Use of Multiple-Angle Snow Camera (MASC) Observations as a Constraint on Radar-Based Retrievals of Snowfall Rate

Abstract:

We use a combination of Ka-band Zenith Radar (KaZR) and Multiple-Angle Snow Camera (MASC) observations at the ARM North Slope Alaska Climate Facility Site at Barrow to quantify snowfall. The optimal-estimation framework is used to combine information from the KaZR and MASC into a common retrieval scheme, where retrieved estimates of snowfall are compared to observations at a nearby NWS measurement site for evaluation. Modified from the operational CloudSat algorithm, the retrieval scheme returns estimates of the vertical profile of exponential PSD slope parameter with a constant number density. These values, in turn, can be used to calculate surface snowrate (liquid equivalent) given knowledge of snowflake microphysical properties and fallspeeds.

We exploit scattering models for a variety of ice crystal shapes including aggregates developed specifically from observations of snowfall properties at high-latitudes, as well as more pristine crystal shapes involving sector plates, bullet rosettes, and hexagonal columns. As expected, initial retrievals suggest large differences (300% for some events) in estimated snowfall accumulations given the use of the different ice crystal assumptions. The complex problem of how we can more quantitatively link MASC snowflake images to specific radar scattering properties is an ongoing line of research. Here, however, we do quantify the use of MASC observations of fallspeed and PSD parameters as constraint on our optimal-estimation retrieval approach. In terms of fallspeed, we find differences in estimated snowfall of nearly 50% arising from the use of MASC observed fallspeeds relative to those derived from traditional fallspeed parameterizations. In terms of snowflake PSD, we find differences of nearly 25% arising from the use of MASC observed slope parameters relative to those derived from field campaign observations of high-altitude snow events. Of course, these different sources of error conspire to make the estimate of snowfall non-unique. Our initial results, however, do suggest use of the MASC as a means to reduce some of this ambiguity in the snowfall retrieval problem.