1. Introduction
Severe winter storms are significant natural hazards. They cause disruptions to transportation, wind damage to buildings, loss of retail sales, closure of schools and businesses, loss of electricity, and hazards to human health causing morbidity and mortality (e.g., Rooney 1967; Babin 1975; Changnon and Changnon 1978; Graff and Strub 1975; Helburn 1982; McKay 1981). Significant impacts from severe winter weather in North America can occur in northern latitudes where snow and subfreezing weather occur for several months each winter (Stommel 1966; Babin 1975; Graff and Strub 1975) and in southern latitudes where snow and extreme cold are rare events (Conner et al. 1973).
The most extreme form of winter storm is the blizzard, combining strong winds and falling or blowing snow to cause low visibility, deep snowdrifts, and extreme wind chill. This combination of strong winds and blowing snow causes a general cessation of routine societal activities making the blizzard the most dangerous of winter storms. Blizzards can cause closures of highways, railroads, and airports, potentially stranding thousands of travelers. This is a hazard to human health and may cause death due to exposure, asphyxiation, or suffocation to those trapped in cars, or lead to death indirectly through heart attacks as people overexert while walking in the blizzard or attempting to clear deep snowdrifts (Changnon and Changnon 1978; Rooney 1967). The northeastern United States blizzard of January 1996 killed 154 in 17 states (Lott et al. 1996); the northern Great Plains blizzard of January 1975 caused 80 deaths (Graff and Strub 1975); the Midwest blizzard of January 1978 killed 73 (Schwartz 2001); 29 died in the Buffalo blizzard of January 1977 (Dewey 1977); and 27 people, mostly stranded motorists, died in the Midwest blizzard of March 1966 (Haines 1966). Strong winds of a severe blizzard may cause damage to buildings, signs, and other structures. Extensive drifting snow on roofs of buildings can lead to collapse of structures due to extreme snow loads, as occurred in the northeastern United States following the 6–8 January 1996 blizzard (DeGaetano et al. 1997). Utility lines may be blown down in a blizzard resulting in loss of electric, telephone, and cable television service. Reduction in retail sales and absenteeism cause financial losses to business and industry (Burrows et al. 1979). A 5-day blizzard in Buffalo, NY, in January 1977 caused losses of $250 million in storm damage, snow removal costs, lost wages, and lost production (Dewey 1977). Schools and government offices are forced to close due to difficulties and dangers of travel and, in some cases, due to loss of electricity and heat. Isolation of families and communities occurs when highway travel and electric and telephone service are curtailed by a blizzard. Those with acute health problems must be transported to a hospital through extraordinary means, such as snowmobile or helicopter, or risk loss of life. Livestock deaths during a blizzard may total in the tens of thousands (Stommel 1966; Changnon and Changnon 1978; Graff and Strub 1975; Steppuhn 1981) and wildlife may be killed through cold or starvation (Graff and Strub 1975). Strong winds and blowing snow during blizzards in mountainous terrain can cause rapid loading of snow on slopes and increase the risk of avalanches (Barry 1992, p. 357; Fredston and Fesler 1994, p. 71).
The term “blizzard” was first used by Henry Ellis along Hudson Bay in 1746 (Black 1971) although Ellis's meaning at the time was ambiguous (Wild 1997). Ludlum (1968) reported that the term was first used in the United States to describe the winter storm of March 1870 in Iowa, however Wild (1997) reported use in South Dakota newspapers in 1867. The term was in general use by the 1880s (Wild 1997). The current definition of a blizzard used by the National Weather Service (NWS) is winds over 16 m s−1 [35 miles per hour (mph)] and falling or blowing snow causing visibility of less than 400 m (0.25 miles) lasting for at least 3 h. An additional criterion for temperatures below 20°F (−7°C) was used for many years but was abandoned in the 1970s (M. L. Branick, 2000, personal communication). There remains some confusion about the definition of a blizzard in the United States. The glossary of the “Encyclopedia of Weather and Climate” (North 1996, p. 855) gives the definition as used by the NWS: “Falling snow over a sustained duration with speeds greater than 15 meters per second (35 miles per hour) and visibility less than 400 meters (0.25 miles) over an extended period.” However, in the same publication on p. 700, Cohen (1996) states: “The blizzard is defined by three criteria: significant accumulation—usually 25 centimeters or more—or blowing of snow that reduces visibility to less then 400 meters (one-quarter mile); sustained winds of at least 56 kilometers (35 miles) per hour; and temperatures of −7°C (20°F) or less” and states that “this last criterion has recently been dropped in some locations.” A recent textbook in physical geography defines a blizzard as a “heavy snowstorm” with winds more than 16 m s−1 and visibility less than 150 m (Scott 1996, p. 194). A blizzard in Canada is defined as wind speed over 40 km h−1 (11 m s−1), wind chill greater than 1600 W m−2, visibility less than 1 km in snow, with a duration of at least 4 h (6 h prior to 1991; Stewart et al. 1995). The Met Office defined a blizzard as wind speed over 56 km h−1 (15 m s−1) causing drifting snow and reduction of visibility to 200 m (Wild 1997).
Recent climatologies of winter storms in the United States include Black's (1971) synoptic climatology of blizzards in the northern Great Plains during 1957–67, and Branick's (1997) study of significant winter weather events during the period 1982–94. Other studies have focused on the meteorology of blizzards (Fox 1952; Salmon and Smith 1980) and other winter storms (e.g., Kocin and Uccellini 1990) or impacts of individual winter storms, as cited above. There has been no modern climatology of blizzards for the conterminous United States. The purpose of this research was to determine the occurrence and location of blizzards in the conterminous United States over recent decades, summarize the data into a modern climatology of blizzards, and analyze the data for time trends, spatial patterns, and relationships with the El Niño–Southern Oscillation (ENSO) teleconnection. The area and population of counties affected by blizzards was also analyzed. The results may have utility in the study of climate change, impacts of natural hazards, emergency management and preparedness, winter weather forecasting, insurance, transportation planning, and the study of concurrent or resultant hazards, such as severe thunderstorms or snowmelt flooding.
2. Methods
A climatology of blizzards could be developed from a review of historical weather conditions at official weather stations and noting the simultaneous occurrence of wind speed over 35 mph, visibility less than one-quarter mile, and the presence of blowing snow. The existence of these conditions for 3 or more hours would constitute a blizzard and these occurrences could be summarized and analyzed for a period of time over the United States. This approach was not taken here because the weather stations recording wind speed and visibility were generally limited to first-order NWS offices that, even prior to NWS modernization and closure of many offices, were located 150–300 km apart. Thus, the spatial resolution of blizzard conditions was poor. In addition, the finest time resolution was hourly and for a portion of the record the finest resolution was every 3 h.
We chose instead to use the description of blizzards in the publication Storm Data. Storm Data is widely used in the climatology of severe weather (Curran et al. 2000; Gaffin and Hotz 2000; Lopez et al. 1995), including severe winter weather (Branick 1997). The Warning Coordination Meteorologist (WCM) at each NWS office submits storm reports for publication in the monthly Storm Data, specifying the type of storm event, counties or regions of a state impacted by the event, date and times of impact, and estimates of damages, injuries, and deaths. Storm Data is available in paper copy since January 1959 and online with an event database since 1993. The quality of data in Storm Data may vary with the WCM preparing the report, the time available to spend investigating each event, and care taken in the written summary. Although there may be some spatial and temporal variability in the quality of the reports in Storm Data (Branick 1997), the publication is an official archive of storm events in the United States.
Monthly issues of Storm Data were scanned over the 41 winters from October 1959 to May 2000 for each storm report of a blizzard or “blizzard conditions” in the 48 conterminous states, as described in the “character of storm” column or within the text of the storm description. We recorded the date of the storm, counties reported to have had blizzard conditions, the estimate of monetary damages, and number of injuries and deaths caused by the blizzard. Reports of “near-blizzard conditions” were not counted as blizzards. Some reports did not list counties affected, but just regions, such as “northwest Iowa” or “western Nebraska.” In those cases, the map of forecast zones was used to assign counties when available, otherwise a consistent and logical method was used to assign the counties affected in the regions specified. The task of the WCM to determine the borders of the area that had a blizzard rather than a severe winter storm that did not meet blizzard criteria is difficult and there may be differences among NWS offices in how precise or carefully these distinctions were made and recorded in Storm Data.
For each blizzard identified in Storm Data, each county affected was recorded and entered into a database. This allowed a tally of blizzards by date and county for the United States and presentation of annual and monthly patterns of blizzards. The area (km2) affected by each blizzard was calculated with ArcView software and using the ArcView Geographic Information System (GIS) database, as compiled from the U.S. Census Bureau. Annual probability of a blizzard was calculated empirically for each county by dividing the number of winters with a blizzard by the number of years in the database (41). The annual number of blizzards, the annual area affected by blizzards, and total population in affected counties each winter were analyzed for time trends with linear regression in the statistical analysis software SPSS. The annual number of blizzards was checked for correlation with ENSO using the Climate Prediction Center database (CPC 1999). ENSO was characterized for the October–December and January–March periods each winter. The warm-phase El Niño events were coded as strong, moderate, or weak (and assigned values of 3, 2, or 1, respectively). The cold-phase La Niña events were coded as strong (−3), moderate (−2), or weak (−1). Neutral events were coded as zero. The population in counties affected by each blizzard was tallied from the U.S. Census Bureau Web site (2001), using 1960 census data for the winters 1959/60 to 1968/69, 1970 census data for the winters 1969/70 to 1978/79, and so on.
3. Results
a. Annual blizzard frequency
There were 438 blizzards identified for the 41 winters 1959/60 to 1999/2000. The annual number of blizzards (by winter) averaged 10.7 and ranged from 1 blizzard in 1980/81 to 27 blizzards in 1996/97 (Fig. 1). A linear regression showed a positive trend in the number of blizzards during 1959–2000 (slope = 0.21 blizzards yr−1, p = 0.001). This yielded a modeled increase from 6.6 blizzards at the beginning of the study period to 15.2 at the end. The standard error of the estimate was 4.7. Whether the increase in the number of blizzards is a true climatological increase or is due to increased reporting efficiency in Storm Data, or both, is not known.
The frequency of blizzards by county over the period 1959–2000 is shown in Fig. 2. Forty of 48 states reported a blizzard during the period. Blizzards were most common in the northern Great Plains where 17 counties in North Dakota and 8 counties in South Dakota had more than 60 blizzards during 1959–2000. Traill County, North Dakota, had 74 blizzards in the 41-yr period, the most of any county in the conterminous United States. A region that we are calling the U.S. “blizzard zone” occupied all of North Dakota and South Dakota and 34 counties in western Minnesota. Each county in the blizzard zone had 41 or more blizzards in the 41 winters, for an average of one or more blizzards per winter.
Outside of the blizzard zone, there were 21–40 blizzards per county during 1959–2000 in the rest of Minnesota, northern and central Iowa, most of Nebraska, northwest Kansas, eastern Colorado, southeast Wyoming, and eastern Idaho. The blizzard frequency decreased to 11–20 per county in the northern and central Rockies, central Great Plains, and western Midwest. Other areas with 11–20 blizzards during the period were in Erie County (Buffalo), New York; Garrett County, Maryland; all of Maine; and portions of southeastern New England. Hancock and Washington counties, Maine, each had 18 blizzards, the largest number in the eastern United States. There were 4–10 blizzards per county during 1959–2000 across the remainder of the northeastern United States, the Great Lakes region, the High Plains of Texas and New Mexico, and the Sierra Nevada Mountains.
The depiction of blizzard occurrences in the mountainous western states is hampered by the large size of counties and the wide variety of topography and weather within large western counties. For example, Los Angeles County, California, had a blizzard on 21–22 December 1996, but the effects were limited to a small area in the less-populated elevations above 1830 m (6000 ft) elevation. The entire county is depicted in our county-level maps (Figs. 2, 4, and 5) and the entire county was used for the calculation of affected population, as described below. Therefore, the area and population affected by blizzards may be overestimated in the western states.
b. Area affected by blizzards
The average area affected by the 438 blizzards was 150 492 km2, about the size of Iowa. The area affected by the blizzards ranged from 1482 km2 in the Cape Cod, Massachusetts, blizzard of 26 January 1987 to 1 054 779 km2 in the 24–27 January 1978 Midwest storm that brought blizzard conditions to 9 states. Although the number of blizzards increased through the period, the total area affected by blizzards each winter (Fig. 3) showed no linear trend through time (slope = 10 105 km2 yr−1, p = 0.45). If these data represent the true number and areas of blizzards, then the average size of blizzards must be getting smaller. It may also be that the NWS is recording smaller, weaker blizzards in recent years that went unrecorded earlier in the period, as occurred also in the official record of tornadoes in the United States (Fujita 1987; Brooks and Doswell 2001).
c. Annual probability of a blizzard by county
Annual probability of a blizzard (Fig. 4) exceeded 50% through all counties of North Dakota and Minnesota, most of South Dakota, western Nebraska, northwest Iowa, and northeastern Colorado. The largest cities in this region are Minneapolis, St. Paul, St. Cloud, Duluth, Sioux Falls, Rapid City, Fargo, Grand Forks, and Greeley. Annual probability of a blizzard was greatest, 63%–76%, across all of North Dakota and adjacent counties in the Red River Valley of northwest Minnesota. The greatest probability, 76%, occurred in six North Dakota counties; Rolette, Towner, and Cavalier along the Canadian border, and Traill, Steele, and Cass along the Red River near Fargo.
The annual probability of a blizzard was 25% (1 yr in 4) to 50% across the remainder of Nebraska and Iowa, southeastern Wisconsin, western Kansas, central and eastern Colorado, eastern Wyoming, portions of Montana, most of Maine, southern New Hampshire, and Garrett County, Maryland. Annual probability of a blizzard occurrence was 12% (1 yr in 8) to 25% in the northern Sierra Nevada Mountains; the remainder of the northern and central Rocky Mountains; central Kansas; most of Wisconsin; in the lake-effect snow regions of Michigan, Ohio, Pennsylvania, and New York; in western Maryland and central Pennsylvania; and the remainder of New England.
There are discontinuities in blizzard probabilities along state boundaries that cannot be easily explained climatologically. Sharp changes in blizzard probability across the Minnesota/Wisconsin border and North Dakota/Montana border suggest historical differences in reporting of blizzards between NWS offices with responsibilities on either side of those borders. The blizzard patterns in Figs. 2 and 4 suggest that blizzards may be underreported in western Wisconsin, eastern Montana, and perhaps in other regions, also. Blizzard frequency may be underreported in this database for the higher elevations of the western mountains, including the Cascades and Wasatch Range.
An application of this blizzard climatology may be seen in the annual probability of a blizzard within a NWS Forecast Office (NWSFO) county warning area. For example, the Grand Forks, North Dakota, NWSFO county warning area (CWA) encompasses 17 counties in eastern North Dakota and 18 counties in northwestern Minnesota. The 35 counties in this CWA reported 99 blizzards during 1959–2000 for an average of 2.4 annually. A blizzard occurred somewhere in the Grand Forks CWA in 35 of the 41 winters in this study, giving an empirical annual probability of 85% for a blizzard occurrence within the CWA. Ten blizzards occurred in the Grand Forks CWA during 1996/97, the greatest of any winter during the study period. These annual probabilities and extremes could be used in training and preparations within the NWS office and in communication of risks to state and county emergency managers.
d. Monthly blizzard frequency
Monthly frequency of blizzards (Fig. 5) peaked during January in the blizzard zone and most other parts of the United States that experienced blizzards. However, blizzards were most frequent during December in the Sierra Nevada; during March in the central High Plains of western Nebraska, eastern Colorado, and Kansas; and during April in Montana. The blizzard frequency in western South Dakota showed a second peak during April, matching the January peak frequency.
There were few blizzards in October and they were limited to the northern and central Great Plains. A peak October frequency of four blizzards occurred in southern Wyoming. Blizzard frequency increased in November with a maximum of 6–9 from eastern Colorado to northern Minnesota. November blizzards in the East were limited to Maine, New Hampshire, western New York, and southern West Virginia where 1–2 occurred. Counties in the blizzard zone in the northern Great Plains had 10–18 December blizzards and blizzard frequency increased across the western and central United States. Blizzards were absent during December in most of the mid-Atlantic states and Appalachians. Blizzard frequency showed a strong peak during January across the Midwest, Great Lakes, and Northeast, but then decreased in those regions during February. Blizzard activity during March shifted from the blizzard zone southward into the central Great Plains where “Colorado low” cyclogenesis also peaks in March (Whittaker and Horn 1981). Much of the East had only one blizzard in March (1993) and blizzards were absent during March in southern Michigan, Indiana, and Illinois. April blizzard frequency was centered in the western Dakotas and was largely absent from the Great Lakes, Northeast (except New England), and the far western mountains. Blizzard conditions were rare during May. No county had more than one May blizzard during 1959–2000 and these were limited to portions of Colorado, Idaho, and Montana.
Hare and Thomas (1979, p. 103) reported that blizzards on the Canadian prairies are most common in February and most likely in southwestern Saskatchewan and eastern Alberta, about 900 km west of the region of greatest frequency shown in this study. Stewart et al. (1995) reported that Canadian prairie blizzards were most common over southwestern Saskatchewan and Wheaton (1998) wrote that “Southern Saskatchewan is the blizzard capital of Canada.” These comments further support our contention that blizzard frequency is underestimated in Storm Data for northeastern Montana.
e. Blizzards and ENSO
The annual number of blizzards and area affected by blizzards was compared to the ENSO teleconnection. Correlations between seasonal ENSO phase and the number of blizzards over the entire winter were weak and marginally significant for the October–December ENSO phase (r = −0.20, p = 0.10) and the January–March ENSO phase (r = −0.24, p = 0.07). The negative correlation indicates blizzards are less likely with El Niña warm phase and more likely with the La Niño cold phase. When the ENSO phase was a moderate or strong El Niño during October–December (n = 9), the mean number of blizzards that winter was 10.3, but when there was a moderate or strong La Niña (n = 7), the mean number of blizzards was 12.7. The average number of blizzards for a strong or moderate El Niño during January–March (n = 8) was 10.0, while the average was 13.6 blizzards when there was a moderate or strong La Niña during January–March (n = 5). Although the relationship is weak, there tends to be 2–3 more blizzards nationally during La Niña than El Niño winters. A regional focus on the blizzards that affected North Dakota or South Dakota showed no significant relationship with October–December (r = −0.11, p = 0.24) or January–March (r = −0.12, p = 0.22) ENSO phases. Smith and O'Brien (2001) found regional patterns and seasonal (early, mid-, and late winter) patterns in the relationships between ENSO and snowfall in the United States. Further study on blizzards may reveal regional patterns with ENSO or other atmospheric teleconnections. The weak relationships found here between blizzards and ENSO should not be used to attribute the occurrence of any particular blizzard to ENSO (see also Barsugli et al. 1999).
f. Population in affected counties
The total population in counties affected by the 438 blizzards during 1959–2000 was 1 078 771 741, an average of 26 311 506 per winter, and 2 462 949 per blizzard (Fig. 6). The smallest affected population was 741 people in the Clark County, Idaho, blizzard of February 1971. The largest number of people affected by a blizzard was 72 864 107 in the March 1993 Superstorm in the East. A linear regression of the population affected each winter showed no significant trend (slope = 168 646 people yr−1, p = 0.63), in spite of an upward trend in the number of blizzards and a 39% increase in United States population from 178 million in 1960 to 247 million in 1990. This irony may be explained by the fact that the population in the blizzard zone of North Dakota, South Dakota, and western Minnesota actually decreased by 30 737 people between 1960 and 1990. So, although U.S. population increased during the period, most of the increase was in the southern and western states where blizzards were rare or did not occur.
4. Summary
This 41-yr county-level climatology of blizzards in the conterminous United States was developed using Storm Data. There were 438 blizzards for an average of 10.7 annually and an increasing linear trend in the number of blizzards. Blizzards annually affect an average of 26 million people over an area of 1.6 million km2. Regional patterns in the annual and monthly frequencies of blizzards and the analyses of areas and populations affected by blizzards allow assessment of risk and vulnerability from this natural hazard. The database could be used for more specialized analyses, such as risk of a blizzard within NWSFO county warning areas, within emergency management districts, within watersheds, along transportation corridors, and so on. Further studies on regional patterns in blizzard frequency and trends and associations with other atmospheric teleconnections will help in understanding these dangerous and costly storms. Efforts should be made to ensure the high quality and consistency of Storm Data to maintain a reliable record of storms and their effects in the United States.
Acknowledgments
Appreciation is extended to the National Weather Service office in Cleveland for use of their archives for early issues of Storm Data and to Patrick Francis for assistance with ArcView. This research was completed with a University Fellowship awarded to Robert Schwartz at Kent State University. Appreciation is extended to Connie McOmber of Cartographic Services, Department of Geography, Ball State University, for assistance with the graphics.
REFERENCES
Babin, G., 1975: Blizzard of 1975 in western Canada. Weatherwise, 28 , 70–75.
Barry, R. G., 1992: Mountain Weather and Climate. 2d ed. Routledge, 402 pp.
Barsugli, J. J., J. S. Whitaker, A. F. Loughe, P. D. Sardeshmukh, and Z. Toth, 1999: The effect of the 1997/98 El Niño on individual large-scale weather events. Bull. Amer. Meteor. Soc., 80 , 1399–1411.
Black, R. E., 1971: A synoptic climatology of blizzards on the north-central Plains of the United States. NOAA Tech. Memo. NWS CR-39, 38 pp.
Branick, M. L., 1997: A climatology of significant winter-type weather events in the contiguous United States, 1982–94. Wea. Forecasting, 12 , 193–199.
Brooks, H., and C. A. Doswell III, 2001: Some aspects of the international climatology of tornadoes by damage classification. Atmos. Res., 56 , 191–201.
Burrows, W. R., R. A. Treidl, and R. G. Lawford, 1979: The southern Ontario blizzard of January 26 and 27, 1978. Atmos.–Ocean, 17 , 306–320.
Changnon, S. A. Jr,, and D. Changnon, 1978: Winter storms and the record-breaking winter in Illinois. Weatherwise, 31 , 218–225.
Cohen, J., 1996: Snowstorms. Encyclopedia of Weather and Climate, S. H. Schneider, Ed., Vol. 2, Oxford University Press, 700–703.
Conner, W. C., K. C. Crawford, and E. L. Hill, 1973: Snow forecasting in the South: Occasional blizzards over the southern Plains to rare Gulf Coast snows. Weatherwise, 26 , 244–248.
CPC, cited 1999: Cold and warm episodes by season. [Available online at http://www.cpc.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.html.].
Curran, E. B., R. L. Holle, and R. E. Lopez, 2000: Lightning casualties and damages in the United States from 1959 to 1994. J. Climate, 13 , 3448–3464.
DeGaetano, A. T., T. S. Schmidlin, and D. S. Wilks, 1997: Evaluation of East Coast snow loads following January 1996 storms. J. Perform. Constr. Facil., 11 , 90–94.
Dewey, K. F., 1977: Lake-effect snowstorms and the record breaking 1976–77 snowfall to the lee of Lakes Erie and Ontario. Weatherwise, 30 , 228–231.
Fox, R. L., 1952: Blizzards of the Northern Plains. Weatherwise, 5 , 123–126. 132.
Fredston, J., and D. Fesler, 1994: Snow Sense: A Guide to Evaluating Snow Avalanche Hazard. 4th ed. Alaska Mountain Safety Center, 116 pp.
Fujita, T. T., 1987: U.S. tornadoes, part 1, 70-year statistics. Satellite and Mesometeorology Research Project Res. Paper 218, University of Chicago, 122 pp.
Gaffin, D. M., and D. G. Hotz, 2000: A precipitation and flood climatology with synoptic features of heavy rainfall across the southern Appalachian Mountains. Natl. Wea. Dig., 24 , 3–15.
Graff, J. V., and J. H. Strub, 1975: The Great Upper Plains blizzard of January 1975. Weatherwise, 28 , 66–69. 83.
Haines, D. A., 1966: The late March blizzard of 1966 in the Middle West. Weatherwise, 19 , 194–197.
Hare, F. K., and M. K. Thomas, 1979: Climate Canada. 2d ed. John Wiley and Sons, 230 pp.
Helburn, N., 1982: Severe winter storms as natural hazards. Great Plains–Rocky Mountain Geogr. J., 10 , 86–95.
Kocin, P. J., and L. W. Uccellini, 1990: Snowstorms along the Northeastern Coast of the United States: 1955 to 1985. Meteor. Monogr., No. 44, Amer. Meteor. Soc., 280 pp.
Lopez, R. E., R. L. Holle, and T. A. Heitkamp, 1995: Lightning casualties and property damage in Colorado from 1950 to 1991 based on Storm Data. Wea. Forecasting, 10 , 114–126.
Lott, N., T. Ross, and M. Sittel, 1996: The winter of '95–'96: A season of extremes. National Climatic Data Center, Tech. Rep. 96-02, Asheville, NC, 15 pp.
Ludlum, D. M., 1968: Early American Winters II 1821–1870. Historical Monogr., Amer. Meteor. Soc., 257 pp.
McKay, G. A., 1981: Snow and living things. Handbook of Snow: Principles, Practices, Management, and Use, D. M. Gray and D. H. Male, Eds., Pergamon Press, 3–31.
North, G. R., Ed.,. 1996: Glossary. Encyclopedia of Weather and Climate, S. H. Schneider, Ed., Vol. 2, Oxford University Press, 853–874.
Rooney, J. F. Jr,, 1967: The urban snow hazard in the United States: An appraisal of disruption. Geogr. Rev., 57 , 538–559.
Salmon, E. M., and P. J. Smith, 1980: A synoptic analysis of the 25–26 January 1978 blizzard cyclone in the central United States. Bull. Amer. Meteor. Soc., 61 , 453–460.
Schwartz, R. M., 2001: Geography of blizzards in the conterminous United States, 1959–2000. Ph.D. dissertation, Kent State University, Kent, OH, 176 pp.
Scott, R. C., 1996: Introduction to Physical Geography. West Publishing, 560 pp.
Smith, S. R., and J. J. O'Brien, 2001: Regional snowfall distributions associated with ENSO: Implications for seasonal forecasting. Bull. Amer. Meteor. Soc., 82 , 1179–1191.
Steppuhn, H., 1981: Snow and agriculture. Handbook of Snow: Principles, Practices, Management, and Use, D. M. Gray and D. H. Male, Eds., Pergamon Press, 60–125.
Stewart, R. E., and Coauthors. 1995: Winter storms over Canada. Atmos.–Ocean, 33 , 223–247.
Stommel, H. G., 1966: The Great Blizzard of '66 on the Northern Great Plains. Weatherwise, 19 , 189–193. 207.
U.S. Census Bureau, cited 2001: County population estimates. [Available online at http://www.census.gov/population/www/estimates/countypop.html.].
Wheaton, E., 1998: But It's a Dry Cold. Fifth House, 185 pp.
Whittaker, L. M., and L. H. Horn, 1981: Geographical and seasonal distribution of North American cyclogenesis, 1958–1977. Mon. Wea. Rev., 109 , 2312–2322.
Wild, R., 1997: Historical review on the origin and definition of the word blizzard. J. Meteor., 22 , 331–340.
Number of blizzards identified each winter.
Citation: Journal of Climate 15, 13; 10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2
Number of blizzards recorded by county during the 41 winters 1959/60 to 1999/2000.
Citation: Journal of Climate 15, 13; 10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2
Total area (km2) affected by blizzards each winter.
Citation: Journal of Climate 15, 13; 10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2
Annual probability of a blizzard by county, based on the frequency of blizzards during 1959/60 to 1999/2000.
Citation: Journal of Climate 15, 13; 10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2
Number of blizzards recorded by county during the 41 winters 1959/60 to 1999/2000 for the months of (a) Oct, (b) Nov, (c) Dec, (d) Jan, (e) Feb, (f) Mar, (g) Apr, and (h) May
Citation: Journal of Climate 15, 13; 10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2
Population in counties affected by blizzards each winter.
Citation: Journal of Climate 15, 13; 10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2
