In recent years, the study of aircraft icing has become focused on the effects of supercooled large drops (SLD) on aircraft performance in flight. Although research aircraft have gathered in-situ data on SLD in a few locations (e.g. Colorado, Sierra Mountains, Newfoundland), there is little understanding of where these conditions tend to occur geographically. Since regular observations of the occurrence of this phenomenon in the atmosphere do not exist, we must look for indirect indicators of their existence that are regularly recorded. One such indicator is the occurrence of freezing drizzle (ZL), freezing rain (ZR) or ice pellets (IP) at the surface. In order for these precipitation types to occur at the surface, ZL and/or ZR must exist in the lower troposphere immediately above the observation point. Research data on ZL and ZR has indicated that these conditions typically extend several thousand feet upward from the surface (Jeck, 1996). Thus, a climatology of the occurrence of ZL, ZR and IP at the surface may serve as a good indicator of how often to expect ZL and ZR in the lower troposphere, and to help identify where it tends to occur geographically. Regional differences in the amount and type of freezing precipitation which occur are further studied via examination of sounding data and surface analyses for cases when ZL, ZR and/or IP events spanned periods of at least 5 hours. These analyses reveal the typical synoptic situations in which SLD occurs aloft, what thermodynamic structures accompany them and how deep and cold the SLD layers tend to be.
B. Bernstein and B. Brown used a National Climatic Data Center (NCDC) database of encoded hourly and 3-hourly surface observations from 207 sites across the contiguous United States for the years 1961 to 1990 to conduct this study. Code was written to count occurrences of ZL, ZR and IP, then break them down by year, month, hour, temperature, wind direction, wind speed and simultaneous occurrence with other forms of precipitation. Ideally, data would have been available from each station for every hour of every day for all 30 years, giving a total of 262,968 observations. Only about 15 stations met this criteria, primarily because only 3-hourly data were saved for most stations during the years 1965-1981. To account for times when stations only recorded 3-hourly data, normalizing factors were used to estimate the number of observations of each meteorological field which would have been recorded if hourly observations were available for the entire period. This approach assumes that long periods of missing data were not present in the dataset. Close examination of the data have shown that such gaps did not exist.
Maps of the annual frequency of occurrence of the all 3 freezing precipitation types combined "ZZ" (see Figure 1) provided clear indications of where freezing precipitation was most common. Overall, a broad maximum in ZZ extends from the Texas panhandle to the Great Lakes, while smaller maxima exist along the east slope of the Appalachian Mountains from New England to North Carolina and within the Columbia River Basin of Oregon and Washington. Recent studies have shown that ZL, ZR and IP occur most frequently to the northeast, north and northwest of surface low pressure centers, especially ahead of warm fronts. The combination of this with typical wintertime storm tracks which run from the Texas panhandle to the Great Lakes, along the east coast shoreline and off the coast of British Columbia, Washington and Oregon matches the locations of the ZZ maxima quite well.
Regional differences in the percentage of freezing precipitation which fell as ZL, ZR or IP are clearly visible in Figures 2a, 2b and 2c. ZL is the most common form of freezing precipitation in the High Plains, especially in eastern Colorado, where values exceed 80%. The percentage of ZL decreases gradually toward the east and drops to less than 20% in the southeast and mid-Atlantic coastal regions. ZR and IP become more common in the midwest and east, with ZR percentages maximizing in the southeast (50-60%) and IP maximizing along the mid-Atlantic coast (50-80%). Distinct maxima in ZR and IP are also evident in the northwest states. IP can be formed when snow does not melt completely while falling through a layer of air with temperature above 0oC before falling into subfreezing air below, allowing an ice nucleus to survive within the raindrop and for that raindrop to freeze into an ice pellet upon supercooling sufficiently. Since ZR and IP are both most commonly formed by the melting process, the locations where IP dominates are locations where incomplete melting is common, while locations were ZR dominates are those where complete melting occurs. The maximum in IP along the Atlantic coast seems to point to the common occurrence of such incomplete melting there, perhaps due to the cooler moisture source offshore, whereas the dominance of ZR near the gulf coast may be due to the location relative to a much warmer moisture source, the Gulf of Mexico. Examination of sounding data and surface maps for ZR and IP events at Greensboro, North Carolina confirm that this is likely the case. The dominance of ZL in the High Plains seems to relate to the lack of deep moisture during freezing precipitation events. Examination of sounding data and surface maps for ZL events at Denver indicate that most events occur in shallow, anticylonic upslope situations, where subfreezing, saturated conditions are limited to the lowest 5000 ft of the atmosphere and dry air exists above. This setup is conducive to the formation of ZL via collision coalescence.
The development of a geographical distribution of the occurrences of ZL, ZR and IP and examination of surface maps and sounding data taken during these events has been very useful for the identification of processes tending to drive the formation of SLD by region and how they related to the location of storm tracks, major topographic features and bodies of water. With this understanding in hand, researchers, forecasters and perhaps even dispatchers and pilots, will have a better idea of what type of SLD to expect in flight, how often and when to expect it as well as how to avoid it.
