive index.
With the increasing use of polarization as an added dimension in imagery for a variety of scientific, defense, and civilian
applications comes a need for better understanding of how the natural environment affects polarization signatures. In the
visible and near-infrared spectral range, the most important environmental component is polarized skylight. To provide
data to help improve understanding of how atmospheric polarization varies with aerosols, clouds, and surface
reflectance, an all-sky polarization imager has been designed, built, calibrated, and operated in a variety of field
experiments. This paper describes modifications made to that instrument to enable continuous, unattended outdoor
operation. The primary modifications were development of a weather-proof housing and an automated sun occulter
incorporating an on-board microcontroller that continually calculates solar position and moves an occulting disk on a
thin metal band to prevent direct sunlight from falling on the polarimeter lens. This occulter is designed to not obstruct
the principal scattering plane, defined as the plane containing the zenith, the Sun, and the observer.
Previous visible-band laboratory measurements have shown that polarization data can be used to determine optical
properties of materials such as the index of refraction with controlled illumination sources. For outdoor measurements,
the complex illumination formed by the polarized sky for visible wavelengths makes this process considerably more
difficult. This paper reports polarization measurements for horizontal painted-metal and PVC plates and the background
atmosphere from a quickly changeable dual-field imaging polarimeter which provides polarization of ground-based
objects nearly concurrently with full-sky polarization. A microfacet model has been developed which accounts for the
polarized sky illumination and solar-reflecting and flat-reflecting microfacets. Data from this model have been used to
explain the primary features of the polarization observed when viewing painted metal and PVC plates outdoors with
clear skies. Future work will attempt to use this model with polarimeter data to retrieve the index of refraction of the
An all-sky imaging polarimeter was deployed in summer 2008 to the Mauna Loa Observatory in Hawaii to study
atmospheric skylight polarization. This paper describes the Mauna Loa deployment and presents an initial comparison of
our data to those observed by Coulson with a zenith-slice polarimeter in the late 1970s and early 1980s. We show how
the all-sky imaging technique yields additional insight to the nature of skylight polarization beyond what is observed in a
single zenith scan.
An imaging spectro-polarimeter was used to measure polarization signatures in five visible and near infrared channels
(450, 490, 530, 630, 700 nm) from four painted steel plates throughout the day as the solar angle and sky conditions
changed. The primary objective of this study was to characterize the influence of variable sky conditions on the observed
polarization signatures. Smooth plates exhibited higher degree of polarization than rough plates and black plates
generally exhibited higher degree of polarization than tan plates. Changes in cloudiness caused large variations in the
polarization, usually reducing the observed degree of polarization.