Environment Environnement

Canada Canada

Extreme
Weather
and
Climate Change
by

David Francis
Lanark House Communications
Toronto

and

Henry Hengeveld
Atmospheric Environment Service
Environment Canada

This document contributes to state of the environment reporting

This paper contains a minimum of 50% recycled fibres,

including 10% post-consumer fibres.
 INTRODUCTORY COMMENTS AND ACKNOWLEDGEMENTS

Extreme Weather and Climate Change was prepared in response to questions often posed by
policy makers and the general public about whether or not perceived changes in weather behaviour in
recent years, particularly with respect to extreme weather events and related disasters, are real, and, if
so, whether such changes are linked to climate change. It is being published as the second in a series of
“special” Climate Change Digest reports aimed at explaining and assessing our current understanding (or
lack thereof) of some of the more complex and controversial aspects of climate change science. This
series complements the regular CCD series focused on scientific studies relating to impacts of climate
change.

This report was prepared by David Francis of Lanark House Communications (Toronto), in collaboration
with Henry Hengeveld, Senior Science Advisor on Climate Change with the Atmospheric Environment
Service of Environment Canada.

The authors wish to acknowledge with appreciation the valuable input, through review and critique, of the
following individuals: Bill Hogg, Doug Whelpdale, and Francis Zwiers (AES Climate Research Branch); Pam
Kertland, David Phillips, and John Stone (AES Policy, Program and International Affairs Directorate); David
Etkin and Roger Street (AES Environmental Adaptation Research Group); Eric Taylor (Environment Canada’s
Pacific and Yukon Region); Jim Abraham and Paul Galbraith (Environment Canada’s Atlantic Region); and
Heather Johannesen (Halifax).

The French text of this report (available separately) was translated by Marie-France Guéraud of the
Translation Bureau, Public Works and Government Services (Montreal) and was edited by Gilles Tardif
of Enviroedit (Keswick, Ontario). Graphic design, artwork, and technical production were provided by
BTT Communications (Toronto).

Photo credits: Canadian Press (pages 7 and 16); Environment Canada, Ontario Region (page 24); Water
Resources Branch, Manitoba Department of Natural Resources (page 25).

Copies of this publication may be obtained, free of charge, by writing to:

Climate and Water Products Division

Atmospheric Environment Service
4905 Dufferin Street
Downsview, Ontario
M3H 5T4

(416) 739-4328

Published by Authority of the Minister of the Environment

© Minister of Supply and Services

Canada 1998

Cataloque No. En57-27/1998-01E

ISBN 0-662-26849-0
ISSN 0835-3980
 TABLE OF CONTENTS

IS EXTREME WEATHER BECOMING MORE COMMON? 1

NATURAL VARIABILITY 9

GREENHOUSE WARMING AND WEATHER EXTREMES 13

IMPLICATIONS OF AN INCREASE IN WEATHER EXTREMES 21

RESPONSES 25

DRAWING CONCLUSIONS 26

BOXES

WHAT IS EXTREME WEATHER? 2

EL NIÑOS AND CLIMATE CHANGE 12

FLOODING IN CANADA 16

THE 1998 ICE STORM 24

THE POSSIBILITY OF SURPRISES 27

i Extreme Weather and Climate Change

 Extreme Weather and Climate Change
Is the world’s weather becoming more extreme? become more common, the costs to society will
So far, during the 1990s alone, the world has wit- increase enormously. If these extremes are an
nessed at least half a dozen floods of epic proportions inevitable consequence of greenhouse warming, then
in Canada and the United States, central Europe, and our current estimates of the impacts of climate
southern China as well as intense droughts in north- change, serious as they are, will have been far too
ern China, northern Vietnam, North Korea, and optimistic, and the need to make decisive cuts in
southern Europe. In 1993, the northeast coast of the greenhouse gas emissions will become even more
United States received its biggest snowstorm in more pressing, as will the need to undertake increasingly
than a century. At the end of 1996, it was the turn of expensive adaptive measures.
Victoria, which was paralyzed by the biggest snowfall
in its recorded weather history. Then, in January Is Extreme Weather Becoming
1998, Canada’s worst-ever ice storm left the Montreal More Common?
area and eastern Ontario without power for weeks. The number of extraordinarily severe floods,
Western Europe, usually noted for the moderation of storms, and other weather calamities that have
its climate, was pounded occurred within the past
by four major storms in 15 to 20 years would seem
the winter of 1990. In Hurricanes, floods, droughts, to suggest that such
1987, southern England events are becoming more
was hit by its worst and other extreme weather common. Figures com-
storm since 1705. In piled by the world insur-
1995, heat waves killed events have the potential
ance industry, for exam-
more than 500 people in
to cause death and destruction ple, show a dramatic
northern and central increase in losses from
India and more than 550 on a catastrophic scale. If they weather-related disasters
in Chicago, numbers that in recent decades. For
pale in comparison, how- become more common, all of the 1960s insured
ever, to the estimated losses from windstorms
5,000–10,000 heat-relat-
the costs to society will
amounted to $2.0 billion
ed deaths that occurred increase enormously. (in 1990 U.S. dollars)
in the central and eastern worldwide. By the 1980s
U.S. in the summer of that figure had crept up to
1980. $3.4 billion for the decade. In just the first three years
The sheer number of such events within the past of the 1990s, however, it leapt to $20.2 billion. Before
two decades raises some serious questions about the 1987 a billion-dollar insurance loss from climate
current and future state of the global climate. Are events was rare, but between January 1988 and
these events part of a long-term trend towards more January 1997 there were 23 such events in the United
extreme weather, or are they just a temporary aberra- States alone. Canada has not yet had a billion dollar
tion? Are they the result of purely natural forces? Or insurance loss from a weather disaster, although total
could they be linked in some way to climate change costs (including insured and non-insured losses) for a
caused by the buildup of greenhouse gases in the few events, such as the 1996 Saguenay flood and the
atmosphere? The answers to these questions are vital- 1998 eastern ice storm, have exceeded this amount.
ly important. Hurricanes, floods, droughts, and other These figures certainly suggest a major rise in the
extreme weather events have the potential to cause number of destructive weather events, but cost alone
death and destruction on a catastrophic scale. If they is not necessarily an accurate indicator of climate

1 Extreme Weather and Climate Change

 WHAT IS EXTREME WEATHER?
Extreme weather, in the most obvious sense, is weather that lies outside a locale’s
normal range of weather intensity. It is therefore, by definition, infrequent or rare.
Extreme weather is also potentially destructive, although not all extreme weather
events end in disasters.

For some weather events, the idea of what constitutes an extreme can vary from
place to place. It often depends on what a region is used to experiencing and what it
is prepared for. A 20-cm snowfall would be an extreme event for Washington, D.C.,
for example, but not for Montreal. In Washington such an event would come close to
an emergency. In Montreal it would be merely an inconvenience.

Extreme events such as hurricanes, tornadoes, and ice storms often require the pres-
ence of a number of special circumstances before they can take place. Many extreme
events also come about as a result of a combination of factors, such as the merging
of two weather systems or the occurrence of a severe weather event in
tandem with some other factor that intensifies its impact. Hurricane Hazel, for exam-
ple, was a weakening tropical storm when it merged with a deep low pressure
system northwest of Toronto in October 1954, producing torrential rains and the
deadliest flood in Canadian history. In the case of the Saguenay flood, water levels in
the Saguenay basin were already at unusually high levels when the largest rainstorm
in the region’s recorded weather history struck on July 19, 1996. Some flooding
would still have occurred if water levels had been normal, but the results might not
have been as catastrophic.

trends. As well as being influenced by the number and Apply the same kind of analysis to world losses
severity of such events, costs also reflect the size and from natural disasters as a whole, however, and the
wealth of the population affected by them, and these results are quite different. Data from Munich Re, one
numbers have been increasing as well. Two American of the world’s largest re-insurance firms, show that
researchers, Roger Pielke, Jr. and Christopher direct economic losses (in 1992 U.S. dollars) from
Landsea, for example, have suggested that increased natural disasters worldwide increased by a factor of
damage costs from hurricanes in the U.S. can be 43 between the last half of the 1960s and the first half
attributed to three factors: inflation, population of the 1990s. Global wealth (as measured by GDP),
growth in vulnerable coastal areas, and the increasing on the other hand, increased by a factor of 2.5 and
prosperity of the people affected. When these factors population by 25%. That means that, with inflation
are taken into account, they argue, the economic already adjusted for by the use of constant dollars,
impact of hurricanes in the U.S. has actually declined economic growth and population increase account
in recent decades. for less than a fourfold rise in these losses. Other

2
Extreme Weather and Climate Change
 Economic losses from
natural catastrophes, 1965–1994

250

200
$US Billion

150

100

50

1965–69 1970–74 1975–79 1980–84 1985–89 1990–94

Time Period

Source: Adapted from Munich Re (1996)

Dramatic increases in economic losses from natural catastrophes, most of them

weather-related, imply an increase in weather extremes. These trends must be inter-
preted cautiously, however, since they are also heavily influenced by population
increases and economic growth.

population factors, such as migration to vulnerable more than 1% of the country’s population, or to cause
areas, might well account for further losses, but it is more than 100 deaths. These criteria partially filter
unlikely that they could explain all of the remaining out distortions arising from population and economic
increase. Since by far the largest part of the increase growth but not totally. Nevertheless, since trends for
in these losses was due to weather-related events, an both earthquakes and weather disasters would be
increase in severe weather is a possibility that has to affected more or less equally by these factors, there is
be looked at seriously. some reason to believe that the data reflect an actual
The likelihood that weather-related disasters are increase in severe weather events.
on the rise is also supported by an analysis done by However, insurance losses and disaster trends are
the Geneva Secretariat for the International Decade at best an indirect barometer of climate change.
for Natural Disaster Reduction. It looked at the rates Historical climate records should provide much more
of change for the four largest categories of major nat- direct evidence of change, but teasing out trends and
ural disasters – floods, tropical storms, droughts, and probabilities for rarely occurring weather extremes in
earthquakes. Between the mid-1960s and the early a body of highly variable data is a tricky proposition.
1990s the number of all of these disasters increased, Problems with data quality and irregularities, espe-
but the weather-related disasters increased at a much cially in older records that provide the benchmark
higher rate. To qualify as a major disaster, an event for change, make it even more difficult. Add in the
had either to cause damage equal to at least 1% of the fact that relatively little statistical analysis has been
affected country’s gross domestic product, to affect directed specifically towards extreme events, and it is

3 Extreme Weather and Climate Change

 Rate of increase of natural disasters
(floods, tropical storms, drought, and
earthquakes), 1963–1967 to 1998–1992
25

20
Number of Disasters

15

10

0
1963–67 1968–72 1973–77 1978–82 1983–87 1988–92
Time Period

Source: Adapted from McCulloch and Etkin (1993)

Data compiled by the Geneva Secretariat of the International Decade for Natural
Disaster Reduction show increases in all of the four leading natural disaster cate-
gories over the past three decades. Weather-related categories, however, show the
highest rate of increase. Since trends for all categories are likely to be more or less
equally affected by social and economic factors, there is reason to believe that the
data reflect an actual increase in severe weather events.

not surprising that the climate record is still far from have increased more than summer temperatures and
shedding as much light as it could on trends in weath- for overnight lows to have warmed more than daytime
er extremes. highs.
Nevertheless, work completed in the last few Temperature, therefore, has actually shown a
years has shown the emergence of some significant lessening of extremes, at least so far, but a tendency
regional trends, although no consistent pattern of towards more extreme precipitation is apparent
change in weather extremes is yet apparent globally. across much of the land area of the Northern
The most reliable trends are those for temperature Hemisphere. Heavy rainfalls have increased in
and precipitation (not surprisingly, since these are the Japan, the United States, the former Soviet Union,
most widely measured climate variables). Many parts China, and countries around the North Atlantic rim.
of the world have shown a decrease in the occurrence Canadian records also reveal a trend towards heavier
of low temperature extremes, as would be expected in precipitation since 1940, although the increase has
a warming climate. Surprisingly, though, there has been mainly confined to the North.
not yet been a noticeable increase in high temperature Drought, on the other hand, has become more
extremes. The reason appears to be related to the ten- common since the 1970s in parts of Africa as well
dency in many regions for winter temperatures to

4
Extreme Weather and Climate Change
 Extreme precipitation trends in
Canada and the United States

Sum of upper 10 percentile daily precipitation events / Sum of all events

55
Canada, 1940–1995

In both Canada and the

United States, the
percentage of annual 50
Per cent
precipitation coming
from the heaviest 10%
of the year’s precipitation
45
events has increased
during the second half of
the century. Canadian
trends, however, have
40
been heavily influenced by 1910 1920 1930 1940 1950 1960 1970 1980 1990
precipitation increases in
the North. In southern 50
Canada, extreme precipi- U.S., 1910–1995

tation events have 45

actually declined over
the course of the
40
century.
Per cent

35

30

25
1910 1920 1930 1940 1950 1960 1970 1980 1990

Source: Environment Canada

as along the coasts of Chile and Peru and in north- statistical evidence is mixed. Canadian researcher
eastern Australia. The North American prairies also Steven Lambert recently examined winter storm
saw an increase in drought during the 1980s, although activity in the extratropical Atlantic and Pacific since
these years were not as dry as either the 1930s or the beginning of the present century. Using intense
the 1950s. low pressure systems as a marker for unusually severe
Severe storms would also appear to be on the storms, he saw little change in the number of these
increase in some regions since the mid-1980s, but the storms before 1970. After 1970, however, severe

5 Extreme Weather and Climate Change

winter storms became considerably more frequent, nique known as downscaling, which estimates local
particularly in the Pacific. Other researchers using conditions on the basis of relationships with larger-
similar methods to analyze storms in the North scale weather patterns. Applied to weather data for
Atlantic have also noted an increase in storm activity. the eastern North Atlantic, the method has shown no
In addition, a study of storms along the eastern coast statistically significant trend in recent storm behav-
of North America found an increase since the mid- iour for the region.
1970s, although storms in this area were still not as Thunderstorms are another important category,
frequent as they had been before 1965. What was because they are frequently associated not only with
different about the post-1965 storms though was their high winds and intense rainfalls but also with hail and
destructiveness. Seven of the region’s eight most tornadoes. Thunderstorm activity is difficult to mea-
destructive storms of the past half century had sure on a broad scale, however, because thunder-
occurred within the last 25 years. storms are highly localized and brief in duration.
Other researchers have questioned these conclu- Consequently, they are unlikely to be recorded unless
sions. The quality of the weather records on which they occur near a weather station. Still, there
they have been based, they argue, is too uneven. is evidence that thunderstorms have become more
Instead, they have preferred using a statistical tech- frequent in some areas. In the United States, for

Frequency of winter storms

in the Northern Hemisphere

100

80
Storms per Winter

60

40

20

0
1900 1920 1940 1960 1980 2000

Source: Adapted from Lambert (1996)

This graph, from an analysis of severe winter storms in the extratropical Atlantic and
Pacific by Steven Lambert of Environment Canada, shows a striking increase in storm
activity after 1970. Other studies of extratropical storms, however, have given varied
results. Some are consistent with Lambert’s findings, while others have not found a
statistically significant trend in storm frequency.

6
Extreme Weather and Climate Change
 The Edmonton tornado
of July 31, 1987, left
27 dead and 200 injured
and caused more than
$250 million in property
damage. It has been
shown that the monthly
frequency of tornadoes
in the Prairie Provinces
corresponds closely to
the average monthly
temperature. This
suggests that warmer
spring and summer
months could bring an
increase in tornado
activity to the region.

example, the fact that most of the increase in heavy in tornado frequency on the Prairies if seasonal
rainfall has occurred during the summer suggests an temperatures rise beyond present normal values.
increase in the number of severe thunderstorms. Hurricanes are the most destructive storms, but
Other evidence comes from northern Australia, earlier records of these are often incomplete. Until the
where there has also been an increase in heavy rain- advent of satellites, storms that did not touch land in
falls during the summer, and France, where severe populated areas often went unrecorded. Records for
hail falls have become more common during the sum- the tropical Atlantic have been reasonably good since
mer months. Analysis of cloud patterns also suggests 1970, however, and these, interestingly enough, show
a general increase in thunderstorm activity in the a declining trend in annual hurricane frequency,
tropical regions of the world. although both 1995 and 1996 saw a larger than aver-
Tornadoes are even more difficult to measure age number of storms. Annual average maximum
than thunderstorms, since they are usually very short- wind speeds of Atlantic hurricanes have also
lived and do not always occur in populated areas decreased for much of the past half century, though
where they are certain to be observed. In the United there has been no trend in the highest wind speeds of
States, where tornadoes occur more frequently than individual storms from year to year. Hurricane activity
anywhere else, studies have shown no increase in the in the Pacific, on the other hand, appears to have
occurrence of strong tornadoes, although reports of increased, but the data are not as reliable as for the
less severe tornadoes have increased. When Atlantic.
Environment Canada researcher David Etkin looked In addition to looking for trends in individual
at tornado activity on the Canadian prairies, how- weather phenomena, climate researchers are also
ever, he found that tornadoes were more frequent in beginning to develop tools that will indicate a
warm springs and summers. As warm springs and tendency towards extremes across a spectrum of
summers would become more common as a result of weather events. The advantage of this approach is
climate change, his results imply an eventual increase that it provides a more direct answer to the question

7 Extreme Weather and Climate Change

 Average annual maximum sustained
wind speed of Atlantic hurricanes, 1940–1993

50
Wind (Metres per Second) 48
46
44
42
40
38
36
34
32
30
1940 1950 1960 1970 1980 1990 2000

Source: Adapted from Landsea et al. (1996)

Average maximum wind speeds of Atlantic hurricanes have generally decreased over
the past 50 years. The number of Atlantic hurricanes has also declined over the same
period. The number of hurricanes in the Pacific, however, may have increased.

of whether the climate in general is becoming more are not as yet available for other countries and
extreme. The Climate Extremes Index of the U.S. regions, it is impossible to say whether this repre-
National Climate Data Center provides an interest- sents a broader hemispheric or global trend.
ing illustration of this approach. It combines several Overall, then, the weather record is inconclusive,
measures of the area covered by extreme tempera- though occasionally suggestive. The world could be in
tures and precipitation, drought, and soil moisture the early phases of a fundamental shift towards a
surpluses into a single value representing the relative climate in which extremes of many kinds are more
predominance of extreme weather events in a given prevalent. Or the present cluster of extreme events
year. Beginning in 1910, it shows an almost cyclical could be a temporary phenomenon. To what extent
waxing and waning of extreme events, with pro- either of these explanations is correct depends very
nounced peaks for the mid-1930s and mid-1950s much on what has caused extreme events to occur so
when the human influence on the climate was much often within the past 15 to 20 years.
less than it is today. The index rises to peak levels
There are three possible causes to consider. First,
again in the mid-1970s, but this time it no longer
if the clustering of extremes proves to be temporary, it
subsides to the same extent as it did previously. In
could be explained entirely as the result of the natural
fact, it remains above the average through the 1980s
variability of the climate system. The peaks on the
and 1990s. The transformation of the peak into a
graph of the U.S. Climate Extremes Index in the
plateau could indicate that more severe weather con-
1930s and 1950s, for example, illustrate what is likely
ditions are becoming a permanent part of the
a natural surge in the frequency of extreme weather.
American climate at least, but since similar indices

8
Extreme Weather and Climate Change
 The U.S. Climate Extremes Index
32 32

28 28

24 24

20 20

16 16

12 12
1900 1920 1940 1960 1980

Source: U.S. National Climate Data Center

The U.S. Climate Extremes Index combines a variety of measures of temperature and
precipitation extremes to give a single annual measure of the frequency of extreme
events. Although it does not track all types of extremes – tornadoes, for example, are
not included – it does provide a useful approximation of trends in weather extremes on
a regional scale. The index shows pronounced but brief peaks in the 1930s and
1950s and a more sustained period of severe weather activity since the mid-1970s.

If, however, the climate is undergoing a fundamental occurs within a complex, quasi-chaotic system such
shift in which extremes become more common, then as the climate system because of the almost infinite
we must look for some basic change in the forces number of forces acting on it. Still, there are clear the-
acting on the climate system. That raises two addi- oretical limits to this variability, and these are set by
tional possibilities. The change could be the result of large-scale controls and feedback processes that
an entirely natural process, such as an increase in govern the amount of energy entering and leaving the
solar radiation, or it could be a consequence of human atmosphere. These include such factors as the intensity
actions, most notably the enhancement of the green- of the sun’s radiation, the earth’s orbit and the tilt of
house effect. its axis, and the concentration of greenhouse gases in
the earth’s atmosphere. How the system behaves
Natural Variability within the limits set by these controls, however, is
Variability is a natural feature of the climate much harder to determine. In the case of extreme
system. It may appear as short-term fluctuations that events, this unpredictability can often be much
come and go within the span of a decade or longer- greater because the worst extremes are frequently the
term changes that last for a century or more. Such result of a chance combination of less extreme events,
variations are the net result of a number of factors. such as a storm and a high tide or the merging of two
One of these is simply the random variability that storm systems.

9 Extreme Weather and Climate Change

 Some short-term climatic abnormalities have a surface waters in the eastern half of the equatorial
more identifiable physical basis. Large volcanic erup- Pacific. It usually lasts from 12 to 18 months and
tions can exert a powerful cooling effect on weather in occurs once every two to ten years. Because the
many parts of the world. This happens because warming occurs in tandem with the Southern
sulphur particles shot into the stratosphere by the Oscillation (a reversal of pressure patterns over the
eruption can partially block incoming sunlight for a South Pacific), climatologists commonly refer to it as
number of years. During the summer of 1816, for the El Niño Southern Oscillation or ENSO.
example, there were repeated frosts in Quebec and Normally the equatorial Pacific is swept by
the New England states. These have been linked to strong easterly trade winds, which cause warm sur-
the extremely powerful eruption of Mt. Tambora face waters to pile up on the western side of the
in Indonesia in 1815. More recently, the eruption of Pacific and bring colder, deeper water to the surface
Mt. Pinatubo in the Philippines in the summer of on the eastern side. In an El Niño year, the trade
1991 – the most powerful eruption of the twentieth winds slacken and the warm water accumulated in the
century – brought cooler temperatures to much of the west gradually returns to the eastern Pacific, prevent-
rest of the world during the next two years. It is esti- ing the cooler water from reaching the surface.
mated that Pinatubo reduced the earth’s average sur- Consequently, the surface temperature of the eastern
face temperature in 1992 by somewhere between Pacific begins to rise, changing the pattern of rising
0.3°C and 0.5°C. In Canada, the Pinatubo cooling and falling air masses over the entire equatorial
was most evident in Ontario and Quebec that year, Pacific and ultimately altering the atmospheric circu-
and the Great Lakes–St. Lawrence region recorded lation over much of the rest of the world.
its coolest July since the 1880s.
By distorting atmospheric circulation patterns,
Short-term fluctuations may also be the result ENSOs bring profound changes to customary weather
of more systematic variations within the climate patterns in the tropics and even in the middle latitudes.
system. The severity of winters in western Europe, for Droughts in Australia and Africa, floods in Brazil and
example, tends to follow the ups and downs of the Paraguay, freak snowstorms in the Middle East, and
North Atlantic Oscillation, an alternation in pressure poor monsoon rains in India and Indonesia have at var-
differences between Iceland and the Azores. In years ious times been linked to ENSO conditions. In
when the difference is large, western Europe enjoys Canada, the effects of ENSOs vary considerably, but
milder winters, while western Greenland and the most common result is an unusually warm and
Labrador experience unusually cold weather. When somewhat drier winter in most parts of the country
the difference is small, the situation is reversed. The except the Atlantic provinces and the high Arctic.
North Atlantic Oscillation tends to switch phase
Since the mid-1970s, El Niños have been both
every couple of years, although it may occasionally
more frequent and more persistent. This change in
get stuck in one phase for up to a decade or more. It
ENSO behaviour can account, at least in part, for
also appears to follow a longer cycle in which it is pre-
many of the weather anomalies of the past couple of
dominantly in one phase for 30–40 years and then
decades. Some of the increase in the global average
predominantly in the other for the next 30–40 years.
temperature of the past two decades, for example,
From 1900 until the late 1940s, the positive phase, in
can be linked to ENSO events, as can precipitation
which pressure differences are large, predominated.
declines in North Africa, Southeast Asia, Indonesia,
From then until around 1980, the negative phase
western Central America, and other parts of the
was more common. Since then, the oscillation has
tropics and sub-tropics. Hurricane patterns are also
returned to its positive phase. The cause of the
affected by the ENSO cycle, because changes to the
oscillation is not well understood, but it is clearly a
upper air flow in the tropics in El Niño years tend to
natural phenomenon which affects the severity of
suppress hurricane formation in the Atlantic and
winter weather in different parts of the North Atlantic
enhance it in the southeast Pacific. The decline in the
region.
frequency of Atlantic hurricanes since 1970 and the
A somewhat similar but much better known apparent increase in Pacific hurricanes may well
phenomenon is the El Niño, the periodic warming of

10
Extreme Weather and Climate Change
reflect this relationship. However, the recent intensi- 16 years and averages slightly more than 11 years.
fication of El Niño activity itself raises serious ques- Sunspots tend to appear in polarized pairs, and over
tions. It has no precedent in the climate records of the a double cycle of about 22 years, on average, the
past 120 years and, according to some climate polarity of the pair will reverse. Although sunspots
researchers, no precedent within the past 1000 years. are actually cooler than the surrounding solar surface,
Why has it happened now? Is it a natural occurrence, they are associated with hotter areas known as facu-
or is it a response to human interference with the lae. Consequently an increase in the number of
climate system? sunspots indicates an increase in radiation from the
Some short-term climatic fluctuations may also be sun. The change in the sun’s energy output over a
related to minor changes in the intensity of the sun’s single cycle, however, is quite small. Recent estimates
radiation that accompany the sunspot cycle. The place it at about 0.1%.
number of sunspots on the sun’s surface increases Many climate records show pronounced cycles
and then decreases over a cycle that varies from 7.5 to that, for a time at least, parallel either the single or the

El Niño frequency and intensity

in the 20th century

Source: James P. Bruce

The occurrence of El Niños is marked by an abnormal warming of the eastern equa-

torial Pacific. As the graph for July sea surface temperature shows, major warming
events (in which the sea surface temperature is 1°C or more warmer than the
1950–1980 average) have not only been more frequent in the second half of the cen-
tury but have also involved a greater degree of warming. A strengthening of the
opposing La Niña phenomenon (related to abnormal cooling of the sea surface) is also
evident in the second half of the century.

11 Extreme Weather and Climate Change

 EL NIÑOS AND CLIMATE CHANGE
What is the driving force behind El Niños? According to a theory proposed by
De-Zheng Sun, a researcher with the U.S. National Oceanographic and Atmospheric
Administration, it’s the temperature difference between the tropical ocean surface
and the ocean depths.

Sun’s findings are based on experiments with a model of the ocean-atmosphere

system in which the temperature of the deep ocean was held steady while the ocean
surface was warmed by solar heat and a gradually increasing greenhouse effect. As
the ocean surface became warmer, a temperature difference between the eastern
and western sides of the model’s ocean began to develop. At 29.2°C an ENSO-like
oscillation of ocean temperatures set in and increased in magnitude as the tempera-
ture of the ocean surface was increased. The oscillations, which go through a com-
plete cycle in four years, closely resemble the behaviour of the tropical Pacific Ocean.
The experiment also showed that the oscillations would not occur in a narrower basin,
such as that of the Atlantic.

If Sun’s theory is correct, then El Niños would initially become stronger in a warming
climate because the surface ocean would warm faster than the deep ocean. Over
time, however, the intensity of El Niños would probably moderate as the deep ocean
gradually became warmer and the temperature difference between it and the surface
diminished. Nevertheless, it would take several hundred years for this process to
take place.

double sunspot cycles. Major droughts in the U.S. mid- Some or all of these mechanisms could account
west, for example, have tended to occur at roughly for a temporary increase in weather extremes. The
20-year intervals. However, attempts to link major longer the extremes persist, however, the less ade-
climatic patterns to the sunspot cycle usually run into quate these explanations become and the more we
difficulty when they attempt to explain how such small must consider the possibility of a long-term shift in
changes in solar output could cause relatively large cli- climate behaviour. An enhanced greenhouse effect is
matic fluctuations. One of the more plausible mecha- one possible reason for such a shift, but climate can
nisms suggested so far links sunspot peaks to warming also change dramatically over the longer term for
in the lower stratosphere, which then affects circula- purely natural reasons.
tion patterns in the rest of the atmosphere below. The great ice age glaciations of the past 2 million
Although the warming still amounts to only a fraction years are the most spectacular examples of such
of a degree, climate model studies by Joanna Haigh of changes, but smaller variations over periods of a few
the University of London suggest that such small varia- hundred to a few thousand years have also occurred
tions in stratospheric temperature might be enough to during our present postglacial climate. The best-
shift winter storm tracks in Europe northward from known example of these variations is the event known
the Mediterranean by an average of about 700 km.

12
Extreme Weather and Climate Change
as the Little Ice Age. During this period, which lasted ence on our present climate. If the surge in extreme
from about 1400 to 1850, the climate in Europe and weather events is the result of a fundamental, long-
many other parts of the world cooled noticeably. By term shift in climate, then it is greenhouse warming
the mid-1600s, when the cooling reached its maxi- rather than a change in solar output that is more
mum extent, the earth’s average surface temperature likely to be driving the change.
had dipped to nearly 1°C below its present value. The
Little Ice Age and other variations in global tempera- Greenhouse Warming and
tures during the past 10,000 years are thought to be Weather Extremes
due primarily to changes in the output of energy from Why would greenhouse warming cause an
the sun, though other factors may also have played increase in weather extremes? One reason is that the
a role. The central part of the Little Ice Age, for exam- additional warming will change the distribution of
ple, coincided with a period of minimal sunspot heat and thus the flow of energy through the climate
activity between 1645 and 1715 known as the system. This will in turn alter the circulation patterns
Maunder Minimum. During that time, the output of of the atmosphere and the oceans, and it will also
solar energy is estimated to have been somewhere modify the hydrological cycle by which water is cir-
between 0.1% and 0.7% lower than it is today. culated between the earth’s surface and the air. As a
The sun’s output has been increasing since about result, the position of many of the world’s major storm
1850, and recent scientific estimates attribute about tracks could shift significantly. To see what the effects
half of the warming of the past century and a third of of such a shift might be, one has only to look at what
that since 1970 to this increase. There is growing evi- happens when circulation patterns are changed by an
dence, however, that the enhanced greenhouse effect El Niño. Some areas would be exposed to more
is largely responsible for the rest of the warming influ- storms and heavier rainfalls, while others might see

Average global temperature changes

since the end of the last ice age

Although earth’s climate

has been remarkably
stable since the last ice
age ended more than
10,000 years ago, small
but significant variations
have occurred on scales
of a few hundred to a few
thousand years. These
fluctuations are thought to
be largely the result of
changes in the output of
energy from the sun.

Source: Adapted from Folland et al. (1990)

13 Extreme Weather and Climate Change

 Estimated solar irradiance, 1874–1988

1368.0

(Watts per Square Metre)

Irradiance Change

1367.5

1367.0

1880 1900 1920 1940 1960 1980

Source: Adapted from Foukal and Lean (1990)

The output of energy from the sun has been increasing since about 1850. About half
of the warming of the earth’s surface over the past century and a third of that since
1970 are thought to be due to this increase. Direct observations of the sun’s
energy output have been available from satellites only since 1978. The estimate
shown here was derived using a computer model and sunspot data, and the results
were calibrated against available satellite observations.

formerly reliable rainfalls give way to prolonged dry outcome of a rise in global temperatures is a wide-
spells. Other areas might actually see improvements spread increase in the amount of water that is moved
in their climates, but if the experience of recent El through the cycle. That is because higher temperatures
Niños is a guide, most localities would encounter at not only increase evaporation and transpiration but
least some weather difficulties that they were poorly also raise the air’s capacity to hold moisture.
prepared to deal with. Over time, communities could Consequently, more moisture will be available in the
adapt to these new conditions, but the costs could be atmosphere to fall as rain and snow. Add to this a more
substantial. unstable atmosphere due to increased convection over
A second and more compelling reason for sus- warmer land and sea surfaces, and the result is an
pecting a link between greenhouse warming and increased potential for major precipitation events in
weather extremes is related to the potential effects many parts of the world. Because of changes in large-
of a warmer climate on the physical processes that scale circulation patterns as well as regional differences
generate different types of weather events. Consider in hydrological processes, the resulting increase in pre-
the example of rainfall. Precipitation is one half of the cipitation will not be spread uniformly around the
hydrological (or water) cycle. Evaporation (and tran- world. In fact, some areas may receive less precipita-
spiration from plants) is the other. A virtually certain tion. However, climate models indicate that a warmer

14
Extreme Weather and Climate Change
atmosphere will increase the amount of moisture number of these events would greatly increase the
transported into the middle and high latitudes of the danger of rapid flooding and related problems such as
Northern Hemisphere. Thus, these areas will be erosion and landslides.
exposed to more rain and snow, and it is there that Paradoxically, an increase in average global pre-
increases in heavy precipitation are likely to be most cipitation could also be accompanied by more fre-
noticeable. quent droughts in many parts of the world. This
This argument, of course, looks at only the most would obviously be true in those areas where circula-
basic (although the most important) physical relation- tion changes cause rainfall to decrease, but droughts
ships between temperature and rainfall. In reality, could also become more frequent even where there
many other factors, such as soil moisture, land- was an increase in rainfall. One reason is that evapo-
scape features, vegetation, season, the physics of ration increases rapidly with temperature. In a
cloud formation, and atmospheric circulation patterns warmer climate, increased evaporation from soils
must also be considered. To estimate the net effect (and transpiration from plants) could therefore offset,
of global warming on precipitation – or any other or more than offset, any input of additional moisture
climate variable for that matter – we must turn to from heavier rainfall. Another reason is that the dis-
elaborate computer models of the climate system. proportionate increase in extremely heavy rainfall
These are known as general circulation models (or means that more of a region’s rain will come from
GCMs). What they do, essentially, is combine what fewer events, with a resulting increase, as several
we know about how the major climate processes work modelling studies have shown, in the number of dry
and then calculate what happens when these process- days. In addition, heavy downpours do a poor job of
es interact for a period of time and key variables, like recharging soil moisture, since much of the water is
the concentration of greenhouse gases, are changed. lost as surface runoff. In areas where precipitation
The results of these exercises contain a number of decreases, these effects could be devastating. In a
uncertainties and sometimes differ considerably from recent study of southern Europe, for example, aver-
one model to another, but they still offer a number of age precipitation declined by 22% with a doubling of
useful insights into the influence of greenhouse warm- carbon dioxide, but the probability of a 30-day dry
ing on climate processes. spell increased two to five times.
One point on which the models agree is that In a midlatitude country like Canada, heavy rain-
warmer climates bring an increase in average global falls (and snowfalls) are usually associated with storms
precipitation. In addition, all of them indicate that the and high winds. If extreme precipitation increases,
greatest increase will take place in the middle and then presumably severe storms would too. Most storm
higher latitudes in winter. Most of them also show an damage comes from three types of storms – tropical
increase in soil moisture in late winter and spring in storms and hurricanes, large extratropical storms, and
the midlatitudes of the Northern Hemisphere. In a thunderstorms. However, these are caused by quite
place like Canada these factors could easily translate different processes, some of which may be more sensi-
into a worst-case scenario for spring flooding – large tive to a warmer climate than others. Consequently,
accumulations of snow, heavy rains, and water-satu- each of these has to be looked at separately.
rated soils. Indeed, the Red River flood of 1997 is a Tropical storms and hurricanes are potentially sensi-
dramatic illustration of such a combination of factors. tive to greenhouse warming in that their formation is
What is even more interesting, when we look restricted to ocean areas where the sea surface tem-
beyond the averages, is that many of the models indi- perature is greater than 27°C. Since greenhouse
cate a substantial increase in heavy precipitation. Two warming may cause a greater area of ocean to reach
studies, for example, show rainfall increasing by this temperature more often, it follows that the zone of
about 10–30% in most latitudes when the amount of hurricane activity could expand and the number of
carbon dioxide in the atmosphere is doubled. hurricanes could increase. There does indeed appear
However, the worst rainfall events – the most extreme to be a good correlation between sea surface temper-
10% – increased by 50% in some regions. A rise in the atures and hurricane frequency – according to one

15 Extreme Weather and Climate Change

 FLOODING IN CANADA
• According to the International Red Cross, floods accounted for the largest num-
ber of disasters globally between 1967 and 1991. In 1996, a “normal” year for
catastrophes according to the reinsurance firm Munich Re, floods were respon-
sible for nearly $35 billion in economic losses or about 57% of natural disaster
losses globally. In Canada, federal payouts for flood losses have exceeded those for
any other type of weather disaster.

• In Canada, flood risks vary with place and season. In early spring, rapidly melting
snowpacks and ice jams pose a risk of flooding on rivers in almost every part of the
country. In summer, thunderstorms are the most frequent cause of floods in all
inland areas of southern Canada. These floods are the result of brief but intensive
downpours and are usually restricted to a small area. Floods caused by
prolonged rainstorms associated with large weather systems are most common
in the spring and fall. Over a two- or three-day period, these storms can drop mas-
sive amounts of water over a large area. The rainstorm that caused the 1996
Saguenay flood, for example, dropped an average of nearly 126 mm of rain over a
100,000 km2 area in 48 hours – enough to keep the St. Lawrence River flowing
for nearly half a month at its average rate of 10,100 m3 per second. The 1997
Red River flood, in contrast, was caused by a combination of factors – a wet
autumn that left soils saturated with moisture, heavier-than-normal snowfall during
the winter, and an early April blizzard that dropped another 50–70 cm of snow and
freezing rain shortly after runoff had begun.

Emerson, Manitoba, April 23, 1997

16
Extreme Weather and Climate Change
 • Overall, the risk of flooding in Canada would increase as a result of a warmer cli-
mate. This increase would come mainly from rainstorm floods, with heavier rainfall
expected to come from more (and possibly more severe) thunderstorms and from
fewer but larger rainstorms associated with large-scale weather systems. Shorter
winters, however, may reduce the risk of snowmelt and ice-jam floods in some
areas, although heavier snowfalls could add to the risk in others.

• Increased flooding risks will require more effective adaptation measures if damage is
to be minimized. Such measures can be expensive initially, but can prevent much
higher costs when heavy floods eventually occur. Manitoba, for example, has invest-
ed heavily in flood protection, mainly because the Red is naturally flood-prone in spring
(due to the fact that it is a north-flowing river and its exit into Lake Winnipeg often
remains frozen after upstream sections to the south have begun to thaw). The Red
River Floodway, built to divert floodwaters around Winnipeg, was completed in 1968
at a cost of $68 million. Without it, the 1997 flood would have left nearly 80% of
the city underwater and forced the evacuation of more than 550,000 residents. In
addition, eight towns and approximately 700 rural homes are protected by dikes.

study, record sea surface temperatures in 1995 The current generation of general circulation
accounted for 61% of the very large number of hurri- models can simulate some of the characteristics of hur-
canes that occurred that year. However, nearly half a ricane behaviour, but the amount of information they
dozen other conditions have to be met before a hurri- can provide is limited because hurricanes are too small
cane can develop, and it is not known whether these for the models to simulate in detail. High resolution
would become more or less common in a warmer cli- regional models or special hurricane models might
mate. Projections from model studies of tropical storm provide some additional insight, but at the moment
behaviour carried out at the Max Planck Institute in there are few results to go on. One of the things that
Germany, though, show a significant reduction in GCMs can tell us, though, is whether there might be a
hurricane activity, especially in the Southern change in the large circulation features that determine
Hemisphere, as a result of a warmer climate. the world’s major prevailing wind patterns and set lim-
Hurricane intensity could also be affected by a its to the poleward range of hurricanes. The evidence
warmer climate. That is because the theoretical limit available from current models indicates little change in
for hurricane strength depends upon the extent of the these features and hence little change in the latitudes
local energy imbalance between the atmosphere and affected by hurricanes.
the ocean. Experiments with climate models suggest What happens to hurricane behaviour in a
that this imbalance will increase in a warmer world, warmer world might also depend on what happens
thus significantly increasing the potential intensity of to El Niños. Because they tend to suppress hurricane
hurricanes. A team led by Thomas Knutson of formation in the Atlantic and support it in the
the U.S. National Oceanic and Atmospheric Pacific, stronger and more frequent El Niños could
Administration, for example, used a regional hurri- cause a geographical shift in hurricane activity, with
cane prediction model to look at how hurricane more hurricanes occurring in the Pacific and fewer
behaviour in the western Pacific might be affected by in the Atlantic. A weakening of El Niño activity, on
a warmer climate. Their model showed a 5–10% the other hand, could shift the balance in the other
increase in hurricane wind speeds as a result of a direction.
2.2°C warming of the sea surface.

17 Extreme Weather and Climate Change

 Regardless of how global warming affects either Global warming could affect this exchange of
the frequency or the intensity of hurricanes, the risk of heat in a variety of ways. If the poles warm more than
accompanying flood damage will almost certainly be the tropics, as they are expected to do, the tempera-
accentuated by rising sea levels. Over the past century, ture differences that drive these storms would dimin-
global warming has caused the average sea level to rise ish and the number of severe storms would likely
by 10–25 cm, and another half-metre increase is decrease. In spite of that, however, some regional tem-
expected by the end of the 21st century. Since flooding perature differences might become greater. The conti-
from hurricane storm surges usually causes more dam- nents, for example, are expected to warm more than
age than wind, the risk of high death tolls and heavy the oceans, and the increasing contrast in temperature
property damage would actually rise even without an between the two could increase the power of coastal
increase in hurricane intensity or frequency. storms. Yet another factor to be considered is water
Extratropical storms are the large low pressure vapour. The heat given off by condensing water
areas or depressions that commonly form and travel vapour is a source of energy for storms, and the extra
along the boundaries between contrasting air masses water vapour held by a warmer atmosphere could
in the middle and high latitudes. The energy for these provide the additional energy needed to sustain more
storms comes mainly from the transfer of heat between powerful storms.
the sub-tropics and the polar regions, and the paths At the moment, it is unclear which of these effects
they follow are roughly determined by the position would predominate, and experiments with models of
and contours of the sub-tropical and polar jet streams. warmer climates have not resolved the dilemma.

Changes in rainfall intensity with

a doubling of carbon dioxide

200
Per Cent Change in Frequency

150

100

50

-50
0.2–0.4 0.8–1.6 3.2–6.4 12.8–25.6
0.4–0.8 1.6–3.2 6.4–12.8 25.6+
Rainfall Intensity (mm/day)

Source: Adapted from Gordon et al. (1992)

Climate models indicate that a warmer climate will bring more precipitation. Some
studies also indicate that heavy rainfalls will become more frequent while light rainfalls
will occur less often. Results from an Australian study, shown here for central North
America, project a striking increase in the heaviest rainfall categories.

18
Extreme Weather and Climate Change
Experimenters using two models developed by the thunderheads. Since a warmer, moister climate will
United Kingdom Meteorological Office found that favour this process, thunderstorm activity can be
the major storm tracks in the Northern Hemisphere expected to increase. Hotter temperatures at the sur-
shifted northward and storm activity intensified, par- face and higher humidity levels will also make the air in
ticularly in the eastern Atlantic and western Europe. thunderstorm cells rise faster and higher, creating more
A Canadian model, on the other hand, showed fewer dynamic storm clouds and more powerful storms.
storms and no significant change in the position of the A significant byproduct of this process would be
storm tracks. However, it did project a substantial an increase in lightning. Observational evidence shows
increase in the number of severe winter storms. a connection between lightning activity around the
Thunderstorms form when heating of the ground world and variations in global surface temperature.
causes warm, humid air to rise and condense in colder Studies using the output from climate models indicate
air above. The heat from the condensing moisture a similar linkage. An experiment at the Goddard
causes the warm air to rise further and leads to the Institute for Space Studies, for example, showed a 6%
rapid buildup of towering cumulonimbus clouds or increase in lightning activity for every 1°C rise in the

Change of frequency and intensity of

winter storms in a 2xCO2 climate

30

25
Frequency of Storms (%)

20

15

10

0
1 2 3 4 5 6 7 8+

Storm Intensity Category

Source: Lambert (1995)

General circulation models have been used to compare storm frequencies under pre-
sent climate conditions and under warmer conditions resulting from a doubling of
greenhouse gases. These results, from a Canadian study, show the number of severe
storms north of 30°N latitude increasing in the warmer climate while the number of
less intense storms remains constant or decreases.

19 Extreme Weather and Climate Change

earth’s average temperature. Since lightning is a fre- an analysis of tornadoes in the Prairie Provinces has
quent cause of forest fires, a warmer climate would shown a tendency for tornado frequency to increase
likely increase the fire hazard, especially if summer dry in the spring and early summer in step with increases
spells become more prevalent as well. in average monthly temperature. It is therefore rea-
Hail is also likely to occur more often. Statistical sonable to assume that tornado outbreaks would
analysis of hailstorms in France shows a close corre- occur more often on the prairies as a result of a
lation between average summer nighttime tempera- warmer climate. Techniques that link local weather
tures (which have been increasing in France as well as phenomena to larger-scale climate patterns may pro-
in many other parts of the world) and hail activity. vide further insight into thunderstorm and tornado
The link between nighttime temperatures and hail- behaviour in a warmer world.
storms, which generally occur in the late afternoon, is Finally, what about temperature? As average
at first hard to fathom, but it depends on a relation- temperatures rise, it would be logical to expect the
ship that is well known to weather forecasters. number of extremely hot days to increase. However,
Minimum nighttime temperatures are influenced by there are good reasons for believing that the increase
the amount of water vapour in the air, and for that could be considerable and out of all proportion to the
reason they are a good predictor
of the following afternoon’s dew-
point – the temperature at which
the air is saturated with water
vapour. The afternoon dewpoint,
Mean summer minimum
in turn, is a useful predictor of temperature and hail
unsettled weather, since it indi- severity in France
cates how much water vapour is
available to fuel a developing
storm. The rise in summer mini- 1
mum temperatures is therefore
0.8
Hail Severity Index

an indicator that the atmosphere

is more prone to thunderstorm,
0.6
and hence hail, activity.
Tornadoes and powerful 0.4
supercell thunderstorms require
special additional conditions for 0.2
their formation, particularly an
inflow of cold, dry air at higher 0
altitudes. Whether or not these 12 13 14 15 16
conditions would become more Mean Minimum Temperature (°C)
prevalent in a warmer climate is
unknown. The large general cir-
culation models used for studies
Source: Adapted from Dessens (1995)
of climate change tell us little
about the behaviour of thunder-
Research into hailstorms in France shows hail activ-
storms because individual thun-
derstorm cells are much smaller
ity increasing when average overnight low tempera-
than the smallest details these tures are higher. This linkage suggests that hail-
models can reproduce. However, storms (and severe thunderstorms) would occur
more often in a warmer climate.

20
Extreme Weather and Climate Change
increase in average temperature. One of these reasons low value in the left-hand curve becomes a very high
involves the mathematical behaviour of frequency dis- value in the right-hand curve. The actual difference,
tributions. The left-hand curve in the graph below measured by comparing the areas under each of the
shows a typical distribution pattern for measurements curves to the right of the 17.3° line, amounts to a fac-
of average summer temperatures. This particular tor of 25. What this means is that an extremely warm
graph is for a location in central England where summer that could be expected to occur only once in
weather records have been kept for 300 years, but every 75 years could be expected once every 3 years in
graphs for other places would show much the same the warmer climate. This is a hypothetical example,
pattern. The plot fits the familiar bell curve, with val- but it does show that a small change in the average can
ues near the average occurring most often and more result in very large changes in the extremes.
extreme values occurring less. In a warmer climate Another reason for expecting a substantial
where temperatures vary according to the same pat- increase in high temperature extremes comes from
tern, the curve keeps the same shape but moves to the experiments with climate models, which show a large
right, with some interesting results. In this example, increase in the number of extremely warm days and
the average temperature has increased by 1.6°C, but if heat waves as a result of global warming. One
we compare the number of summers with average Canadian study suggests that an increase of 4°C in
temperatures above 17.3°C, we see an enormous Toronto’s average temperature would likely increase
change between the two graphs. What had been a very the risk of summer days with temperatures exceeding

Changes in the frequency of

extreme temperatures with a change in
the mean temperature

Source: Adapted from Munich Re (1996)

Small changes in the average can result in large changes in extremes. In this exam-
ple taken from central England, an increase of 1.6°C in the average summer
temperature causes the probability of a very warm summer (exceeding 3 standard
deviations) to increase from once in 75 years (a probability of 1.3%) to once in 3 years
(a probability of 33.3%).

21 Extreme Weather and Climate Change

30.5°C from 1 in 10 to almost 1 in 2. Estimates for storm of January 1998, which left 25 dead, stranded
other parts of the world show similar results. An nearly 3 million people without heat or electricity,
Australian study, for example, concluded that a 0.5°C forced the mobilization of 15,000 troops, and caused
rise in average temperature would increase the num- an estimated $2 billion in damage. It should be
ber of extremely hot days (above 35°C) in the state of remembered, though, that even relatively small-scale
Victoria by 25%. With a rise of 1.5°C, the number of events can result in impressively large repair bills. A
hot days would increase by 50–100%. 90-minute downpour in Ottawa in August 1996, for
But if a warmer climate increases extremes at the example, caused more than $20 million in insured
high end of the temperature scale, it should also logi- losses plus additional costs for repairs to roads and
cally reduce extremes at the low end. Simulations sewers.
with models tend to bear this out. In the Australian At the other end of the hydrological spectrum,
study, for example, the possibility of a 5-day cold spell drought can cause agricultural losses in the billions.
below 0°C was reduced by 20–40%, depending on the In 1988, Canada’s grain production was reduced by
region considered. Similarly, calculations by nearly a third because of drought, and the resulting
Canadian climate modeller Francis Zwiers showed export losses were estimated at $4 billion. Hot, dry
that the lowest temperature that Winnipegers might summers also provide ideal conditions for forest fires,
expect to face in a twenty-year period would rise from and the difference between fire management costs in
the present –45°C to a balmier –33°C in the kind of a good forest fire year as opposed to a bad one can
climate that could be expected if carbon dioxide con- amount to as much as $200 million or more.
centrations in the atmosphere were to double. Internationally, the costs of extreme weather
are often higher. In 1996 a series of floods in China
Implications of an Increase in resulted in more than 3000 deaths, left 5 million
Weather Extremes homeless, closed 8000 factories, and killed tens of
A possible connection between global warming thousands of animals, resulting in economic losses
and extreme weather is of more than academic inter- estimated at US$20 billion. The North Korean
est. The economic, human, and ecological costs of famine, which reached catastrophic proportions in
even a small long-term increase in extremes would be 1997, was in part the result of floods in 1995 and 1996
substantial. Just how substantial can be visualized if followed by drought in 1997. In Jamaica, losses from
we consider what extreme weather already costs us. Hurricane Gilbert in 1988 amounted to about one
In Canada, individual floods have caused as many third of the country’s gross national product, and
as 79 deaths (Hurricane Hazel in 1954), displaced more than 40% of the island’s housing was destroyed.
more than 25,000 people (the Red River flood of Recovering from such disasters places a major strain
1997), and resulted in property damage and other on countries with limited economic means and may
costs exceeding $1 billion (the Saguenay flood of intensify the demand on resource bases and ecosys-
1996). A single hailstorm caused $120 million worth tems that are already overstressed. Aid from industri-
of damage in Winnipeg in July 1996, and insured alized countries such as Canada also plays an essential
losses from other hailstorms have gone as high as role in disaster recovery. If weather disasters become
$400 million, as they did in Calgary in 1991. The mas- more frequent, the demand for aid will increase
sive snowstorm that struck Victoria and B.C.’s lower accordingly. If an increase in weather extremes is per-
mainland at the end of 1996 resulted in an estimated suasively linked to greenhouse warming and climate
$200 million in losses, caused 1700 people to seek change, the moral obligation of industrialized coun-
emergency food and shelter, left 150,000 homes with- tries to provide disaster relief will also be considerably
out power, and blocked some roads for as long as 10 greater.
days. The storm also contributed to 500 avalanches The escalating costs of weather-related property
and 65 pollution spills. The most destructive single losses over the past couple of decades are already
weather event to date, however, was the eastern ice straining the resources of the world’s insurance indus-

22
Extreme Weather and Climate Change
try. Since 1987 it has had to deal with at least one damage from droughts, floods, and storms could
$1 billion disaster almost every year. Some large make famines not only more frequent but far more
insurance and re-insurance companies have already difficult to deal with.
been forced into bankruptcy by weather-related
claims, and some, such as Lloyd’s of London, have Responses
faced solvency crises at least in part because of losses It is clear that weather extremes are becoming an
from weather disasters. If these trends continue, increasingly serious problem for our society. There is
insurers will have to raise premiums, limit their liabil- also a reasonable probability that global warming will
ity, or even back out of existing markets that have make the problem worse. The difficulty is that we
become too risky, as many hurricane insurers have in don’t know for sure the extent to which the present
the Carribean. Increasingly, governments may have wave of extremes is a natural climatic phenomenon,
to provide protection against risks that insurers will nor do we know the real potential for the intensifica-
no longer cover. Governments, and hence taxpayers, tion of extreme weather in a warmer climate. To
will also face increasing costs for emergency services, diminish these uncertainties, we have to devote much
disaster relief, and disaster-related foreign aid. more scientific effort to the study of severe weather as
Other sectors of society and the economy would a feature of climate. This effort will have to focus on
also face rising costs in order to counter the risks of three areas in particular.
more dangerous and destructive weather. Structures First, we need to have a much better grasp of pre-
such as buildings, dams, bridges, and hydro towers sent and emerging trends in severe weather. One of
would have to be built to withstand heavier stresses, the chief difficulties here is that extreme weather
transportation services would face more frequent inter- events, by definition, occur infrequently. They are also
ruptions and the possibility of greater damage to their usually limited in area and often short in duration.
equipment and infrastructure, and farmers would have Consequently, they tend to leave a weak statistical trail
to improve irrigation and alter cropping practices to or none at all. However, the occurrence of many of
deal with the increased occurrence of drought. Failure these events can be inferred through analysis of the
to take these precautions would only result in higher larger-scale weather patterns to which they are relat-
economic costs, greater social disruption, and ed. The existence of storms and the intensity of winds,
increased loss of life when a disaster eventually strikes. for example, can be estimated from regional pressure
The ecological impacts of extreme weather on patterns. There are also problems with the compara-
natural and agricultural ecosystems should not be bility or quality of historical climate records.
overlooked either. Events such as the 1998 ice storm Instruments change over time and so do the physical
in eastern Canada or the 1987 wind storm in southern characteristics of weather stations and their surround-
England, for example, weaken or destroy millions of ings. These factors can introduce false trends into
trees which, in their healthy state, provide food and long-term climate analyses, but careful analysis can
habitat for wildlife and a sink for absorbing carbon detect these problems and it is often possible to intro-
dioxide. Droughts increase the risk of forest fires and duce statistical corrections to compensate for them.
hasten the dieback of forests weakened by other The second requirement is for a better under-
stresses. Floods wash away scarce topsoil and often standing of natural climatic variability. Some of this
diminish the productivity of agricultural lands. Losses can be derived from the instrumental climate record –
from individual events may not be an insurmountable that is, the record of weather observations taken with
problem, but over time and in combination with other properly calibrated instruments under controlled con-
impacts, the cumulative effect of severe weather on ditions. For some places, this information may extend
ecosystems could be considerable. back a century and a half or more, but for much of the
Food security presents a particularly worrisome world it is considerably shorter. To detect longer-term
problem, given the continuing growth of the world’s swings in climate, it is necessary to turn to palaeo-
population and the levelling off in production of key climatic data, the information that comes from tree
foodstuffs such as wheat and rice. Increased crop rings, sediments, ice cores, fossils, and other natural

23 Extreme Weather and Climate Change

 THE 1998 ICE STORM
Freezing rain is a common experience in most parts of Canada. Although Victoria
averages only 2 hours of it per year, places such as Yellowknife, Regina, and Toronto get 35
hours in the course of an average year, while St. John’s endures a formidable 148 hours.
Episodes of freezing rain are generally short, but from time to time they can develop into major
ice storms that are notable both for their sparkling beauty and for the crushing weight of ice
they leave on power lines and trees. Damaging ice storms have occurred as recently as 1986
in Ottawa and 1984 in St. John’s, but the storm that struck much of eastern Canada in
January 1998 was remarkable for its persistence, its extent, and, ultimately, the magnitude
of its destruction. Over a period of six days, up to 100 mm of freezing rain fell intermittently
over an area that at one point extended from central Ontario to Prince Edward Island. Downed
power lines left nearly three million people without electricity. The Montreal area was worst
affected, and some localities south of the city were without power for as long as five weeks.
The storm was blamed for 25 deaths, and with initial estimates of damages in the range of
$1–2 billion, it is by far the costliest weather catastrophe in Canadian history.

Freezing rain occurs when warm, moist air that is above the freezing point overlies cold air
at ground level that is below the freezing point. Because rain falling from the warmer air is
supercooled as it passes through the colder layer, it freezes instantly on contact with the
ground. This situation commonly occurs along fronts where warm and cold air masses on
opposite sides of the freezing point meet, but the fronts generally move along quickly and the
freezing rain conditions are short-lived. The 1998 ice storm, however, was the product of
quite unusual conditions – an El Niño–related sub-tropical jet stream that brought warm,
moist air out of the American south, a steady flow out of the northeast that maintained a
shallow layer of cold air in the lowlands of the Ottawa and St. Lawrence valleys, and finally a
stagnant ridge of high pressure over the Atlantic that kept the whole system in place.

In a warming climate, milder winter temperatures could possibly cause an increase in freez-
ing rain in places where average daily temperatures begin to meander around the freezing
point instead of remaining firmly
below it. However, this does not
necessarily translate into the more
frequent occurrence of events such
as the ice storm of 1998. This was
the result of the simultaneous
occurrence of at least three special
factors, and there is no compelling
reason at the moment to assume
that a warmer climate would make
it significantly more likely for such
an event to happen again. Ottawa, January 1998

24
Extreme Weather and Climate Change
evidence that may bear an imprint of past climate on the outcome. To determine the net effect of the
changes. Quite a bit can be learned in this way about interactions between these processes, we must rely on
previous variations in temperature and rainfall, climate models, but our present general circulation
although little can be found out about other extremes models cannot produce the fine detail needed to study
such as storms. Historical documents, such as mem- regional climate behaviour and small-scale features
oirs, narratives, ships’ logs, and insurance records, such as thunderstorms and hurricanes. To get around
can also provide direct, though qualitative, evidence this problem, climate modellers are developing tech-
of severe weather in the past. Indirect historical evi- niques for estimating the occurrence of weather
dence, such as records of grain prices, can also point extremes from the larger-scale patterns that the
to the occurrence of droughts and wet spells and models now produce. More can also be learned by
other climatic factors that affect the supply of agri- using physical process models to study the dynamics
cultural commodities. of events such as hurricanes and thunderstorms under
Finally, the possible connection between global conditions that might prevail as a result of global
warming and various types of weather extremes needs warming. Yet another line of inquiry involves the use
to be explored more thoroughly. We know of mecha- of high resolution regional models to get a better look
nisms by which global warming could increase the at how the climates of specific areas would behave
frequency or intensity of severe weather, but other under different kinds of warming scenarios. Such
physical processes could also have an important effect regional studies are particularly important, because

Climate adaptation measures, such as Winnipeg’s Red River Floodway, shown here in April
1997, are one way of dealing with weather extremes. Although these measures may be
expensive, the return in damages avoided is usually much greater. If adaptive measures are
overwhelmed by unexpectedly extreme impacts, however, communities can be faced with
sudden and catastrophic costs. Improved adaptive measures – such as the expansion of
flood control systems, a tightening of land use regulations, and additional reinforcement of
buildings and other structures – would be an essential response to any increase in weather
extremes.

25 Extreme Weather and Climate Change

the effects of global warming on regional climates and warming could increase both the frequency and the
particularly on extreme weather is likely to vary sub- intensity of some kinds of extreme weather events.
stantially from one part of the world to another. Our understanding of climate processes, for example,
Regardless of the outcome of these studies, it is suggests very strongly that heavy rainfalls and thun-
already clear that modern societies are becoming derstorms will be more common in a warmer climate.
more vulnerable to economic damage from weather Heat waves and hot days would also increase. As for
extremes as their infrastructures become more elabo- other types of extreme events, such as extratropical
rate and their populations grow larger and more con- storms and hurricanes, the arguments are conflicting.
centrated. In fact, one of the lessons of the 1998 ice Finding conclusive proof that a connection
storm and the recent Red River and Saguenay floods between greenhouse warming and weather extremes
is that we in Canada have underestimated both the does or does not exist will not be easy, at least for
risks and the consequences of at least some of the another decade or two. For one thing, we simply do
major weather hazards not know enough about
that we now face. It is natural climate variability
important, therefore, for Certainly there is evidence of to have a reliable bench-
us to reassess our esti- mark of comparison. For
increasing extremes in some parts
mates of the risks of another, global warming
weather disasasters and of the world. Although these is still in its early stages.
our ability to cope with Most scientific studies of
them. That means review- increases could simply be a result global warming are based
ing our calculations of on a doubled-CO2 atmos-
the probabilities of dif-
of natural climate variability, they phere – one in which
ferent kinds of weather are also consistent with many of atmospheric concentra-
extremes to determine tions of carbon dioxide
whether the event that the changes expected as a result of are double the pre-indus-
now has an estimated trial level of approxi-
probability of occurring greenhouse warming. mately 280 parts per mil-
once in 100 years should lion (ppm) or some more
be upgraded to, say, a recent value. The present
1 in 50 probability. That kind of change would, in concentration of about 360 ppm is only about 30%
turn, necessitate a serious second look at many of our above the pre-industrial base. If we add in the effects
zoning laws, building codes, and engineering stan- of other greenhouse gases, we are still only about
dards as well as our plans and capabilities for dealing halfway to the equivalent of a CO2 doubling. In addi-
with emergencies. Evidence of a significant link tion, the effect of the increase in greenhouse gases is to
between global warming and extreme weather events some extent offset by the sunlight-scattering effects of
would make it even more imperative to conduct such high concentrations of sulphate aerosols over and
a review, and it would set the bar even higher when it downwind of the industrialized regions of the globe.
came to defining acceptable levels of protection. Another problem is that natural climate variability will
sometimes reinforce and sometimes moderate any
Drawing Conclusions effects that greenhouse warming might have on
Is global warming loading the dice in favour of extreme weather formation. As a result, changes in the
extreme weather? Certainly there is evidence of occurrence of extreme weather events will be irregular
increasing extremes in some parts of the world. and the detection of trends will be complicated.
Although these increases could simply be a result of Carbon dioxide concentrations are expected to
natural climate variability, they are also consistent reach the doubling point within the next 50–75 years,
with many of the changes expected as a result of and they will continue rising unless decisive action is
greenhouse warming. We also know that there are taken to stabilize or reduce emissions. Studies with
plausible physical mechanisms by which global transient climate change models that simulate the

26
Extreme Weather and Climate Change
 THE POSSIBILITY OF SURPRISES
Models and other scientific tools can indicate some of the more likely outcomes of cli-
mate change, but we cannot be sure that they will identify all of them. Climate is the
product of a bafflingly complex array of processes, and the possibility of major sur-
prises is always present. Climates of the past, rather than changing gradually, have
sometimes shown a tendency, at least regionally, to switch abruptly to a radically dif-
ferent mode. A surprise of this sort could have devastating consequences, both for
natural ecosystems and human societies.

The course of climate change could veer into unanticipated territory, for example, if
extensive thawing in the world’s subpolar regions released the vast amounts of
methane now locked up as frozen hydrates in the permafrost of these areas. This
would result in a rapid intensification of the greenhouse effect that would not only
increase extremes of heat but would greatly distort other climate patterns in many
parts of the globe.

Another possibility is related to the processes that drive the “Atlantic conveyor” – the
sub-surface flow of warm water into the North Atlantic that gives western Europe its
mild winters. The warm water is drawn northward to replace cold salty water that
plunges to the ocean floor at various locations in the northern North Atlantic. This
process of deep water formation is the essential driving mechanism for the Atlantic
conveyor, but it only works when the water is very salty and cold. Either of these fac-
tors could be easily altered by a warming climate. Were the Atlantic conveyor to weak-
en or shut down entirely, the climate of western Europe would cool dramatically and
the climate of the Northern Hemisphere generally would become much more variable.
Such a change could take place within decades.

Some recent studies with combined ocean and climate models show the Atlantic con-
veyor weakening and then returning to its former strength as the carbon dioxide con-
centration of the atmosphere gradually increases to double the pre-industrial value. At
four times the pre-industrial concentration, however, the conveyor remains weak for
up to 500 years. Other studies suggest that the Atlantic conveyor could become less
stable and more variable in a warmer world.

27 Extreme Weather and Climate Change

gradual evolution of an enhanced greenhouse effect kinds of extremes that we are seeing now could
suggest that its impact on global climate will become become increasingly common. In the meantime, we
increasingly apparent somewhere between 2010 and must do as much as we can, not only to increase our
2030 – well within the lifetime of children born today. understanding of future risks but also to improve our
If greenhouse warming is forcing an increase in adaptation to present hazards. An ounce of preven-
weather extremes, it should be clear by then, and the tion is still worth a pound of cure.

Bibliography
Is Extreme Weather Becoming Karl, T.R., R.W. Knight, D.R. Easterling, and R.G.
More Common? Quayle, 1996. Indices of climate change for the
United States. Bulletin of the American Meteorological
Born, K., 1996. Tropospheric warming and changes Society, 77, 279-292.
in weather variability over the northern hemisphere
during the period 1967–1991, Meteorology and Lambert, Steven J., 1996. Intense extratropical
Atmospheric Physics, 59, 201–215. northern hemisphere winter cyclone events:
1899–1991. Journal of Geophysical Research, 101,
Davis, R.E. and R. Dolan, 1993. Nor’easters. 21219–21325.
American Scientist, 81, 428–439.
Landsea, C.W., N. Nicholls, W.M. Gray, and L.A.
Dessens, J., 1995. Severe convection weather in the Avila, 1996. Downward trend in the frequency of
context of a nighttime global warming. Geophysical intense Atlantic hurricanes during the past five
Research Letters, 22, 1241–1244. decades. Geophysical Research Letters, 23, 1697-1700.

Environment Canada, 1995. The state of Canada’s cli- London, S.J., S.G. Warren, and C.J. Hahn, 1991.
mate: monitoring variability and change. Environment Thirty-year trend of observed greenhouse clouds over
Canada, Downsview, Ontario. the tropical oceans. Advances in Space Research, 11(3),
45–49.
Etkin, D. Beyond the year 2000: More tornadoes in
western Canada? Implications from the historical McCulloch, J. and D. Etkin, eds., 1993. Proceedings of
record. Natural Hazards, 12, 19–27. a workshop on improving responses to atmospheric extremes:
The role of insurance and compensation. Environment
Kaas, E., T-S. Li, and T. Schmith, 1996. Statistical Canada, Downsview, Ontario.
hindcast of wind climatology in the North Atlantic
and northwestern European region. Climate Research, Munich Re, 1997. Topics: Annual review of natural cata-
7, 97–110. strophes. Munich Re, Munich.

Karl, T.R., R.W. Knight, and N. Plummer, 1995.

Trends in high-frequency climate variability in the
twentieth century. Nature, 377, 217–220.

28
Extreme Weather and Climate Change
Nicholls, N., G.V. Gruza, J. Jouzel, T.R. Karl, L.A. Natural Variability
Ogallo, and D.E. Parker, 1996. Observed climate
variability and change. In Climate change 1995: The sci- Burroughs, W.J., 1992. Weather cycles: Real or imagi-
ence of climate change. Contribution of Working Group nary? Cambridge University Press, Cambridge.
I to the Second Assessment Report of the
Intergovernmental Panel on Climate Change. J.T. Environment Canada, 1995. The state of Canada’s cli-
Houghton, L.G. Meira Filho, B.A. Callander, N. mate: monitoring variability and change. Environment
Harris, A. Kattenberg, and K. Maskell, eds. Canada, Downsview, Ontario.
Cambridge University Press, Cambridge.
Folland, C.K., T.R. Karl, and K.Ya. Vinnikov, 1990.
Ostby, F.P., 1993. The changing nature of tornado cli- Observed climate variations and change. In Climate
matology. In Preprints, 17th conference on severe local change: The IPCC scientific assessment. J.T. Houghton,
storms, October 4–8, 1993, St. Louis, Missouri, pp. 1–5. G.J. Jenkins, and J.J. Ephraums, eds. Cambridge
University Press, Cambridge.
Pearce, D.W., W.R. Cline, A.N. Achanta, S.
Fankhauser, R.K. Pachauri, R.S.J. Tol, and P. Foukal, P. and J. Lean, 1990. An empirical model of
Vellinga, 1996. The social costs of climate change: total solar irradiance variation between 1874 and
Greenhouse damage and the benefits of control. In 1988. Science, 247, 556–558.
Climate change 1995: Economic and social dimensions of
climate change. Contribution of Working Group III to Haigh, J.D., 1994. The role of stratospheric ozone in
the Second Assessment Report of the Intergovern- modulating the solar radiative forcing of climate.
mental Panel on Climate Change. J. Bruce, H. Lee, Nature, 370, 544–546.
and E. Haites, eds. Cambridge University Press,
Cambridge. Haigh, J.D., 1996. The impact of solar variability on
climate. Science, 272, 981-984.
Pielke, Roger A., Jr. and Christopher W. Landsea,
1997. Normalized hurricane damages in the United Hansen, J.E., A. Lacis, R. Ruedy, M. Sato, and H.
States: 1925–1995. Draft paper. National Center for Wilson, 1993. How sensitive is the world’s climate?
Atmospheric Research, Boulder, Colorado. National Geographic Research and Exploration, 9,
142–158.
Schmidt, H. and H. von Storch, 1993. German Bight
storms analysed. Nature, 365, 791. Lean, J., J. Beer, and R. Bradley, 1995. Reconstruc-
tion of solar irradiance since 1610: Implications for
Suppiah, R. and K.J. Hennessy, 1996. Trends in the climate change. Geophysical Research Letters, 22,
intensity and frequency of heavy rainfall in tropical 3195–3198.
Australia and links with the southern oscillation.
Australian Meteorological Magazine, 45, 1–18. Morton, Oliver. The storm in the machine. New
Scientist, 157 (2119), 22–27.
Tsonis, A.A., 1996. Widespread increases in low-fre-
quency variability of precipitation over the past cen- Shine, K.P., Y. Fouquart, V. Ramaswamy, S.
tury. Nature, 382, 700–702. Solomon, and J. Srinivasan, 1995. Radiative forcing.
In Climate change 1994. J.T. Houghton, L.G. Meira
Von Storch, H., J. Guddak, K.A. Iden, T. Jùnsen, J. Filho, J. Bruce, H. Lee, B.A. Callander, E. Haites, N.
Perlwitz, M. Reistad, J. de Ronde, H. Schmidt, and Harris, and K. Maskell, eds. Cambridge University
E. Zorita, 1993. Changing statistics of storms in the North Press, Cambridge.
Atlantic? Report No. 116. Max-Planck-Institut für
Meteorologie, Hamburg.

29 Extreme Weather and Climate Change

Greenhouse Warming and itation over Europe in the current and 2 x CO2 cli-
Weather Extremes mates using the UKMO climate model. Quarterly
Journal of the Royal Meteorological Society, 121,
Bengtsson, L., M. Botzet, and M. Esch, 1996. Will 1451–1476.
greenhouse gas–induced warming over the next 50
years lead to higher frequency and greater intensity of Hall, N.M.J., B.J. Hoskins, P.J. Valdes, and C.A.
hurricanes? Tellus, 48A, 57–73. Senior, 1994. Storm tracks in a high resolution GCM
with doubled CO2. Quarterly Journal of the Royal
Broccoli, A.J., S. Manabe, and J.F.B. Mitchell, 1995. Meteorological Society, 120, 1209–1230.
Comments on “Global climate change and tropical
cyclones:” Part II. Bulletin of the American Meteorological Hennessey, K.J. and A.B. Pittock, 1995. Greenhouse
Society, 76, 2244–2245. warming and threshold temperature events in
Victoria, Australia. International Journal of Climatology,
Carnell, R.E., C.A. Senior, and J.F.B. Mitchell, 1996. 15, 591–612.
An assessment of measures of storminess: simulated
changes in northern hemisphere winter due to Kattenberg, A., F. Giorgi, H. Grassi, G.A. Meehl,
increasing CO2. Climate Dynamics, 12, 467–476. J.F.B. Mitchell, R.J. Stouffer, T. Tokioka, A. J.
Weaver, and T.M.L. Wigley, 1996. Climate models –
Cubasch, U., J. Waszkewitz, G. Hegerl, and J. projections of future climate. In Climate change 1995:
Perlwitz, 1995. Regional climate changes as simulated The science of climate change. Contribution of Working
in time-slice experiments. Climate Change, 31, Group I to the Second Assessment Report of the
273–304. Intergovernmental Panel on Climate Change. J. T.
Houghton, L. G. Meira Filho, B.A. Callander, N.
Dessens, J., 1995. Severe convection weather in the Harris, A. Kattenberg, and K. Maskell, eds.
context of a nighttime global warming. Geophysical Cambridge University Press, Cambridge.
Research Letters, 22, 1241–1244.
Knutson, T.R., R.E. Tuleya, and Y. Kurihara, 1998.
Emanuel, K.A., 1995. Comments on “Global climate Simulated increase of hurricane intensities in a CO2-
change and tropical cyclones:” Part I. Bulletin of the warmed climate. Science, 279, 1018–1020.
American Meteorological Society, 76, 2241–2243.
Lambert, S.J., 1995. The effect of enhanced green-
Etkin, D. Beyond the year 2000: More tornadoes in house warming on winter cyclone frequencies and
western Canada? Implications from the historical strengths. Journal of Climate, 8, 1447–1452.
record. Natural Hazards, 12, 19–27.
Lighthill, J, G. Holland, W. Gray, C. Landsea, G.
Fowler, A.M. and K.J. Hennessey, 1995. Potential Craig, J. Evans, Y. Kurihara, and C. Guard, 1994.
impacts of global warming on the frequency and mag- Global climate change and tropical cylones. Bulletin of
nitude of heavy precipitation. Natural Hazards, 11, the American Meteorological Society, 75, 2147–2157.
283–303.
O’Brien, J.J., T.S. Richards, and A.C. Davis, 1996.
Gordon, H.B., P.H. Whetton, A.B. Pittock, A.M. The effect of El Niño on U.S. landfalling hurricanes.
Fowler, and M.R. Haylock, 1992. Simulated changes Bulletin of the American Meteorological Society, 77,
in daily rainfall intensity due to the enhanced green- 773–774.
house effect: implications for extreme rainfall events.
Climate Dynamics, 8, 83–102. Price, C. and D. Rind, 1994. Possible implications of
global climate change on global lightning distribu-
Gregory, J.M. and J.F.B. Mitchell, 1995. Simulation tions and frequencies. Journal of Geophysical Research,
of daily variability of surface temperature and precip- 99, 10823–10831.

30
Extreme Weather and Climate Change
Saunders, M.A. and A.R. Harris, 1997. Statistical evi- Implications
dence links exceptional 1995 Atlantic hurricane sea-
son to record sea warming. Geophysical Research Letters, Etkin, D., 1997. Climate change and extreme events:
24, 1255–1258. Canada. Draft paper. Environmental Adaptation
Research Group, University of Toronto.
Sun, De-Zheng, 1997. El Niño: A coupled response
to radiative heating? Geophysical Research Letters, 24, Pearce, D.W., W.R. Cline, A.N. Achanta, S.
2031–2034. Fankhauser, R.K. Pachauri, R.S.J. Tol, and P.
Vellinga, The social costs of climate change:
Whetton, P.H., A.M.Fowler, M.R. Haylock, and A.B. Greenhouse damage and the benefits of control. In
Pittock, 1993. Implications of climate change due to Climate change 1995: Economic and social dimensions of cli-
the enhanced greenhouse effect on floods and mate change. Contribution of Working Group III to the
droughts in Australia. Climate Change, 25, 289–317. Second Assessment Report of the Intergovernmental
Panel on Climate Change. J. Bruce, H. Lee, and E.
Zwiers, F.W., 1998. Changes in the extremes of the Haites, eds. Cambridge University Press, Cambridge.
climate simulated by CCC GCM2 under CO2 dou-
bling. Journal of Climate (in press). Street, R., 1997. Weather impacts in Canada. Paper
presented at the Workshop on the Social and
Economic Impacts of Weather, Boulder, Colorado,
2–4 April.

31 Extreme Weather and Climate Change