Some Thoughts on the Societal and Economic
Impact of Winter Storms

Paul J. Kocin
National Centers for Environmental Prediction
NOAA/National Weather Service
Camp Springs, MD

The National Weather Service produces summaries of the deaths, injuries and damages attributed to a variety of weather events. From the figures provided (Figure 1), (Figure 2), it would appear that winter storms, including both snowstorms and icestorms, have much less an impact on society than such phenomena as convective storms (tornadoes, severe thunderstorms, hail; 100 to 200 deaths annually and damages of 1 to 2 billion dollars), heat waves (1000 fatalities in 1995), tropical cyclones (20 to 30 billion dollars for Andrew, for example), floods (21 billion dollars estimated during the Mississippi floods of 1993), drought (4.5 billion dollars in 1988), and cold (1.9 billion dollars in 1989). During the period from 1988 thru 1995, the contribution from winter storms to the total damage estimate of $114 billion was a little over $3 billion, less than 3% of the total. In contrast, the damage estimates from just 2 events, 1992's Hurricane Andrew and the Mississippi River flooding of 1993 yielded about half the total. During this same 8-year period, 4101 deaths and 26,441 injuries were attributed to weather-related causes. Winter storms accounted for 372 of the deaths and 5690 of the injuries. These numbers work out to about 9% of all weather-related deaths and nearly 20% of the injuries, with the bulk of the injuries associated with icestorms. These figures would seem to indicate that winter storms, with emphasis on snowstorms, appear to have a relatively small effect on the economy and society, especially when compared to other weather threats such as hurricanes, floods and tornadoes.

However, events during recent years indicate that the impact of a major snowstorm on society could be more significant than the above figures imply. For example, the March 1993 Superstorm and the January 1996 Blizzard of 1996 were 2 events that impacted the daily routine of 50 to 100 million people over the course of several days, and in some instances, for more than a week. These 2 events were predicted with relative accuracy, and emergency managers, private industry, the media and the general population seemed to respond to the warnings in a multitude of ways. In the March 1993 Superstorm, newspaper "reports" showed damage from $1 billion to as much as $6 billion and some 200 to 300 deaths. In the January 1996 blizzard, the NWS disaster survey report just released cited insurance losses over $500 million and fatalities, primarily attributed to overexertion, at 60. The economic costs were provided by the American Insurance Association, primarily for wind and water damage. What wasn't included in these numbers were public cleanup costs and the economic impact of lost business. While these numbers were not included, the likelihood is that these numbers could be very high. According to the Liscio Report, a private financial newsletter that documents national economic trends, the impact of just 1 storm, such as the March 1993 Superstorm or the January 1996 blizzard, could be critical since payroll surveys, part of the Bureau and Labor statistics that reflect the state of the national economy, can be influenced significantly by snowstorms. According to the January 9, 1996 report, "Weather is a significant and all but neglected factor in economic performance". The newsletter cites the example that during the cold and snowy winter of 1977/1978, the GDP was moving at a 6% to 7% rate through the first 3 quarters of 1977 and the last 3 quarters of 1978. The index slowed to -1% in the last quarter of 1977 and less than 3% in the first quarter of 1978 during a winter that was characterized by several major snow events in heavily populated regions of the nation. The newsletter states that winter storms will impact retail sales and housing activity but the impact on payrolls may be particularly influential as many people are temporarily put out of work by a major snow event. While it is not clear exactly what these impacts are, it is clear that there are significant impacts that are not consistently measured. These effects may have national and global economic implications.

Thus, the economic effects of winter storms may be more indirect than the damage-driven effects of weather phenomena such as hurricanes, flooding, tornadoes and hail. Perhaps, even the physical effects of snowstorms may also be somewhat minimized. Death tolls from other phenomena tend to be concentrated and obvious (for example, from physical trauma or drowning), such as in hurricanes and tornado outbreaks. Traffic accidents and deaths and injuries that result from attempts to clear snow, overexertion, the combination of cold temperatures and strong winds can be anecdotal and occur in many different areas, unlike the concentrated nature of other weather events. It also raises a number of other questions not only on the impacts of such storms and how we measure them, but how do forecasts alter the impact? What would have been the effects of the March Superstorm and the Blizzard of 1996 if there were no forewarning? What impact does a relatively accurate forecast several days in advance have over a forecast of say 1 day in advance. What would be the impact if there were no forecasts?

The meteorological community has seen tremendous improvements in recent years in its ability to predict cyclogenesis several days in advance with significant gains in numerical weather prediction techniques. In recent years, certain winter storms that had once been barely predictable a day in advance are now being forecast with some degree of accuracy as much as 5 to 7 days in advance and with confidence 1 to 2 days in advance. As a result, the March 1993 and January 1996 snowstorms, 2 of the most significant winter storm events to affect the most heavily populated region of the nation during the present decade, were predicted with a remarkable degree of success. Both storms resulted in loss of life and societal and economic impact that were very significant. How significant? How do we measure the societal and economic benefits such forecasts may have entailed? We know how warnings were issued to keep the population off the roads, retailers were able to provide some supplies that would otherwise be unavailable, and airlines moved their jets to locations where planes could fly. On the other side of the coin, not every storm is predicted with such a high degree of skill. For example, on 17 March 1997, weather forecast models provided information indicating that the New York City area could receive a major snowstorm in the following 48 hours, prompting the issuance of a "Winter Storm Watch". However, the later nighttime model simulations had a considerably different character, with New York to receive no snow, which subsequently verified. While it is humbling to see that such forecast dilimmas can still exist, it still raises questions about what kinds of costs are involved when forecasts are wrong, what savings occur when they're correct, and how society benefits from forecasts of winter storms. Given the wrenching budgetary times when it is probably critical to come up with hard numbers of how such improvements may have mitigated some of the impact of these storms seems to be of paramount importance.

Given that winter storms may have more impact on the population than the Weather Service numbers indicate, there must be a more objective means of quntifying that impact. Therefore, it is not simple to address the issue of economic and societal trends directly, but winter storm trends in the Northeast United States can shed some light on the impact of winter storms on a part of the nation that is home to 1/4 to 1/3 of the nation's population. In Fig. 2, mean seasonal snowfall is depicted during the 20th century for an ensemble of approximately 30 sites shown in Fig. 3. This figure shows the widely varying seasonal nature of snowfall along the Northeast urban corridor. These figures can be viewed as a measure of the periodic nature of winter storm activity. To summarize some general findings, it appears that the 1990's represent a period of great fluctuations with some very snowfree seasons competing with some of the snowiest seasons of the 20th century, in particular, 1992-93, 1993-94 and 1995-96, which was the snowiest season of the 20th century. The 1990's are characterized by a significant number of extreme winter weather events that had considerable societal and economic impact. There may have been an increase in the number of intense coastal cyclones, including the Halloween Gale of 1991, the coastal storm of January 1992, the December 1992 "Storm of the Century", the March 1993 "Superstorm", the onslaught of numerous ice and snowstorms during the winter of 1993-94, the February 1995 "Nor'easter", the January 1996 Blizzard of 96 and a host of lesser events during the course of the winter of 1995-96. While the 1990's have produced 3 winters at least as severe as any in the 20th century, it could be kept in perspective that the snowy winter of 1995/96 was surrounded by the winters of 1994-95 and 1996-97, 2 seasons that were relatively gentle with respect to winter storms. This period contrasts with much of the 1970's and 1980's, which saw a large number of relatively snowfree seasons, mild winters, but a few significant exceptions, in particular, the late 1970's from 1977 thru 1979, 1982-83 and 1986-87. This period of subdued activity followed the most active period of seasonal snowfall this century, the period from the late 1950's through the late 1960's. To put the last half of the 20th century in perspective, several trends with Northeast snowfall become obvious:

The 1950's began as a period of minimal snowfall, having small impact from 1950 through the middle of the decade. Starting the late 1950's, lasting through the 1960's and extending into 1970-71 in portions of northern New England, the Northeast experienced perhaps the most extended period of winter storm and snowfall effects this century with numerous major winter storms. The period of snowy winters ended in the early 1970's and with some exceptions continued through the early 1990's. The major exceptions were the late 1970's and a couple winters during the 1980's, when a few major events occurred. Since the early 1990's, the winters in the Northeast have been characterized by more frequent and debilitating winter storm events and some particularly debilitating winters, interspersed with some remarkably quiet seasons, as well. The limitation of examining seasonal snowfall is that it only tells part of the story with respect to the impact of winter weather. The effects of bitter cold, high winds and freezing rain are also part of the winter storm problem, and it isn't clear how to measure the costs of these phenomena when comparing them to snowstorms.

The apparent trends in seasonal snow, which is probably similar in nature to other regions of the United States, points to cyclical nature of severe winter weather and how variable the impact can be from one season to the next. During the 1996-97 season, the East was quiet, for the most part, and much of the winter weather activity shifted to the West, the Pacific Northwest, the Northern Plains and the Great Lakes, which experienced above-normal winter weather activity, with all the attendant economic and social disruption. The problem may not be that these cyclical periods keep occuring but more that we may not have a reliable indicator of what the real costs of these phenomena are to society. These costs appear to be real and their impact is substantial. How we measure these costs is a topic open to debate.

Societal Aspects of Weather

Workshop's Main Page

Table of Contents