A Climatology of Freezing Rain in the Great Lakes Region of North America

John Cortinas Jr. Cooperative Institute for Mesoscale Meteorological Studies, University of Oklahoma, and National Oceanic and Atmospheric Administration/Oceanic and Atmospheric Research/National Severe Storms Laboratory, Norman, Oklahoma

Search for other papers by John Cortinas Jr. in
Current site
Google Scholar
PubMed
Close
Full access

We are aware of a technical issue preventing figures and tables from showing in some newly published articles in the full-text HTML view.
While we are resolving the problem, please use the online PDF version of these articles to view figures and tables.

Abstract

A 15-yr climatology of freezing rain surrounding the Great Lakes region of North America has been constructed using data from rawinsondes, surface stations, and gridded reanalyses from the National Centers for Environmental Prediction. This climatology reveals that there is a general increase in the freezing-rain frequency from west to east; however, the distribution shows large spatial and temporal variability. Most freezing-rain events are short lived and occur near sunrise between the months of December and March. Continuous freezing rain typically lasts less than 1 h, with 7% of events lasting longer than 5 h.

Most freezing rain is associated with extratropical cyclones, occurring northeast of the cyclone center in the presence of midlevel upward vertical motion, and air that is nearly saturated at low and midlevels, subfreezing near the surface, and warm (>0°C) at low levels (≈850 hPa). The location of the mean extratropical cyclone track during freezing-rain events in the Great Lakes region east of 87°W and the climatological cyclone track during January partially explains the eastward increase in freezing-rain frequency. Additionally, it appears that the proximity of the Atlantic Ocean to the area east of 81°W provides a large moisture source for these events, while western events appear to receive moisture from the distant Gulf of Mexico. The Great Lakes appear to have some effect on reducing the occurrence of freezing rain very near the western shores.

Corresponding author address: Dr. John Cortinas Jr., University of Oklahoma–NOAA Cooperative Institute for Mesoscale Meteorological Studies, 1313 Halley Circle, Norman, OK 73069.

Abstract

A 15-yr climatology of freezing rain surrounding the Great Lakes region of North America has been constructed using data from rawinsondes, surface stations, and gridded reanalyses from the National Centers for Environmental Prediction. This climatology reveals that there is a general increase in the freezing-rain frequency from west to east; however, the distribution shows large spatial and temporal variability. Most freezing-rain events are short lived and occur near sunrise between the months of December and March. Continuous freezing rain typically lasts less than 1 h, with 7% of events lasting longer than 5 h.

Most freezing rain is associated with extratropical cyclones, occurring northeast of the cyclone center in the presence of midlevel upward vertical motion, and air that is nearly saturated at low and midlevels, subfreezing near the surface, and warm (>0°C) at low levels (≈850 hPa). The location of the mean extratropical cyclone track during freezing-rain events in the Great Lakes region east of 87°W and the climatological cyclone track during January partially explains the eastward increase in freezing-rain frequency. Additionally, it appears that the proximity of the Atlantic Ocean to the area east of 81°W provides a large moisture source for these events, while western events appear to receive moisture from the distant Gulf of Mexico. The Great Lakes appear to have some effect on reducing the occurrence of freezing rain very near the western shores.

Corresponding author address: Dr. John Cortinas Jr., University of Oklahoma–NOAA Cooperative Institute for Mesoscale Meteorological Studies, 1313 Halley Circle, Norman, OK 73069.

1. Introduction

Freezing rain is a type of hazardous weather that has a significant impact on people, commerce, and property in many parts of the world. A tragic reminder of this fact occurred 5–9 January 1998, when a large ice storm struck portions of southeast Canada and the northeast United States. The National Oceanic and Atmospheric Administration (NOAA) estimates that the storm killed 44 people and caused nearly $4 billion of damage in the United States and Canada (NOAA 1998). Although the significant damage and fatalities occurred east of the Great Lakes of North America, these devastating storms occur in the Great Lakes region as well. One such ice storm moved across southern Ontario on 13–14 January 1968 and produced ice accumulations up to 4 cm over 2.5 × 104 km2. The effects from this storm included widespread power outages, disrupted mail and fire service, and collapsed buildings throughout the area (McKay and Thompson 1969).

In other parts of the Great Lakes, Storm Data reports that on 2–3 March 1976 one of the region’s worst ice storms coated extreme western New York with ice up to 10 cm thick. The storm caused an estimated $86 million worth of damage, and the President of the United States declared six counties major disaster areas (NOAA 1976). Despite the severity of this type of hazardous weather, relatively little research has been done to study and document the climatology of freezing rain events over the Great Lakes even though some areas surrounding the Great Lakes are major population centers for Canada and the United States and are severely impacted by these events.

The primary cause of freezing rain has been known since the early 1900s. It occurs in the presence of a thermal stratification that melts frozen precipitation (i.e., snow) completely in an elevated warm (>0°C) layer, which, because of an insufficient number of active ice nuclei, become supercooled through conduction and evaporation as they continue to descend through a subfreezing layer of air near the surface. These supercooled droplets freeze immediately upon impact with subfreezing objects at the surface (e.g., Okada 1914; Frankenfield 1915; Bennett 1959; Stewart 1985; Stewart and King 1990; Cyzs et al. 1996; Zerr 1997). The synoptic-scale environment that produces this thermal stratification often is caused by strong baroclinic systems that advect low-level warm, moist air over a shallow layer of subfreezing air at the surface (e.g., Stewart et al. 1990; Rauber et al. 1994; Keeter et al. 1995; Holle and Watson 1996). In this study, freezing drizzle was not considered since other research (Bocchieri 1980; Huffman and Norman 1988; Cober et al. 1996) suggests that the physical processes associated with freezing drizzle may be different than those associated with freezing rain, thereby making the results for freezing rain ambiguous if both phenomena are combined.

Case studies of Great Lakes events show that freezing rain can occur with extratropical cyclones that move across the Great Lakes soon after a polar air mass has established itself across the region. As these cyclones move east or northeast across or south of the Great Lakes, warm and moist air between 850 and 700 hPa is advected northward over the cold subfreezing surface air. The upward vertical motion associated with this warm advection pattern produces precipitation that freezes upon contact with subfreezing objects on the ground (Stewart and King 1987, 1990; Martner et al. 1993). Other notable observations during Great Lakes events include strong vertical shear near the bottom of the warm layer, precipitation organized in linear bands, and the occurrence of other types of precipitation with freezing rain (e.g., rain, snow, and ice pellets). Aside from these case studies, the author is unaware of any studies that have analyzed the distribution of many freezing rain events and summarized the synoptic-scale processes associated with them in the Great Lakes region.

The purpose of this study is to determine and document the spatial and temporal characteristics of freezing-rain events in the Great Lakes region, as well as the prevalent synoptic-scale processes that accompany them. Section 2 describes the data and the methodology used in the study; section 3 describes the observed climatology of these events; section 4 discusses some reasons for the observed spatial distribution and provides suggestions for future research; and section 5 lists the conclusions from this study.

2. Data

Hourly surface observations in the region of the Great Lakes were obtained from the National Climatic Data Center (NCDC) for 1976–90. This time period was selected because of the data availability at NCDC. Data prior to 1976 were not used because some station data were available for one out of every three hours, and data after 1990 were not used to assure that present weather observations were taken by human observers and not by the Automated Surface Observing System in the United States. The surface dataset that was used in this study was created by merging surface data in two different formats (DATSAV2 and TD-3280) into one consistent format. The quality control procedures for each format are documented in reports published by the United States Air Force (1986) and the United States Department of Commerce (1994). No additional quality control was performed on the data. To ensure reliable statistical results and that the annual cycle was sampled adequately for most years, only locations that had at least 80% of the present weather reports for 10 of the 15 years were used.1 After applying this criterion, 89 stations were available for analysis (Fig. 1). Synoptic-scale features were examined using the National Centers for Environmental Prediction (NCEP) Grid Point Data Set, redistributed on CD-ROM by Mass et al. (1987). This dataset2 contains meteorological data grids at 1000, 850, 700, 500, and 250 hPa, with a horizontal grid spacing of roughly 250 km.

The thermodynamic structure of freezing rain events was examined using rawinsonde data from the NOAA Radiosonde Data of North America CD-ROM (NOAA 1996). Data taken at 0000 and 1200 UTC were extracted for sites near the Great Lakes, if, at that location, freezing rain was occurring at 1100 or 2300 UTC, regardless of intensity or duration. This 1-h offset takes into account the actual time that the rawinsonde was released, typically 45 min to 1 h before 0000 or 1200 UTC. Temperature, dewpoint, and wind data were extracted and linearly interpolated to 25-hPa levels for analysis. Rawinsonde data from Sault Ste. Marie, Michigan; Green Bay, Wisconsin; Buffalo, New York; and Flint, Michigan, were used in this study.

3. Climatology

The climatological analyses in this study are designed to focus on the spatial and temporal distributions of freezing rain and the prevalence of particular synoptic-scale processes during these events. The distribution and frequency of the hourly surface observations provide a reasonable dataset to document and analyze the mesoscale spatial and temporal distributions of freezing rain. The lack of an upper-level dataset with similar station spacing, however, precludes a thorough climatological examination of the mesoscale processes that control these events. Instead, synoptic-scale processes are examined with rawinsonde data at particular stations in the Great Lakes region and rawinsonde data that have been objectively analyzed to a synoptic scale.

a. Spatial and temporal distribution of events

Hourly surface observations were used to construct the temporal and spatial distribution of freezing rain over the Great Lakes region. A relative frequency map of freezing rain, based on present weather reports, shows that in the region of the Great Lakes it is an infrequent mesoscale event, comprising less than 1% of all hourly weather observations each year at any one station (Fig. 2). As is expected with infrequent mesoscale events, there is a large variability in the frequency across the region. Over the western region of the domain, freezing rain occurs less frequently and the distribution is more homogeneous than over the eastern region. The annual frequency of freezing rain ranges from 0.06% in the western region to 0.45% in the eastern region. Between these extremes lies a distribution that suggests the Great Lakes play a role in the regulation of freezing rain that will be discussed in more detail later in this paper.

Across southern Ontario, a large area of high relative frequency extends from the St. Lawrence Valley to extreme southwestern Ontario, where it decreases rapidly. Because of inadequate station spacing, caution must be applied when interpreting the results in southeastern Ontario. This area is bounded by Lakes Erie and Ontario on the south side and Lake Huron on the west side. This distribution is similar to that observed by McKay and Thompson (1969), who report over 50 h of freezing precipitation3 per year in southeastern Ontario and southern Québec, and Stewart and Isaac (1999), who report over 25 h of freezing rain per year across extreme southeastern Ontario. West of this region, a low relative frequency exists along the western shore of Lake Huron. A similar drop in frequency occurs west of Lake Michigan. Baldwin (1973) also reports relative minima west of Lakes Huron and Michigan, where he reported that fewer than 12 days of freezing rain are observed annually within each of these areas.

An examination of the frequency of simultaneous freezing-rain reports across the region shows that the horizontal extent of freezing rain events exhibit mesoscale variability, with a linear arrangement usually oriented east–west (Fig. 3). To illustrate this spatial pattern, two sites were selected that were surrounded by several nearby stations: Buffalo and Flint. At each station, the number of times that Buffalo and that station reported freezing rain simultaneously was divided by the number of times that freezing rain was reported at Buffalo (and multiplied by 100%). A value of 50 indicates that freezing rain was reported at that station 50% of the time freezing rain was reported at Buffalo. A similar calculation was done for Flint.

These results indicate that broad areas of freezing rain occur infrequently and that even over mesoscale areas freezing rain usually occurs sporadically. The mesoscale organization of these events resembles an ellipse and is characterized by a large frequency gradient surrounding each site. For example, as one moves away from Buffalo, the frequency at which freezing rain occurs simultaneously at Buffalo and other locations decreases rapidly. From Buffalo to Niagara Falls, New York (roughly 10 km north of Buffalo), the frequency decreases to 63%. In central Michigan, a large frequency gradient also surrounds Flint. The frequency decreases to 52% at Lansing (roughly 80 km to the west of Flint). This elongated frequency pattern, characterized by a large spatial gradient, was found for other locations as well (not shown). This pattern suggests that mesoscale areas of continuous freezing rain typically are associated with phenomena that exhibit an east–west linear, or banded, arrangement; one of which may be a warm front.

Over the entire region, freezing rain occurs between November and April, with very few reports during the months of October and May (Fig. 4). Freezing rain is reported most often between the months of December and March; during each of these months, the average number of freezing rain reports across the entire region is roughly 200 per year. Although there is a slight decrease in the median number of these reports during January and February, there is little variability in the median monthly frequency from December to March. The predominant intensity of freezing rain is light, reported for 98% of all freezing-rain observations. Additionally, freezing rain often does not occur concurrently with other types of precipitation at a particular location (Table 1), although different types of precipitation may occur before and after periods of freezing rain.

Since the predominant intensity of freezing rain is light, the event duration usually defines the severity. Before examining the distribution of event durations, the definition of an event must be established. Previous studies differ on the definition of an event and some include the nonprecipitation observations as part of the event duration. Gay and Davis (1993), for example, defined a freezing-rain event as the number of hours between the first report of freezing precipitation to the last report within a 24-h period. This definition, however, includes hours with or without any precipitation, potentially overestimating the duration of freezing rain. The current study examined four distributions of event duration based on a nonevent time interval (i.e., the number of intervening nonfreezing-rain observations between freezing-rain observations): 0, up to 6, up to 12, and up to 24. The event duration is the number of hourly freezing rain observations during a particular period that may have an unlimited number of predefined time intervals of no freezing rain, which are not counted as part of the duration. For example, if the nonevent time interval is zero, than only sequential hourly observations would be counted as one event. This also means that two 1-h events would be counted if an observation of freezing rain during hour one is followed by a non-freezing-rain observation in hour two and then another freezing rain observation in hour three.

All distributions of the event duration indicate that most events are short lived, regardless of how the event is defined (Fig. 5). When the nonevent time interval is zero (i.e., no nonevent observations between successive freezing rain observations), the majority of freezing-rain events (47%) last long enough to be reported during only one standard reporting time, and only 7% of all events are observed during six or more hourly reports. When the nonevent reports are increased to six, then the percentage of events that consist of only one hourly report decreases to 34% and the number of six or more hourly observations during the event increases to 15%. A systematic increase in nonevent observations from zero to six shows that the duration distribution changes little after a 3-h nonevent interval (not shown). This means that, for most events, if the freezing rain is discontinuous, the number of intervening hourly observations of no freezing rain is usually three or less.

One can also speculate that the short-lived nature of freezing rain at a particular location may be the warm advection at the surface associated with the passage of a warm front. After frontal passage, the subfreezing surface layer may warm above freezing and change the freezing rain to rain. When the horizontal advection is weak, Stewart (1985) suggests that the lack of long-lived freezing rain events is because they are self-limiting. In the absence of condensation, evaporation, adiabatic heating or cooling, and horizontal advection, melting in the warm layer lowers the air temperature to 0°C. Additionally, any refreezing in the low-level subfreezing layer will warm the air and ultimately create an isothermal layer near 0°C. Using a simple analytical model and a rainfall rate of 1 mm h−1, Stewart shows that a shallow (≈500 m) elevated warm layer can be eliminated within 2h. Observations from previous studies of freezing-rain environments (Stewart and King 1987, 1990; Martner et al. 1993; Rauber et al. 1994) suggest that both of these processes can affect event longevity.

The long-lived events during which freezing rain is reported for six or more sequential observations are expected to be more severe than the short-lived events and would be of great concern to the public.4 At many locations, these severe events occur two or more times every five years, with the two highest frequencies at Ottawa, Ontario, and Altoona, Pennsylvania (Fig. 6). The spatial distribution of these significant events shows that they occur most frequently in eastern two-thirds of the region, with some mesoscale variability. Although the number of severe events at each location is statistically significant at the 1% level, it is difficult to evaluate the statistical significance of the spatial distribution because of the irregular station spacing. The two longest events occurred at Rome, New York, and Val d’Or, Québec, where continuous freezing rain was reported for 21 h on 2–3 March 1976 and 18 h on 16 November 1989, respectively.

Both short- and long-lived events can occur at any time of the day. Across the Great Lakes region, the maximum frequency occurs at 1200 UTC and the minimum frequency occurs at 1900 and 2000 UTC, with a frequency range of slightly more than 2% (Fig. 7). A nonparametric significance test shows that the frequency for most hours, including the hours of maximum and minimum frequency, are statistically significant at the 1% level. This observed cycle is similar to the diurnal solar cycle and suggests that it plays a minor role in creating surface conditions conducive to freezing rain. Near sunrise, typically the coldest time of the day, freezing rain occurs most frequently, whereas the lowest frequency is near the warmest time of the day, near sunset. The reasons for the occurrence of a secondary maximum at 0400 UTC, which is also statistically significant at the 1% level, are not apparent.

b. Meteorological conditions

An examination of the surface conditions associated with freezing rain across the entire region shows that the largest percentage of reports occurs when the surface dry-bulb temperature is between 0° and −2°C, the dewpoint temperature is between 0° and −2°C (Fig. 8), and the dewpoint depression is between 1° and 2°C (not shown). The existence of subsaturated surface conditions during many freezing rain observations suggests that evaporational cooling could be important in creating subfreezing air near the ground. Freezing rain rarely occurs (less than 1.5% of freezing-rain reports) when the dry-bulb temperature is less than −10°C. The highest and lowest dry-bulb temperatures associated with freezing rain were 4° and −16°C, respectively. Freezing rain can occur when the air is above freezing if it became that way from rapid warming and if the ground and other objects are still below freezing. Rain that impacts objects that remain subfreezing freezes rapidly. The increase in the number of active ice nuclei at temperatures less than −10°C is a possible explanation for the low frequency of freezing rain at those temperatures (Pruppacher and Klett 1980, p. 264). Light easterly winds (2–4 m s−1) at the surface, often associated with high pressure over Hudson Bay, also accompany freezing rain over most of the region (Fig. 9).

As discussed earlier, the existence of a deep warm layer over a subfreezing (or recently subfreezing) surface layer provides the necessary thermal stratification to create freezing-rain conditions. The synoptic-scale processes that produce this stratification over the Great Lakes were examined by creating a series of composite maps with the NCEP dataset. To create the composite maps, all reports of freezing rain at 1100 or 2300 UTC were stratified by longitude and organized into three regions to identify any meridional differences in the synoptic-scale organization of the prevalent physical processes. Freezing rain observations from 94° to 87°W, 87° to 81°W, and 81° to 74°W were classified as western, central, and eastern events, respectively (see Fig. 1 for region boundaries). The selection process produced 167, 198, and 297 observations in the western region, central region, and eastern region, respectively. Each composite map was created by extracting the gridded data for each freezing rain time and computing the mean and the standard deviation for each variable of interest at each grid point.

An examination of the synoptic-scale conditions associated with events in each region shows that most events are associated with synoptic-scale upward motion associated with strong low- and midlevel warm advection and increasing positive vorticity advection with height. These conditions occur as a result of a shortwave trough moving eastward across the Great Lakes region. Because the synoptic-scale organization is similar in each region, although the location of the features relative to the freezing-rain region is different, only the details of central region events will be discussed.

The composite of surface conditions shows an extratropical cyclone southwest of the freezing-rain region and a strong anticyclone northeast of the area over Maine and southern Québec (not shown). Although surface wind data were not available in the NCEP grid point dataset, the sea level pressure pattern suggests that the strength and the movement of the polar high pressure center may be important in establishing the cold air near the surface before the event (Fig. 10a). As a result of the high pressure, surface winds across the freezing rain region are easterly or southeasterly, while winds south of the region are southerly. The low- to midlevel (1000–666 hPa) relative humidity5 indicates that the air is moist (≥75%; standard deviation of 18%–21%). The median direction of the wind, the position of the extratropical cyclone relative to the freezing rain area, and the orientation and the shape of the climatological areas of simultaneous reports of freezing rain (Fig. 3) suggest that such an arrangement may be caused by a warm front. Indeed, studies of freezing rain events in this region have shown that freezing rain can occur along and north of warm-frontal boundaries (Stewart and King 1987, 1990; Martner et al. 1993; Rauber et al. 1994).

Above the subfreezing and moist surface conditions, warm advection and positive differential vorticity advection associated with a midlevel shortwave trough creates a warm moist layer and implies that upward vertical motion is occurring over the freezing rain (central) region (Fig. 10b). At 850 hPa, a closed circulation and a shortwave trough are located 400 km to the west-northwest of the region, producing strong low-level warm advection over the entire region (≈2 × 10−1 °C h−1). A short-wave trough is also found at 700 and 500 hPa, where warm advection associated with southwesterly winds also occurs (Figs. 10c and 10d).

Upper-level composites from central region events were compared with composites maps during non-freezing-rain events, hereafter referred to as nonevents, to determine if there were any features unique to the freezing-rain events. Nonevents occurred when a surface cyclone was within 100 n mi of the freezing-rain composite location (Fig. 10a) and freezing rain had not occurred anywhere in the Great Lakes region 1 h before and 1 h after the rawinsonde observation time (1100 and 2300 UTC). These nonevents were selected by examining the 0000 and 1200 UTC surface weather charts between December and March for all years between 1976 and 1990. This selection process produced 45 nonevents.

A comparison of the event and nonevent climatologies in the central region of the lakes shows several differences. During freezing rain events, the mean sea level pressure near the center of the high pressure center over the eastern part of the domain is higher than during nonevents, while the pressure magnitude near the centers of the surface cyclones are similar (not shown). A warm anomaly at 850, 700 (Figs. 10b and 10c), and 500 hPa of up to 8°C also distinguishes freezing-rain events from nonevents. Although the event and nonevent climatologies both indicate an area of maximum warm advection that is similar in magnitude, the area associated with nonevents is south of that associated with freezing-rain events (not shown). The short-wave troughs at 850, 700, and 500 hPa had greater cyclonic curvature during nonevents than during freezing-rain events, which created a greater magnitude of vorticity in the main short-wave trough during nonevents. Other differences between events and nonevents were less significant and suggest that antecedent conditions (12–24 h before the event), such as the strength of the cold advection associated with the surface high pressure center and the magnitude of the midlevel warm advection prior to the passage of the short-wave trough, may be important in determining whether or not freezing rain will occur.

Given the existence of strong warm advection during freezing rain and the importance of upward vertical motion in generating precipitation, it is of interest to diagnose the processes that force quasigeostrophic adiabatic ascent (e.g., Bluestein 1992): positive differential vorticity advection and warm advection. These processes are represented by the two terms on the right-hand side of the quasigeostrophic omega equation:
i1520-0493-128-10-3574-e1
where f and σ are the Coriolis and static stability parameters, ζg and vg are the geostrophic absolute vorticity and the geostrophic wind, R is the universal gas constant, p is pressure, and T is temperature. The relative contributions of these processes were assessed by using the composite gridpoint data to calculate the differential vorticity advection [first term on the right-hand side of (1)] between 850 and 500 hPa and the thermal advection [second term on the right-hand side of (1)] at 850 hPa during freezing-rain events.

Over the entire central region, positive differential vorticity advection and warm advection both contribute nearly equally to ascent, with the maximum combined forcing east of the region (Fig. 11a). The analysis shows that the maximum contribution to ascent by thermal advection is located east of the region and the maximum contribution from differential vorticity advection is located in the southern portion of the region (Fig. 11b). Holle and Watson (1996) also found that both processes contributed to ascent during a freezing-rain event in the central United States. A comparison of these processes during events and nonevents shows that upward motion over the region is greater during nonevents because of a larger contribution from positive differential vorticity advection (not shown).

The standard deviation of the term on the left-hand side of (1) indicates that there is significant variability in the upward motion that occurs over the entire region during these events, including some events when downward motion exists over the region (Fig. 11a). Since the composite maps are composed of individual events that occurred at different locations across the entire central region, Fig. 11a does not provide an accurate description of the vertical motion at each freezing-rain location. Therefore, the distribution of vertical motion was determined by computing the left-hand side of (1) at each freezing-rain location in the central region. It was found that upward motion was associated with most freezing-rain events, although there was no correlation between the strength of the vertical motion and the duration or the intensity of the precipitation. For all central region events, the mean value of the left-hand side of (1) was 20 × 10−13 Pa s−1 m−2 and the standard deviation was 16 × 10−13 Pa s−1 m−2. Downward motion was associated with only 3% of the events, which had a median duration that included two hourly observations.

In addition to the importance of dynamical processes that cause upward motion during freezing rain, the thermal stratification also plays a critical role. Between 1976 and 1990, 15 rawinsondes were launched at Flint during freezing rain. These data (except for two times when data were unavailable) were used to create a median thermodynamic profile during freezing rain (Fig. 12).

The profile at Flint indicates a stable stratification below 500 hPa that is nearly saturated, similar to the composite sounding found by Bocchieri (1980) using 48 U.S. stations. Within the lowest saturated subfreezing layer, a temperature inversion exists above 950 hPa and extends from the ground up to near 850 hPa. Strong vertical wind shear and veering winds indicate that warm advection is occurring near the top of the subfreezing layer. Above roughly 900 hPa, this warm advection produces a warm layer that extends up to 775 hPa (≈1.5 km deep layer). Above the warm layer, a nearly saturated layer exists in which the lapse rate is roughly 4°C km−1 to 500 hPa. The largest variability in the dry-bulb and dewpoint temperatures exists near the base of the warm layer at 900 hPa (Fig. 12b), suggesting that the top of the cold layer typically varies between 950 and 850 hPa. Because the temperatures within the nearly saturated layer above the warm layer are lower than −10°C, many ice nuclei are active, suggesting that most freezing precipitation begins as snow and melts as it descends through the warm layer. The thermodynamic structure at Flint during freezing rain was similar to that at Sault Ste. Marie, Green Bay, and Buffalo (not shown).

The results of the synoptic-scale composite maps and the rawinsonde data show that five conditions appear frequently during freezing-rain events over the Great Lakes region: subfreezing air near the surface, a deep layer (2–3 km) of nearly saturated air above the ground, upward vertical motion, a deep warm layer (depth of 1 km) above the subfreezing layer, and midlevel air that is near saturation and cold enough to support active ice nuclei. These conditions also are similar to those found in studies of individual freezing rain events near Lakes Erie and Ontario (Stewart and King 1987, 1990; Martner et al. 1993) and other locations in the United States (Rauber et al. 1994; Holle and Watson 1996).

4. Discussion

So far, the observed climatological distributions of freezing rain have been shown with little reference to the causes of these distributions. Although the data used in this study preclude a thorough examination of all these causes, it is still possible to provide a basic understanding of the spatial distribution, particularly the general increase in freezing rain from west to east and the relative minimum frequency of freezing rain on the western shores of the Great Lakes (Fig. 2). As shown by the composite charts, freezing-rain conditions typically are associated with midlevel short-wave troughs and are found in the northeast quadrant of extratropical cyclones (Fig. 10). To determine whether the location of the cyclone track has an effect on the eastward increase in freezing rain frequency, the average cyclone locations before, during, and after freezing rain were computed by averaging the mean sea level pressure in 12-h intervals from 24 h before (−24 h) to 24 h after (+24 h) each freezing-rain report used in creating the composite maps (Fig. 13).

For western and central events, the mean cyclone center at −24 h is in the Texas panhandle. At −24 h a cyclone center is not discernable for eastern region events. Over the next 12 h, the positions of the low centers diverge as the mean cyclone position shifts northeastward prior to freezing rain across the central and eastern regions, while the low center position moves little 12 h prior to freezing rain in the western region. For eastern region events, a cyclone center is present in central Illinois at −12 h. The cyclone centers continue to move northeast over the next 12 h and are usually southwest of the freezing rain region when freezing rain occurs (0 h). Twelve hours later (12 h), the cyclone centers continue to move northeastward toward southeastern Canada. Based on the average positions of the cyclone centers, some cyclones associated with freezing rain in all regions decelerate as they cross the Great Lakes, which may contribute to some prolonged periods of freezing rain in the central and eastern regions. A similar deceleration was observed by Angel and Isard (1997) for strong cyclones crossing the Great Lakes, as well as those that occurred during February and formed poleward of 43°N.

A comparison of the average cyclone centers at 0 h with the climatological distribution of cyclones for January (Zishka and Smith 1980) shows that the average cyclone location for central and eastern events occurs in a climatologically favored location for winter cyclones (Fig. 14). The relative frequency of cyclones also decreases toward the west, similar to the decrease in freezing-rain frequency across the Great Lakes. This suggests that the eastward increase in the frequency of freezing-rain events may be associated with the high climatological frequency of cyclones that follow a favorable track for freezing rain in the central and eastern regions, when other conditions, such as a warm temperature anomaly over a subfreezing surface layer, are present. Moreover, since freezing-rain conditions may exist 200–300 km ahead of a cyclone, it is possible that these conditions exist over the eastern and central regions during western events as well. The western region, however, would receive less freezing rain than the other regions since it would not be in a favorable location for freezing rain during central and eastern events.

Given the proximity of the Atlantic Ocean and the topography of the eastern Great Lakes region, one may speculate that these two features also contribute to the relatively high frequency of freezing rain in that region. In the warm sector of the cyclone, warm air at the surface is advected northward and rises over the dome of cold air north of the warm front. As indicated by the surface isobaric pattern, the circulation associated with western region events appears to advect low-level moisture from the distant Gulf of Mexico and not from the Atlantic Ocean, providing limited moisture for continuous, widespread precipitation (Fig. 15a). In contrast, during central and eastern region events, surface winds advect a continuous source of moist air from the nearby Atlantic Ocean into the freezing-rain region, where it rises above the cold dome of air at the surface and creates precipitation (Fig. 15b). In addition to the proximity of a moisture source, the Appalachian Mountain Range across Pennsylvania, New York, and southeastern Canada provides numerous valleys in which subfreezing air settles and remains during freezing rain events. Moreover, the St. Lawrence Valley in southern Québec provides a conduit through which subfreezing air from eastern Canada can reach the eastern Great Lakes and create freezing-rain conditions in the presence of an elevated, nearly saturated warm layer.

Besides the general increase of freezing rain from west to east across the Great Lakes, the areas immediately to the west of Lake Huron, Lake Erie, Lake Michigan, and southwest of Lake Ontario experience a lower frequency of freezing rain than areas farther inland. This climatological distribution suggests that the Great Lakes may play a role in regulating the production of freezing rain. One may speculate that this distribution is the result of fall and spring events during which subfreezing air moves across an unfrozen lake and warms to a temperature equal to or greater than 0°C because of sensible heat flux from the lake surface. During these conditions, the predominant type of precipitation along the coast would be rain; although, freezing rain may occur briefly if the temperature of exposed objects is equal to or less than 0°C from a prior period of subfreezing temperatures. Stewart and King (1990) similarly describe the influence of a lake on a rain–snow transition zone.

A comparison of freezing rain observations at two western lakeshore and two inland locations illustrates the possible influence of the lakes on the frequency of freezing rain at some downwind locations. Freezing rain observations from 1976 to 1990 at Oscoda and Lansing, Michigan, and Milwaukee and Madison, Wisconsin, were stratified by month (Fig. 16). It appears that the influence of the lakes causes the frequency at some lakeshore locations (Oscoda and Milwaukee) to be less than at nearby inland locations (Lansing and Madison) in the fall, early winter, and spring. During these periods, the unfrozen lake increases the low-level temperature of air masses from slightly less than 0°C to greater than 0°C as they cross the lake and arrive at the downwind lakeshore. Although upper-level conditions may be favorable for freezing rain, rain will occur near the lakeshore because of the warm surface temperature. Farther inland, diabatic cooling may reduce the temperature to 0°C or less and freezing rain will occur.

Because of the lake’s influence on freezing rain, one would expect few freezing-rain events at lakeshore locations when the lake is unfrozen and the surface wind is from the lake. An examination of the wind direction at Oscoda during freezing rain observations in December—a month when the average water temperature is between 1° and 2°C (Irbe 1992)—shows that 80% of the observations were associated with winds that were not from the lake or were calm (not shown).

During the winter, the temperature of the lakeshore water decreases to near 0°C and may eventually freeze. When this happens the water no longer provides a heat source for subfreezing air masses that cross it; therefore, the subfreezing air crosses the lake and arrives subfreezing at the lakeshore. Given favorable low-level and midlevel environmental conditions, freezing rain will occur at the lakeshore and inland. Indeed, by February the frequency of freezing rain at the lakeshore and inland locations are nearly equal (Fig. 16). Detailed studies of individual lakeshore events obviously are necessary to confirm this hypothesis and provide more information about the extent and the magnitude of the lake’s influence on the occurrence of freezing rain.

This study provides information on where and when freezing rain occurs, as well as the primary meteorological conditions that accompany these events in the Great Lakes region of North America. There still remain, however, many unanswered questions about these events: Which factors determine the longevity of these events? Why are most short lived and a few devastatingly long? Which processes determine the intensity of freezing rain? What are the important influences of the Great Lakes on freezing-rain events near the lakeshore? Which factors cause these events to be so infrequent and so variable each year? Given the current state of numerical modeling, can these events be simulated accurately? Future research that examines these questions will broaden the current knowledge of freezing-rain events over the Great Lakes and hopefully improve the forecasts of these devastating events.

5. Conclusions

A climatology of freezing-rain events in the Great Lakes region was created, and these are the key findings.

  • Freezing rain occurs infrequently at individual locations primarily between December and March, usually is short-lived, often is not mixed with other types of precipitation, and occurs most frequently near sunrise.

  • There is large spatial and temporal variability in the annual frequency of freezing rain, with a general increase in frequency from west to east that appears related to the climatological track of surface cyclones, the proximity of the Atlantic Ocean to these parts of the lakes, and the presence of numerous valleys. This track favors freezing rain across the central and eastern regions of the Great Lakes.

  • A relative minimum in frequency is observed west of Lakes Huron, Michigan, Ontario, and Erie, suggesting that the lakes may have a moderating influence on the frequency of freezing rain near the shoreline during some freezing-rain events.

  • The meteorological conditions observed during freezing rain are those usually associated with the region downstream of an upper-level short-wave trough. These conditions create low-level temperatures that are greater than 0°C, warm temperature anomalies at 850 and 700 mb, and ascent over the freezing rain region. Surface conditions include nearly saturated subfreezing air and light easterly winds.

  • Compared to winter surface cyclones that do not produce freezing rain, freezing-rain events are characterized by a stronger surface anticyclone, a midlevel warm anomaly, and a weaker midlevel short-wave trough. The lack of other significant differences suggests that antecedent processes, such as the intensity of the midlevel warm advection and cold advection near the surface, may determine whether or not freezing rain will occur.

Acknowledgments

The help of many individuals is greatly appreciated: Joan O’Bannon skillfully assisted with some of the figures; Steve Fletcher and Dan Fauchier created the initial surface observation dataset; Mr. Greg Carbin, Drs. Dave Schultz and Harold Brooks, as well as three anonymous reviewers, provided insightful reviews of this work; and Neal Lott at the NOAA/National Climatic Data Center provided the surface dataset.

REFERENCES

  • Angel, J. R., and S. A. Isard, 1997: An observational study of the influence of the Great Lakes on the speed and intensity of passing cyclones. Mon. Wea. Rev.,125, 2228–2237.

  • Baldwin, J., 1973: The climates of the United States. National Oceanic and Atmospheric Administration, U.S. Department of Commerce, 113 pp. [NTIS COM-74-11708/6.].

  • Bennett, I., 1959: Glaze: Its meteorology and climatology, geographical distribution, and economic effects. Environmental Protection Research Division Tech. Rep. EP-105, Headquarters, U.S. Army Quartermaster, Research and Engineering Command, Natick, MA, 217 pp. [NTIS AD-216668.].

  • Bernstein, B. C., and B. G. Brown, 1997: A regional climatology of freezing precipitation for the contiguous United States. Preprints, 10th Conf. on Applied Climatology, Reno, NV, Amer. Meteor. Soc., 176–180.

  • Bluestein, H. B., 1992: Synoptic–Dynamic Meteorology in Midlatitudes. Vol. 1, Principles of Kinematics and Dynamics, Oxford University Press, 431 pp.

  • Bocchieri, J. R., 1980: The objective use of upper air soundings to specify precipitation type. Mon. Wea. Rev.,108, 596–603.

  • Cober, S. G., J. W. Strapp, and G. A. Isaac, 1996: An example of supercooled drizzle drops formed through a collision-coalescence process. J. Appl. Meteor.,35, 2250–2260.

  • Czys, R. R., R. W. Scott, K. C. Tang, R. W. Przybylinski, and M. E. Sabones, 1996: A physically based, nondimensional parameter for discriminating between locations of freezing rain and ice pellets. Wea. Forecasting,11, 591–598.

  • Frankenfield, H. C., 1915: Sleet and ice storms in the United States. Mon. Wea. Rev.,43, 608.

  • Gay, D. A., and R. E. Davis, 1993: Freezing rain and sleet climatology of the southeastern USA. Climate Res.,3, 209–220.

  • Holle, R., and I. Watson, 1996: Lightning during two central U.S. winter precipitation events. Wea. Forecasting,11, 599–614.

  • Huffman, G. I., and G. A. Norman Jr., 1988: The supercooled warm rain process and the specification of freezing precipitation. Mon. Wea. Rev.,116, 2172–2182.

  • Irbe, G. J., 1992: Great Lakes Surface Water Temperature Climatology. Climatological Studies 43, Environment Canada, 215 pp. [Available from Canadian Government Publishing Centre, Supply and Services Canada, Ottawa, ON K1A 0S9, Canada.].

  • Kanamitsu, M., 1989: Description of the NMC global data assimilation and forecast system. Wea. Forecasting,4, 335–342.

  • Keeter, K. K., S. Businger, I. G. Lee, and J. S. Waldstreicher, 1995:Winter weather forecasting throughout the eastern United States. Part III: The effects of topography and the variability of winter weather in the Carolinas and Virginia. Wea. Forecasting,10, 42–60.

  • Koch, S. E., M. DesJardins, and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor.,22, 1487–1503.

  • Martner, B. E., J. B. Snider, R. J. Zamora, G. P. Byrd, T. A. Niziol, and P. I. Joe, 1993: A remote-sensing view of a freezing-rain storm. Mon. Wea. Rev.,121, 2562–2577.

  • Mass, C. F., H. J. Edmon, H. J. Friedman, N. R. Cheney, and E. E. Recker, 1987: The use of compact discs for the storage of large meteorological and oceanographic data sets. Bull. Amer. Meteor. Soc.,68, 1556–1558.

  • McKay, G. A., and H. A. Thompson, 1969: Estimating the hazard of ice accretion in Canada from climatological data. J. Appl. Meteor.,8, 927–935.

  • NOAA, 1976: Storm Data: A Composite of Outstanding Storms. Vol. 18, 24 pp. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • ——, 1996: Radiosonde data of North America 1946–1995. Vol. 1. National Environmental Satellite, Data and Information Service/National Climatic Data Center. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • ——, Cited 1998: Eastern U.S. flooding and ice storm. [Available online from http://www.ncdc.noaa.gov/ol/reports/janstorm/janstorm.html.].

  • Okada, T., 1914: Notes on the formation of glazed frost. Mon. Wea. Rev.,42, 284–286.

  • Pruppacher, H. R., and J. D. Klett, 1980: Microphysics of Clouds and Precipitation. D. Reidel, 714 pp.

  • Rauber, R. M., M. K. Ramamurthy, and A. Tokay, 1994: Synoptic and mesoscale structure of a severe freezing rain event: The St. Valentine’s Day ice storm. Wea. Forecasting,9, 183–208.

  • Stewart, R. A., and G. A. Isaac, 1999: Freezing precipitation in Canada. Atmos.–Ocean,37, 87–102.

  • Stewart, R. E., 1985: Precipitation types in winter storms. Pure Appl. Geophys.,123, 597–609.

  • ——, and P. King, 1987: Freezing precipitation in winter storms. Mon. Wea. Rev.,115, 1270–1279.

  • ——, and ——, 1990: Precipitation type transition regions in winter storms over southern Ontario. J. Geophys. Res.,95, 22 355–22 368.

  • ——, R. W. Crawford, N. R. Donaldson, T. B. Low, and B. E. Sheppard, 1990: Precipitation and environmental conditions during accretion in Canadian east coast winter storms. J. Appl. Meteor.,29, 525–538.

  • United States Air Force, 1986: Climatic database users handbook. No. 4, DATSAV2 Surface Rep. USAFETAC/UH-86/004, 52 pp. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • United States Department of Commerce, 1994: Hourly surface airways observations. TD-3280. 40 pp. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • Zerr, R. J., 1997: Freezing rain: An observational and theoretical study. J. Appl. Meteor.,36, 1647–1661.

  • Zishka, K. M., and P. J. Smith, 1980: The climatology of cyclones and anticyclones over North America and surrounding ocean environs for January and July, 1950–77. Mon. Wea. Rev.,108, 387–401.

Fig. 1.
Fig. 1.

Topographical map showing the locations of upper-air (dots) and surface observations (pluses and dots): AOO = Altoona, PA; BUF = Buffalo, NY; FNT = Flint, MI; GRB = Green Bay, WI; IAG = Niagara Falls, NY; LAN = Lansing, MI; MKE = Milwaukee, WI; MSN = Madison, WI; OSC = Oscoda, MI; RME = Rome, NY; SSM = Sault Ste. Marie, MI; YOW = Ottawa, ON; and YVO = Val D’Or, PQ. Solid dark lines indicate the boundaries of the western, central, and eastern regions used in this study

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 2.
Fig. 2.

Frequency (×100%) of hourly freezing-rain reports relative to the number of available hourly observations in the research dataset from 1976 to 1990. To estimate the expected number of freezing rain observations per year, multiply the values by 0.876. Contours were drawn objectively using an objective analysis scheme described by Koch et al. (1983)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 3.
Fig. 3.

Frequencies (%) of freezing-rain reports at individual stations when freezing rain was reported at (a) Buffalo and (b) Flint from 1976 to 1990. Areas of frequencies greater than 40% are indicated by hatching

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 4.
Fig. 4.

Monthly distributions of freezing-rain relative frequencies over the Great Lakes region from 1976 to 1990. Top and bottom of each box represent 0.75th and 0.25th quantiles; line in the middle of the box represents the 0.50th quantile (median); upper and lower whiskers represent the upper and lower inner fences (1½ times the distance of the interquartile range from the 0.75th and the 0.25th quantiles); star indicates mean. Circles indicate outliers, data points outside the inner fences. Second y axis indicates the expected number of hourly freezing-rain reports based on a 30-day month

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 5.
Fig. 5.

Frequencies of freezing-rain reports per event from 1976 to 1990 across the Great Lakes region for different event definitions. Each definition of a freezing-rain event is based on the number of nonevent hourly reports separating reports of freezing rain. The number of events for each definition is shown in the inset table

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 6.
Fig. 6.

Average number of 5-h or longer events per 5-yr period from 1976 to 1990. Areas that experience two or more events per 5-yr period are hatched

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 7.
Fig. 7.

Frequency of freezing-rain reports by hour from 1976 to 1990

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 8.
Fig. 8.

Surface dry-bulb and dewpoint temperatures during freezing rain from 1976 to 1990. Each open circle indicates an observed dry-bulb–dewpoint temperature pair. Points along the solid line indicate saturated conditions. Histograms show the distributions of dry-bulb and dewpoint temperatures during freezing rain

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 9.
Fig. 9.

The mode of the surface wind direction distribution during freezing rain. Arrows point in the direction toward which the wind is blowing

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 10.
Fig. 10.

Mean synoptic conditions during freezing-rain events and nonevents over the central region of the Great Lakes from 1976 to 1990 (see Fig. 1 for boundaries of the central region): (a) sea level pressure (hPa, solid line) and 1000–666-hPa relative humidity (%, dashed line) during freezing-rain events; (b) 850-hPa geopotential height (m, dark solid line), temperature (°C, solid/dashed lines), temperature anomaly (°C, shaded), and wind (full barb and half-barb denote 5 and 2.5 m s−1, respectively); (c) same as (b) except for 700-hPa geopotential height; and (d) 500-hPa geopotential height (m, dark solid line) and geostrophic wind (full barb and half-barb denote 5 and 2.5 m s−1, respectively) during freezing-rain events. See the text for an explanation of the anomaly calculation

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 11.
Fig. 11.

(a) Mean (solid lines) and standard deviation (dashed lines) of the combined forcing of vertical motion by differential 850–500-hPa vorticity advection and 850-hPa thermal advection (×10−13 Pa s−1 m−2) during 198 central region freezing-rain reports from 1976 to 1990 (see Fig. 1 for region boundaries); (b) contributions of 850–500-hPa vorticity advection (dashed lines) and 850-hPa thermal advection (solid lines) to vertical velocity (×10−13 Pa s−1 m−2) during the same events used in (a). Positive (negative) values in (a) and (b) are associated with upward (downward) motion

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 12.
Fig. 12.

(a) Median values of dry-bulb temperature (°C, dark solid line) and dewpoint temperature (°C, dark dashed line) from a distribution of freezing-rain soundings at Flint. Data are plotted on a skew T–logp diagram, with pressure (vertical axis) given in hPa and winds shown by full barb = 5 m s−1 and half barb = 2.5 m s−1. (b) Median absolute deviation of the dry-bulb temperature (°C, dark solid line) and the dewpoint temperature (°C, dark dashed line) at Flint

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 13.
Fig. 13.

Mean locations of extratropical cyclones associated with 167 western (W), 198 central (C), and 297 eastern (E) freezing-rain reports at 24 and 12 h before the report (−24, −12), during the report (0), and 12 and 24 h after the report (12, 24) between 1976 and 1990. All times are relative to freezing rain reports at 1100 or 2300 UTC

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 14.
Fig. 14.

The 1950–70 areal distributions of (a) events and (b) relative variability with preferred propagation tracks superimposed for Jan cyclones. Values represent 28-yr totals (Zishka and Smith 1980)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 15.
Fig. 15.

Mean sea level pressure (hPa, solid line) and standard deviation (dashed line) during (a) 167 freezing-rain reports over the western region and (b) 297 freezing-rain reports over the eastern region of the Great Lakes from 1976 to 1990 (see Fig. 1 for region boundaries)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Fig. 16.
Fig. 16.

Monthly distributions of hourly freezing-rain reports at (a) Oscoda and Lansing, and (b) Milwaukee and Madison between 1976 and 1990. The stippled bars indicate lakeshore locations and the solid bars indicate inland locations

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3574:ACOFRI>2.0.CO;2

Table 1.

Percentage of time other precipitation type or thunder occurs with freezing rain over the Great Lakes region

Table 1.

1

Although Bernstein and Brown (1997) found the relative frequency to be insensitive to the strict data availability criteria used in this study, other statistics, such as duration, depend heavily upon a complete dataset.

2

Kanamitsu (1989) describes the details of how the gridpoint data were created.

3

Freezing precipitation includes freezing rain and freezing drizzle.

4

The subjective definition of a severe event was selected based on what the author expects most people would consider a severe long-lived freezing rain event.

5

The only moisture variable available in the gridpoint dataset was 1000–666-hPa relative humidity.

Save
  • Angel, J. R., and S. A. Isard, 1997: An observational study of the influence of the Great Lakes on the speed and intensity of passing cyclones. Mon. Wea. Rev.,125, 2228–2237.

  • Baldwin, J., 1973: The climates of the United States. National Oceanic and Atmospheric Administration, U.S. Department of Commerce, 113 pp. [NTIS COM-74-11708/6.].

  • Bennett, I., 1959: Glaze: Its meteorology and climatology, geographical distribution, and economic effects. Environmental Protection Research Division Tech. Rep. EP-105, Headquarters, U.S. Army Quartermaster, Research and Engineering Command, Natick, MA, 217 pp. [NTIS AD-216668.].

  • Bernstein, B. C., and B. G. Brown, 1997: A regional climatology of freezing precipitation for the contiguous United States. Preprints, 10th Conf. on Applied Climatology, Reno, NV, Amer. Meteor. Soc., 176–180.

  • Bluestein, H. B., 1992: Synoptic–Dynamic Meteorology in Midlatitudes. Vol. 1, Principles of Kinematics and Dynamics, Oxford University Press, 431 pp.

  • Bocchieri, J. R., 1980: The objective use of upper air soundings to specify precipitation type. Mon. Wea. Rev.,108, 596–603.

  • Cober, S. G., J. W. Strapp, and G. A. Isaac, 1996: An example of supercooled drizzle drops formed through a collision-coalescence process. J. Appl. Meteor.,35, 2250–2260.

  • Czys, R. R., R. W. Scott, K. C. Tang, R. W. Przybylinski, and M. E. Sabones, 1996: A physically based, nondimensional parameter for discriminating between locations of freezing rain and ice pellets. Wea. Forecasting,11, 591–598.

  • Frankenfield, H. C., 1915: Sleet and ice storms in the United States. Mon. Wea. Rev.,43, 608.

  • Gay, D. A., and R. E. Davis, 1993: Freezing rain and sleet climatology of the southeastern USA. Climate Res.,3, 209–220.

  • Holle, R., and I. Watson, 1996: Lightning during two central U.S. winter precipitation events. Wea. Forecasting,11, 599–614.

  • Huffman, G. I., and G. A. Norman Jr., 1988: The supercooled warm rain process and the specification of freezing precipitation. Mon. Wea. Rev.,116, 2172–2182.

  • Irbe, G. J., 1992: Great Lakes Surface Water Temperature Climatology. Climatological Studies 43, Environment Canada, 215 pp. [Available from Canadian Government Publishing Centre, Supply and Services Canada, Ottawa, ON K1A 0S9, Canada.].

  • Kanamitsu, M., 1989: Description of the NMC global data assimilation and forecast system. Wea. Forecasting,4, 335–342.

  • Keeter, K. K., S. Businger, I. G. Lee, and J. S. Waldstreicher, 1995:Winter weather forecasting throughout the eastern United States. Part III: The effects of topography and the variability of winter weather in the Carolinas and Virginia. Wea. Forecasting,10, 42–60.

  • Koch, S. E., M. DesJardins, and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor.,22, 1487–1503.

  • Martner, B. E., J. B. Snider, R. J. Zamora, G. P. Byrd, T. A. Niziol, and P. I. Joe, 1993: A remote-sensing view of a freezing-rain storm. Mon. Wea. Rev.,121, 2562–2577.

  • Mass, C. F., H. J. Edmon, H. J. Friedman, N. R. Cheney, and E. E. Recker, 1987: The use of compact discs for the storage of large meteorological and oceanographic data sets. Bull. Amer. Meteor. Soc.,68, 1556–1558.

  • McKay, G. A., and H. A. Thompson, 1969: Estimating the hazard of ice accretion in Canada from climatological data. J. Appl. Meteor.,8, 927–935.

  • NOAA, 1976: Storm Data: A Composite of Outstanding Storms. Vol. 18, 24 pp. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • ——, 1996: Radiosonde data of North America 1946–1995. Vol. 1. National Environmental Satellite, Data and Information Service/National Climatic Data Center. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • ——, Cited 1998: Eastern U.S. flooding and ice storm. [Available online from http://www.ncdc.noaa.gov/ol/reports/janstorm/janstorm.html.&rsqb;.

  • Okada, T., 1914: Notes on the formation of glazed frost. Mon. Wea. Rev.,42, 284–286.

  • Pruppacher, H. R., and J. D. Klett, 1980: Microphysics of Clouds and Precipitation. D. Reidel, 714 pp.

  • Rauber, R. M., M. K. Ramamurthy, and A. Tokay, 1994: Synoptic and mesoscale structure of a severe freezing rain event: The St. Valentine’s Day ice storm. Wea. Forecasting,9, 183–208.

  • Stewart, R. A., and G. A. Isaac, 1999: Freezing precipitation in Canada. Atmos.–Ocean,37, 87–102.

  • Stewart, R. E., 1985: Precipitation types in winter storms. Pure Appl. Geophys.,123, 597–609.

  • ——, and P. King, 1987: Freezing precipitation in winter storms. Mon. Wea. Rev.,115, 1270–1279.

  • ——, and ——, 1990: Precipitation type transition regions in winter storms over southern Ontario. J. Geophys. Res.,95, 22 355–22 368.

  • ——, R. W. Crawford, N. R. Donaldson, T. B. Low, and B. E. Sheppard, 1990: Precipitation and environmental conditions during accretion in Canadian east coast winter storms. J. Appl. Meteor.,29, 525–538.

  • United States Air Force, 1986: Climatic database users handbook. No. 4, DATSAV2 Surface Rep. USAFETAC/UH-86/004, 52 pp. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • United States Department of Commerce, 1994: Hourly surface airways observations. TD-3280. 40 pp. [Available from National Climatic Data Center, Federal Building, 151 Patton Avenue, Asheville, NC 28801-5001.].

  • Zerr, R. J., 1997: Freezing rain: An observational and theoretical study. J. Appl. Meteor.,36, 1647–1661.

  • Zishka, K. M., and P. J. Smith, 1980: The climatology of cyclones and anticyclones over North America and surrounding ocean environs for January and July, 1950–77. Mon. Wea. Rev.,108, 387–401.

  • Fig. 1.

    Topographical map showing the locations of upper-air (dots) and surface observations (pluses and dots): AOO = Altoona, PA; BUF = Buffalo, NY; FNT = Flint, MI; GRB = Green Bay, WI; IAG = Niagara Falls, NY; LAN = Lansing, MI; MKE = Milwaukee, WI; MSN = Madison, WI; OSC = Oscoda, MI; RME = Rome, NY; SSM = Sault Ste. Marie, MI; YOW = Ottawa, ON; and YVO = Val D’Or, PQ. Solid dark lines indicate the boundaries of the western, central, and eastern regions used in this study

  • Fig. 2.

    Frequency (×100%) of hourly freezing-rain reports relative to the number of available hourly observations in the research dataset from 1976 to 1990. To estimate the expected number of freezing rain observations per year, multiply the values by 0.876. Contours were drawn objectively using an objective analysis scheme described by Koch et al. (1983)

  • Fig. 3.

    Frequencies (%) of freezing-rain reports at individual stations when freezing rain was reported at (a) Buffalo and (b) Flint from 1976 to 1990. Areas of frequencies greater than 40% are indicated by hatching

  • Fig. 4.

    Monthly distributions of freezing-rain relative frequencies over the Great Lakes region from 1976 to 1990. Top and bottom of each box represent 0.75th and 0.25th quantiles; line in the middle of the box represents the 0.50th quantile (median); upper and lower whiskers represent the upper and lower inner fences (1½ times the distance of the interquartile range from the 0.75th and the 0.25th quantiles); star indicates mean. Circles indicate outliers, data points outside the inner fences. Second y axis indicates the expected number of hourly freezing-rain reports based on a 30-day month

  • Fig. 5.

    Frequencies of freezing-rain reports per event from 1976 to 1990 across the Great Lakes region for different event definitions. Each definition of a freezing-rain event is based on the number of nonevent hourly reports separating reports of freezing rain. The number of events for each definition is shown in the inset table

  • Fig. 6.

    Average number of 5-h or longer events per 5-yr period from 1976 to 1990. Areas that experience two or more events per 5-yr period are hatched

  • Fig. 7.

    Frequency of freezing-rain reports by hour from 1976 to 1990

  • Fig. 8.

    Surface dry-bulb and dewpoint temperatures during freezing rain from 1976 to 1990. Each open circle indicates an observed dry-bulb–dewpoint temperature pair. Points along the solid line indicate saturated conditions. Histograms show the distributions of dry-bulb and dewpoint temperatures during freezing rain

  • Fig. 9.

    The mode of the surface wind direction distribution during freezing rain. Arrows point in the direction toward which the wind is blowing

  • Fig. 10.

    Mean synoptic conditions during freezing-rain events and nonevents over the central region of the Great Lakes from 1976 to 1990 (see Fig. 1 for boundaries of the central region): (a) sea level pressure (hPa, solid line) and 1000–666-hPa relative humidity (%, dashed line) during freezing-rain events; (b) 850-hPa geopotential height (m, dark solid line), temperature (°C, solid/dashed lines), temperature anomaly (°C, shaded), and wind (full barb and half-barb denote 5 and 2.5 m s−1, respectively); (c) same as (b) except for 700-hPa geopotential height; and (d) 500-hPa geopotential height (m, dark solid line) and geostrophic wind (full barb and half-barb denote 5 and 2.5 m s−1, respectively) during freezing-rain events. See the text for an explanation of the anomaly calculation

  • Fig. 11.

    (a) Mean (solid lines) and standard deviation (dashed lines) of the combined forcing of vertical motion by differential 850–500-hPa vorticity advection and 850-hPa thermal advection (×10−13 Pa s−1 m−2) during 198 central region freezing-rain reports from 1976 to 1990 (see Fig. 1 for region boundaries); (b) contributions of 850–500-hPa vorticity advection (dashed lines) and 850-hPa thermal advection (solid lines) to vertical velocity (×10−13 Pa s−1 m−2) during the same events used in (a). Positive (negative) values in (a) and (b) are associated with upward (downward) motion

  • Fig. 12.

    (a) Median values of dry-bulb temperature (°C, dark solid line) and dewpoint temperature (°C, dark dashed line) from a distribution of freezing-rain soundings at Flint. Data are plotted on a skew T–logp diagram, with pressure (vertical axis) given in hPa and winds shown by full barb = 5 m s−1 and half barb = 2.5 m s−1. (b) Median absolute deviation of the dry-bulb temperature (°C, dark solid line) and the dewpoint temperature (°C, dark dashed line) at Flint

  • Fig. 13.

    Mean locations of extratropical cyclones associated with 167 western (W), 198 central (C), and 297 eastern (E) freezing-rain reports at 24 and 12 h before the report (−24, −12), during the report (0), and 12 and 24 h after the report (12, 24) between 1976 and 1990. All times are relative to freezing rain reports at 1100 or 2300 UTC

  • Fig. 14.

    The 1950–70 areal distributions of (a) events and (b) relative variability with preferred propagation tracks superimposed for Jan cyclones. Values represent 28-yr totals (Zishka and Smith 1980)

  • Fig. 15.

    Mean sea level pressure (hPa, solid line) and standard deviation (dashed line) during (a) 167 freezing-rain reports over the western region and (b) 297 freezing-rain reports over the eastern region of the Great Lakes from 1976 to 1990 (see Fig. 1 for region boundaries)

  • Fig. 16.

    Monthly distributions of hourly freezing-rain reports at (a) Oscoda and Lansing, and (b) Milwaukee and Madison between 1976 and 1990. The stippled bars indicate lakeshore locations and the solid bars indicate inland locations

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 1523 910 168
PDF Downloads 613 160 17