Development of a daytime multispectral technique for detecting supercooled liquid water-topped mixed phase clouds

Satellite remote sensing of clouds via passive radiometers offers a power means for detecting and characterizing the properties of cloud cover globally. Many previous researches have demonstrated the general utility of multispectral information across the optical part of the electromagnetic spectrum (i.e., 0.4 to 14 µm) for determining cloud occurrence, classifying cloud type, and retrieving cloud top height/pressure, integrated liquid/ice water content, cloud emissivity, and cloud top microphysics. Cloud phase information, deduced conventionally from measurements in near infrared (~3.9 µm) and more recently via the infrared window (using a combination of 8.5, 11, and 12 µm bands), is of practical value to the aviation community. Aircraft flying through mixed-phase clouds in which supercooled water droplets coexist with ice particles below 0˚C can result in rapid ice accretion on the wings and frames, which can directly cause adverse impacts to flight performance and aircraft crashes. However, owing to the limitations of most contemporary satellite observing systems (passive instruments), the phase information is limited to near cloud top. Since recent in situ observations have often shown instances of altostratus clouds having supercooled liquid at their tops but a predominantly ice phase residing below, this may be particularly an important limitation for assessing these cloud phase characteristics. The physical mechanisms responsible for this phase structure, and the frequency/scale/distribution of occurrence, are not well understood.

In this study, we present a daytime multispectral technique for detecting the special class of mixed-phase cloud. The approach takes advantage of differential absorption between liquid and ice particles in the near infrared part of the spectrum. We define a liquid-normalized reflectance ratio for distinguishing between pristine liquid and liquid-topped ice clouds. Spectral reflectance characteristics of supercooled liquid water-topped mixed-phase clouds in near infrared channels are investigated via radiative model simulations. Cloud vertical inhomogeneity is also examined using reflectance and effective radius ratios from various channel combinations. Based on the a-priori database of simulated reflectance ratios, a particular threshold that gives indication of a detectable signal for liquid-over ice clouds is determined. Preliminary case study results are validated using aircraft in situ measurements. Demonstrable with current low-earth-orbiting satellite observations from MODIS, the algorithm is developed for the future high-temporal/spectral resolution geostationary of the GOES-R series Advanced Baseline Imager.