On 21 August 2017 we measured skylight polarization during a total solar eclipse in Rexburg, Idaho, using two all-sky polarimetric imagers. The all-sky polarization images were recorded using three simultaneously operating digital singlelens-reflex (DSLR) cameras with good low-light sensitivity. Each camera was equipped with a 180° field-of-view fisheye lens to view the entire sky and each lens contained a fixed linear polarizer orientated at 0° , 60° , and 120° , respectively, to recover the first three Stokes parameters. Skylight polarization was measured from sunrise to sunset in the cameras’ blue, green, and red channels. Before and after totality, the maximum sky polarization occurred in its usual pattern with a band of maximum polarization positioned 90° from the sun. However, during totality skylight polarization became nominally symmetric about the zenith. This was observed clearly in the blue and green channels and less obviously in the red channel, which had a greatly diminished signal. At and near the observation site, we also operated an infrared cloud imager, a hand-held spectrometer to measure surface reflectance, and an AERONET solar radiometer to characterize the atmospheric aerosols. This ancillary data set provided a complete characterization of the conditions of the surrounding atmosphere and underlying surfaces.
Knowing the thermodynamic phase of a cloud–whether it is composed of spherical water droplets or polyhedral ice crystals–is critical in remote sensing applications and in climate studies. We recently showed that we can determine cloud phase with visible-wavelength sky polarimetry, and in this presentation we extend that method to shortwave infrared wavelength bands near 1.6 microns. We describe the instrument, a passive, three-channel polarimeter with spectral bands at 1550 nm, 1640 nm, and 1700 nm with approximate width of 40 nm and how we are using it in experiments to discriminate between liquid-water and ice clouds. This portable polarimeter measures scattered sunlight using polarizers orientated at 0° , 45‡ , and 90° with respect to the solar vertical scattering plane. It has a 4.9° field-of-view and a motorized, computer-controlled pan-and-tilt mount that controls the positioning of the polarimeter so that it can measure any point in the sky.
Getting students interested in science, specifically in optics and photonics, is a worthwhile challenge. We developed and implemented an outreach campaign that sought to engage high school students in the science of polarized light. We traveled to Montana high schools and presented on the physics of light, the ways that it becomes polarized, how polarization is useful, and how to take pictures with linear polarizers to see polarization. Students took pictures that showed polarization in either a natural setting or a contrived scene. We visited 13 high schools, and presented live to approximately 450 students.
At any given time, clouds cover approximately 60% of the Earth’s surface and they strongly influence weather and climate; however, they are one of the largest sources of uncertainty in climate models and predictions of atmospheric effects on remote sensing measurements. Knowing the cloud thermodynamic phase – whether a cloud is composed of ice crystals or liquid particles – is critical in these applications. Knobelspiesse et al. (Atmos. Meas. Tech., 8, 1537–1554, 2015) showed theoretically that the sign of the S1 Stokes parameter can be used to detect cloud thermodynamic phase when observed with a ground-based passive polarimeter and demonstrated this principle with a zenith-viewing polarimeter. In this theory, a positive S1 value indicates a liquid cloud, while a negative S1 value indicates an ice cloud. In this paper, we report the use of our all-sky polarimeter, operating at 450 nm (10 nm band) to detect ice, liquid, and multi-layered clouds. The cloud thermodynamic phase was independently verified with a dual-polarization lidar pointed at the zenith.
A solar eclipse provides a rare opportunity to observe skylight polarization during conditions that are fundamentally different than what we see every day. On 21 August 2017 we will measure the skylight polarization during a total solar eclipse in Rexburg, Idaho, USA. Previous research has shown that during totality the sky polarization pattern is altered significantly to become nominally symmetric about the zenith. However, there are still questions remaining about the details of how surface reflectance near the eclipse observation site and optical properties of aerosols in the atmosphere influence the totality sky polarization pattern. We will study how skylight polarization in a solar eclipse changes through each phase and how surface and atmospheric features affect the measured polarization signatures. To accomplish this, fully characterizing the cameras and fisheye lenses is critical. This paper reports measurements that include finding the camera sensitivity and its relationship to the required short exposure times, measuring the camera’s spectral response function, mapping the angles of each camera pixel with the fisheye lens, and taking test measurements during daytime and twilight conditions. The daytime polarimetric images were compared to images from an existing all-sky polarization imager and a polarimetric radiative transfer model.
Polarization can be used to detect manmade objects on the ground and in the air, as it provides additional information beyond intensity and color. Skylight can be strongly polarized, so the detection of airplanes in flight requires careful consideration of the skylight degree and angle of polarization (DoLP, AoP). In this study, we detect poorly resolved airplanes (≥ 4 pixels on target) in flight during daytime partly cloudy and smoky conditions in Bozeman, Montana. We used a Polaris Sensor Technologies SWIR-MWIR rotating imaging polarimeter to measure the polarization signatures of airplanes and the surrounding skylight from 1.5 to 1.8 μm in the short-wave infrared (SWIR). An airplane flying in a clear region of partly cloudy sky was found to be 69% polarized at an elevation angle of 13° with respect to the horizon and the surrounding skylight was 4-8% polarized (maximum skylight DoLP was found to be 7-14% at an elevation angle of 50°). As the airplane increased in altitude, the DoLP for both airplane and surrounding sky pixels increased as the airplane neared the band of maximum sky polarization. We also observed that an airplane can be less polarized than its surrounding skylight when there is heavy smoke present. In such a case, the airplane was 30-38% polarized at an elevation angle of 17°, while the surrounding skylight was approximately 40% polarized (maximum skylight DoLP was 40-55% at an elevation angle of 34°). In both situations the airplane was most consistently observed in DoLP images rather than S0 or AoP images. In this paper, we describe the results in detail and discuss how this phenomenology could detect barely resolved aircrafts.
Knowledge of the polarization state of natural skylight is important to growing applications using polarimetric sensing. We previously published measurements and simulations illustrating the complex interaction between atmospheric and surface properties in determining the spectrum of skylight polarization from the visible to near-infrared (1 μm).1 Those results showed that skylight polarization can trend upward or downward, or even have unusual spectral discontinuities that arise because of sharp features in the underlying surface reflectance. The specific spectrum observed in a given case depended strongly on atmospheric and surface properties that varied with wavelength. In the previous study, the model was fed with actual measurements of highly variable aerosol and surface properties from locations around the world. Results, however, were limited to wavelengths below 1 μm from a lack in available satellite surface reflectance data at longer wavelengths. We now report measurement-driven simulations of skylight polarization from 350 nm to 2500 nm in the short-wave infrared (SWIR) using hand-held spectrometer measurements of spectral surface reflectance. The SWIR degree of linear polarization was found to be highly dependent on the aerosol size distribution and on the resulting relationship between the aerosol and Rayleigh optical depths. Unique polarization features in the modeled results were attributed to the surface reflectance and the skylight DoLP generally decreased as surface reflectance increased.