NOAA Technical Memorandum NWS SR-193, Section 5

Section 5 Table of Contents:

5. Year-to-year Variations
A. National by Year
B. National by Decade
C. Regional by Decade
D. States by Decade

5. YEAR-TO-YEAR VARIATIONS

A. NATIONAL BY YEAR

The number of lightning victims by year from 1959 to 1994 is shown in Figure 16. Only during one year, 1963, were there more deaths (210) than injuries (209). While the number of deaths have decreased, the number of injuries has increased during the period, as shown by the ratio of injuries to deaths in Figure 17.

When the growth in population is taken into account during this period, the trends can be separated into the following factors considered in more detail by López and Holle (1995, 1996):

• A 30% linear reduction in the rate of casualties per population during the period is attributed to improved forecasts and warnings, better awareness of the lightning threat, better quality of buildings available for shelter, and/or other socioeconomic changes.
• An additional 40% reduction in the rate of deaths per population may be due to improved emergency communications and medical care, such as cardiopulmonary resuscitation (CPR).
• The rate of injuries per population has decreased by only 8% because of the transfer of what may have become deaths into injuries due to better medical attention and other factors.
• Additional fluctuations on the scale of one or two decades broadly parallel national-scale changes in frequencies of thunderstorm days, cyclones, and surface temperatures (López and Holle, 1996).

Other long-term series of lightning-caused deaths include a decrease in the rate of deaths per million people from 1922 to 1979 in Singapore (Pakiam et al., 1981). A greater decrease was found in England and Wales from 1852 to 1990 (Elsom, 1993). In Australia, Coates et al. (1993) found an increase in the absolute number of deaths from 1824 to 1918, then a decrease through 1991; however the fatality rate per 100,000 people decreased through the record. When the state of Colorado was considered by López et al. (1995), no steady trend in the number of casualties was apparent from 1950 to 1991.

	Figure 16: Number of lightning deaths, injuries, and casualties (deaths and injuries) by year from 1959 to 1994 for the US.
	Figure 17: Ratio of the number of lightning injuries to deaths by year from 1959 to 1994 for the US.

Figure 18 shows the annual numbers of damage reports through the entire period of record. Holle et al. (1996) showed that Storm Data damages due to lightning are vastly underreported. Even though so many events are missed, the sample appears to indicate a systematic increase through time that could be due to the increase in population.

There is a substantial number of lightning victims and damages every year. Lightning entries have the least year-to-year variability of all convective-weather causes in Table 1, and less variability than nearly all other phenomena in Table 2. For example, some years have no hurricane deaths, and tornado casualties vary widely among years due to individual storm events. But the vulnerability to lightning is a constant and widespread threat to people and property during every thunderstorm season.

Figure 18: Number of lightning damage reports by year from 1959 to 1994 for the US.

B. NATIONAL BY DECADE

Time series by decade of reported frequencies of casualties and damages for the US are in Figure 19, and population-weighted time series are in Figure 20. Entries for the 1990s were made comparable to other decades by doubling the number of frequencies reported from 1990 to 1994. There was one large event due to an aircraft crash in 1963 (Maryland, 81 deaths).

Reported damage frequencies for the 1990s in Figure 19 are greater than during previous decades, while casualties are also somewhat more than in recent decades. Since only the first half of the 1990s has been included, it remains to be seen whether these trends will continue through the rest of the decade.

Population-weighted casualties by decade in Figure 20 show a decrease until the 1990s, when the rate increases again. The damage report rate also was decreasing until it increased sharply in the 1990s. Again, the changes in trends are largest in the 1990s, which includes 1990 to 1994 in this analysis.

	Figure 19: Number of reported US lightning casualties and damage reports by decade.
	Figure 20: Population-weighted US lightning casualties and damage reports by decade.

C. REGIONAL BY DECADE

Eight regions of the United States were used for many of the subsequent analyses in this paper, as shown by Figure 21. The decadal trends in both casualties and damages are graphed for the eight regions in Figures 22 and 23.

Most regions have as many or more damage reports as casualties in both the actual and population-weighted time series. The only exceptions are in the southern Rockies and the southeast, where relatively few damages are reported.

Some general conclusions from reported frequencies in Figure 22 are the following:

• Casualties increased during the last two decades in the southeast, southern Rockies, and west coast states, but not elsewhere.
• Damage reports increased during the last two decades in all regions except in the west coast, midwest and northeast regions.

Population-weighted results in Figure 23 indicate that:

• Recent trends in casualties and damages for the west coast region are not as strong when population is taken into account as the actual reports indicated.
• Weighting by population made no difference in the other six regions.

Figure 21: US regions for analysis of Storm Data entries.

Figure 22: Decadal variations of number of reported lightning casualties (solid line with squre) and damages (dashed line with x) by region of the US.

Figure 23: Decadal variations of lightning casualties per million people per year (solid line with square) and damages per million people per year (dashed line with x) by region of the US.

D. STATES BY DECADE

Another way to show changes through the decades is to plot reported frequencies for each state by decade, as shown in Figures 24 and 25, and Tables 22 and 23. Before weighting by population, casualties for the entire 36-year period were shown by state in Figures 1, 2 and 3, and damage reports in Figures 8 and 9. Decadal maps in Figure 24 and data in Table 22 for reported frequencies show the following:

Casualties

• Florida, Michigan, and Pennsylvania are always in the top ten.
• Highest numbers tend to be in the southeast US and most populous eastern states.
• There is a shift towards the southwest US with time.
• Most other states stay within one group of ten from the previous decade.

Damage reports

• Pennsylvania is always in the top ten.
• Highest numbers tend to be in the plains and most populous eastern states.
• States sometimes change by two groups of ten from the previous decade.
• These variations in damage reports do not necessarily represent actual changes in damages. For example, Georgia increased from 83 reports in the 1960s to 321 from 1990 to 1994, while Mississippi decreased from 58 in the 1960s to six reports in the 1990s.
• Table 22 shows more of these large variations in damage reports.

After weighting by population, casualty rates for the entire 36-year period were shown by state in Figures 11 and 12, and the damage rates in Figure 15. Decadal maps of the rates in Figure 25 and data in Table 23 show:

Casualty rate

• New Mexico and Arkansas are always in the top ten.
• Highest casualty rates are most consistent in the Rocky Mountain and southeastern states.
• Wyoming drops from a very high rate in the 1960s to a very low rate in the 1990s.

Damage rate

• Nebraska and Oklahoma are always in the top ten.
• Highest rates tend to be in the northern plains, southeast, and northeast states.
• Idaho maintains a high to very high rate during all decades.
• However, as noted earlier, there are variations in damage reports through time that are not likely to represent actual changes in damages, as shown in Table 23.

In summary, the decadal maps of numbers and population-weighted rates of casualties and damages show broad agreement with the maps for the 36-year Storm Data record. However, there are fairly large variations among decades in casualties and their rates, and very large variations in damages and their rates.

Figure 24: Decadal variations of lightning casualties and damages by state.

Figure 25: Decadal variations of the rates of lightning casualties and damages by state.