Meteorologische Zeitschrift, Vol.10, No.2, 901-908(April 2001)
Meteorologisches Institut, Freie Universität Berlin, Berlin, Germany
Based on the data of 4 solar cycles, the global structure
of the signal of the 11-year sunspot cycle (SSC) in the stratosphere and
troposphere was examined in earlier studies, using correlations
between the solar cycle and heights and temperatures at different pressure
Here, this work is expanded to show the differences of
geopotential heights and temperatures
between maxima and minima of the SSC. The global solar signal is shown
and the differences between the hemispheres are stressed. It is pointed
out that during the northern winter (January/ February) the QBO
modulates the global solar signal
This study puts the earlier work on a firmer ground and gives the community
of modellers, who simulate the solar signal with GCMs,
quantitative values for comparison.
Bisher haben wir die globale Struktur des Signals des 11-jährigen
Sonnenfleckenzyklus (SFZ) in der Stratosphäre und Troposphäre an
Hand von Korrelationen zwischen dem SFZ und Höhen oder Temperaturen
in verschiedenen Druckniveaus untersucht. Dazu hatten wir Daten
von 4 Zyklen des SFZ zur Verfügung, 1958 - 1998.
Hier wird diese Untersuchung ausgedehnt, indem die
Differenzen der geopotentiellen Höhen zwischen den Maxima und
Minima im SFZ berechnet werden. Es werden das globale Signal gezeigt,
die Unterschiede zwischen den Hemisphären betont, und es wird besonders
darauf hingewiesen, daß während des Nordwinters (Januar/Februar)
das globale solare Signal in der unteren Stratosphäre von der QBO
Diese Untersuchung liefert quantitative Werte des solaren Signals
im 11-jährigen Zyklus, die für Vergleiche mit Modellergebnissen
The search for a signal of the 11-year sunspot cycle (SSC) in the heights and
temperatures of the lower stratosphere was previously successfully conducted
for the Northern Hemisphere with a data set from the Freie Universität
Berlin (FUB), covering four solar cycles. The historic analyses from the
FUB start with the IGY (International Geophysical Year) in July 1957 when
new radiosonde stations were established and existing ones were improved. The
analyses were always carried out as a research project with continuity in
time and space and there is general agreement that this data set is very
reliable also during the early years in the series.
This work has been extended to the whole globe by
means of the NCEP/NCAR re-analyses (Kalnay et al., 1996) for the period
1958 - 1998. Correlations based on the re-analyses
show that the solar signal exists in the Southern Hemisphere too, and
that it is of
nearly the same size and shape as on the Northern Hemisphere (van Loon and
Labitzke, 1998 and 1999).
It is the purpose of this paper to show the size of the changes of the
stratospheric heights and temperatures which can be attributed to the SSC.
The period of 1968 - 1998 is used for the global studies as the re-analyses
are less reliable in the early part of the period, before 1968,
mainly because of the lack of
radiosonde stations over the Southern Hemisphere, the lack of high reaching
balloons in the early years and the scarce satellite information before 1979.
As a measure of the SSC the
monthly mean values of the 10.7cm solar flux are used. This is an
objectively measured radiowave, highly and positively correlated
with the SSC. No causal relationship is implied by the use of this parameter.
For the range of the SSC the mean difference of
the 10.7cm solar flux between solar minima (about 70 units) and solar
maxima (about 200 units) is used, i.e. 130 units. Any linear correlation
can be represented also
by a regression line with y=a+bx, where x in this case is the 10.7cm
solar flux and b is the slope. This slope is
used here, multiplied by 130, in order to get the differences between
solar minima and maxima, as presented in this paper.
For the analysis of the solar signal over the Northern Hemisphere, the
FUB 30-hPa height data are used for the period 1958 - 1998. To show the
solar signal on a global scale, the NCEP/NCAR re-analyses
are used for the period 1968 - 1998, because of the scarce data before this
date, particularly over the Southern Hemisphere.
Linear correlations and the calculated differences of stratospheric elements
show directly the synoptic structure of the global solar signal. Except for
the northern winters it is not necessary to involve other
influences like the QBO and the SO, as can be judged from Fig.32.
Figure 31 shows on the left hand side the correlations between the 10.7cm
solar flux and the annual means of the 30-hPa heights for the period
from 1958 to 1998, for the Northern Hemisphere (FUB data). The areas with
than 0.5 are shaded for emphasis. (The statistical significance of the
correlations cannot be determined because
we have only 4 solar cycles and the degrees of freedom are therefore
limited). Clearly, there are large areas with correlations higher than 0.5,
and the axis of highest correlations is at about 30N. Over the polar region
and into middle latitudes the correlations are weak in the annual
The differences of the annual mean 30-hPa heights between solar maxima
and minima are shown on the right hand side of Fig.31. In the areas of
highest correlations the annual mean
30-hPa heights are about 70 geopotential meters higher in the maxima
than in the minima of the 11-year solar cycle.
Figure 32 shows the time series of the 10.7cm solar flux and of the annual
mean 30-hPa heights (and their three-year running means)
at a gridpoint near Hawaii (30N, 150W) where the correlations exceed 0.7;
where half the interannual variance of the heights is associated with the
solar cycle. Obviously, this signal is larger than any other influence
controlling the interannual variability in the subtropics.
Figure 31: Left: Correlations between the 10.7cm
solar flux (the 11--year solar cycle) and the annual mean 30-hPa heights,
1958 - 1998 (FUB data); shaded for emphasis where the correlations are
above 0.5. Right: Height differences (geopot.m) between solar maxima and
Figure 32: Time series of the 10.7cm solar flux (dashed line)
and of the annual mean heights (thin solid line) and their three-year
running means (heavy solid line) in geopot.km for the gridpoint 30N/150W.
The NCEP/NCAR re-analyses of the stratosphere are less reliable in the early part of
the period because of the lack of radiosonde stations, especially over the
Southern Hemisphere, the lack of high reaching balloons in the early
years, and the scarce satellite information before 1979. Therefore, the
shorter period from 1968 til 1998 is used here to analyze the global
The vertical structure of the correlations of the zonal mean temperatures
from 1000 to 10-hPa is shown in Fig.33, together with the
respective temperature differences between solar maxima and minima.
Again, correlations larger than 0.5 are shaded. The largest
correlations (up to 0.6) and the
largest temperature differences (up to 1.5K) are confined to the height
regime between 200 and 20-hPa, and to the latitude range from 45N to 45S
(r above 0.4).
Figure 33: Vertical meridional sections for the period
1968--1998 of (top): the correlations between the zonally
averaged annual mean temperatures and the 11-year solar
cycle (shaded for emphasis where the correlations are above
than 0.5); (bottom): the zonally averaged temperature
differences (K) between solar maxima and minima of the
annual means. (NCEP/NCAR data)
From hydrostatic considerations the temperature differences
lead to height differences further up, and these differences as well as the
correlations are shown in Fig.34. Correlations larger than
0.5 start at 150 hPa/50N (Fig.34, upper part), grow with
height to more than 0.7 around 20 hPa/30N, and above 50 hPa the
correlations larger than 0.5 reach from 60N to 50S.
The zonal mean height differences between solar maxima and minima
(Fig.34, lower part) grow to values larger than 100
geopot.m near 10 hPa.
Figure 34: Vertical meridional sections for the period
1968 - 1998 of (top): the correlations
between the zonally averaged annual mean heights and the
11-year solar cycle (shaded for emphasis where the correlations
are larger than 0.5);
(bottom): the zonally averaged height differences (geopot.m)
between solar maxima and minima of the annual means.
The 70-hPa level is chosen here for the analysis of the horizontal
structure of the temperature signal, because the signal starts to
be well pronounced at this level over both hemispheres, cf.
The horizontal distributions of the correlations with the solar cycle
and of the temperature differences between solar maxima and minima
are shown in Fig.35 for the 70-hPa level. The banded
structure of the correlations (Fig.35, upper panel) with
clear maxima (up to 0.7), centered near 30N and 20S,
respectively, is similar to results shown earlier for the 30-hPa heights
(e.g. van Loon and
Labitzke, 1998 and 1999).
In the annual mean the temperature differences are above 1K from about
40N to 30S, reaching 1.8K in the areas of largest correlations,
Fig.35, lower panel.
Figure 35: Correlations between the 10.7cm solar flux and the
annual mean 70-hPa temperatures, 1968 - 1998; shaded for emphasis
where the correlations are above 0.5; (bottom): temperature
differences (K) between solar maxima and minima;
shaded where the differences are above 1 K. (NCEP/NCAR data)
The horizontal structure of the correlations between the annual mean
30-hPa heights and the 10.7cm solar
flux is shown in Fig.36, together with the height differences between
solar maxima and minima. The extensive belts
with high correlations over both hemispheres stand out clearly (upper panel),
as well as the maxima of the height differences. On this projection the
global range of the solar signal is easily recognized: from about
50N to about 45S, where the correlations are above 0.5, the stratospheric
heights are between 60 and 80 meters higher in maxima than in minima of
the solar cycle. The belts of maxima (up to 80m) lie at about 30N and 20S,
and there is a clear minimum over the equator (about 70m). This implies
an anomalous eastwind in the annual mean over the tropics during solar
maxima (and vice versa in minima) and points to an influence of the
SSC on the QBO, as discussed by Salby and Callaghan (2000).
Figure 36: Same as Fig.35, but for the 30-hPa heights;
upper part: correlations; lower part: height differences; shaded
where the differences are above 75 geopot.m.
In addition to the results in the annual mean it is of interest to follow
the solar signal through the year. This is shown in
Fig.37 with the correlations and with the differences
between solar maxima and minima of the monthly mean zonal
mean 70-hPa temperatures. Most impressive are the
very consistent positive temperature differences at the 70-hPa level:
temperature differences of about 1K are observed through the year from
about 40N to 40S, with maxima over the tropics (up to 1.5K between 30N
and 30S) during the
northern (May through September) as well as the southern (November through
January) summers. As in our earlier results, there exists always a
symmetry around the equator and a second maximum (in the correlations or
height/temperature differences) is found over the other hemisphere,
This hints to a connection to the two branches of the Hadley
circulation (Labitzke and van Loon, 1995). Over both polar regions
correlations are weak and therefore
the temperature differences are not discussed here.
Figure 37: Two-monthly running (top): correlations between the
10.7cm solar flux and the zonal
mean 70-hPa temperatures; (bottom) temperature differences (K) between
solar maxima and minima. (NCEP/NCAR data)
The temperature differences at the 30-hPa level (Fig.38,
upper panel) are very similar to the differences at the 70-hPa level.
That means that throughout the year between
about 40N and 40S and from about 100 hPa up to 20 hPa the whole lower
stratosphere is warmer (between 1 and 2K) during maxima than minima of
the SSC. Such a value
of temperature change (1 to 2K) has been required by
Salby and Callaghan (2000) to explain their results. The strong signal in the temperatures
at the 30-hPa level implies a continuation of the solar signal in the
heights further up in the middle stratosphere, above the 10-hPa level
which is the highest level available from the NCEP/NCAR re-analyses.
The structure of the differences of the zonal mean 30-hPa heights between
solar maxima and minima, Fig.38 (lower panel) is also
very clear: except for the polar regions where the correlations
(not shown) are weak, the heights are always higher during maxima
of the solar cycle.
Between 60N and 40S the zonal mean 30-hPa heights are practically 75 to
100 geopot. meters higher during the maxima, except for the late
northern winter when the dynamics of the arctic polar vortex are
disturbing the solar signal from the northern subtropics to the
Antarctic (Labitzke and van Loon, 2000), Section 6.5.
The temperature (Fig.37) and height differences
(Fig.38) over the subtropics agree with our earlier work
using radiosonde stations (Labitzke and van Loon, 1995).
We showed that an increase in the temperatures from minimum
to maximum in the 11-year solar cycle is found already in the
upper troposphere and that the height increases observed in the stratosphere
must follow from the hydrostatic relationship.
We suggested that the positive temperature differences could be
explained to some extend by an intensified Hadley circulation, i.e. intensified
downward motion in the upper troposphere during solar maxima
in the ring of largest positive correlations.
The concurrent correlations and height differences over the
Southern Hemisphere agree
with this idea, as there are always two cells (downward branches)
of the Hadley circulation, moving meridionally as the sun with the seasons.
Figure 38: Two-monthly running (top): 30-hPa temperature
differences (K) and (bottom): 30-hPa height differences
(geopot.m) between solar maxima and minima. (NCEP/NCAR data)
We have shown several times that during the northern winters the solar
signal is weak if one uses a full
series of stratospheric heights for correlation with the 11-year solar
cycle (Labitzke, 1987; Labitzke and van Loon, 1988; van Loon and Labitzke, 1994). The QBO modulates
the solar signal and it is necessary to group the data into years when
the QBO in the lower stratosphere (about 45 hPa) was in its west phase
and years when it was in its east phase.
Figure 39 displays the correlations for the Northern
Hemisphere winters (FUB data), grouped
into the east and west phases of the QBO (left hand panels). During the east
phase (upper left) the correlations are similar to those of
the annual mean (Fig.31) and of the northern summer
(not shown). Over the Aleutian Islands the height differences
between maxima and minima of the solar cycle for the northern winters in
the east phase (upper right hand panel) are larger than 150m.
This suggests a dynamical response
which during solar maxima leads to an intensification of the
stratospheric Aleutian high, i.e. the planetary height wave 1.
The Aleutian high, on the other hand results from warming through sinking
motions on the poleward side of the strong Asian tropospheric jet stream.
Therefore, one may speculate that this jet stream is also influenced by
the solar cycle.
Over the rest of the
hemisphere a belt of positive anomalies is found around 30N,
similar to the pattern in the maps discussed before. The positive height
anomalies intensify the polar night jet during solar maxima.
An intensification of the Aleutian high is connected
with an intensification of the polar vortex through teleconnections
(Shea et al., 1992), therefore
the negative correlations and the negative differences over the Arctic
can be understood in this context. The negative differences over the north
polar region are also associated with an intensification of the
stratospheric polar night jet during solar maxima, as far as the
winters during the east phase of the QBO are concerned. Labitzke and van Loon (2000) showed that this effect is most pronounced in the latter part of the
winter, namely in February.
During the west phase of the QBO, the correlations are positive
and the height differences large, up to +450 geopot.m., over the
Arctic in January/February, due to the fact that in this phase of
the QBO major stratospheric warmings tend to take place during
solar maxima (Fig.39, lower panels). These strong
stratospheric warmings over the Arctic are connected with widespread
downward motion over the Arctic and middle high latitudes. At the
same time upward motion and cooling takes place outside the high
latitudes over the rest of the Northern Hemisphere and far into the
Southern Hemisphere, as observed first by Fritz and Soules (1970). This
cooling acts against the usual warming (positive correlations,
e.g., Fig.31) during solar maxima and therefore the correlations
are weak and the differences are small outside the high northern
Figure 39: Same as Fig.31, but for January/February and grouped
into the years in the east phase of the QBO, upper two panels, and
the years in the west phase of the QBO, lower two panels. No shading
is applied here for the height differences. (FUB data, 1958 - 1998)
The differences between east and west phase in the QBO are also seen
over the Southern Hemisphere, although it is summer there,
the east phase of the QBO the pattern of the solar signal is similar to
that of the northern summer, (not shown). Correlations larger than 0.5
cover most of the Southern Hemisphere, with the largest correlations
in a belt between 10 and 30S, (Fig.40, upper left panel).
From Australia into the Pacific and from South America across the
Atlantic, there are large areas with correlation coefficients higher
During the east phase the height differences over the Southern Hemisphere in
summer, Fig.40 (upper right panel), are of the same
magnitude as over the northern summer hemisphere, (not shown).
As discussed already above, during the west phase of the QBO
(Fig.40, lower panels), the solar signal has nearly
disappeared in the Southern Hemisphere, as the dynamic interaction
from the Northern (winter) Hemisphere counteracts the solar signal
over the Southern (summer) Hemisphere
during this phase of the QBO.
Figure 40: Same as Fig.39, but for the Southern Hemisphere;
shaded where the differences are above 70 geopot.m. (NCEP/NCAR data, 1968 - 1998)
For the annual mean the magnitude of the solar signal is analyzed on the
time-scale of 11-years.
solar maxima and minima are calculated and it is shown that over a
large vertical and
latitudinal range the heights and temperatures in the stratosphere
are higher during solar maxima than during solar minima.
In most months (and in the annual mean) the signal is larger over the
Northern Hemisphere, where
it extends farther poleward und also farther down into the upper troposphere.
We suggest that the solar effect influences the diabatic meridional
over the tropics and subtropics, and particularly the Hadley circulation in
the sense that the solar signal
intensifies the Hadley circulation in the maximum phase of the solar cycle.
During the northern winters the QBO modulates the global solar signal.
These results support our earlier work and give the community
of modellers who are simulating the solar signal with GCMs
quantitative values for comparison (e.g., Haigh, 1996;
Balachandran et al., 1999; Shindell et al., 1999).
We thank the members of the Stratospheric Research Group, FUB for
professional support and Ms. S. Leder for doing the computations and
graphics. Thanks are due to Harry van Loon for comments on the manuscript.
Drs. A. Matthews and G. Bodeker
provided a scientific haven for K.Labitzke during her sabbatical stay in
NIWA/Lauder, New Zealand. The project was partly funded by the
BMBF (Bundesministerium für Bildung und Wissenschaft) within KIHZ.
The 10.7cm solar flux data are from the World Data Center A, Boulder,
Balachandran, N., Rind, D., Lonergan, P. and D. Shindell, 1999:
Effects of solar cycle variability on the lower stratosphere and the
troposphere. J. Geophys. Res., 104, 27321-27339.
Fritz, F. and S. D. Soules, 1970: Large--scale temperature changes
in the stratosphere observed from Nimbus 3. J. Atmos. Sci., 27,
Haigh, J. D., 1996: The impact of solar variability on climate.
Science, 272, 981-984.
Kalnay, E., Kanamitsu, R., Kistler, R., Collins, W., Deaven, D.,
Gandin, L., Iredell, M., Saha, S., White, G., Zhu, Y., Chelliah, M.,
Ebisuzaki, W., Higgins, W., Janowiak, J., Mo, K.C., Ropelewski, C., Wang, J.,
Reynolds, R., Jenne, R., Joseph, J., 1996: The NCEP/NCAR 40-year reanalysis
project. Bull. Am. Meteor. Soc., 77, 437-471.
Labitzke, K., 1987: Sunspots, the QBO, and the stratospheric
temperature in the north polar region. Geophys. Res. Lett., 14, 535--537.
Labitzke, K. and H. van Loon, 1988: Associations between the 11-year solar
cycle, the QBO and the atmosphere. Part I: The troposphere and stratosphere
in the northern hemisphere winter. J.A.T.P., 50, 197-206.
Labitzke, K. and H. van Loon, 1995: Connection between the
troposphere and the stratosphere on a decadal scale. Tellus, 47 A,
Labitzke, K. and H. van Loon, 2000: The QBO effect on the global
stratosphere in northern winter. J.A.S.-T.P., 62, 621--628.
Salby, M. and P. Callaghan, 2000: Connection between the solar cycle
and the QBO: The missing link. J.Clim., 13, 2652--2662.
Shea, D.J., van Loon, H. and K.Labitzke, 1992: Point correlations
of geopotential height and temperature at 30mb and between 500mb and 30mb.
NCAR/TN-368+STR, 6pp plus 285pp maps.
Shindell, D., Rind, D., Balachandran, N., Lean, J., and J.
Lonergan, 1999: Solar cycle variability, ozone, and climate. Science, 284, 305-308.
van Loon, H. and K. Labitzke, 1994: The 10--12 year atmospheric
oscillation. Met. Zeitschr., N.F., 3,259--266.
van Loon, H. and K. Labitzke, 1998: The global range of the
stratospheric decadal wave. Part I: Its association with the sunspot cycle
in summer and in the annual mean, and with the troposphere. J.Clim., 11,
van Loon, H. and K. Labitzke, 1999: The signal of the 11--year
solar cycle in the global stratosphere. J.A.S.-T.P., 61,
Stratospheric Research Group
Last modified: Fri Mar 14 16:21:21 MET 2003