The Impact of Mount Pinatubo on World-Wide Temperatures

Abstract

We monitor and model the effects on world-wide temperatures of the June 1991 volcanic eruption of Mount Pinatubo in the Philippines. Global mean air temperatures were reduced, by up to 0.5°C at the surface and 0.6°C in the troposphere, for some months in mid-1992, in approximate accord with model predictions. Differences from these predictions occurred in the Northern Hemisphere winters of 1991-1992 and 1992-1993, as a result of atmospheric circulation changes that yielded continental surface warmings not fully reproduced by the model. The effects of the eruption were less evident by 1994. A superposed-epoch composite for five major tropical eruptions shows significant global post-eruption cooling at the surface when the effects of the El Ni7o-Southern Oscillation are removed from the data. Stratospheric warmth following Pinatubo lasted until early 1993 according to Microwave Sounding Unit data.

Full text

INTERNATIONAL JOURNAL
OF
CLIMATOLOGY,
VOL.
16,487497
(1996)
THE IMPACT
OF
MOUNT PINATUBO ON WORLD-WIDE
TEMPERATURES
D.
E.
PARKER
Hadley Centre, Meteorological Ofice, London Road, Bmcknell RGI2 2SI:
UK
H.
WILSON
Columbia University and NASA, Goddard Institute for Space Shtdies, New York.
USA
I?
D. JONES
Climatic Research Unit, University
of
East Anglia,
UK
J.
R.
CHRISTY
Department of Atmospheric Science, University of Alabama
in
Hunlsville, Huntsville, Alabama. USA
AND
C.
K.
FOLLAND
Hadley Centreo Meteorological Ofice, London Road, Bracknell RG12 2SL
UK
Received I4 May
I995
Accepted
3
October
I995
ABSTRACT
We monitor and model the effects on world-wide temperatures of the June 1991 volcanic eruption of Mount Pinatubo in the
Philippines. Global mean air temperatures were reduced, by up to
0.5"C
at the surface and
0.6"C
in the troposphere, for some
months in mid-I 992, in approximate accord with model predictions. Differences from these predictions occurred in the Northern
Hemisphere winters of 1991-1992 and 1992-1 993, as a result of atmospheric circulation changes that yielded continental surface
warmings not
fully
reproduced by the model. The effects of the eruption were less evident by 1994.
A
superposed-epoch
composite for five major tropical eruptions shows significant global post-eruption cooling at the surface when the effects
of
the
El
Nifio-Southern Oscillation are removed from the
data.
Stratospheric warmth following Pinatubo lasted until early 1993 according
to Microwave Sounding Unit data.
KEY
WORDS:
volcanic eruptions; temperature variations; atmospheric circulation; Mount Pinatubo.
1.
INTRODUCTION
Many observational studies have been carried out to assess the climatic impacts
of
major volcanic eruptions. We do
not review these here, but refer readers to recent reviews by Mass and Portman (1989), Robock (1991) and
McCormick
et
al.
(1
995), and to the assessments made
in
the IPCC Reports (Folland
et
al.,
1990, section 7.2.1
;
Folland et al., 1992; Shine
et
al.,
1990). Most, but not all,
of
the studies reviewed find post-eruption surface cooling
of
several tenths "C on a world-wide average, despite difficulties encountered in making these assessments,
including the choice
of
radiatively effective eruptions, lack
of
statistical significance owing to small samples, and the
complicating influence
of
other atmospheric and oceanic variations such
as
the El Nifio-Southern Oscillation
(ENSO). Another widely recognized sequel to major eruptions is stratospheric warming (Parker and Brownscombe,
1983; Labitzke, 1994; McCormick
et
al.,
1995).
In this note we examine the changes of global surface temperature and tropospheric (Microwave Sounding Unit
(MSU)) air temperature observed since the June 199
1
eruption
of
Mount Pinatubo in the Philippines, in the light of
model predictions (Hansen
et
al.,
1992; Kolomeev
et
al.,
1993)
of
a global
or
hemispheric surface and tropospheric
cooling
of
about 0.5"C, at its strongest about a year
after
the eruption. We also use global average surface
temperatures around the times of Pinatubo and earlier eruptions to calculate superposed epoch global time series:
CCC 0899-8418/96/050487-11
0
Controller, HMSO, Norwich, England, 1996

488
D. E.
PARKER
ET
AL
these series are compensated for ENSO, using an index
of
the Southern Oscillation, to isolate the volcanic influence
more clearly.
An
influence of volcanic eruptions on Northern Hemisphere atmospheric circulation has already been discussed
by Robock and Mao (1992, 1995), Kirchner and Graf (1 995) and Graf
et al.
(1994), and is implied by the results of
Groisman (1 992). We examine this further with particular emphasis on Pinatubo.
Finally, MSU data are used to monitor stratospheric temperature changes following Pinatubo. The results are used
to conh both observational findings (Labitzke and McCormick, 1992; Labitzke, 1994) and modelling studies
(Graf
et al.,
1993; Kirchner and Graf, 1993). Linkages between the stratospheric temperature changes and those of
the Northern Hemisphere atmospheric circulation have already been proposed (Robock and Mao, 1992, 1995; Graf
et al.,
1993, 1994; Kirchner and Graf, 1993; Kelly
et al.,
1996): we briefly summarize their implications for the
analysis of volcanic effects.
2. RESULTS
2.1.
Changes
in
global surface temperature following Pinatubo
In Figure l(a) we show, in the solid curve, 3-month running mean global surface air temperature anomalies
following the June 1991 eruption of Pinatubo. Anomalies were calculated from a base-period 1961-1990, then
expressed relative
to
anomalies (not actual values) for the effectively pre-eruption 3-month period April to June
199
1.
Results relative to the 6-month period January to June 1991 and the 12-month period July 1990 to June 199
1
(not shown) were almost identical. The anomalies are based on land station air temperatures (Jones, 1994) and an
updated data base of the night-time marine air temperature (NMAT) measurements made fiom ships and buoys
described in Bottomley
et al.
(1 990). Day-time marine air temperatures were not used, because they are affected by
solar heating of ships' fabric. Air rather than sea-surface temperatures (SST) were used for compatibility with the
model results. A global cooling of about 0~5°C is observed between the eruption and the northern summer of 1992,
interrupted, however, by relative warmth between January and March 1992. Another period of relative warmth in
January to March 1993 was followed by a smaller cooling in the Northern Hemisphere summer of 1993. Early 1994
did not show a corresponding period of warmth, but a renewed warm period was seen in early 1995. If NMAT is
replaced by SST, the maximum observed cooling is reduced by about 0.1"C (not shown), owing to the greater
thermal inertia of the ocean.
Also included in Figure l(a) as dashed lines are corresponding global surface air temperature changes following
Pinatubo as predicted by the Goddard Institute for Space Studies (GISS) global coupled atmosphere-ocean climate
model (Hansen
et al.,
1988, 1992). This model has
8"
by 10" horizontal resolution, nine atmospheric layers, an
oceanic mixed layer with specified horizontal heat transports, and vertical diffusion to the deep ocean. The dashed
sequences in Figure l(a) were derived using randomly different initial atmospheric conditions for runs 'P1
'
and 'P2'
of the model
as
described by Hansen
et al.
(1 992). In both these
ms,
greenhouse gases increased linearly,
as
in
'scenario B'described in figure 2 of Hansen
et
al.
(1
988). The imposed aerosol burdens were also the same: the time-
dependence of global optical depth followed Hansen
et al,'s
(1 988) simulations of the effects of El Chichon, but with
an overall enhancement by a factor of 1.7, and with aerosols restricted uniformly to the belt 20"s to 30"N in the first
three months after the eruption, followed by a poleward spread into both hemispheres giving uniform cover in
January 1992. The implied net radiative forcings may, according to reanalysed aerosol data, be overestimated by up
to 20 per cent (Hansen
et al.,
1996). Also, the true aerosol distribution, although widespread, was not uniform by
January 1992 (Grant
et al.,
1994). Hitchman
et
al.
(1 994) present
a
climatology of stratospheric aerosol and show
that the combination of sporadic volcanic injections and stratospheric circulation is unlikely to result in a uniform
distribution in practice. Nonetheless, volcanic influences on global tropospheric temperature have been fitted well by
a simple exponential-decay formula
(Christy
and McNider, 1994), and, despite its shortcomings, the GISS model
simulates successfully the observed overall cooling, although it gives only weak indications of the temporary
warmings affecting the Northern Hemisphere winter.
Figure I@) shows that the observed temporary warmings took place almost entirely over land and were not
reproduced by the model, at least for Northern Hemisphere winter 1991-1992. The model also slightly under-
estimated the maximum cooling over land, but slightly overestimated the observed overall changes
of
global-ocean

The effect
of
the
Mt.
Pinatubo eruption, June
1991
M
M
(*ti*
to
Apd1-J~
1991)
1.0
-
Obscned
Land
Surface
Air
Temperahre
(rrl.ti~
to
April-J~~
1991)
0
--I--
MOadpreatc.rioaS
0..
-
V
f
03-
OJ-
-
*
0
0,
*
LI
$
0.0
-
4
45-
h-
h.
-
B
*
-1.0
4.6
08
“91
AM192
AMJ93
Am94
W9S
M96
AMJW
AMlI
AN91 AM192
AM193
AMJPI
AN95
AMJ%
AUJW
--
I-,
Modd
pmdkliws
-
-
-
-
...........I...).......1.1.../.....1....
s
0
0.4
fi
-
Observed
MSU
2R
(relalive
to
April-June
1991)
0.4
A
-
-
Obwed
Night
Marine
Air
Temperature
-
--
I--
Moddledloaertropospherepredlction
V
(dative
to
April-Joar
1991)
v-
--
I-
-
ModclprrdieW
-
m
0’:
-
0
2:
c-
8
P-
m
0.2
-
-
b
Figure 1. (a) Observed (solid line) and modelled (dashed lines) global surface land with night-time marine air temperatures following the eruption of Mount Pinatubo. Values are 3-month
nmaing
mean anomalies relative to the period April to June 1991, The series
run
from this period through to the period October to December 1995.
@)
As (a) but
for
land only. (c) As (a) but for
$
\o
ocean
only.
(d)
As
(a) but for the global lower troposphere. Observed are Microwave Sounding Unit (MSUZR)
data:
modelled are layer tempemtures, weighted to match MSUZR
TI
-
P
- -

490
D.
E.
PARKER
ET
AL.
surface air temperature (Figure l(c)). Global lower to middle tropospheric temperature changes (Figure l(d)) were
well simulated: cooling was slightly overestimated except in early 1993. See also section 2.3.
The reasons for the observed global surface warmth in the Northern Hemisphere winter of 1991-1992 are
apparent from the temperature and mean-sea-level pressure anomalies in Figures 2(a) and 3. Anomalies in these
fields are taken relative to the previous five Northern Hemisphere winters, for compatibility with superposed-epoch
analyses (see below) in which this was done to eliminate longer-term climatic changes. Sea-surface temperatures
(SSTs) are used in Figure 2(a) because they have better coverage than night-time marine air temperature, which
yielded noisier results. The SSTs came from the Global sea-Ice and Sea-Surface Temperature (GISST1.l) data set
with some unreliable Southern Ocean and Arctic SSTs deleted (Parker
et
al.,
1994). The procedure used to blend the
land air temperatures with the SSTs was simpler than that used by Parker
et al.
(1 994) because of the new grid-box
analysis of the land air temperatures (Jones, 1994), but data from small islands were still given enhanced weighting
relative to the proportion of land in their grid-box. Figure 2(a) shows that the enhanced warmth was concentrated
over northern Eurasia and over central North America (see also Halpert
et
al.,
1993)
as
a result of anomalies of
atmospheric circulation.
A
strong (1 1 hPa) anticyclonic mean-sea-level pressure anomaly centred over the British
Isles (Figure 3) resulted in the advection of warm air across Scandinavia: positive temperature anomalies further east
were ensured by the continued, albeit indirect, advection of some of this air, along with enhanced southerly flow in
central Asia. The enhanced warmth over North America was associated with northward advection east of a deeper
and eastward-displaced Aleutian low, relative to the previous 5-winter average (Figure 3). This circulation anomaly,
in
turn,
may have resulted from the El Niiio event then in progress, as evidenced by the anomalous warmth of the
central and eastern equatorial Pacific in Figure 2(a) and by the El Niiio-atmospheric circulation relationships
reviewed by Glantz
et
al.
(1 99 1). These results, along with the well known tropical warmth during El Niiio events
(Pan and Oort, 1983), suggest that compensation of observed series, such as those in Figure 1, for ENS0 (see also
Robock and Mao, 1992, 1995) may yield a clearer picture of the overall cooling influence of major volcanic
eruptions. We do this in an elementary way in section 2.2.
The GISS model, having generally only 1-2 layers in the stratosphere, may not simulate realistically the
dynamical interactions between the troposphere and stratosphere (Hansen
et al.,
1996) which appear to underlie the
observed atmospheric circulation changes in the Northern Hemisphere in winter following tropical eruptions (see
section 3).
So,
scenario ‘Ply (Hansen
et
al.,
1992) shows mixed winter surface temperature anomalies over Northern
Eurasia whereas ‘P2’ has anomalous warmth over central and eastern Siberia but greater coldness than observed in a
belt from Europe through central Asia to China.
Also, the GISS model cannot produce an El Niiio (Robock and Liu, 1994), because it has specified horizontal heat
transports in the oceanic mixed layer. Without this influence for winter warmth over central North America, the
model’s internal variability appears to have dominated the results there, with anomalous coldness
in
scenario ‘P1
’
(Hansen
et al.,
1992) but warmth in scenario ‘P2’,
in
the 1991-1992 Northern Hemisphere winter. Overall,
therefore, neither simulation fully reproduces the global-average warmth
in
this season. Furthermore, additional
simulations would, owing to internal variability, probably yield very disparate results on a regional scale, as would
additional volcanic eruptions of similar magnitude, location, and seasonal timing in the real world.
The model was in better agreement with observations regarding overall global coolness in the Northern
Hemisphere summer
of
1992 (Figure l(a)), when natural and modelled internal variability are smaller than in
winter. Observations (not shown) included strong cold anomalies (down to
-
3°C) over North America and central
northern Asia, and warm anomalies (up to
+
2°C) over Europe. The latter are not a consistent feature of post-
eruption summers (e.g. Groisman, 1992). The observed anomalies are consistent with anomalous advection by the
atmospheric circulation (not shown). There were also warm anomalies (up to
+
2°C) over the eastern North Pacific,
in connection with the El Niiio.
2.2.
Comparison with previous tropical eruptions
We chose Krakatau (August 1883), Pelee (May 1902), Soufiiere (May 1902) and Santa Maria (October 1902)
combined
as
a single event (October 1902), Agung (March 1963), and El Chich6n (April 1982). These eruptions are
associated with the largest estimated stratospheric aerosol optical depths (Sat0
et
al.,
1993). We restricted our
selection to the tropics, because Robock and Mao (1992, 1995) found that Northern Hemisphere winter warmth

MOUNT
PINATUBO
IMPACT ON TEMPERATURES
49
1
Surface tem erature anomalies
relative to average
of
DJF 198d7
-
1990/1
December 19
6
1
to Februar 1992
Figure 2. (a) Land-surface air and sea-surface temperatures, December 1991 to February 1992, relative to the average
for
the same season
of
the
previous
5
years. Isopleths every
0.5"C.
Negative values shaded.
(?J)
As
(a) but
MSUZR
data
for
the lower troposphere. lsopleths every 0.5T
with zero omitted. Negative contours dashed

492
D.
E.
PARKER
ET
AL.
Figure
3.
Northern Hemisphere mean-sea-level pressures, December 1991
to
February 1992, relative
to
the average for the same season of the
previous
5
years. lsopleths at
1
hPa intervals. Data source: blended analyses of the Meteorological Office, NCAR, and Scripps Institute of
Oceanography
tended to occur in the first winter after tropical eruptions, but in the second winter after higher-latitude eruptions., We
carried out a superposed-epoch analysis of monthly global land-surface air and sea-surface temperature anomalies
around the time of these eruptions and Pinatubo. Sea-surface temperatures were used because they have better
coverage than night marine air temperatures: global changes
of
these are expected to be very similar (Bottomley
et
al.,
1990). The anomalies were first referenced to the 5-year pre-eruption period, to remove the effects of longer term
climatic changes, and a low-pass smoothing filter, with half-power at
6
months period, was applied. The smoothed
anomalies were then composited; also standard errors were calculated from them (Figure 4(a)). Similar analyses
have been carried out by Kelly
er
al.
(1
996). The smoothed global temperature drop reached nearly 0.2"C about 15
months after the eruptions, effectively in the Northern Hemisphere summer ofthe post-eruption year. However, this
dips only marginally below
two
standard errors, largely because a major El Nifio, giving global warmth, followed the
eruption of El Chichon.
So
we then adjusted all the original monthly global surface temperatures for
ENS0
using
linear regressions against the standardized monthly Southern Oscillation index (Tahiti minus Darwin mean-sea-level
pressure: see Ropelewski and Jones (1987)) six months previously with a five-term binomial smoothing to eliminate
intraseasonal variations (Jones (1988); see also Robock and Mao (1992)). (Compensation for the Southern
Oscillation on a local basis, before taking global averages, was carried out by Robock and Mao (1995) and
could be more effective because geographically varying lags can be incorporated; but it could also yield noisy results

MOUNT PINATUBO IMPACT
ON
TEMPERATURES
493
in data-sparse areas and in regions where the ENSO signal is not dominant.) Then the composite coolness 15 months
after eruptions was close to 0.25"C (Figure 4@)) with reduced standard errors
so
that the statistical significance
reached the 99 per cent level, assuming five degrees of freedom in a one-tailed t-test, which is applicable given an a
priori expectation of cooling. The adjustment removed most of the warming following El Chichon. The adjusted
cooling following Pinatubo was comparable with, or a little greater than, that following most of the other eruptions
(Figure 4(c)). There was also significant cooling around 30 months after eruptions: because most of the eruptions
were in Northern Hemisphere spring, this was in the Northern Hemisphere autumn. Some of the remaining
variability between eruptions indicated by Figure
40,
and c) may have been caused by sparser data coverage around
the times of the earlier eruptions (Madden
et al.,
1993).
We have thus demonstrated consistency in overall cooling at the global surface after tropical eruptions. Next, we
examined a composite mean-sea-level pressure field (not compensated for the Southern Oscillation) for Northern
Hemisphere winters following the earlier eruptions. This field (not shown; covering an area similar to Figure 3)
indicates a strong similarity to Figure 3 over much of Eurasia: anomalous high pressure centred over France allowed
the advection of warm air across northern Europe towards central Asia. There was a weaker similarity over North
America and the Pacific, with a slightly deepened and eastward-displaced Aleutian low. The correlation with Figure
3, using 5" latitude
x
5" longitude grid-boxes, was
0.67
for the entire data area. Composite surface temperature
anomalies in the Northern Hemisphere winters following these four eruptions, not compensated for ENS0 (not
shown), were somewhat similar to Figure 2(a) (correlations 0.41, 0.34, 0.39
for
Northern Hemisphere, Southern
Hemisphere, globe), with warmth over northern Eurasia following all four eruptions and over North America
following Agung and El Chichon. There was also warmth over the central equatorial Pacific because the 1982-1 983
El Niiio was included in the composite.
As
after Pinatubo, this coincidence with El Nifio is likely to have contributed
to the anomalies in the mean-sea-level pressure field over the North Pacific and probably also therefore to the
temperature anomalies over North America (Glantz
et al.,
199 1).
2.3.
Changes
in
temperature
aloft
In Figure l(d) we present 3-month running mean global temperature anomalies for the lower to middle
troposphere based on 'MSU2R' radiances. These are most sensitive to temperatures near the
750
hPa level, with
substantial influence from the surface to 500 hPa but little sensitivity to temperature at 300 hPa (Spencer and
Christy, 1992). These MSU data indicate a global cooling of about 0.6"C between mid-1991 and mid-1992. This
was also noted by McCormick
et al.
(1 995). The pronounced temporary warmings
in
early 1992 and early 1993 at
the surface (Figure l(a)) were barely evident in the lower to middle troposphere (Figure I(d)). Tropospheric mid-
latitude continental winter temperature anomalies in general tend to be smaller than, although highly correlated with,
those at the surface, according to both the MSU (Spencer
et aE.,
1990) and radiosondes. Monthly averages of
radiosonde temperatures in the data base of Parker and Cox (1995) show a reduction of standard deviation
approaching 50 per cent between the surface and mid-troposphere in these regions in winter, and this decrease of
amplitude with height was evident in the winters of 1991-1992 and 1992-1993.
So
the weaker tropospheric
continental positive anomalies in these winters were more than compensated by negative anomalies in the subtropics
and the Arctic, especially in its North American sector (Figure 2(b)), giving negative global anomalies (Figure l(d)).
A similar balance occurred in March 1990, which was very warm at the surface over the Northern Hemisphere
(Parker and Jones, 1991). Note that this was not preceded by a tropical eruption (see section 3).
Figure 5 is a 17-year monthly temperature series for the global lower stratosphere, based on MSU Channel 4,
which monitors mainly the 150 hPa to
30
hPa layer, with peak sensitivity near
70
hPa (Spencer and Christy, 1993).
This series is dominated by the warmings following
El
Chichon (Parker and Brownscombe, 1983) and Pinatubo
(Labitzke and McCormick, 1992; Labitzke, 1994). These papers show that the major warmings occurred in the
tropics, where most, but not all (Trepte
et
al.,
1993) of the aerosol was concentrated in the first
4-6
months after the
eruptions, before its dispersal by synoptic-scale eddies and the meridional stratospheric circulation. When
El
Chichon erupted, the tropical stratospheric quasi-biennial oscillation (QBO) was in its warm phase (Labitzke and
McCormick, 1992) whereas the QBO was in its cold phase when Pinatubo erupted (Labitzke, 1994). This partially
explains the warmth before El Chich6n and coldness before Pinatubo in Figure 5 (Christy and Drouilhet, 1994). In
addition, the eruption of Cerro Hudson in Chile in August 1991 may have enhanced the post-Pinatubo warming

494
0.2
0
0.0
73
-0.2
D.
E. PARKER
ET
AL.
-
-
-
0.4:"
'
"
''
"
I'
"
"
''
"
~
0.2
I
-100
-50 0
50
100
Months before/cfter
eruption
0.41
'
' ' '
'
' '
'
'
'
'
'
'
'
'
' ' ' '
1
0.2
1
t
-100
-
50
0
50
100
Months
before/cfter eruption
0.41
" "
I'
"
'
" "
'
"
'
'
1
t
V
-0.4L,,
.
I
I I I I
I,
1.1
I
I,,
I
-100
-50
0
50
100
Months
before/after eruption
Figure
4.
(a) Composite sequence
of
low-pass filtered monthly global land-surface air with sea-surface temperatures around the time
of
five
tropical volcanic eruptions
(Krakatau,
Pebe-Soufiere-Santa Maria (composite), Agung, El Chichh, Pinatubo) (solid line). Values are
expressed
as
anomalies relative to the
5-year
pre-eruption periods. The dashed lines
are
f
1
standard
mr.
(b)
As (a) with the influence of the El
NifiWSouthem Oscillation removed (see text). (c) As
(b)
but showing individual sequences
for
Pinatubo (light solid line) and
for
the other
eruptions (dashed lines), along with the composite sequence (heavy solid line)

MOUNT PINATUBO IMPACT ON TEMPERATURES
495
-1.0
3
8
-
-
-
- -
-
-
I
I
I
I
I I
I
I
I
I
I I
I I
I I
Monthly MSU 4 global temperature anomalies for lower stratosphere,
relative to a
1984-90
average.
[
Data
to December
1995
]
1.57
I
I
I
I
1
I
I
I
I
I
I
I
-0.5
1
Figure
5.
Monthly MSU-4
global temperature anomalies
for
the
lower
stratosphere, relative to a
1984-1990
average
(Angel], 1993) and the eruption of Nyamuragira in Ahca in December 198
1
(Dutton and Christy, 1992) may have
contributed to the warming before El Chichon. The solar-cycle modulation of ozone and human-induced ozone
depletion may also have contributed, respectively, to the fluctuations and downward trend of temperature in Figure
5
(Folland
et
al.,
1992; McCormick
et
al.,
1995). The stratospheric warming following Pinatubo was modelled
successfdly by Graf
et
al.
(1 993). The warmings following both El Chich6n and Pinatubo were followed by stronger
coolings, in which the chemistry of ozone depletion in the presence of volcanic aerosols may be implicated
(McCormick
et
al.,
1995).
3. DISCUSSION
Our demonstration of the consistent effects of Pinatubo and other tropical eruptions on Northern Hemisphere winter
atmospheric circulation is in accord with recent observational findings (Groisman, 1992; Robock and Mao, 1992,
1995; Graf
et
al.,
1994). These papers, along with the modelling studies of Graf
et
al.
(1993) and Kirchner
and
Graf
(1995), suggest that the following processes may be in operation. The warming of the tropical stratosphere (e.g.
Labitzke and McCormick, 1992) imposes an anomalous westerly thermal wind at these levels in mid-latitudes thus
strengthening the winter polar stratospheric vortex. As a result, more planetary wave energy is thought to be trapped
in the middle and high-latitude troposphere (Graf
et
al.,
1994), giving enhanced westerlies overall,
as
in Figure 3,
and the positive surface temperature anomalies seen in Figure 2(a). The ‘ECHAM2’ model used by Graf
et
al.
(1 993) and Kirchner and Graf
(
1995) has four levels between 10 hPa and
100
hPa and three levels between 100 hPa
and
200
hPa (Roeckner
et
al.,
1992), probably enabling a better simulation of troposphere-stratosphere interactions
than with the GISS model (see section
2.1),
which has layers representing 10-70 hPa, 70-150 hPa, and 15&
200 hPa (Hansen
et
al.,
1996).

496
D.
E. PARKER
ET AL.
A
similar winter atmospheric circulation pattern in the Northern Hemisphere may, however, also occur without
volcanic triggering (Baldwin
et
al.,
1994; Graf
et
al.,
1994; Kodera and Yamazaki, 1994). There were, for example,
enhanced westerlies over northern Eurasia in the winters of 1988-1989 and 1994-1995 (not shown, but Figure l(b)
indicates warmth in 1994-1995), when there was no volcanically induced differential warming of the tropical
stratosphere.
So
although overall post-tropical-eruption surface and tropospheric cooling appears to be real, it may
be difficult to detect because of the consequential and coincidental changes
of
atmospheric circulation, which may
not be unique to post-eruption seasons. In a given location, especially at high latitudes, there may not be a volcanic
cooling signal (Kelly
et
al.,
1996). Furthermore, volcanic aerosols are not the only influence on tropical stratospheric
temperatures. For example, the impacts of the solar 1 1-year cycle
of
ultraviolet radiation, and the QBO, on both the
stratosphere and the troposphere have been modelled by Rind and Balachandran (1995) whose results showed
qualitative agreement with the observed extratropical response (e.g. Labitzke and van Loon, 1988). In a study of the
aftermath
of
the eruption of
El
Chichon, Dunkerton and Delisi (1991) diagnosed competing influences of volcanic
aerosols, the QBO, and extratropical forcing, on the tropical upper stratosphere. However, the volcanic effects
dominated temperature changes in the tropical lower stratosphere,
as
discussed also in section
2.3.
Kirchner and Graf (1995) found the observational data base on its
own
to be inadequate to clearly resolve the
separate and combined climatic impacts of volcanic eruptions and
El
Niiio events. They therefore also analysed
atmospheric model simulations for Northern Hemisphere winter with
El
Niiio and volcanic forcing, using
eigenvector techniques to isolate and assess the El Niiio and volcanic signals. The results showed some accord
with the limited observational evidence, e.g. in the volcanic case there was warming at the surface and in the lower
troposphere over northern Eurasia, despite the limitation of the simulations to perpetual January.
A
combined
empirical-modelling approach is expected to offer the best way forward in this, as in many other, areas of climatic
research.
ACKNOWLEDGEMENTS
Tracy Basnett, Andrew Colman, Robert Hackett, and Matthew O’Donnell provided valuable support in
data
analysis
and graphics production.
REFERENCES
Angell,
J.
K. 1993. ‘Comparison of stratospheric warming following Agung, El Chich6n and Pinatubo volcanic eruptions’,
Geophys. Res.
Lea,
Baldwin, M.
I?,
Cheng,
X.
and Dunkerton, T. J. 1994. ‘Observed correlations between winter-mean tropospheric and stratospheric circulation
anomalies’,
Geophys. Res. Len.,
21,
1141-1
144.
Bottomley, M., Folland, C. K., Hsiung, J., Newell, R. E. and Parker, D. E. 1990.
Global
Ocean
Surjace
Temperature Atlas (GOSTA),
Joint
Meteorological Office/Massachusetts Institute of Technology Project. Project supported by
US
Department of Energy, US National Science
Foundation and
US
Office of Naval Research. Publication funded by UK Departments of Energy and Environment. 20
+
iv pp. and 3 13
plates. HMSO, London.
20,
715-718.
Christy, J. R. and Drouilhet,
S.
J. 1994. ‘Variability in daily zonal mean lower-stratospheric temperatures’,
1
Climate,
7,
106120.
Christy,
J.
R. and McNider, R. T. 1994. ‘Satellite greenhouse signal’,
Nature,
367,
325.
Dunkerton, T.
J.
and Delisi, D.
I?
1991. ‘Anomalous temperature and zonal wind in the tropical upper stratosphere, 1982/1983’,
1
Geophys.
Res.,
96,
2263 1-22641.
Dutton,
E.
G.
and Christy,
J.
R. 1992. ‘Solar radiative forcing at selected locations and evidence for global lower tropospheric cooling following
the eruptions of El ChichQ and Pinatubo’,
Geophys. Res. Lett.,
19,
2313-2316.
Folland, C.
K.,
Karl,
T. R. and Vinnikov, K. Ya. 1990. ‘Observed climate variations and change’, in Houghton,
J.
T.,
Jenkins,
G.
J.
and
Ephraums, J. J. (eds),
Climate Change, the IPCC scienh>c Assessment,
Section 7, WMO/UNEP, IPCC, Cambridge University Press,
Cambridge, pp. 195-238.
Folland, C. K.,
Karl,
T. R., Nicholls, N., Nyenzi, B.
S.,
Parker,
D.
E. and Vinnikov, K. Ya. 1992. ‘Observed climate variability and change’, in
Houghton, J. T., Callander, B. A. and Vamey,
S.
K. (eds),
Climate Change
1992--The
Supplementary Reporr
to
the IPCC scienh&
Assessment,
Section C, WMO/UNEP, IPCC, Cambridge University
Press,
Cambridge, pp. 135-170.
Glantz, M.
H.,
Katz,
R. W.
and
Nicholls,
N.
(eds)
1991.
Teleconnections Linking Worldwide Climate Anomalies,
Cambridge University
Press,
Cambridge,
x
+
535
pp.
Graf, H.-F.,
Kirchner,
I.,
Robock, A. and Schult,
1.
1993. ‘Pinatubo eruption winter climate effects: model versus observations’,
Climate
m.,
9,
Graf,
H.-F., Perlwitz,
J.
and Kirchner,
I.
1994. ‘Northern Hemisphere tropospheric mid-latitude circulation after violent volcanic eruptions’,
Beitr Phys. Amos.,
67,
3-13.
Grant, W.
B.,
Browell, E.
V,
Fishman,
J.,
Brackett,
V
G.,
Veiga, R. E., Nganga, D., Minga, A., Cros, B., Butler, C. F., Fenn, M. A,, Long, C.
S.
and Stowe,
L.
S.
1994. ‘Aerosol-associated changes in tropical Stratospheric
ozone
following the eruption of Mount Pinatubo’,
1
Geophys.
Res.,
99,
8197-8211.
81-93.

MOUNT PINATUBO IMPACT ON TEMPERATURES
497
Groisman,
F‘.
Ya.
1992.
‘Possible regional climate consequences of the Pinatubo eruption: an empirical approach’, Geophys. Res. Lett.,
19,
1603-
Halpert, M.
S.,
Ropelewski, C. F., Karl, T. R., Angell, J. K., Stowe, L. L., Heim, R. R., Jr., Miller, A. J. and Rodenhuis, D. R.
1993. ‘1992
brings
Hansen, J., Fung,
I.,
Lacis,
A.,
Rind, D., Lebedeff,
S.,
Ruedy, R., Russell,
G.
and Stone,
I?
1988.
‘Global climate changes as forecast by Goddard
Hansen,
J.,
Lacis, A., Ruedy,
R.
and Sato, M.
1992.
‘Potential climate impact of Mount Pinatubo eruption’, Geophys. Res. Lett.,
19,
215-218.
Hansen, J., Sato, M., Ruedy, R., Lacis,
A.,
Asamoah,
K.,
Borenstein,
S.,
Brown, E., Cairns, B., Caliri,
G.,
Campbell, M., Cum, B., decastro,
S.,
Druyan, L., Fox, M., Johnson, C., Lemer, J., McCormick, M.
I?,
Miller, R., Minnis, P., Morrison, A,, Pandolfo,
L.,
Rambem,
I.,
Zaucker,
F., Robinson, M., Russell,
F‘.,
Shah,
K.,
Stone,
I?,
Tegen,
I.,
Thomason,
L.,
Wilder, J. and Wilson, H.
1996.
‘A Pinatubo climate modelling
investigation’, in Fiocco,
G.,
Fua, D. and Visconti,
G.
(eds), The EApects ofMt Pinatubo Eruption on the Aimosphere and Climate, NATO AS1
Series Volume, Subseries
I,
Global Environment Change, Springer-Verlag,
233-272.
1606.
return
to moderate global temperatures’, Eos (Am. Geophys. Union),
74,
433439.
Institute for Space Studies three-dimensional model’,
1
Geophys. Res.,
93,
9341-9364.
Hitchman, M. H., McKay, M. and Trepte, C. R.
1994.
‘A climatology of stratospheric aerosol’,
1
Geophys. Res.,
99,
20689-20700.
Jones, P. D.
1988.
‘The influence of ENS0 on global temperatures’, Climate Monitor,
17,
80-89.
Jones, P. D.
1994.
‘Hemispheric surface air temperature variations: a reanalysis and an update to
1993’,
1
Climate,
7,
1794-1802.
Kelly,
P.
M., Jia,
F‘.
and Jones,
I?
D.
1996.
‘The spatial temperature response to large explosive volcanic eruptions’, Int.
1
Climatol., in press.
Kirchner,
1.
and Graf, H.-F.
1995.
‘Volcanoes and El Niiiesignal separation in winter’, Climate Dyn.,
11,
341-358.
Kodera, K. and Yamazaki,
K.
1994.
‘A possible influence of recent polar stratospheric coolings on the troposphere in the Northem Hemisphere
Kolomeev, M.
P.,
Nikonov,
S.
A,, Sorokovikova,
0.
S.
and Khmelevtsov,
S.
S.
1993.
‘Simulation of the Northern Hemisphere climate response to
Labitzke, K.
1994.
‘Stratospheric temperature changes after the Pinatubo eruption’,
1
Amos. Ter,: Phys.,
56,
1027-1034.
Labitzke, K. and McCormick, M. P.
1992.
‘Stratospheric temperature increases due to Pinatubo aerosols’, Geophys. Res. Lett.,
19,
207-210.
Labitzke, K. and van Loon, H.
1988.
‘Associations between the
1
I-year solar cycle, the QBO and the atmosphere. Part
I:
the troposphere and
Madden, R. A., Shea, D. J., Branstator,
G.
W., Tribbia, J. J. and Weber, R.
0.
1993.
‘The effects of imperfect spatial and temporal sampling on
Mass, C. F. and Portman, D. A.
1989.
‘Major volcanic eruptions and climate: a critical evaluation’,
1
Climate,
2,
566593.
McCormick, M.
P.,
Thomason, L. W. and Trepte, C. R.
1995.
‘Atmospheric effects of the Mt Pinatubo eruption’, Nature,
373,
399404.
Pan,
Y.
H. and Oort, A. H.
1983.
‘Global climate variations connected with sea surface temperature anomalies in the eastem equatorial Pacific
Parker, D. E. and Brownscombe, J.
L.
1983.
‘Stratospheric warming following the
El
Chichon volcanic eruption’, Nature,
301,
406-408.
Parker, D. E. and Cox, D.
1.
1995.
‘Towards a consistent global climatological rawinsonde data-base’, Int.
1
Climatol.,
15,
473496.
Parker, D. E. and Jones,
F‘.
D.
1991.
‘Global warmth in
1990’,
Weather,
46,
302-31
1.
Parker, D. E., Jones,
P.
D., Folland, C.
K.
and Bevan,
A.
1994.
‘Interdecadal changes of surface temperature since the late nineteenth century’,
1
Rind,
D.
and Balachandtan,
N.
K.
1995.
‘Modelling the effects of UV variability and the QBO on the troposphere-stratosphere system. Part
11:
Robock, A.
1991.
‘The volcanic contribution to climate change of the past
100
years’, in Schlesinger, M. E. (ed.), Greenhouse-gas-induced
Robock, A. and Liu,
Y.
1994.
‘The volcanic signal in Goddard Institute for Space Studies three-dimensional model simulations’,
1
Climate,
7,
Robock, A. and Mao,
J.
1992.
‘Winter warming from large volcanic eruptions’, Geophys. Res. Let?.,
19,
2405-2408.
Robock, A. and Mao, J.
1995.
‘The volcanic signal in surface temperature observations’,
1
Climate,
8,
10861
103.
Roeckner, E., Arpe, K., Bengtsson, L., Brinkop,
S.,
Diimenil, L., Esch, M., Kirk,
E.,
Lunkeit, F., Ponater, M., Rockel, B., Sausen, R., Schlese,
U.,
Schubert,
S.
and Windelband, M.
1992.
‘Simulation of the present-day climate with the ECHAM model: impact of model physics and
resolution’, Max-Planck-Institut
fiir
Meteorologie Report No.
93,
Hamburg,
172
pp.
Ropelewski, C.
F.
and Jones, P. D.
1987.
‘An
extension of the Tahiti-Darwin Southern Oscillation Index’, Mon. Wea. Rev.,
115,
2161-2165.
Sato, M., Hansen, J. E., McCormick, M. P. and Pollack, J. B.
1993.
‘Stratospheric aerosol optical depths,
185&1990’,
1
Geophys. Res.,
98,
Shine, K. P., Derwent, R.
G.,
Wuebbles, D. J. and Morcrette, J.-J.
1990.
‘Radiative forcing of climate’, in Houghton, J. T., Jenkins,
G.
J. and
Ephraums, J.
J.
(eds), Climate Change, the
IPCC
Scientific Assessment, Section
2,
WMO/UNEP,
IPCC, Cambridge University Press,
Cambridge, pp.
41-68.
Spencer, R. W. and Christy, J. R.
1992.
‘Precision and radiosonde validation of satellite gridpoint temperature anomalies. Part
11:
a tropospheric
retrieval and trends during
1979-90’,
1
Climate,
5,
858-866.
Spencer, R. W. and Christy, J. R.
1993.
‘Precision lower stratospheric temperature monitoring with the MSU: technique, validation and results
1979-1991’,
1
Climate,
6,
1194-1204.
Spencer, R. W., Christy, J.
R.
and Grody, N. C.
1990.
‘Global atmospheric temperature monitoring with satellite microwave measurements:
method and results
1979-84’,
1
Climate,
3,
11
11-1
128.
Trepte, C. R., Veiga, R. E. and McCormick, M.
P.
1993.
‘The
poleward dispersal of Mount Pinatubo volcanic aerosol’,
1
Geophys.
Res.,
98,
winter’, Geophys. Res. Lett.,
21,
809-812.
the eruption of Mount Pinatubo’, Russian Meteorol. Hydrol.,
4,
9-13.
stratosphere in the Northern Hemisphere in winter’,
1
Atmos. Terr Phys.,
50,
197-206.
estimates of the global mean temperature: experiments with model data’,
1
Climate,
6,
1057-1066.
Ocean for the
1958-73
period’,
Mon.
Wea. Rev,,
111,
12441258.
Geophys. Res.,
99,
14373-14399.
The troposphere’,
1
Climate,
8,
2080-2095.
Climatic Change: a Critical Appmisal
of
Simulations and Observations, Elsevier, Amsterdam, pp.
429-443.
4455.
22987-22994.
18563-18573.