Content uploaded by Gregory J. Mccabe
Author content
All content in this area was uploaded by Gregory J. Mccabe on Jun 25, 2015
Content may be subject to copyright.
Content uploaded by Gregory J. Mccabe
Author content
All content in this area was uploaded by Gregory J. Mccabe on Jun 25, 2015
Content may be subject to copyright.
Content uploaded by Gregory J. Mccabe
Author content
All content in this area was uploaded by Gregory J. Mccabe on Jun 25, 2015
Content may be subject to copyright.
15 J
UNE
2001 2763MCCABE ET AL.
q2001 American Meteorological Society
Trends in Northern Hemisphere Surface Cyclone Frequency and Intensity
G
REGORY
J. M
C
C
ABE
U.S. Geological Survey, Denver Federal Center, Denver, Colorado
M
ARTYN
P. C
LARK AND
M
ARK
C. S
ERREZE
Cryospheric and Polar Processes Division, Cooperative Institute for Research in Environmental Sciences, University of Colorado,
Boulder, Colorado
(Manuscript received 27 July 2000, in final form 20 November 2000)
ABSTRACT
One of the hypothesized effects of global warming from increasing concentrations of greenhouse gases is a
change in the frequency and/or intensity of extratropical cyclones. In this study, winter frequenciesand intensities
of extratropical cyclones in the Northern Hemisphere for the period 1959–97 are examined to determine if
identifiable trends are occurring. Results indicate a statistically significant decrease in midlatitude cyclone fre-
quency and a significant increase in high-latitude cyclone frequency. In addition, storm intensity has increased
in both the high and midlatitudes. The changes in storm frequency correlate with changes in winter Northern
Hemisphere temperature and support hypotheses that global warming may result in a northward shift of storm
tracks in the Northern Hemisphere.
1. Introduction
Variations in extratropical cyclone frequency and in-
tensity have a direct influence on surface climate
through effects on cloud cover, winds, and precipitation
frequency, duration, and magnitude. Thus, the frequen-
cies and characteristics of cyclones (as well as anticy-
clones) have been examined by a number of studies
(e.g., Colucci 1976; Sanders and Gyakum 1980; Parker
et al. 1989; Agee 1991; Serreze et al. 1997; Key and
Chan 1999). Assessments of changes in cyclone activity
that may result from global warming are important for
understanding regional climate change, which, in turn,
is necessary to evaluate impacts on ecological systems,
socioeconomic sectors (including agriculture, fisheries,
water resources, and human settlements), and human
health (Watson et al. 1998).
A number of studies have assessed the effect of global
warming on cyclone activity through sensitivity exper-
iments with general circulation models (GCMs). Using
simulations of doubled-CO
2
climatic conditions from
the Canadian Centre for Climate Modeling and Analysis
GCM, Lambert (1995) found a significant reduction in
the total number of cyclones in both hemispheres, but
an increased frequency of intense cyclones, most pro-
nounced in the Northern Hemisphere. However, storm
Corresponding author address: Dr. Gregory J. McCabe, U.S. Geo-
logical Survey, Denver Federal Center, MS 412, Denver, CO 80225-
0046.
E-mail: gmccabe@usgs.gov
tracks showed little change in position. Lambert (1995)
showed the reduction in cyclones as consistent with a
reduction in baroclinicity in the lower troposphere in a
warmed climate. In a simulation of future climatic con-
ditions using the European Centre for Medium-Range
Weather Forecasts (ECMWF) Hamburg GCM, that in-
cluded gradually increasing atmospheric CO
2
, Konig et
al. (1993) also found a slight reduction in cyclone fre-
quency for doubled-CO
2
climate conditions. However,
contrasting with Lambert (1995), Konig et al. also iden-
tified a poleward shift in cyclone activity over theNorth
Atlantic Ocean during autumn and winter, with a similar
tendency over extended areas of the Southern Hemi-
sphere. In addition, over the North Pacific Ocean, an
eastward displacement of cyclone activity was noted,
especially for winter and spring. Other modeling studies
have examined regional (e.g., over the Atlantic and Pa-
cific Oceans and North America) changes in storm
tracks and cyclone frequencies associated with potential
global warming (Held 1993; Stephenson and Held 1993;
Hall et al. 1994; Watson et al. 1998).
These investigations point to considerable uncertainty
in the response of both storm track locations and cyclone
frequencies to global warming. As discussed by Held
(1993) the issue is complex. While on the one hand,
weakening of low-level baroclinicity in midlatitudes
may lead to a weakening of storm tracks, it may be
possible for some systems to draw on the increased
supply of energy represented by strengthening of the
baroclinicity aloft. The response of midlatitude eddies
2764 V
OLUME
14JOURNAL OF CLIMATE
is also strongly influenced by increased levels of at-
mospheric moisture expected in a warmer climate. One
effect is a direct enhancement of eddy energy through
the additional release of latent heat [this mechanism was
invoked by Hall et al. (1994) to explain the intensifi-
cation of the Atlantic storm track in a warmed climate].
However, a competing effect exists in that when the
atmosphere becomes moister, horizontal latent heat
transport increases, meaning the eddies are more effi-
cient in transporting energy poleward (Held 1993).
Since smaller eddies are required to maintain the same
temperature gradient, one may anticipate that eddy am-
plitudes will decrease as temperatures increase. These
effects are relevant to changes in storm activity in a
zonal mean sense.
Zonal asymmetries in storm track locations introduce
additional complexities. Depressions tend to be orga-
nized into two primary tracks (Atlantic and Pacific).
These tracks can be altered by regional changes in the
mean stationary wave pattern in a warmed climate, as
well as regional differences in continental versus oce-
anic warming (Stephenson and Held 1993). These
changes are depicted differently in various GCMs. Fur-
thermore, the relationship between regional baroclin-
icity and storm activity is not completely linear. For
example, in an observational study of storm activity in
the Northern Hemisphere, Nakamura (1992) demon-
strated that baroclinic wave activity is actually sup-
pressed in the Pacific basin during winter when the low-
level regional baroclinicity and the strength of the Pa-
cific jet is maximized. While the exact mechanisms are
not clear, Nakamura (1992) shows that in a statistical
sense, baroclinic wave activity in the Pacific is limited
when the Pacific jet exceeds 45 m s
21
, as is common
in midwinter. This midwinter suppression of baroclinic
wave activity is not evident in the Atlantic basin, where
the strength of the jet rarely exceeds 45 m s
21
.
Given the large differences between GCM predictions
of storm activity under global warming, it is warranted
to examine the extent to which changes have occurred
within the period of instrumental records. Agee (1991)
studied trends in the annual frequency of surface cy-
clones and anticyclones for the Northern Hemisphere,
and their relationships with temperature. Agee found
that in the Northern Hemisphere, warming and cooling
trends are accompanied, respectively, by increases and
decreases in both cyclone and anticyclone frequencies.
These results are at odds with those modeling studies
that indicate that warming will be associated with de-
creased cyclone activity in the midlatitudes. In a more
recent study, Serreze et al. (1997) examined trends in
winter (Oct–Mar) cyclone frequencies for the midlati-
tudes (308–608N) and the high latitudes (608–908N) of
the Northern Hemisphere. They found that for the study
period 1966–93 high-latitude cyclone frequencies in-
creased and midlatitude cyclone frequencies decreased.
Serreze et al. (1997) did not compute correlations
between winter cyclone frequencies and Northern Hemi-
sphere winter temperatures. However, their results sug-
gest an inverse relation between midlatitude cyclone
frequency and winter Northern Hemisphere tempera-
tures (which have been increasing), and a positive re-
lation between high-latitude cyclone frequencies and
winter Northern Hemisphere temperatures. This is con-
trary to the results reported by Agee (1991). The dif-
ferences between these two studies may be due to 1)
the regions analyzed (Serreze et al. divided their analysis
of the Northern Hemisphere into midlatitude and high-
latitude cyclone frequencies) and 2) the method of com-
puting the number of cyclones during a season.
Key and Chan (1999) examined trends in both 1000-
(surface) and 500-hPa cyclone frequencies for the
Northern and Southern Hemispheres for the period
1958–97. They found that for winter months (Dec–Feb),
both the 1000- and 500-hPa cyclone frequencies in the
midlatitudes of the Northern Hemisphere decreased,
whereas cyclone frequencies for the high latitudes of
the Northern Hemisphere increased. However, for
spring, summer, and autumn, Key and Chan found op-
posite trends in 1000-hPa and 500-hPa cyclone fre-
quencies for both the mid- and high latitudes of the
Northern Hemisphere.
In summary, there is considerable uncertainty as to
how cyclone activity will change in a warmed climate.
The purpose of this study is to examine observed secular
changes and trends in cyclone frequency and intensity
in the Northern Hemisphere, and assess relationships
between those changes and variations in Northern Hemi-
sphere temperature.
2. Data and methods
The cyclone data used in this study were obtained
from a 39-yr (1959–97) record of 6-hourly cyclone sta-
tistics for the Northern Hemisphere. The detection al-
gorithm is essentially that described by Serreze (1995)
and Serreze et al. (1997) except it has been modified
for application to the more temporally consistent sea
level pressure (SLP) fields from the National Centers
for Environmental Prediction–National Center for At-
mospheric Research (NCEP–NCAR) reanalysis project
(Kalnay et al. 1996). Cyclone detection relies on the
identification of gridpoint SLP values surrounded by
gridpoint values at least 1 hPa higher than the central
point being tested. Intensity is based on the local La-
placian of pressure at each cyclone center. For this study,
only cyclones that existed for two or more consecutive
observation periods (i.e., at least 12 h) were used. In
addition, we focus on the winter season (Nov–Mar) for
which Northern Hemisphere temperature trends have
been largest. Winter Northern Hemisphere temperature
data for the years 1959–97 were obtained from the Jones
et al. (1999) dataset (available online at http://
cdiac.esd.ornl.gov/ftp/trends/temp/jonescru/global.dat).
15 J
UNE
2001 2765MCCABE ET AL.
F
IG
. 1. Standardized departures of winter (Nov–Mar) cyclone
counts for 58latitudinal bands in the Northern Hemisphere, 1959–
97.
F
IG
. 2. Standardized departures of winter (Nov–Mar) cyclone
counts in the Northern Hemisphere, 1959–97, for (a) high latitudes
(608–908N), and (b) midlatitudes (308–608N).
3. Results and discussion
a. Trends in cyclone frequency
Figure 1 illustrates cyclone counts for 58latitudinal
bands (expressed as standardized departures and illus-
trated on an equal area basis). The standardized depar-
tures (zscores) were computed for each 58latitudinal
band by subtracting the respective 1959–97 mean from
each value and dividing by the respective 1959–97 stan-
dard deviation. The standardized departures indicate a
decrease with time in cyclone frequency for the region
between 308and 608N and an increase in cyclone fre-
quency for the region poleward of 608N.
Following the approach used by Serreze et al. (1997),
the standardized departures of cyclone counts were av-
eraged for the 308–608N (midlatitude) and 608–908N
(high latitude) bands. The data were subdivided into
these groups because previous studies (i.e., Serreze et
al. 1997; Key and Chan 1999) found significant differ-
ences in cyclone characteristics for these latitudinal
bands. Application of linear regression (and accounting
for the effects of lag-1 correlation in the cyclone fre-
quency time series on the effective number of obser-
vations) indicates a statistically significant increase in
high-latitude winter cyclone frequency with time (Fig.
2a, correlation with time is 0.37, significant at a 95%
confidence level), and a statistically significantdecrease
in midlatitude winter cyclone frequency (Fig. 2b, cor-
relation with time is 20.50, significant at a 90% con-
fidence level). These results are similar to those reported
by Serreze et al. (1997). An obvious caveat is that the
observed increase in high-latitude cyclone frequency
may be due, in part, to improvements in the quantity
and quality of assimilation data with time. However, it
is unlikely that improved observations can account for
the simultaneous decrease in midlatitude cyclone fre-
quency.
In examining these time series more closely, it is
apparent that the trends, especially for the midlatitudes,
reflect a regime shift during the mid-1970s (Trenberth
1990). This corresponds to the mid-1970s climate tran-
sition noted in other studies. Miller et al. (1993) iden-
tified an abrupt shift in the basic state of theatmosphere–
ocean climate system over the North Pacific Ocean dur-
ing the 1976–77 winter season. This was associated with
a pronounced change in storm tracks across a large part
of the Northern Hemisphere (Folland and Parker 1990;
Trenberth 1990; McCabe et al. 2000).
Another interesting feature in Figs. 2a and 2b is a
pronounced increase in cyclone frequency in high lat-
2766 V
OLUME
14JOURNAL OF CLIMATE
F
IG
. 3. Standardized departures of the mean winter (Nov–Mar) AO
index, 1959–97.
itudes and a pronounced decrease in midlatitudes around
1989. These changes occur coincident with a significant
change in the Arctic oscillation (AO) index (Fig. 3).
The AO is defined as the leading empirical orthogonal
function of Northern Hemisphere area-weighted sea lev-
el pressure anomalies poleward of 208N (Thompson and
Wallace 1998). The positive polarity of the AO can be
interpreted as a strengthening and shrinking of the polar
vortex. The North Atlantic oscillation (NAO) represents
the primary component of the AO. During positive AO
conditions cyclone activity in the Northern Hemisphere
shifts poleward (Serreze et al. 1997; Clark et al. 1999).
When the AO is negative, the polar vortex is in a weak-
ened state and cyclone activity shifts south. The cor-
relations between the AO and the time series of high-
latitude and midlatitude cyclone frequencies are 0.69
and 20.56, respectively, both statistically significant at
the 95% confidence level. The changes in cyclone dis-
tributions seen here also are consistent with previous
studies illustrating recent decreases in SLP over the Arc-
tic with compensating pressure increases over the mid-
latitudes (Walsh et al. 1996; Serreze et al. 1997, 2000).
Hurrell (1996) suggests that almost 50% of the winter
(Dec–Mar) temperature variance over the Northern
Hemisphere (north of 208N) since 1935 is due to the
combined effects of variability in atmospheric circula-
tion forced by the NAO (a component of the AO) and
the Southern Oscillation. Thompson et al. (2000) sub-
sequently show that over the period 1968–97 approxi-
mately half of the observed winter warming over Siberia
and all of the winter cooling over eastern Canada and
Greenland is linearly congruent with the monthly time
series of the AO.
Another important mode of global climate variability
is the El Nin˜o–Southern Oscillation (ENSO; Kiladisand
Diaz 1989). Some studies have suggested a relation be-
tween ENSO and global warming (Trenberth and Hoar
1996, 1997). One index of ENSO is the sea surface
temperature (SST) anomaly in the Nin˜o-3.4 region. The
Nin˜o-3.4 region encompasses the area 58S–58N, and
1208–1708W. Positive (negative) Nin˜o-3.4 SST anom-
alies correspond to El Nin˜o (La Nin˜ a)conditions. Winter
(Nov–Mar) Nin˜o-3.4 SST values were averaged to com-
pute a time series for the years 1959–97. Correlations
between the Nin˜o-3.4 time series and cyclone frequen-
cies for the high and midlatitudes are not found to be
significant (0.01 and 20.18, respectively).
The conclusion that anthropogenic greenhouse warm-
ing has been a primary driver of changes in northern
high-latitude climate must be treated cautiously (Serreze
et al. 2000). Although climate models and observations
generally agree in terms of changes in Arctic temper-
atures and atmospheric circulation, there are some dis-
crepancies. For example, models generally estimate the
strongest Arctic warming to occur during autumn and
winter, whereas observations indicate the greatest warm-
ing during winter and spring (Kattenberg et al. 1996;
Serreze et al. 2000).
b. Trends in cyclone intensity
For both the high latitudes and the midlatitudes, cy-
clone intensity has increased over the 1959–97 period
(Figs. 4a and 4b). The linear trend (accounting for lag-
1 correlations in the intensity time series) for high lat-
itudes (r50.53, significant at a 99% confidence level)
is much stronger than that for midlatitudes (r50.39,
significant at a 90% confidence level). In addition, high-
latitude cyclone intensity exhibits a pronounced increase
around 1989 similar to the change observed in cyclone
frequency, and coincident with the increase in the AO
index. Again, these results must be viewed with rec-
ognition that the assimilation database available for the
NCEP–NCAR reanalysis has improved through time.
c. Relations with Northern Hemisphere temperature
Time series of winter cyclone frequency andintensity
were correlated with the time series of mean Northern
Hemisphere winter temperature. Results indicate statis-
tically significant correlations for both the high and mid-
latitudes (Table 1, Figs. 5a and 5b). The correlation
between winter temperature and cyclone frequency for
the midlatitudes (r520.58) is much stronger than that
for the high latitudes (r50.38). The negative corre-
lation between midlatitude cyclone frequency and win-
ter temperature, and the positive correlation between
high-latitude cyclone frequency and winter temperature
lend support to the idea that global warming may result
in increased cyclone activity in high latitudes and de-
creased cyclone activity in midlatitudes (Held 1993; Ste-
phenson and Held 1993; Konig et al. 1993; Hall et al.
1994; Lambert 1995). In contrast to the results for cy-
clone frequency, the correlations between winter tem-
perature and cyclone intensity are small and are not
significant for either latitudinal band (Table 1).
One interpretation of these results is that increasing
15 J
UNE
2001 2767MCCABE ET AL.
F
IG
. 4. Standardized departures of winter (Nov–Mar) cyclone in-
tensity in the Northern Hemisphere, 1959–97, for (a) high latitudes
(608–908N), and (b) midlatitudes (308–608N).
T
ABLE
1. Correlations between mean Northern Hemisphere winter
temperature and winter cyclone frequency and intensity for the high
latitudes (608–908N) and the midlatitudes (308–608N).
Cyclone characteristic Correlation with
winter temperature
High-latitude cyclone frequency
Midlatitude cyclone frequency
High-latitude cyclone intensity
Midlatitude cyclone intensity
0.38*
20.58**
0.25
20.13
* Significant at a 95% confidence level.
** Significant at a 99% confidence level.
F
IG
. 5. Comparisons of standardized departures of mean Northern
Hemisphere winter (Nov–Mar) temperature and (a) high latitude (608–
908N) cyclone frequency, and (b) midlatitude (308–608N) cyclone
frequency (1959–97).
levels of atmospheric CO
2
have forced increases in
Northern Hemisphere temperature, which in turn force
changes in circulation and cyclone activity. This inter-
pretation is consistent with model sensitivity studies.
However, a more straightforward interpretation is that
cyclone activity and temperature simply covary on in-
terannual timescales, and the trends we see are not re-
lated to greenhouse forcing in any way. In this context,
recall from section 3a that interannual variations in
Northern Hemisphere temperature may themselves be
forced by interannual variations in the dominant modes
of low-frequency atmospheric variability (Hurrell 1996;
Thompson et al. 2000; Serreze et al. 2000). In this ob-
servational study we cannot objectively determine
which of these opposing interpretations is true. Our in-
tent in presenting these changes in cyclone activity and
temperature is to provide a more complete picture of
observed changes in the climate system.
4. Conclusions
This study shows that for the Northern Hemisphere,
winter cyclone frequency has increased in high latitudes
and has decreased in midlatitudes. However, for both
latitudinal bands, winter cyclone intensity has increased.
The changes in cyclone frequency correspond to sig-
2768 V
OLUME
14JOURNAL OF CLIMATE
nificant climate transitions: the mid-1970s climate tran-
sition and the 1989 shift in the AO index. Winter cy-
clone frequency is significantly correlated with Northern
Hemisphere winter temperatures. Variability in winter
cyclone intensity is not significantly correlated with
Northern Hemisphere temperatures. These results pro-
vide some support for climate modeling studies (e.g.,
Konig et al. 1993) suggesting that global warming may
be associated with an increase in high-latitude cyclone
frequency and a decrease in midlatitude cyclone fre-
quency.
REFERENCES
Agee, E. M., 1991: Trends in cyclone and anticyclone frequency and
comparison with periods of warming and cooling over the North-
ern Hemisphere. J. Climate, 4, 263–267.
Clark, M. P., M. C. Serreze, and D. A. Robinson, 1999: Atmospheric
controls on Eurasian snow extent. Int. J. Climatol., 19, 27–40.
Colucci, S. J., 1976: Winter cyclone frequencies over the eastern
United States and adjacent western Atlantic, 1964–1973. Bull.
Amer. Meteor. Soc., 57, 548–553.
Folland, C. K., and D. E. Parker, 1990: Observed variations of sea
surface temperature. Climate–Ocean Interaction, M. E. Schles-
inger, Ed., Kluwer, 21–52.
Hall, N. M. J., B. J. Hoskins, P. J. Valdes, and C. A. Senior, 1994:
Storm tracks in a high-resolution GCM with doubled carbon
dioxide. Quart. J. Roy. Meteor. Soc., 120, 1209–1230.
Held, I. M., 1993: Large-scale dynamics and global warming. Bull.
Amer. Meteor. Soc., 74, 228–241.
Hurrell, J. W., 1996: Influence of variations in extratropical winter-
time teleconnections on Northern Hemisphere temperatures.
Geophys. Res. Lett., 23, 665–668.
Jones, P. D., T. J. Osborn, K. R. Briffa, and D. E. Parker, 1999: Global
monthly and annual temperature anomalies (degrees C), 1856–
1998. Climatic Research Unit, School of Environmental Sci-
ences, University of East Anglia, Norwich, United Kingdom.
[Available online at http://cdiac.esd.ornl.gov/ftp/trends/temp/jo-
nescru/global.dat.]
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Re-
analysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
Kattenberg, A., and Coauthors, 1996: Climate Models: Projections
of future climate. Climate Change 1995: The Science of Climate
Change, Contribution of Working Group I to the Second As-
sessment of the Intergovernmental Panel on Climate Change, J.
T. Houghton et al., Eds., Cambridge University Press, 289–357.
Key, J. R., and A. C. K. Chan, 1999: Multidecadal global and regional
trends in 1000 mb and 500 mb cyclone frequencies. Geophys.
Res. Lett., 26, 2053–2056.
Kiladis, G. N., and H. F. Diaz, 1989: Global climatic anomalies as-
sociated with extremes of the Southern Oscillation. J. Climate,
2, 1029–1090.
Konig, W., R. Sausen, and F. Sielman, 1993: Objective identification
of cyclones in GCM simulations. J. Climate, 6, 2217–2231.
Lambert, S. J., 1995: The effect of enhanced greenhouse warming
on winter cyclone frequencies and strengths. J. Climate, 8, 1447–
1452.
McCabe, G. J., A. G. Fountain, and M. Dyurgerov, 2000: Variability
in winter mass balance of Northern Hemisphere glaciers and
relations with atmospheric circulation. Arct. Antarct. Alp. Res.,
32, 64–72.
Miller, A. J., D. R. Cayan, T. P. Barnett, N. E. Graham, and J. M.
Oberhuber, 1993: The 1976–77 climate shift of the PacificOcean.
Oceanography, 7, 21–26.
Nakamura, H., 1992: Midwinter suppression of baroclinic wave ac-
tivity in the Pacific. J. Atmos. Sci., 49, 1629–1642.
Parker, S. S., J. T. Hawes, S. J. Colucci, and B. P. Hayden, 1989:
Climatology of 500-mb cyclones and anticylones, 1950–85.
Mon. Wea. Rev., 117, 558–570.
Sanders, F., and R. Gyakum, 1980: The synoptic-dynamic climatology
of the Bomb. Mon. Wea. Rev., 108, 1589–1606.
Serreze, M. C., 1995: Climatological aspects of cyclone development
and decay in the Arctic. Atmos.–Ocean, 33, 1–23.
——, F. Carse, R. G. Barry, and J. C. Rogers, 1997: Icelandic low
cyclone activity: Climatological features, linkages with the NAO
and relationships with recent changes in the Northern Hemi-
sphere circulation. J. Climate, 10, 453–464.
——, and Coauthors, 2000: Observational evidence of recent change
in the northern high-latitude environment. Climatic Change, 46,
159–207.
Stephenson, D. B., and I. M. Held, 1993: GCM response of Northern
winter stationary waves and storm tracks to increasing amounts
of carbon dioxide. J. Climate, 6, 1859–1870.
Thompson, D. W. J., and J. M. Wallace, 1998: The Arctic Oscillation
signature in the wintertime geopotential height and temperature
fields. Geophys. Res. Lett., 25, 1297–1300.
——, ——, and G. C. Hegerl, 2000: Annular models in the extra-
tropical circulation. Part II: Trends. J. Climate, 13, 1018–1036.
Trenberth, K. E., 1990: Recent observed interdecadal climate changes
in the Northern Hemisphere. Bull. Amer. Meteor. Soc., 71, 988–
993.
——, and T. J. Hoar, 1996: The 1990–1995 El Nin˜o–Southern Os-
cillation event: Longest on record. Geophys. Res. Lett., 23, 57–
60.
——, and ——, 1997: El Nin˜o and climate change. Geophys. Res.
Lett., 24, 3057–3060.
Walsh, J. E., W. L. Chapman, and T. L. Shy, 1996: Recent decrease
of sea level pressure in the central Arctic. J. Climate, 9, 480–
486.
Watson, R. T., M. C. Zinyowera, and R. H. Moss, 1998: The regional
impacts of climate change: An assessment of vulnerability. A
Special Report of IPCC Working Group II, Cambridge Univer-
sity Press, 517 pp.
