Abstract
Snow sublimating in dry air is a forecasting challenge and can delay the onset of surface snowfall and affect storm-total accumulations. Despite this fact, it remains comparatively less studied than other microphysical processes. Herein, the characteristics of sublimating snow and the potential for nowcasting snowfall reaching the surface are explored through the use of dual-polarization radar. Twelve cases featuring prolific sublimation were analyzed using range-defined quasi-vertical profiles (RDQVPs) and were compared with environmental model analyses. Overall, reflectivity Z significantly decreases, differential reflectivity ZDR slightly decreases, and copolar-correlation coefficient ρhv remains nearly constant through the sublimation layer. Regions of enhanced specific differential phase Kdp were frequently observed in the sublimation layer and are believed to be polarimetric evidence of secondary ice production via sublimation. A 1D bin model was initialized using particle size distributions retrieved from the RDQVPs using numerous novel polarimetric snow retrieval relations for a wide range of forecast lead times, with the model environment evolving in response to sublimation. It was found that the model was largely able to predict the snowfall start time up to 6 h in advance, with a 6-h median bias of just −18.5 min. A more detailed case study of the 8 December 2013 snowstorm in the Philadelphia, Pennsylvania, region was also performed, demonstrating good correspondence with observations and examples of model fields (e.g., cooling rate) hypothetically available from such a tool. The proof-of-concept results herein demonstrate the potential benefits of incorporating spatially averaged radar data in conjunction with simple 1D models into the nowcasting process.
Significance Statement
The goals of this work are to comprehensively survey the dual-polarization radar characteristics of snow evaporating in dry air and to investigate whether information gleaned from polarimetric radars can be used with a predictive model to help make short-term predictions about when snow will overcome dry air and reach the ground. We found that by using this radar information and a simple model we could predict the start time of snow up to 6 h in advance with reasonable accuracy. In conjunction with other available data, this proof of concept could help forecasters to make short-term predictions about when snowfall impacts will begin.
© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).
