Microphysical characterization of boundary layer ice particles: results from a 3-year measurement campaign in interior Alaska

Abstract

Abstract A three-winter study has been conducted to better understand the relationship between atmospheric conditions and ice fog or diamond dust microphysics. Measurements were conducted east of downtown Fairbanks in the Interior Alaska during non-precipitating conditions. Atmospheric conditions were measured with several weather stations around the Fairbanks region and two Meteorological Temperature Profiler instruments (ATTEX MTP-5HE and MTP-5PE). Near surface ice particle microphysical observations were conducted with the Particle Phase Discriminator (PPD-2K) instrument which measures particles from 8 to 112 μm (sphere-equivalent). Panoramic camera images were captured and saved every ten minutes throughout the campaign for visual assessment of atmospheric conditions. Machine Learning was used to classify both cloud particle microphysical characteristics from the PPD-2K data as well as to categorize boundary layer conditions using the panoramic camera images. For panoramic camera images, data were categorized as: cloudy, clear, fog, snowing, and a nearby powerplant plume was characterized. For the PPD-2K Machine Learning study the scattering pattern images were used to identify: rough surface, pristine, sublimating, and spherical particles. Three additional categories were used to identify indeterminant or saturated images. These categories as well as categories derived from weather station data (e.g., temperature ranges) are used to quantify ice microphysical properties under different conditions. For the complete microphysical dataset, pristine plates or columns accounted for 15.5%, 16.3% appeared to be sublimating particles and 43.4% were complex particles with either rough surfaces or multiple branches. Although the temperature was as warm as −20°C during measurements, only 1.3% of particles were classified as liquid.