413
Bulletin of the American Meteorological Society
An Introduction to Trends in Extreme
Weather and Climate Events:
Observations, Socioeconomic
Impacts, Terrestrial Ecological
Impacts, and Model Projections*
Gerald A. Meehl,
a
Thomas Karl,
b
David R. Easterling,
b
Stanley Changnon,
c
Roger Pielke Jr.,
a
David Changnon,
d
Jenni Evans,
e
Pavel Ya. Groisman,
b
Thomas R. Knutson,
f
Kenneth E. Kunkel,
c
Linda O. Mearns,
a
Camille Parmesan,
g
Roger Pulwarty,
h
Terry Root,
i
Richard T. Sylves,
j
Peter Whetton,
k
and Francis Zwiers
l
*This is the first of five papers in the “Understanding Changes in
Weather and Climate Extremes” series.
a
National Center for Atmospheric Research, Boulder, Colorado.
b
National Climatic Data Center, Asheville, North Carolina.
c
Illinois State Water Survey, Champaign, Illinois.
d
Northern Illinois University, De Kalb, Illinois.
e
The Pennsylvania State University, University Park, Pennsyl-
vania.
f
Geophysical Fluid Dynamics Laboratory, Princeton, New Jersey.
g
University of Texas at Austin, Austin, Texas, and University of
California, Santa Barbara, Santa Barbara, California.
ABSTRACT
Weather and climatic extremes can have serious and damaging effects on human society and infrastructure as well
as on ecosystems and wildlife. Thus, they are usually the main focus of attention of the news media in reports on cli-
mate. There are some indications from observations concerning how climatic extremes may have changed in the past.
Climate models show how they could change in the future either due to natural climate fluctuations or under conditions
of greenhouse gas–induced warming. These observed and modeled changes relate directly to the understanding of so-
cioeconomic and ecological impacts related to extremes.
h
University of Colorado, Boulder, Colorado.
i
University of Michigan, Ann Arbor, Michigan.
j
University of Delaware, Newark, Delaware.
k
CSIRO, Aspendale, Victoria, Australia.
l
CCCMA, Victoria, British Columbia, Canada.
Corresponding author address: Dr. Gerald A. Meehl, Climate and
Global Dynamics Division, NCAR, P.O. Box 3000, Boulder, CO
80307-3000.
E-mail: meehl@ncar.ucar.edu
In final form 9 August 1999.
©2000 American Meteorological Society
Understanding Changes in Weather and Climate Extremes and Their Impacts
The following series of five articles was motivated by a need to develop a more comprehensive assess-
ment of changes in weather and extreme climate events. We were interested not only in the impact of extreme
weather and climate events, but whether these events were changing in frequency or intensity along with their
impacts. Impacts were viewed in terms of loosely managed ecosystems where wildlife flourishes, as well as
socioeconomic systems and more heavily managed ecosystems such as agriculture. From a climate perspec-
tive, this included a focus both on the historical record and projections for future change.
During the summer of 1998 a group of nearly 30 climate scientists, social scientists, and biologists met for
10 days at the Aspen Global Change Institute to discuss what we now know, and how we could reduce some
of our major uncertainties. These articles summarize much of the work during that meeting and new informa-
tion since the meeting.
—Tom Karl, Stan Changnon, Dave Easterling, Jerry Meehl, and Camille Parmesan
414
Vol. 81, No. 3, March 2000
1. Introduction
Though weather and climate extremes can have
negative effects on society and ecosystems in many
obvious ways (floods, droughts, damaging high winds,
extreme heat, and cold, etc.), for some systems in some
areas, extreme events are beneficial. Bird breeding in
wetlands in the arid zone of Australia may only occur
after, say, a one in 5-yr rainfall event (i.e., the biota
could be poorer without sporadic extreme rainfall
events and perhaps even better off with more of them).
The record warm and unusually dry winter in the
northern United States resulting from the influence of
El Niño 1997–98 brought major gains in reduced en-
ergy costs ($5.6 billion), record retail and home sales
($5 billion), and a reduction of about 800 lives nor-
mally lost to winter conditions. Conversely, losses
in the southern and western states amounted to over
$4 billion with 200 lives lost due to storm activity.
Parts of the northwest coast of Australia receive most
of their rainfall from sporadically occurring tropical
cyclones, and even human systems (e.g., water stor-
age) would be damaged without them. Therefore, we
need a comprehensive understanding not only of what
has happened and what may happen with changes in
weather and climate extremes, but also what those
changes could mean in a variety of different contexts
in human and natural systems.
2. Defining changes of extremes
To understand how changes in weather and climate
extremes could influence society and ecosystems, it
is useful first to conceptually address how such ex-
tremes could change in a statistical sense. Figure 1
presents a typical distribution of a climate variable that
is normally distributed, such as temperature. The solid
curve represents the present-day frequency distribu-
tion of a weather phenomenon (such as the daily maxi-
mum temperature). Shading indicates the extreme
parts of the distribution, representing events in the tails
of the distribution that occur infrequently (i.e., values
that are far from the mean or median value of the dis-
tribution). If there is a simple shift of the distribution
in a future climate, there will be an increase in extreme
events on one end and a decrease at the other (Fig. 1a).
This can occur through a change of the mean where,
for example, if the temperature at a location warms by
a certain amount, this will almost certainly produce an
increase in the number of extreme hot days and a de-
FIG. 1. Schematic diagram depicting how changes in mean and
variance can affect extreme weather and climate events.
415
Bulletin of the American Meteorological Society
crease in the number of extreme cold days. It is im-
portant to note that the frequency of extremes changes
nonlinearly with the change in the mean of a distribu-
tion, that is, a small change in the mean can result in a
large change in the frequency of extremes (Mearns
et al. 1984).
Other aspects of the distribution may also change.
For example, the standard deviation in a future climate
may increase, producing changes in extreme events at
both ends of the frequency distribution (Fig. 1b). A
change in the variance of a distribution will have a
larger effect on the frequency of extremes than a
change in the mean (Katz and Brown 1992), though
these events must be “extreme” enough (i.e., more than
one standard deviation from the mean) for this result
to hold. Relatively speaking, a 1°C change in the stan-
dard deviation of the distribution will have a greater
impact on the frequency of an extreme temperature
than a 1°C change in the mean of the distribution. To
complicate matters, the mean, standard deviation, and
even the symmetry of the distribution may change at
the same time, consequently altering the occurrence
of extremes in several different ways (Fig. 1c, show-
ing changes in both mean and variance). Not only can
the parameters of the distribution change as noted
above (mean, variance, etc.) but for variables like pre-
cipitation, which is not normally distributed but bet-
ter represented by a gamma distribution, a change in
the mean also causes a change in the variance
(Groisman et al. 1999). This helps to explain why in-
creases in total precipitation are disproportionately
expressed in the extremes as discussed by the
Easterling et al. paper in this series (Easterling et al.
2000). However, the relationship between the mean
and variance arises because each statistic depends on
both the shape and scale parameters of the gamma dis-
tribution. It is possible to change those parameters in
a way that adjusts the mean while holding the variance
constant.
Fortunately, it is often possible to estimate changes
in infrequent extremes, such as those that might oc-
cur once every 10–100 years, without detailed knowl-
edge of the parent distribution. This is because
statistical science provides a well-developed asymp-
totic theory for extreme values (see, e.g., Leadbetter
et al. 1983), which predicts that the largest observa-
tion in a large sample, such as the annual maximum
temperature or 24-h precipitation amount, will tend to
have one of only three extreme value distributions de-
pending only upon the shape of the upper tail of the
parent distribution. One of these three distributions is
the familiar Gumbel distribution (Gumbel 1958). This
approach to extreme value analysis, together with re-
lated techniques based on the study of the crossing of
high thresholds (e.g., Davison and Smith 1990), has
been used extensively and reliably in meteorology, cli-
matology, and hydrology to predict the rare extremes
of phenomena such as precipitation, streamflow, tem-
perature, and wind speed. Statistical science continues
to make advances in extreme value theory (e.g., Smith
et al. 1997; Dupuis and Field 1998).
3. Consequences
The frequency and/or intensity of extremes can
change, and both can cause major problems. For ex-
ample, during the 1990s the number of insured catas-
trophes (weather events causing more than $5 million
in losses) in the United States doubled over prior fre-
quencies. Losses soared, but the average loss per event
did not increase.
When we discuss extremes we must consider them
from the point of view of the statistical characteristics
described above and from the socioeconomic or eco-
logical effects of the event. The latter can be thought
of in terms of thresholds of the physical systems be-
yond which serious impacts occur. For example, there
is an effect on human mortality and morbidity if there
is an occurrence of a series of days in summer when
minimum temperatures exceed 30°C (Kalkstein and
Smoyer 1993). This relationship can also apply to
domesticated animals. For cattle in the central Great
Plains, THI values (a combined temperature and hu-
midity index) greater than 84 for a series of three days
results in significant cattle deaths from heat stress
(Hahn and Mader 1997).
These are seemingly straightforward conse-
quences, but often the relationship between human
society, natural ecosystems, domesticated animals,
wildlife, and weather and climate extremes is not al-
ways linear and not intuitive. Thus, the effects of these
thresholds are not universal since the vulnerability of
the human and natural systems contributes to how
severe the impact will be, independent of the physi-
cal system itself. In this sense, the impact of climate
on society and ecosystems could change due to
changes in the physical climate system (including both
natural and anthropogenic causes) or due to changes
in the vulnerability of society and ecosystems (even
if the climate does not change; e.g., Kunkel et al.
1999).
416
Vol. 81, No. 3, March 2000
For example, if the level of hurricane activity of
the more active hurricane seasons in the 1940s and
1950s were to occur today, societal impacts would be
substantially greater than in earlier decades. The im-
pact of a higher tropical storm frequency could be
made worse by population increases and density, more
people living near the coast, greater wealth, and other
factors (Pielke and Pielke 1997; Diaz and Pulwarty
1997), though human choices have produced this situ-
ation. Thus, it is worth noting in this context that the
vulnerability of systems to particular extremes also
may change as a result of adaptation to changing cli-
matic means. Consider the following case: reduced
frost frequency may be interpreted as meaning a re-
duced risk of frost damage to crops. But farmers sim-
ply may use the opportunity this presents to plant
earlier in the season, and effectively maintain their
current or even greater risk of a damaging frost. One
could also imagine some similar adaptations in natu-
ral ecosystems that would have the effect of reducing
the impact of the change in the frequency of extremes
due to the change in the mean.
4. Aspects of weather and climate
extremes
The series of papers that follows describes the cur-
rent state of our knowledge concerning the effects of
weather and climate extremes from several points of
view, all of which are interrelated. First we examine
statistical changes in extremes already observed. Then
we address the socioeconomic and ecological impacts
that are closely tied to changes in extremes, and look
at projections from climate model experiments con-
cerning how weather and climate extremes could
change in the future.
Acknowledgments. We would like to thank the National Sci-
ence Foundation, National Oceanic and Atmospheric Adminis-
tration, National Aeronautics and Space Administration, and the
U.S. Department of Agriculture/Forest Service for providing sup-
port to the Aspen Global Change Institute to host the Climate Ex-
tremes Workshop, August 1998. This workshop provided the
impetus for this and the accompanying papers.
References
Davison, A. C., and R. L. Smith, 1990: Models for exceedances
over high thresholds. J. Roy. Stat. Soc., 52B, 393–442.
Diaz, H. F., and R. S. Pulwarty, 1997: Hurricanes, Climate and
Socioeconomic Impacts. Springer-Verlag, 292 pp.
Dupuis, D. J., and C. A. Field, 1998: Robust estimation of ex-
tremes. Can. J. Stat., 26, 199–216.
Easterling, D. R., J. L. Evans, P. Ya. Groisman, T. R. Karl, K. E.
Kunkel, and P. Ambenje, 2000: Observed variability and
trends in extreme climate events: A brief review. Bull. Amer.
Meteor. Soc., 81, 417–425.
Groisman, P. Ya., and Coauthors, 1999: Changes in the probabil-
ity of heavy precipitation: Important indicators of climatic
change. Climatic Change, 42, 243–283.
Gumbel, E. J., 1958: Statistics of Extremes. Columbia University
Press, 375 pp.
Hahn, G. L., and T. L. Mader, 1997: Heat waves in relation to
thermoregulation, feeding behavior, and mortality of feedlot
cattle. Proceedings of the Fifth International Livestock Envi-
ronment Symposium, R. W. Bottcher and S. J. Hoff, Eds., Vol.
1, American Society of Agricultural Engineers, 563–571.
Kalkstein, L. S., and K. E. Smoyer, 1993: The impact of climate
change on human health: Some international perspectives.
Experientia, 49, 969–979.
Katz, R. W., and B. G. Brown, 1992: Extreme events in a chang-
ing climate: Variability is more important than averages. Cli-
matic Change, 21, 289–302.
Kunkel, K. E., R. A. Pielke Jr., and S. A. Changnon, 1999: Tem-
poral fluctuations in weather and climate extremes that cause
economic and human health impacts: A review. Bull. Amer.
Meteor. Soc., 80, 1077–1098.
Leadbetter, M. R., G. Lindgren, and H. Rootzen, 1983: Extremes
and Related Properties of Random Sequences and Process.
Springer-Verlag, 336 pp.
Mearns, L. O., R. W. Katz, and S. H. Schneider, 1984: Extreme
high temperature events: Changes in their probabilities with
changes in mean temperature. J. Climate Appl. Meteor., 23,
1601–1613.
Pielke, R. A., Jr., and R. A. Pielke Sr., 1997: Hurricanes: Their
Nature and Impacts on Society. John Wiley and Sons, 298 pp.
Smith, R. L., J. A. Tawn, and S. G. Coles, 1997: Markov chain
models for threshold exceedances. Biometrika, 84, 249–268.
