 Storm Clouds gathering over the Coastal Plain Photo By The National Park Service |
Hurricanes |
The long term trends in the climate considered in the previous section, particularly the seasonal precipitation change, may have a major impact on people, but they do not consciously notice them or their impact. Extreme and unusual weather events are another matter. They impose themselves on life, creating at the very least some inconvenience. Throughout recorded history North Carolina has been visited by hurricanes, winter storms, tornados, floods, and similar violent events, along with more silent killers such as heat waves and droughts. The last few years have been no exception. There have been, as Figure 2 suggests, a range of spectacular events, and indeed it seems that the weather has been unusually tempestuous during this period. Is this really the case? What can it suggest about future weather and climate? Estimating the number of events which have occurred over the past century is difficult. Observations have improved over time, while the number of people in the state to see, record and be affected by them has been continually increasing. So events seem to have become more frequent as the 20th century progressed. This has been particularly the case for local events such as flash floods or tornados, but even for major features like hurricanes it is necessary to make some assumptions and estimates about their frequency. Nevertheless, the main trends are clear and they can be used to put recent series of phenomena into the long-term context of climatic variability and change.
Hurricanes have always been a major, highly visible, component of the weather and climate of North Carolina, but their number has differed dramatically from decade to decade. Table 5 includes all hurricanes and hurricane remnants which appear to have had an impact somewhere in the state. (The category is expressed as the Simpson-Saffir scale, which is shown in the Technical Appendix-Table B). The actual numbers, especially early in the 20th century, include some estimates. Before satellite tracking arrived in the 1960s, for example, it was not clear whether a period of coastal erosion or flooding was the result of an unseen and unreported hurricane offshore or of a frontal system passing off the coast. The general results, however, are likely to be reliable. A rather small number of storms influenced the state in the 1911-1930 period, and they were not very intense. Then there was a slow increase, culminating in the 1950s with a whole series of storms, Hazel being especially notable as the only storm of the century to be a category 4 storm while over the state. Then came another quiet period from 1961 until 1980. In fact the present active period did not really get going until the late 1980s, especially with Hugo in 1989.
| Table 5. The number and intensity of hurricanes affecting North Carolina by decade in the 20th Century | ||||||||||
| Category | 1901-1910 | 1911-1920 | 1921-1930 | 1931-1940 | 1941-1950 | 1951-1960 | 1961-1970 | 1971-1980 | 1981-1990 | 1991-2000 |
| 4 | -- | -- | -- | -- | -- | 1 | -- | -- | -- | -- |
| 3 | 2 | -- | -- | 1 | 1 | 4 | -- | -- | 2 | 2 |
| 2 | -- | -- | -- | 2 | -- | 2 | -- | -- | 1 | 3 |
| 1 | 6 | 3 | 5 | 3 | 5 | 2 | 4 | 1 | 1 | 2 |
| Tropical | -- | 3 | -- | 1 | 3 | 3 | 3 | 6 | 4 | 5 |
| Total | 8 | 6 | 5 | 7 | 9 | 12 | 7 | 7 | 8 | 12 |
| *The decadal frequency and intensity of hurricanes influencing North Carolina. The numbers for tropical storms (TS) are estimated in the earlier years. | ||||||||||
None of the hurricanes in the active period from Hugo onwards have been unprecedented in their size, or in the amount of wind and rain they bring. Many have given around 10 inches of rain in some small areas, and up to 6 inches over a much broader region. The locations affected, and the type of impact, depend strongly on the track of the eye of the storm. Hurricane Isabel is a typical example (Figure 10). The storm came ashore near Cape Lookout in the early afternoon of September 18, 2003. It arrived as a category 2 storm, but decreased to a category 1 over land as the eye tracked north-northwest. The storm left the state over Northampton County before 8:00 p.m. that evening. Much of the Piedmont received somewhat less than 4 inches of rain from the storm, the Coastal Plain getting between 4 inches and 7 inches. Some wind-related damage and river-based flooding ensued, but greater problems arose from the interaction of wind and water. As with most hurricanes, the strongest winds were to the east of the eye, being above 90 mph from the southeast over most of the Tidewater region, compared to less than 70 mph from a northerly direction over the Piedmont. The strong winds over the Tidewater pushed the waters of the sounds northwards, causing severe flooding on the north shores.
Isabel is but the latest - as of this writing - in the series of similar hurricanes of the last decade or so (as indicated in Figure 2.19 in The North Carolina Atlas). Slightly different tracks or configurations give different, sometimes very different, impacts. Bonnie, at the end of August 1998, was another storm arriving as a category 2 hurricane. This storm crossed Cape Fear and gave 12 inches of rain on the southeast coast, 6 inches over much of the Coastal Plain. Actual damage away from the coastline was rather light. More recently, tropical storm Allison, on June 15-16, 2001, gave between 3 and 6 inches of rain throughout the Coastal Plain, with a maximum in the northeast of over 10 inches. There was some flooded rivers and swamps in the area of maximum rainfall, but little impact elsewhere. Much more significant was the one-two punch of Dennis (August 28-September 8) and Floyd (September 7 - 19) in 1999. Dennis, slow-moving and wavering off-shore before a direct pass across the state from Cape Lookout to the northwest Piedmont (Figure 2), gave 15 inches of rain on the Outer Banks, decreasing to 6 inches in the northern Piedmont. Floyd's track from the Brunswick beaches to the Virginia border at Currituck produced 15 inches over the inner Coastal Plain, slightly less to the right of its track. Coming so close together, they gave a tremendous total rainfall. The consequences were the major and devastating floods over most of the Coastal Plain. But the situation was not unique, since sequences of two or even three hurricanes in quick succession, often giving much flooding, have commonly occurred in the past. But the long quiescent period of the second half of the 20th century seems to have put them into the distant past, while the development of the state during this quiet period has intensified the likely impact of these hurricanes when they do occur.
Back To TopThe floods associated with the tremendous amounts of rainfall from hurricanes Dennis and Floyd were truly unusual events, with widespread flooding over almost all of the Coastal Plain. The previous major flood was in 1964 (Figure 11) and even that was confined mainly to the southern mountains and the Neuse river basin. The last time most of the Coastal Plain was inundated was back in 1945, when it was also back-to-back hurricanes which were responsible for the rain that created the floods. Indeed, it is usual for major floods to be created by two events in sequence. At least one is usually a hurricane. As a result, it is not coincidence that the frequency of major flooding reflects the distribution of hurricane frequency. And these widespread floods usually occur in the late summer or Fall. The only ones not involving hurricanes were the 1928 flood, which affected the mountains and southern Piedmont and was the result of a very wet period soaking the ground followed by a slow-moving, vigorous frontal storm, and the flood of 1901. This latter, which affected the Catawba and the valleys farther west, was caused by a pair of what were then called squall lines, having the characteristics of what we would now identify as weather fronts. Hurricanes do not seem to have been involved.
Most floods, of course, are much less spectacular and widespread than those associated with Dennis and Floyd, although they can be equally destructive in the area affected. Many are flash floods created when a summer thunderstorm pours a great deal of water onto a single river basin in a short period of time (see figure 2.17 in The North Carolina Atlas). Also common, especially in the mountains, are winter flash floods that occur when intense rain falls on frozen ground. In general, it is extremely difficult to determine whether these are increasing or decreasing in frequency. They are only recorded when someone sees them, or when a river-flow gauge is operating. So, population changes influence the reports and one would perhaps expect an increasing number of floods simply because there are now more people to see them. In addition, development in the basins should, theoretically, increase the tendency to flood, while flood control structures have been built to minimize the impact of sudden downpours of rain. As a result, the actual causes of any trends that might occur are not clear. Nevertheless, from time to time there are certainly sets of flash floods similar to those which occurred in the mountains in January, 1998 (Figure 2). Soils at high elevations were frozen, those lower down were saturated with water. A slow-moving cyclonic storm put about 4 inches of rain over much of the area, with almost 15 inches falling in isolated high elevations. This created what would have been a fairly common set of floods for the mountains in winter. Then a line of thunderstorms came through, adding even more water, so that several places in Avery, Mitchell and Yancey counties reported that a wall of water rushed downstream. High water levels and the force of the water created much destruction in the valleys. Even away from the valley floors, winds associated with the storms uprooted many trees from the sodden ground, increasing the damage. The result was, for many areas, a storm which was the "worst since 1977".
Back To TopWinter storms are commonly associated with snow, ice and freezing rain. Memorable ones usually involve a mixture of all three. Simple snowstorms are most frequent in the mountains, and several are expected each year, with snow lying for several days in the valleys, several weeks on the peaks. But snow can occur anywhere. One memorable event was the storm of 22-24 December, 1989, which left more than 15 inches of snow on the ground on Christmas Day in Wilmington, and left parts of the Outer Banks with drifts several feet deep. This was the greatest snowfall on record in Wilmington, the 12 inches in February, 1973 being the only other time that values exceeded 10 inches there in the 20th century. Even in the higher parts of the Piedmont amounts over 10 inches occur only once every few years or so. In most cases, storms which produce this amount of snow are highly disruptive for travel, and cause many closings of businesses. But rarely are they more than a nuisance, while many people see them as an opportunity to enjoy winter to the full.
But when ice rather than, or in conjunction with, snow, is involved, the story is very different. Most encounters with ice come early on a winter morning when overnight cooling causes freezing conditions which turn what would have been a light dew into a sheet of ice - with the increased cooling likely on bridges making them the most susceptible sites. But the true ice storms are the real problem. These involve freezing rain. The temperature at which rain lands on the surface depends on the atmospheric temperatures above the surface. Most of the time the atmosphere is above freezing and it simply rains. Sometimes it is much colder and snow develops. Occasionally it is in between and freezing rain occurs. Then a raindrop falling through a cold atmosphere is cooled below the freezing point but is still liquid when it arrives at the ground. As soon as this super-cooled rain drop hits an obstacle, be that a tree limb, a power line, or the ground surface, it will freeze. So ice can rapidly accumulate on the obstacle, bending or breaking limbs and lines and coating the surface with ice. To get the freezing rain, the atmospheric temperatures have to be just right (or just wrong). A bit warmer and rain falls, a bit colder and its snow. Usually in any ice storm all three occur. The events of December 4, 2002 were a fine example.
For a couple of days prior to the event cold, Arctic air sat over the state. Then, late on December 4th , warmer, moist air streamed in from the west, rode up over the cold air and created the precipitation. The type of precipitation was arranged in a series of bands oriented more or less parallel to the mountains (Figures 12a & 12b). The mountains and western Piedmont had snow. Moving southeastward, the amount of freezing rain increased, reaching a maximum just north of the line between Charlotte and Raleigh. Farther southeast rain became the main feature. As the whole weather system evolved and moved eastward, one precipitation type often replaced another. For most of the Piedmont, for example, the initial snow gave way to freezing rain. This allowed great ice accumulations on wires and limbs, and the weight brought many crashing down. The result, typical of an ice storm in modern times, was difficulty of movement over roads, much damage, especially roof damage, to individual properties and a lack of electricity. In this case well over one million customers were affected, some being without power for over a week.
A similar event occurred during January 25-27, 2004. A blast of bitter cold air from the north spread over most of the central part of the state. Meanwhile, warmer, humid air from the southwest overode it, producing a mixture of snow, sleet, and freezing rain. In many areas of the Piedmont and western foothills it covered roads with several inches of glaze, resulting in numerous accidents and at least seven fatalities. The low-level cold air could not rise and escape westward over the mountains, a phenomenon called "cold air damming". This meant that the ice and sub-freezing temperatures persisted for several days, causing dozens of school systems to close for three days.
North Carolina's tornados are not able to compete with those of the central USA in frequency or intensity, Rarely exceeding F2 on the Fujita scale (shown in Technical Appendix-Table A). But sometimes they are features of weather which have major impacts. Most tornados are associated with isolated thunderstorms, and they tend to occur as small, relatively weak and isolated events. So their impact is usually minimal and local, with the amount of damage or injury caused depending greatly on exactly where the event occurs. So, in general, there are between 10 and 20 tornado days per year, most days having one or perhaps two tornados (Table 6). At irregular intervals there are more severe outbreaks, with several tornados spawned by a single weather system, and with more widespread impacts. Hurricane Floyd, for example, created 20 tornados along the coast on September 15, 1999.
| Table 6. Summary of annual tornado occurence in North Carolina between 1998 and 2002 | |||||
| Category | Year | ||||
| F5 | 1998 | 1999 | 2000 | 2001 | 2002 |
| F4 | 1 | -- | -- | -- | -- |
| F3 | 2 | -- | -- | -- | -- |
| F2 | 6 | 9 | -- | -- | 1 |
| F1 | 27 | 8 | 1 | 2 | 4 |
| F0 | 30 | 21 | 23 | 11 | 6 |
| Water Spout | 14 | 3 | 1 | 3 | -- |
| Other | 4 | 8 | 2 | -- | 4 |
| Total | 84 | 49 | 27 | 16 | 15 |
| Deaths | 2 | 1 | 0 | 0 | 0 |
| Injuries | 53 | 38 | 1 | 3 | 0 |
| Property Damage($M) | 92.732 | 13.672 | 0.930 | 0.316 | 3.081 |
| Crop Damage($) | 75000 | 0 | 0 | 0 | 0 |
| # of Days | 29 | 9 | 16 | 10 | 9 |
| *Other = dust devils & funnel clouds | |||||
The two major outbreaks of the last few years occurred during 1998. Of the 29 days with tornados that year, 20 of them had a single event. None of these had an intensity greater than F1, most did less than $25,000 in damage and only one produced any injuries. On the other hand, 10 storms occurred on March 20 when a large, well-organized line of thunderstorms drifted northeastward across the northern Piedmont. Even then, only one, an F3 storm, did serious damage as it traveled 12 miles from southwest of Mayodan to just northwest of Eden in Rockingham County. Approximately 600 residences and numerous businesses were destroyed or damaged, including one industrial building where damage was estimated at $25M. Two deaths and numerous injuries were also caused. The second and larger outbreak occurred a couple of months later, this time associated with a series of thunderstorms generated in eastern Tennessee. They brought hail, high winds and rain on a swath from Madison County through Forsyth County, with at least 20 individual tornadoes involved. The severest was an F4 in rural Caldwell county, which caused two injuries but did relatively little damage. In contrast, an F3 storm went through a housing subdivision near Clemmons, Forsyth County, destroying several houses and damaging hundreds more, giving property damage totaling over $50M and at least five injuries.
No part of the state is immune from tornados, although the main events of 1998 affected only parts of the mountains and the Piedmont. However, the Coastal Plain, as Figure 2.18 in The North Carolina Atlas indicates, is the major tornado location. That region was certainly the location of the largest single outbreak on record. And this was an outbreak worthy of the mid-west (Figure 13). The system started in South Carolina, entered the state over Union County around 5 p.m. on the afternoon of March 28, 1984. At 9:20 p.m. that night, it died away as it moved over Chowan and Perquimans Counties into Virginia. In North Carolina, 25 tornados between them did over $325 million dollars worth of damage, killed 40 citizens and injured 400 others. In terms of human life, this was one of the greatest natural disasters the state has ever known.
Everybody knows that heat waves are extended periods when temperature and humidity are high throughout the day and it doesn't cool down much at night. Beyond that, there is no definition, meteorological or legal, of how extended the periods have to be, or how hot and humid it has to get before a heat wave is said to have occurred. So it is difficult to indicate when one actually occurs, let alone realistically examine trends. Nevertheless, the National Weather Service has some criteria for issuing health-related heat stress warnings that can be used as a guide. These warnings use the Heat Index, a combination of temperature and humidity measurements which approximately represent how hot it actually feels - assuming that one is lightly clothed and doing only light work in the shade. The Index is actually expressed as a temperature. Two consecutive days when the daytime values exceeds 105 and the night does not dip below 80 are the criteria, the heat wave actually starting on the second day. Raleigh met such criteria on 22-24 July, 1998. Not since 1952 had there been such an intense and prolonged hot and humid spell. In the intervening years air conditioning has become much more common and this helped many people survive without too much stress. But it is suspected that the sudden change between indoor and outside conditions may have increased the physical stress for many people. There were numerous cases of health problems, including a few fatalities directly associated with the heat.
There are very few stations in North Carolina with long-term records of both temperature and humidity, so that it is difficult to judge the frequency of heat waves or even how big an area they cover when the do occur. The 1998 heat wave was centered on Raleigh, and probably extended across most of the Coastal Plain. It certainly reached the Marine Corps Air Station at Cherry Point, inland of Morehead City towards New Bern, the only other long-term observing station in the area. It does not seem to have reached the coastline itself. Indeed, the shorter record at Wilmington suggests that the almost perpetual breezes blowing off the ocean cool things off somewhat.
Also, the 1998 event did not go westward as far as Greensboro or Charlotte. Both of these cities are at a higher altitude than Raleigh, and slightly cooler. They were both certainly hot and humid, but not sufficiently hot and humid to meet the National Weather Service heat wave criteria. From the rather scanty records available, it appears that the Coastal Plain, represented by the Cherry Point observations, is the area most prone to heat waves. During the 1950s through 1970s there was little difference between the Coastal Plain and the Piedmont in frequency, but in the last two decades of the 20th century they have become much more prevalent on the Coastal Plain (Figure 14). This may be the result of a recent, and very minor, temperature or humidity increase there, since the National Weather Service criteria seem to put North Carolina right on the border between having a heat wave or not. Small differences in conditions then have a seemingly great consequence in the number of heat waves. Rarely do Raleigh, Greensboro and Charlotte record heat waves at the same time. Sometimes it seems that extra warmth caused by Charlotte's more southerly position is most important, at other times its location at a relatively high altitude seems to be the key. In general, the Piedmont in the 1950s through the '70s had one or perhaps two heat waves per decade, and the Coastal Plain not many more. In the 1980s the number increased substantially, particularly on the Coastal Plain. Now Cherry Point seems to dominate, with Raleigh a clear second. The only station in the mountains having long records of temperature and humidity is in Asheville. No heat wave as defined here has ever been recorded there.
Drought may be, rather obviously, the opposite of floods, but it is much more difficult to define and detect. It is something which seems to creep up on people and they hardly recognize it until it is already well established. There is no real nation-wide definition since drought depends not only on the average rainfall expected in any area, but also the common year-to-year rainfall variability in the area. Further, whether or not a drought occurs depends on whether one is worried about agriculture, construction, energy generation, water supply or forest-fire danger. Consequently, several ways of measuring drought have been developed, none of them completely satisfactory for all regions or all impacts. One measure for which there is a long, detailed record, is the Palmer Drought Severity Index (PDSI, shown in Technical Appendix-Table C). Although nationally this emphases agricultural drought, in North Carolina it is a good general measure for most impacts. This Index indicates that drought was virtually absent during the first decade of the 20th century, to be followed by a slow increase into the 1920s (Figure 15). That decade was much drier than the following one, the Dust Bowl years which devastated the mid-section of the nation. The droughts of the 1930's may have been less severe here, but nevertheless both decades had significant impacts on a rural, agricultural state, particularly since there was no Federal safety net for the farmers affected. Frequent droughts continued throughout the '40s and '50s and then almost vanished for the 1960s and '70s. The 1980s again had droughts while the '90s were wet. There was, however, another trend during those 100 years. Early in the period the mountains were much less susceptible to drought than the rest of the state. By the end they had become the most likely place to get drought.
The recent major drought (Figure 2) was most marked in the mountains. A long period of below normal rainfall began there in the summer of 1998. This expanded to include parts of the Piedmont the following winter, and parts of the Coastal Plain by the summer of 1999. It was never a severe drought outside the mountains, and the rains of hurricanes Dennis and Floyd easily broke it in the Piedmont and almost literally washed it away on the Coastal Plain. Meanwhile, in the mountains the drought continued and got worse. In October 2000 the PDSI fell below -3, signaling the onset of a severe drought. Stream flow was extremely low, groundwater was becoming scarce and there was a growing need to reduce water use. This continued almost unchanged for a year or so. Starting near the end of 2001 the drought area expanded to include virtually the whole of the state. By the summer of 2002 much of it was in extreme drought (PDSI <-4, and precipitation for the last 6 months being less than 60 percent of normal). Virtually everywhere had problems obtaining water for human and agricultural use (Figure 16). In September, 2002 abundant rain finally came to the mountains after a period of over four years with low rainfall and water problems. The Coastal Plain was also wet at this time, but over the Piedmont the extreme drought persisted from another month before abundant rain ameliorated the water problems.
This recent drought was one of the top three in the last 100 years. Previous to this was one in 1986 which was of shorter duration but even more intense. It started in the southern mountains early in the year, expanded to affect the whole state between May and October, and then died away rather rapidly. Between June and September the southern mountains had PDSI <-5 (an exceptional drought, associated with rainfall in the previous 12 months being no more than 65 percent of normal). In July, 1986 the whole of the southern mountains and Piedmont were below -5, the most intense drought month of the century. There was a rather shorter, and perhaps even more intense, drought in the winter of 1933-34, when the whole of the Piedmont had PDSI below -5 for a month. But probably the severest drought in our record was that between September, 1926 and June, 1927. All of the state outside the mountains had PDSI below -3 in this period, with much of the Coastal Plain below -4. For much of the southeast of the state the severe drought persisted almost to the end of the year. And this drought came hard on the heels of one lasting from July, 1925 until January, 1926.
There appears to be no pattern in the timing of these major droughts, but it is common in the North Carolina climate for late summer and early fall to be the time when agricultural drought and water resource problems are quite likely in most years. This results from a combination of two factors. First, because rainfall is fairly evenly distributed throughout the year while evaporation reaches a maximum in the heat of summer, there is a natural tendency for more water loss in summer. So the water available, whether in the soil, deeper in the ground, or in stream flow, is least around summer's end. Second, summer rainfall is much more variable from year to year than is that of winter. This can be attributed mainly to the action of the Bermuda High, a local feature of the global atmosphere. This is an anticyclone, a more-or-less circular region of high pressure with surface winds spiraling out in a clockwise direction. The weather of an anticyclone is the opposite to that which would be expected from a typical rain-bringing storm, having light winds, no weather fronts, and actively discouraging the formation of clouds and rain. In winter the Bermuda High is over the Atlantic Ocean away to our southeast, but in summer it moves north to a center very roughly over the island of Bermuda. While there it plays a minor role in weather but in most summers it either expands or drifts eastward, or both, to affect more people. While it is overhead there is no rain. If it stays for a week or so there is a nice break from summer rains; if it stays much longer it begins to get dry. Sometimes it stays for much of the summer and, if so, there is a major problems with drought. Often, as in 1999, it takes a hurricane to come and in and push it away to break the drought.
 |
 |
