n March 1888, an epic snowstorm disrupted nor- mal life in much of th - PDF document

I n March 1888, an epic snowstorm disrupted nor- mal life in much of the Northeastern United States. The Blizzard of 1888 dumped more than 3 ft (91 cm) of snow in much of the Hudson Valley and Connecticut, while winds as high as 36 m s created drifts of more than 20 ft (6 m); (Kocin 1983; Cable 1988). One hundred and five years later, another | ones„will aid those who wish to create such a scale. Furthermore, an awareness of such factors benefits all affected by snowstorms, especially those charged with warning government and the general public. This article will introduce the various factors that influence the impact of snowstorms and illustrate their significance. While numerous factors account for variation in snowstorm impacts, most can be clas-sified under four metafactors: meteorological varia-tions, governmental response, actions of the general public, and meteorologists and the media. A sum-mary table grouping the factors into major or minor categories will also be presented. Finally, this article will conclude with a call to action for both forecasters and the community at large. Ultimately, meteorolo-gists and broader society should revise the concept of snowstorms into one of snow events, a richer term that reminds us that not only are meteorological fac-tors responsible for variations in snowstorm impacts, but numerous human-created factors also play a role. In other words, while the Blizzards of 1888 and 1993 were both significant snowstorms, they were dramati-cally different snow events.METHODOLOGY.Snowstorms that occurred in four major cities in Upstate New York„Buffalo, Rochester, Syracuse, and Albany„between 1888 and 2003 were studied to evaluate the significance of factors influencing snow events. Figure 1 shows the locations of the four cities within New York State. These cities were selected because of their similar ABLE 1. Dates, amounts (in.), and a qualitative assessment of the disruption for the case study storms for Buffalo. Note that disruption„a mea-sure of the snow event„is not correlated with amount. Question mark indicates a lack of suffi-cient data.Date(s) of stormAmountDisruption21–22 Jan 190217.4Low?17 Mar 193619.0High14–16 Dec 194536.0Low28–29 Nov 195519.9Medium29-30 Dec 196124.5Low30 Nov–2 Dec 197639.8Low28–30 Jan 1977*Extreme30 Nov–Dec 197922.4Medium10–11 Jan 198228.8Low27–29 Feb 198428.3Medium19–21 Jan 198533.2High9–10 Dec 199541.2Medium18–20 Nov 200038.9High24 Dec 2001–1 Jan 200281.6High*Amount of accumulation unkown. ABLE 2. Dates, amounts (in.), and a qualitative assessment of the disruption for the case study storms for Rochester. Note that disruption„a measure of the snow event„is not correlated with amount.Date(s) of stormAmountDisruption28 Feb–Mar 1900*43.5Medium11–12 Dec 194421.5High15–19 Feb 195830.2Medium14 Feb 196018.4Medium19–20 Feb 196021.6Medium23 Jan 196618.2High30–31 Jan 196626.7High5 Dec 197710.8**High6–7 Feb 197825.0Low8–10 Dec 198125.1None28 Feb–1 Mar 198429.0Medium11–12 Mar 199221.9Low13–14 Mar 199323.2Low3 Jan 199623.0Low12–15 Jan 199929.2Medium4 Mar 199922.3High6 Mar 199918.4Medium*Two-day storm; no 29 Feb 1900.**Another 13.1 in. fell 6–9 Dec 1977.. 1. Locations of the case study cites within New York State. 1784DECEMBER 2005 size and attributes, such as climate, elevation, and economy. Another reason these cities were selected was to examine differences between lake-effect versus synoptic-scale snow events. Largely due to variations in lake-effect snow, average annual snow-fall ranges from 63 in. (160 cm) in Albany to 120 in. (305 cm) in Syracuse. Prior to 1888, these cities had little involvement in snow mitigation and gener-ally waited for warmer weather to alleviate snow problems. Thus, the first case study in the sample was the Blizzard of 1888. However, because sparse news coverage and weak governmental response hampered efforts to study storms early in the sam-pling period, most of the cited examples are from the 1930s through 2003.For each city, a list of the largest snowstorms within the study period was compiled; details of this process are in the appendix. The 10 largest storms and selected storms ranked 11…20 were then selected as the case studies. Lower-ranked storms were chosen either to fill temporal gaps in data or because they oc-curred very close in time to a top 10Ž storm. In total, 59 case studies, for an average of nearly 15 per city, were studied. Complete lists of case study storms are shown in Tables 1…4. Because the largest snowstorms for each city were determined strictly by snowfall for that city and not in consideration of whether other cities received significant accumulation, comparisons ABLE 3. Dates, amounts (in.), and a qualitative assessment of the disruption for the case study storms for Syracuse. Note that disruption„a measure of the snow event„in not correlated with the amount. Question mark indicates a lack of suf-ficient data.Date(s) of stormAmountDisruption6–8 Mar 193218*Medium8–9 Feb 195825.3Medium16–17 Feb 195829.2 High30 Jan–1 Feb 196642.3High5–10 Dec 197724.2Medium?28 Feb–3 Mar 198430.9Low15–17 Dec 198925.2*Low14–21 Jan 199238.6Low11–15 Mar 199231.7Low13–14 Mar 199342.0Medium4-9 Jan 199442.2High30–31 Dec 199725.9Medium*Based on newspaper reports. ABLE 4. Dates, amounts (in.), and a qualitative assessment of the disruption for the case study storms for Albany. Note that disruption„a measure of the snow event„in not correlated with the amount. Question mark indicates a lack of suf-ficient data.Date(s) of stormAmountDisruption11–14 Mar 188846.7High22–25 Feb 189318.2Low?14 Feb 191423.5Medium18–20 Jan 193617.9Low8–12 Mar 194117.8Low8–9 Feb 195821.1Medium15–16 Feb 195817.9Medium24–25 Dec 196618.3Low22 Dec 196912.3Medium25–28 Dec 196926.4High24–25 Nov 197122.5None15–16 Jan 198324.5Low13–14 Mar 199326.6Medium25 Dec 200221.0Medium3–4 Jan 200320.8Medium6–7 Dec 200318.0Lowof how a single storm differentially impacted multiple cities were rarely done.For each case study, newspaper accounts from two days before the storm began until news cover-age ended were read; most storms disappeared from news coverage within a week after the last flakes fell. To learn more about governmental response to snow, the author interviewed Commissioners of SnowŽ in Buffalo, Rochester, and Syracuse, and examined budgets and expense reports for all four cities. Finally, local broadcast meteorologists were asked to respond to a set of interview questions. This was done to get a sense of both their involvement in snow events and their beliefs about the influence of the meteorology community.METEOROLOGICAL VARIATIONS. Meteoro-logical variations are perhaps the most obvious cause for differences in snow events (see Changnon 1969; Kocin and Uccellini 2004a). This is largely because meteorological parameters such as total snow accu-mulation are widely available and easily understood. While the total amount of snow is important, varia-tions in other parameters of a snowstorm, such as snowfall rate (intensity), snow density, air tempera-AMERICAN METEOROLOGICAL SOCIETY ture, wind, and duration can have just as much, if not a greater, effect on the associated snow event. Indeed, the term blizzard implies both windy (�15 m s) and snowy conditions leading to a reduction in visibility (0.4 km; Branick 1997). Timing of a storm is also important, although this is mainly because of varia-tions in human factors, such as traffic volume.For Buffalo, Rochester, and Syracuse, intensity is often more important than total snowfall amount. Because these cities individually average more than 90 in. (229 cm) of snow per season, snow mitigation operations are efficient, members of the general public are experienced in dealing with snow, and fore-casters are skilled at snow prediction. Thus, a large but unexceptional amount of snow may not cause much disruption of the normal routine. However, if the snow falls with extraordinary intensity, normal life can quickly grind to a halt, even if for just a few hours. When 25 in. (63 cm) of snow fell on Gridlock Monday 2000Ž (20 November), thousands of Buffalo commuters and students were stranded downtown in what the Buffalo News (21 November 2000) termed the worst storm since . . . 1977.Ž In Rochester, an intense (and largely underforecast) storm depos-ited more than 22 in. (56 cm) of snow on Thursday 4 March 1999. Local governmental authorities closed all interstate highways because of accidents, causing massive gridlock. Finally, when snow fell at a rate exceeding 4 in. (13 cm) h on Tuesday 4 January 1994 (the Syracuse SnowburstŽ), Syracuse area schools and offices dismissed students and workers early, resulting in widespread gridlock.An equally significant meteorological factor is when a snowstorm occurs, although timing is impor-tant primarily because of variations in traffic volume. Timing as an attribute influencing snow events can be further subdivided into different scales. On the seasonal scale, early or late season snowstorms may lead to less significant snow events because of rela-tively warm ground, higher sun angles, and longer day length; all of these reduce snow accumulations on roadways. Similarly, snowstorms that occur on holidays, weekends, or during school breaks are of-ten lesser snow events than what might be expected based on the meteorological data. For example, the December 2001 SuperstormŽ dumped more than 80 in. (203 cm) of snow at the Buffalo airport. Because it occurred between Christmas and New Years Day„when schools were closed and many workers were on vacation„disruption for most residents of the City of Buffalo and its suburbs was arguably less than that of the Gridlock Monday 2000 storm. Likewise, the Blizzard of 1993 (studied in Rochester, Syracuse, and Albany) had a relatively small impact compared to other record-setting snowstorms, such as the Blizzard of 1977 in Buffalo or the Blizzard of 1996 in Syracuse, simply because it occurred on a weekend. Even when a snowstorm occurs during the hectic shopping week-ends before Christmas (examples include Buffalo in 1995 and Syracuse in 1989), the general public largely chooses to stay home rather than venture out. Conversely, all three aforementioned gridlock storms occurred on weekdays. Additionally, the most intense snow fell between 1000 and 1400 LT. Thus, in contrast to popular opinion, storms that are intense at midday are often more disruptive than those that strike at rush hour. Midday storms cause a super…rush hour with schools and workplaces being dismissed simultaneously. Comparatively, storms that peak in the overnight or early morning hours may compel workers and students to remain at home; for storms occurring late in the day, students are already home or in the process of returning home, and worker dismissals are staggered. In response to the Syracuse Snowburst 1994 storm, the City of Syracuse and downtown businesses created a nonbinding plan to dismiss workers on a gradual basis, with the aim of preventing a super…rush hour and widespread grid-lock (J. Wright 2003, personal communication).If two snowstorms occur close together and the period between them is cold, the second storm is often a more disruptive snow event. In February 1958, two synoptic-scale storms affected Syracuse and Albany in just over a weeks time. Although the first storm dumped more snow on Syracuse„21 in. (53 cm) versus 17 in. (43 cm) with the second„the second storm was more disruptive. In Albany, the second storm was more disruptive than expected when compared with similar storms in 1941 and 1966. Snow remaining from the first storm and equipment failure largely explain why a second storm is a more disruptive event. When two 15-in. (38 cm) lake-effect snowstorms affected Buffalo in 1985, more than half of the citys snow mitigation equipment was out of service the day after the second storm (Buffalo News22 January 1985). Even if no major snowstorms occur, a prolonged period of snow and cold can have increas-Early or late-season snowstorms may be problematic because of the potential for downed power lines and unprepared motorists. However, few reported problems with downed trees and lines were observed in this study, perhaps because of the focus on cities or maybe due to a lack of case-study storms before late November. As for unprepared motorists, their disruptive effect was also negligible, even when a storm was the first of the season. 1786DECEMBER 2005 ingly negative effects. A record 78.1 in. (198 cm) of snow fell in Syracuse during January 2004. Two fatal accidents, on two separate days, occurred late in the month when drivers skidded across a long bridge on Interstate 81, hit packed snow along the guardrail, flew off the highway, and fell 15 m to the ground below. Prior to the accidents, measurable snow had fallen every day for more than 2 weeks, and although plow drivers for New York State had been salting and plowing daily, they had not had time to physically re-move the snow along the guardrail. Similar incidents, thankfully with less deadly results, occurred in Roch-ester following a similar period in January 1999.Air temperature is another meteorological factor that affects snow events. Neither cold nor warm tem-peratures are necessarily worse; instead, each presents a different set of complications for snow mitigation crews. At temperatures below …9°C, road salt is largely ineffective (Moran et al. 1992). However, the drier and thus less dense snow at these temperatures is easier for plow crews to push and cars to drive over (P. Noto 2003, personal communication). While tem-peratures near 0°C increase salts effectiveness, snow has a higher water content and greater density. This heavier and more clumpy snow puts a greater strain on snow mitigation equipment„assuming the equip-ment is able to break through the slush. More than 161 km of Rochester streets were left impassible when seven contractors failed to clear their routes after a storm in February/March 1984. Several interviewed by the newspaper blamed the unusually dense snow for causing equipment failure and their breach of contract. One contactor termed the snow cast ironŽ Rochester Democrat and Chonicle,1 March 1984), an appropriate metaphor for the way dense snow fuses together when cold weather follows a storm. Thus, if a city is slow to respond to a storm with dense snow, removal can become impossible. The March 1936 St. Patricks DayŽ storm dumped 19 in. (48 cm) of snow and sleet on Buffalo; the precipitation mixture had an usually low snow-to-water ratio of ~7:1. Snow removal crews were called late and found frozen slush that was then impossible to clear, and the city was paralyzed for a week.Wind is another meteorological factor that affects the snow events, although its significance is less than might be expected. Within the cities studied, there is little open space for snow to drift. Additionally, such open space is essentially constant; thus, mitigation operations have been adjusted to give drift-prone areas extra attention if needed [P. Noto 2003, personal communication; J. Wright 2003, personal communi-cation; Albany (NY) Bureau of Streets Annual Report 1986]. Nonetheless, exceptionally strong winds, coupled with light snow, can cause widespread drift-ing, and that is a major problem. Such was the case in Buffalo with both the Blizzard of 1977, when winds gusted to 31 m s, and after the January 1985 storm, when winds gusted to 22 m sTo summarize, meteorological variations in snow-storms can greatly affect their impact and thus cause dramatically different snow events. However, snow events are also affected by many nonmeteorological factors, such as governmental response, and problems in this area can create major snow events out of minor snowstorms.GOVERNMENTAL RESPONSE. Although me-teorological variations in snowstorms„especially in-tensity„are important in understanding differences in snow events, social factors are equally important. As shown already, timing is significant largely be-cause of social reasons, such as variations in traffic volume or holiday periods. Behavior of the govern-ment is another important social factor. Smart plan-ning, competent workers, and reliable contractors aid government and minimize disruption, but failure in any of these areas can create a snow disaster.Effective snow mitigation begins with good leader-ship at the highest level of local government. Thus, when political fights spill into snow mitigation opera-tions, minor snowstorms quickly become major snow events. Among the cities studied, Buffalo has experi-enced the most politically enhanced snow events. A budget dispute between Mayor George Zimmerman and Buffalo Common Council was already simmer-ing when the St. Patricks Day 1936 storm hit the city. Because the snow budget was exhausted, Zimmerman did not order crews to begin plowing until the eve-ning, well after the snow was already settling into a dense, slushy mess. Council blamed the lack of funds on fiscal mismanagement by the mayor, while the mayor-appointed public works director accused the council of failing to provide sufficient funds for equipment (Buffalo News, 18 March 1936). Travel throughout the city remained at a standstill for nearly a week until crews from adjacent Erie County and melting from warmer temperatures finally cleared the snow. In 1985, Mayor James JimmyŽ Griffin and Buffalo Common Council were already fighting over who should direct snow removal operations when a severe storm hit. Lacking a snow mitigation plan, and hampered by equipment failure, workers did an uneven and generally poor job of snow mitigation, causing citizens to flood council with complaints, including discriminatory plowing (Buffalo NewsAMERICAN METEOROLOGICAL SOCIETY 26 January 1985). Despite these complaints and similar ones following a snowstorm in January 1984, Mayor Griffin was reelected three times. Although there have been some examples to the contrary, such as Mayor John Lindsey in New York City in 1969 (see McKelvey 1995), problems with snow mitiga-tion are not always fatal to political careers„even in Buffalo.Regardless of political leadership, without compe-tent workers and the proper equipment, snow mitiga-tion operations will fail. Nonetheless, it appears that the workers are rarely a problem„both newspapers and Commissioners of Snow consistently praise them as being dedicated and capable. For example, Joseph N. Giambra, Commissioner of Snow for Buffalo, termed his employees superŽ and noted that theyll work 16 hours for you and not hesitate about coming in the next morningŽ (2003, personal communica-tion). Additionally, the size of the workforce is almost always sufficient. One notable exception occurred in Rochester in December 1944, when a labor short-age due to World War II exacerbated snow removal difficulties initially caused by poor leadership (the workers were called late). Although most cities since then have employed enough workers, having a sufficient quantity of working snow removal equip-ment is another story. In all four cities studied, old equipment was used well beyond its life expectancy, and breakdowns hampered operations. Examples of this include snowstorms in Buffalo 1936, 1945, and 1985; Albany in 1958; and Syracuse in 1958 and 1966. However, thanks to improvements in snow removal technology, equipment problems today are less com-mon and less severe than those in the past.In all four cities, contractors are an important component of the snow mitigation force. Use of con-tractors allows cities to save money on equipment purchases and labor, yet they can still increase their workforce when an extraordinary storm occurs. Unfortunately, contractors sometimes fail to appear or fail to use adequate equipment for clearing heavy, compacted snow. Historically, Rochester has relied on contractors more than the other cities, and in both 1977 and 1984 no-show contractors slowed recov-ery after major storms. In these cases, the problem was exacerbated by the citys lack of trucks small enough to squeeze down narrow residential streets. Consequently, Rochester today has a larger number of small trucks than in the past, and contractors are required to bring their equipment to the Public Works Garage for inspection annually (P. Noto 2003, personal communication). More generally, all of the cities studied are less reliant on contractors today compared to the past (based on city budgets; P. Noto 2003, personal communication; J. Wright 2003, personal communication).Finally, outside aid„in the form of Disaster Declarations and National Guard assistance„is more common today than in the past. While such assistance is rarely available until after a snow event is well underway, the extra labor and governmental funds help people in cities resume normal activities more quickly.ACTIONS OF THE GENERAL PUBLIC.Equally significant to governmental response are the actions of the general public. Most important are car-related actions, both in terms of parking and driving. Parked cars impede plowing operations, while cars in driving lanes„especially if stalled, abandoned, or gridlocked„get in the way. More generally, indi-vidual decisions by members of the general public to remain at home or venture out can collectively affect a snow event.The greatest problem that cities face in dealing with snow is parked cars. One commissioner claimed that parked cars were his citys biggest nemesis,Ž while another twice referred to them as a nightmareŽ (P. Noto 2003, personal communication; J. Wright 2003, personal communication). Unlike adjacent suburbs, which almost always ban nighttime parking during the snow season, the cities studied are ham-pered by old neighborhoods with a shortage of off-street parking. At best, cars parked on the street slow down plows and restrict curb-to-curb clearing opera-tions. At worst, they can block plows from entering narrow streets. Problems are not limited to residential areas, either; illegal downtown parking created many problems encountered in the mid…twentieth century (see storms from Rochester, 1966; Syracuse, 1958; and Albany, 1958). Because of parking problems in the past, the cities studied have instituted large fines, strong enforcement, and consistent parking regula-tions. For example, Syracuse drivers may park on only one side of the street based on whether the date is odd or even; the actual weather is irrelevant. As a result of these parking policy changes, problems with parked cars are less substantial today, although they are still significant.Motorists also disrupt snow removal efforts. Whether cars are moving, idling, stalled, or aban-doned, they slow traffic flow„including snow plows. In worst-case scenarios, traffic can become completely gridlocked (e.g., Buffalo Gridlock Monday 2000 storm). Nonetheless, traffic today seems to have a less visible effect than it did in the past. For example, commuters return to work following major storms in the 1950s and 1960s caused massive traffic jams in every city for every storm studied during this time. Return-to-work traffic jams do not occur today in the case-study cities. The cause for this change is unclear; however, the author believes it is probably a combination of population declines, job losses, and improvements in parking regulation.Finally, the general public influences snow events simply by choosing whether or not to remain at home. People staying home during the Blizzard of 1993 have been credited with assisting mitigation efforts Albany Times-Union, 15 March 1993). On the other hand, police blamed violators of a driving ban for slowing down cleanup and removal operations fol-lowing the January 1985 storm in Buffalo (Buffalo News, 23 January 1985), and sightseers from the sub-urbs hindered operations following the 1969 storms in Albany (Albany Times-Union, 30 December 1969). Finally, looters caused an extra burden by distracting police (who often are needed to ticket illegally parked cars prior to towing) following storms in Syracuse (1966) and Buffalo (1977 and 1985).METEOROLOGISTS AND THE MEDIA.Although meteorologists and the media do not con-trol the weather, drive snowplows, or have a large collective effect as do members of the general public, they do have an indirect influence on snow events by altering the behavior of the government and general public. Meteorologists influence behavior by issuing forecasts and warnings; the media influence behavior through their coverage of storm preparations, the storm itself, and storm aftermath. It should be noted that since many meteorologists are employed by media companies their influences are intertwined; furthermore, additional media that do not employ or contract meteorologists also disseminate weather information.The power of meteorologists to affect behavior has grown with time. For the case study storms that occurred before approximately 1980, it was rare to see an newspaper article about the storm before it actu-ally began. Since then, however, newspaper articles warning of an impending storm or discussing storm preparations have become commonplace. Poststorm evaluations of governmental response and meteoro-logical forecasts are not uncommon either.Meteorologists affect governmental response by providing warnings in advance of a snowstorm. When a well-forecast storm strikes a well-prepared city, the streets may have been presalted, plows are gassed up and ready to go, and city workers are on the job. An early example of this was a lake-effect storm that struck Buffalo in December 1961; the magnitude of the snow event was minimal. The Buffalo Newscredited Weather Service, Inc., with notifying the Streets Division early enough for adequate prepara-tion (30 December 1961). Similarly, good forecasts greatly aided snow mitigation operations in Syracuse during and after the Blizzard of 1993„and this was despite one of the largest snow totals on record (J. Wright 2003, personal communication).Meteorologists also influence actions of the gen-eral public. Again, the Blizzard of 1993 is a prime example of a storm that had a lesser impact because of a well-prepared general public (Albany Times-Union15 March 1993). In Rochester, the February/March 1984 storm, despite occurring during the middle of the workweek, had a relatively minor impact initially because of the general publics preparation (although problems with no-show contractors ultimately made this a very disruptive storm).Conversely, if meteorologists forecasts are incor-rect, ire descends from the public. For example, an early December 1977 storm in Syracuse was much less significant than predicted, prompting the edito-rial board of the Syracuse Post-Standard to blame misleadingŽ weather reports and fearsomeŽ film clips for the unnecessary closing of 50 local school districts (10 December 1977). On the other end of the spectrum, when a forecasted minor storm became a major storm on Thursday 4 March 1999, a massive traffic jam resulted in Rochester.Corresponding to the rise in meteorologists influ-ence is an increase in the general publics knowledge and understanding of meteorology. All five meteo-rologists interviewed agreed with this statement and cited reasons such as advances in technology (e.g., Doppler radar) and easier access to information through sources such as The Weather Channel® and World Wide Web. However, they cautioned that members of the general public have increasingly high expectations due to promotion and use of such tech-nology. Furthermore, most broadcast meteorologists interviewed by the author complained of insufficient on-air time for a complex discussion of weather pos-sibilities and uncertainties. Some speculated that communication problems may stem from closer to home due to an in-house disconnect between them and their respective news departments.The media also reflect and influence public opinion. Newspapers use both editorials and news articles to illustrate their belief in the ability, or ineptitude, of government. If government fails to properly mitigate snow, news coverage takes a nega-AMERICAN METEOROLOGICAL SOCIETY tive tone (some representative case studies include Buffalo 1936 and Rochester 1944). In Albany, persis-tent problems with snow removal in late 1969 turned news coverage from relatively neutral to antagonistic. News headlines best illustrate this change: while initial headlines expressed awe at the magnitude of the snow problem, such as Wow, We Are Snowed!Ž Albany Times-Union, 23 December 1969), later head-lines expressed increasing frustration, such as Area Remains Paralyzed: Another Storm is on the WayŽ (ibid., 29 December 1969).The combined influence of meteorologists and the media is most evident in a phenomenon termed by this author as the mad rush. Mad rushes occur when members of the general public mob grocery and hard-ware stores in search of the essentials, such as bread, milk, toilet paper, and snow shovels, in advance of an impending snowstorm. In this study, the first mad rushes were observed in the 1990s, suggesting that better forecasts and media dissemination are at least partially responsible for such behavior. The mad rush also varies spatially. Although eight case study storms affected Buffalo and Rochester in the 1990s and 2000s, none of them triggered a mad rush. However, mad rushes were common in Albany during this time; mad rushes have also been observed elsewhere in the Northeast corridor and also in connection with hurricanes. Ironically, mad-rush behavior is generally unnecessary in regard to snowstorms. With two ex-ceptions„both in Buffalo„no recent storm has para-lyzed any portion of any city studied for more than few days. Although vulnerable groups of the general public, such as the elderly or residents of rural areas, could be stuck in place for several days, it is unlikely that most urban and suburban residents could face such a predicament from all but the worst of storms. Unanswered questions regarding the spatial variation in the mad rush outside of Upstate New York and more general ques-tions about why this behavior occurs suggest a need for further research. Finally, one reviewer of this paper suggested that the mad rush may actually benefit cities by lessening the need for people to travel during a snowstorm. Future researchers may also want to examine this supposition.CONCLUSIONS. Thinking of snowstorms„meteorologi-cal happenings„in the larger context of snow events„social creations„ allows us to better understand why similar snowstorms often have dramatically differ-ent impacts on society. Tables 5 and 6 summarize the factors that affect snow events; Table 5 shows factors that directly influence a snow event, while Table 6 shows indirect influences. All factors shown in Table 5 are grouped underneath the meta-fac-tors of meteorological variations, governmental response, or actions of the general public. Factors in Table 5 are further separated by whether variations in them generally have a major or minor influence in affecting the magnitude of a snow events dis-ruption. Several factors not previously discussed, such as the clearing of fire hydrants and sidewalks, generally had little effect on snow events in the case study cities (Call 2004), however, these may have more significance in other cities or noncity areas. Indeed, budget-related problems can be especially significant for smaller municipalities, such as towns (ibid., p. 35). Table 6 includes factors relating to the metafactor of meteorologists and the media, as well as several factors that are either relatively constant, such as a citys climate, or difficult to measure, such as residents level of experience. Because of the in-direct nature of the factors in this table, no further grouping was attempted. While climate and city layout and terrain have a small influence, at least for the case study cities (Call 2004), future researchers may wish to determine the importance and relative influence of the other factors shown in Table 6.Two important limitations of this study are its regional focus and its emphasis on cities. Rooney (1967), for example, found significant regional dif-ferences between snow events in the Upper Midwest and those in the western Great Plains, and this study found significant differences even among four similar cities. For example, Buffalo has had the most politi- ABLE. 5 Factors directly affecting snow events in the cities studied.MeteorologicalvariationsGovernmentalresponseActions of thegeneral publicMajor Amount of snow Intensity of snow Timing Temperature (snow density) Equipment Preparation Use of Contractors Politics Parked cars Motorists Choosing to stay home, or notUsually Wind Duration Labor Outside aid Budget Adherence to driving bans Criminal behavior Clearing of hydrants and sidewalks cally enhanced events, while Rochester has had the most problems with contractors. Nonetheless, the author believes that the metafactors introduced here could be applied to cities elsewhere in the United States. An awareness of the nonmeteorological fac-tors affecting the impact of snowstorms could be especially valuable in regions less accustomed to snow, because the actions of the government, general public, meteorologists, and media can significantly reduce the disruption of snow events. Finally, this study focused on cities, and caution should be used when applying the findings to suburban or rural ar-eas. Although the author believes that the factors are transferable, the significance of them changes greatly because of dramatic differences in the suburban and rural landscape. For example, suburban areas near the case study cities have overnight on-street parking bans, a lower street density, large drainage ditches, and other differences that greatly aid plows and lessen problems with cars (Rochester Democrat and Chronicle, 3 March 1984). In rural areas, wind causes significant drifting, and residents can be isolated for long periods, behooving them to purchase extra sup-plies before a storm.To end, this study has shown the importance of considering nonmeteorological factors as well as meteorological ones when studying snow events. Whether preparing forecasts and warnings for clients or the general public, forecasters should consider the role of both social and meteorological factors and adjust the tone of their pronouncements accordingly. Ultimately, improvements in the tone and emphasis of forecasts will increase the ability of the meteorologi-cal community to provide valuable information for the government and the general public.ACKNOWLEDGMENTS. I wish to thank Anne Mosher, Mark Monmonier, and Susan Millar for their guid-ance and suggestions during this study. Without the honest and insightful responses of Tom Atkins, Steve Caporizzo, Tom DiVecchio, Don Paul, and Wayne Mahar, the discus-sion of meteorologists influence on snow events would not have been possible. Similar thanks are in order to the Commissioners of Snow: Joseph N. Giambra, Paul Noto, and Jeff Wright. Thanks also to Steve McLaughlin for his assistance in defining severe snowstorms and perspective on NWS procedures. Finally, appreciation is extended to all six reviewers of this article„their comments greatly strengthened my argument.APPENDIX: CREATING THE LISTS OF CASE STUDY STORMS. Creating the lists of case study storms was a surprisingly problematic task. Although snow events are determined by much more than snow amounts, using daily snow data to develop the lists was easier and faster than more complex methods [such as those used Branick(1997)]. For Albany, the local National Weather Service Forecast Office maintains a list of the largest snowstorms from 1888 through the present on its Web site; this list was sufficient for this study. For the other three cities, lists of storms had to be constructed. Some information was available for Buffalo and Rochester, such as a list of the greatest 24-h snowfalls since 1980 and a list of the greatest daily snowfalls since 1943, respectively. Although these provided a useful starting point, be-cause long duration but less intense storms did not appear on the lists, the daily snow record also had to be examined. For Syracuse, no list of any type was available. Thus, for Buffalo, Rochester, and Syracuse, simple objective criteria defining a severe snowstorm had to be created; then these had to be applied to daily snow accumulation data to create lists of severe snowstorms.Unfortunately, no official criteria define severe snowstorms. Although 6 in. (15.2 cm) of snow is generally the minimum criterion for heavy snow in the Northern United States (Changnon 1967; Branick 1997), that amount is fairly common in the case study region. More problematically, snow falls nearly daily in lake-effect snow regions during winter, and it is difficult to tell if snowfall from consecutive days is from one or multiple snowstorms. Thus, a working definition was created as follows: at least 1 in. (2.5 cm) of snow per day with a minimum of 10 in. (25.4 cm) in total. Because no list of any type was available for Syracuse, this initial definition was used to create the list of severe snowstorms shown in Table 3.After studying these storms, it was discovered that most long-duration storms with relatively low amounts per day (such as 14…22 January 1992) had almost no impact, probably because the snow inten-sity never strained the citys capacity. As a result, ABLE 6. Factors indirectly affecting snow events in the cities studied.Meteorologists and the mediaAdditional factors Accuracy of forecasts Lead time Tone of coverage Amount of coverage Climate City layout and terrain Credibility of government, meteorologists, and media  Experience with past eventsAMERICAN METEOROLOGICAL SOCIETY the definition of a severe snowstorm„at least by Buffalo, Rochester, and Syracuse standards„was revised to include only those storms with at least 3 in. (8 cm) of snow per day and at least 14 in. (36 cm) in total accumulation.The higher amount of snow per day was needed to prevent long-duration low-impact snowstorms (generally lake-effect episodes) from being considered severe. The requirement for total accumulation was made to keep the lists of potential case study storms manageable; as shown in Tables 1…4, virtually every case study storm featured at least 18 in. (46 cm) of accumulation. Exceptions were either studied before the definition was revised or were secondary storms that occurred close in time to top 10Ž storms.Finally, the revised definition was applied to the Syracuse climate data and the new list of severe snowstorms was compared with the original list (see Table A1). Not surprisingly, several storms dropped far in the rankings, while a few did not meet the new criteria. Although some of these storms were insig-nificant, the 1994 Syracuse Snowburst and second February 1958 storm would not have been studied if the revised definition had been used initially. The revised rankings belie the major snow events associ-ated with these storms.In conclusion, because the revised definition favors shorter but more intense storms, the author believes it is a superior way to determine severe snow-storms. This is because intense snowstorms, at least in the case study cities, are often more disruptive than less intense storms with greater snow accumulations. Nonetheless, as shown by the rank order changes of Syracuse storms, significant snow events may be excluded from study by this definition. It is likely that some major snow events associated with low-ac-cumulation storms in the other cities may have been missed in this study. The challenge of determining a severe snowstorm by using a parameter such as total snow accumulation, or even snow intensity, under-scores a central point of this article: the disruption caused by a snow event is based on much more than meteorological parameters.REFERENCESAlbany (NY) Bureau of Streets Annual Report, 1986: Proceedings of the Common Council: Vol. 2: Message of the Mayor and Reports of City Officers. V-B Print-ing Company. [Available from Albany Public Library, 161 Washington Ave., Albany, NY 12210.] ABLE A1. Largest storms, ranked from highest to lowest accumulation, that affected Syracuse based on the original definition of a severe snowstorm as having at least 1 in. (2.5 cm) of snow per day and 10 in. (25.4 cm) in total accumulation. Revised dates, amounts, and rank based on defining a severe snowstorm as having at least 3 in. (7.6 cm) per day with at least 14 in. (36 cm) in total. Numbers in italics indicate storms that were studied but would not have been studied under the revised definition. All amounts are inches.MonthYear Original date(s)Reviseddate(s)OriginalamountRevisedamountOriginalrankRevisedrankJan–Feb196630–130–142.342.311Jan19944–94–542.216.4227Mar199313–1413–1442.042.032Jan199214–211838.619.8414Mar199211–1511–1431.730.154Feb–Mar198428–328–3 30.930.963Feb195815–2016–1729.216.9725Jan192529–3029–3027.527.585Feb 19587–108–525.321.1910 Dec198915–1715–1725.2*25.2*106Dec199730–313–3124.824.8117Dec19775–10—24.2„12„Dec19914–64–524.024.013.58Dec200023–2826–2724.015.513.529* Amount based on newspaper reports. Branick, M. L., 1997: A climatology of significant win-ter-type weather events in the contiguous United States, 1982…94. Wea. Forecasting 12, 193…207.Cable, M., 1988: The Blizzard of 88. Atheneum, 197 pp.Call, D. A. 2004: Urban snow events in Upstate New York: An integrated human and physical geographi-cal analysis. M.A. thesis, Dept. of Geography, Syra-cuse University, 120 pp.Changnon, S. A., Jr., 1969: Climatology of severe winter storms in Illinois. Illinois State Water Survey BulletinNo. 53,45 pp.Hart, R. E., and R. H. Grumm, 2001: Using normal-ized climatological anomalies to rank synoptic-scale events objectively. Mon. Wea. Rev.129,2426…2442.Kocin, P. J., 1983: An analysis of the Blizzard of 88.Ž Bull. Amer. Meteor. Soc64, 1258…1272.„„, and L. W. Uccellini, 2004a: Northeast Snow-storms. Meteor. Monogr., No. 54,Amer. Meteor. Soc., 818 pp.„„, and „„, 2004b: A snowfall impact scale derived from Northeast storm snowfall distributions. Bull. Amer. Meteor. Soc.85, 177…194.McKelvey, B., 1995: Snow in the Cities: A History of Americas Urban Response. University of Rochester Press, 248 pp.Moran, V. M., L. A. Abron, and L. W. Weinberger, 1992: A comparison of conventional and alternative deic-ers: An environmental impact perspective. Chemical Deicers and the Environment, F. M. DItri, Ed., Lewis Publishers, Inc., 341…361.Rooney, J. F., Jr., 1967: The urban snow hazard in the United States: An appraisal of disruption. Geogr. Rev.57, 538…559.Uccellini, L. W., P. J. Kocin, R. S. Schneider, P. M. Sto-kols, and R. A. Dorr, 1995: Forecasting the 12…14 March 1993 superstorm. Bull. Amer. Meteor. Soc.76, 183…199.Zielinski, G. A., 2002: A classification scheme for winter storms in the eastern and central United States with an emphasis on Noreasters. Bull. Amer. Meteor. Soc.83, 37…51.AMERICAN METEOROLOGICAL SOCIETY |