Introduction
There is a wealth of possible sources of natural variability in stratospheric
ozone. The concentration of stratospheric ozone varies as a result of
different factors, some interacting among themselves through their effects on
chemistry and transport. The seasonal variations, given in terms of annual
and semi-annual oscillations (AO, SAO), have been studied for many years
e.g.. On interannual timescales dynamical
feedbacks in the Earth system lead to effects of originally tropical
phenomena, such as quasi-biennial oscillation (QBO) and El Niño–Southern
Oscillation (ENSO), on midlatitude wave structures and wave propagation.
Planetary waves play an essential role in driving the zonal mean transport by
the Brewer–Dobson circulation (BDC) and eddy-mixing processes
. This affects the zonal mean meridional transport of
trace gases from the tropics to midlatitudes and polar latitudes in the
stratosphere and also produces variations in the strength of the polar winter
vortices and stratospheric warming events . On the other hand,
naturally induced ozone variability is also caused by interannual changes in
solar ultraviolet spectral irradiance . Many studies
regarding the interannual variability can be found in the literature
e.g.. Even though a number of mechanisms have been proposed as
interpretations of the natural ozone variations in previously mentioned
analyses,
there are still open questions in the attribution of the causes to the
effects observed in the stratosphere. For instance, discrepancies between
different model simulations e.g. or inconsistencies
between model simulations and observations e.g. in the
response of ozone to a natural oscillation complicate the understanding of
the cause of the observed effect. Moreover, these modes of variability do not
always play a role in isolation; there are numerous examples
e.g. for which the interactions are known to
occur, throwing more complexity into the understanding of this topic. A
better comprehension of the natural oscillations would offer a better
recognition and predictability of stratospheric ozone trends. As a matter of
fact the knowledge of the natural ozone variations is useful to unmask
effects on trends, i.e. to distinguish between natural signals and
anthropogenic signals. In addition, the anthropogenic signal can influence
the natural signal, for example the recently postulated changes in the
Brewer–Dobson circulation (BDC) in response to increasing greenhouse gases
. Consequently, the understanding of ozone variability is
very useful for the detection and attribution of long-term changes.
An analysis of this sort at a single station may offer valuable information,
useful not only for the comprehension at regional levels but also for the
validation of model simulations. In fact, have found that
model simulations of the ozone response to solar variability are inconsistent
with satellite observations. Therefore, observational studies can be very
helpful to shed light on the subject. In addition, our station can contribute
to the understanding of the natural oscillations since there are just a few
observational studies based on ground-based stations e.g. of naturally induced stratospheric ozone
variability in midlatitudes. The GROMOS (GROund-based Millimeter-wave Ozone
Spectrometer) ozone radiometer has been performing continuous observations of
stratospheric ozone profiles since 1994 above Bern (46.95∘ N,
7.44∘ E; 577 m). GROMOS is part in the Network for the Detection of
Atmospheric Composition Change (NDACC). Long-term ground-based measurements
allow the empirical characterisation of natural cycles in stratospheric ozone
as a function of altitude. In we applied the multilinear
parametric trend model to derive the long-term trend in
stratospheric ozone above Bern. Here, we optimise this method to determine
the basic natural oscillations in stratospheric ozone over the past 18 years.
The selection of the time interval was the same as in ,
based on the assumption that the concentration of equivalent effective
stratospheric chlorine (EESC) peaked in 1997 at midlatitudes .
The regression model includes a linear term, the annual and semi-annual
oscillation, QBO, ENSO and the solar activity cycle. Many studies use
regression analyses to evaluate natural variability in the middle atmosphere
e.g..
The layout of this study is as follows: the description of the
data sources employed is provided in Sect 2. Details of the regression
technique are given in Sect. 3. Section 4 presents a short summary of the
most important processes for stratospheric ozone production and destruction.
Section 5 describes the results of the analysis and provides an overview of
the possible causes of the natural oscillations observed in stratospheric
ozone. Finally the conclusions are summarised in Sect. 6.
Data sources
The present study is based on stratospheric ozone profiles observed by
GROMOS. The instrument is a ground-based ozone microwave radiometer which is
part of the NDACC. It has been continuously observing the middle atmosphere
above Bern, Switzerland (46.95∘ N, 7.44∘ E, 577 m above
sea level) since November 1994. It measures the thermal microwave emission of
a rotational transition of ozone at 142.175 GHz. The altitude range of the
retrieved ozone profiles covers 25 to 70 km with a vertical resolution of
8–12 km in the stratosphere. The measurement contribution between 50 and
0.5 hPa (20 to 52 km) is higher than 0.8 (corresponding to an a priori
contribution of less than 20 %); therefore the retrieved ozone values at
these altitudes are primarily based on the measured line spectrum. For
technical details, measurement principle and retrieval procedure on the
instrument, see for example and and
references included therein. The vertical ozone profiles from GROMOS have
been validated by means of nearby ozone sondes, ground-based stations and are
in good agreement with satellite observations. Its data set has been used for
studies of ozone–climate interaction, middle atmospheric dynamics as well as
for long-term monitoring of the stratospheric ozone layer and for the
detection of trends .
The data used to analyse temperature, zonal wind, meridional wind and vertical
wind for natural variability are from the European Centre for Medium-Range
Weather Forecasts (ECMWF) operational analysis for the given location and time
interval. In addition, we have also utilised temperature profiles measured by
TEMPERA (temperature radiometer) microwave radiometer. This instrument was
located at the University of Bern, as well as GROMOS, until the end of 2013,
after which it was moved to Payerne, Switzerland (46.82∘ N,
6.95∘ E, 491 m a.s.l. and 40 km southwest of Bern) in the frame
of a measurement campaign. TEMPERA is a novel ground-based microwave
radiometer that measures the thermal radiation emitted by molecular oxygen in
the microwave spectrum region in a frequency range from 51 to 57 GHz. This
radiation contains information on the atmospheric temperature. This is the
first radiometer that provides temperature profiles in the troposphere and in
the stratosphere at the same time. In this study we only use the
stratospheric temperature profiles with an altitude range of 18–50 km (70
to 0.7 hPa) and a vertical resolution of 15 km, retrieved with a
measurement response higher than 0.6. The TEMPERA radiometer is described in
more detail in and .
Regression analysis
The regression analysis of the time series of ozone monthly means from GROMOS
and the other monthly mean products (temperature, zonal wind, meridional wind
and vertical wind) from ECMWF for the period from January 1997 to January
2015 has been carried out using the following multilinear regression
function:
y^(t)=a+b⋅t+c1⋅qbo1(t)+d1⋅qbo2(t)+e⋅F10.7(t)+f⋅MEI(t)+∑n=12(gn⋅sin(2π⋅tln)+hn⋅cos(2π⋅tln)),
where t is the time, and a and b are the constant term and the linear trend
of the fit. The QBO indices are qbo1 and qbo2. These
terms were implemented by using the normalised Singapore zonal winds at 30
and 50 hPa. These are provided by the Free University of Berlin via
http://www.geo.fu-berlin.de/met/ag/strat/produkte/qbo/index.html. The
F10.7 term is the normalised time series of the solar radio flux at 10.7 cm,
which is a proxy of the solar activity cycle. Moreover, the MEI term
represents the normalised Multivariate ENSO index (MEI) time series used to
identify the ENSO variability during the regression analysis. Both indices
are available from www.esrl.noaa.gov/psd/data/climateindices/list. The
sum term comprises 2 sine and cosine functions with the period length
ln, which represent the annual and semi-annual cycles. The usage of sine
and cosine functions give access to the amplitude and the phase of each
harmonic. In contrast to the harmonics and QBO, the multilinear
regression model has no access to the phase of the solar activity cycle and
ENSO. This results in the detection of the instantaneous response of ozone to
the solar activity cycle and ENSO. Nevertheless, we already know from the
literature e.g. that the impact of the
solar activity cycle and ENSO in ozone have a certain time lag. In order to
avoid this problem the fitting of these terms was made with a time shift
(1-year delay for the solar activity cycle and 1-season delay for the ENSO).
The coefficients a, b, c1, d1, e, f, g1, g2,
h1 and h2 are fitted to the monthly means using the method of
. The uncertainties of the monthly means are also
required for the regression analysis. For more details on this we refer to
.
In the first panel of each column the GROMOS monthly
means in blue and the calculated fit in red is represented. Every column represents a
different pressure level, representative of the lower, middle and upper
stratosphere (23, 10 and 3 hPa respectively). In the second, third and
fourth panel of each column the ozone fitted signals of the proxies
QBO (magenta line), solar F10.7 cm flux (red line) and ENSO (green line) are shown.
The last panel of every column shows the residuals.
Figure shows three examples of the fit in the
lower, middle and upper stratosphere (23, 10 and 3 hPa). The ozone monthly
means measured by GROMOS (blue line) and the calculated fit (red line) are
represented in the first panel of each column. In the second, third and
fourth panels of every column the ozone fitted signals of the
proxies QBO (magenta line), solar F10.7 cm flux (red line) and ENSO (green
line) are shown. Finally in the lowermost panels the residuals are represented. The
residuals are within 0.5 ppm except for some special cases. In the lower
stratosphere the regression model explains about 50 % of the variance
whereas in the middle and upper stratosphere it explains around 80 %.
Stratospheric ozone
In this section, we give a brief overview of the most important processes for
the photochemistry, chemistry and transport of stratospheric ozone. In 1930,
postulated that ozone is formed by the photolysis of
O2 at wavelengths shorter than 240 nm (Reaction R1), immediately
followed by the recombination of atomic oxygen with molecular oxygen and any
mediating air molecule M via the three-body Reaction (R2)
. Ozone is removed locally by both transport and chemical
processes . proposed the photolysis
of ozone (Reactions R3 and R4) and its recombination with
atomic oxygen (Reaction R5) to balance the production of ozone. The
photodissociation of ozone leads to the formation of oxygen atoms in either
their ground state (3P) or in their first excited state (1D)
.
O2+hν→O+OO2+O+M→O3+MO3+hν(λ≥320nm)→O2+O(3P)O3+hν(λ≤320nm)→O2+O(1D)O+O3→O2+O2
Because atomic oxygen and ozone molecules are rapidly interconverted, it is
useful to consider both as a family: the odd-oxygen family. Since a typical
timescale for meridional transport is of the order of months, the relevant
quantity in this context is the concentration of odd oxygen, not that of its
components. The lifetime of odd oxygen in a parcel of air is much longer than
the lifetime of an individual O atom or O3 molecule.
The chemical lifetime of odd oxygen ranges from weeks at 30 km to a year at
20 km. Therefore, odd oxygen has a sufficiently long lifetime to be influenced
by meridional transport processes . Consequently
transport processes, for example the Brewer–Dobson circulation, contribute
to the stratospheric ozone distribution.
The partitioning of odd oxygen depends upon the photolysis rate of ozone, the
O + O2 reaction rate coefficient and the air density. The photolysis
rate generally depends on the absorption cross section of ozone and the
number of incident UV photons. The number of photons in turn depends upon a
number of other parameters: altitude, latitude, season and local time. All of
these parameters implicitly depend on the solar zenith angle. As the seasons
change from winter to summer, the solar zenith angle decreases. Therefore,
the path that UV photons must travel is shorter in summer than in winter.
Consequently, the photolysis rate coefficients at the upper
stratosphere and lower mesosphere have maximum values during summer, when the
path length is shorter, and minimum values during winter, when the path length
is longer. On the other hand, the rate of photolysis depends on both the
number of UV photons and the number of ozone molecules available to interact
with photons.
Results and discussion
Amplitudes of the natural oscillations
Amplitudes of the natural oscillations in stratospheric ozone
derived by multilinear regression from the GROMOS observations at Bern
(1997–2015). The ozone mean profile divided by 10 is shown by the green
line.
The aim of a regression study is to reproduce the evolution through time of
the variable under assessment by means of a linear combination of basic
functions. To achieve this goal the regression model includes basic functions
representing, in our case, the solar activity cycle, the El
Niño–Southern Oscillation (ENSO), the quasi-biennial oscillation (QBO)
and the annual and semi-annual oscillation. As a result of using them to fit
the ozone monthly means during the regression analysis, we may also use them
to quantify the natural variability of ozone. The left panel of
Fig. shows the amplitude of the regression coefficients of these
terms in ppm, along with the ozone mean profile (green line) divided by 10 in
order to plot all these quantities together, whereas in the right panel the
amplitudes are plotted in percent. The magenta line represents the annual
oscillation (AO), the semi-annual oscillation (SAO) is the cyan line, the
orange line is the QBO, ENSO is the red line and the solar radio flux at
10.7 cm (F10.7) is represented by the blue line. The amplitude of the AO
dominates at 10 and 2 hPa whereas the SAO has its maximum at 3 hPa. Near
3 hPa at midlatitudes, the magnitude of the SAO amplitude is larger than
the magnitude of the AO amplitude. This effect is also observed by
. After the annual cycle, the solar variability seems to
be the largest source of variation in ozone, exhibiting its influence around
20 hPa. However, the observed 11-year oscillation in stratospheric ozone
observed by GROMOS microwave radiometer could be influenced by interfering
processes which we will discuss later.
Annual oscillation (AO)
The amplitude of the ozone-AO depending on the months of the year is
represented in % in the left panel. The profile of the amplitude of ozone-AO
in % (blue line) and its phase given as the month of the maximum (red line)
is shown in the right panel.
In the middle stratosphere our observations indicate maximum ozone
concentrations in spring–summer and a minimum during autumn–winter. The amplitude of the maximum at 10 hPa (32 km) is of the order of
16 % (blue line in Fig. , right panel) or around 1 ppm (magenta
line in Fig. , left panel). On the other hand in the upper
stratosphere during the spring–summer period we observe a minimum and a
maximum in autumn–winter. This maximum has a peak amplitude around
0.6 ppm (magenta line in Fig. , left panel) or over 16 % (blue
line in Fig. , right panel) over 2 hPa (42 km). Our results are
in agreement with other observational and modelling studies
. For instance,
obtained a maximum AO-amplitude of the order of 15 %
at 26 km and around 18 % at 40 km with observations between 1995 and 2002
from the Bordeaux microwave radiometer. Further, we are in agreement with the
results of , who used the same regression technique as us,
but data from MIPAS on ENVISAT for the period of July 2002 to April 2012. The
seasonal variation of ozone can be understood through the partitioning of odd
oxygen. Due to the higher flux of shortwave radiation in summer more odd
oxygen would be expected in the upper stratosphere. The main reason for the
minimum ozone in summer in the upper stratosphere is known to be the
temperature-dependent photochemistry. The temperature-induced ozone change is
principally dominated by the NOx, ClOx and HOx catalytic ozone
destruction cycles and also by Reaction (R5)
, which is sensitive to temperature variations. As a result,
ozone depletion is higher in summer than in winter at the stratopause as we
can see in Fig. ; the annual ozone maximum around 2 hPa occurs in
winter. This feature shows the anti-correlation of ozone and temperature in
the stratopause region. After the equinox, when temperatures start to be
warmer the odd-hydrogen and odd-oxygen reactions proceed faster and ozone is
destroyed more rapidly. During wintertime, when temperatures are lower,
these reactions proceed more slowly, causing the ozone density to increase
.
Below the upper stratosphere the anti-correlation of ozone and temperature
breaks down, partly because of the small concentration of atomic oxygen and
the high air density at lower altitudes where atomic oxygen immediately
converts into ozone by Reaction (R2). Indeed, Fig. clearly
shows a phase reversal of the ozone-AO at 3 hPa, where the partitioning of
odd oxygen is balanced by the change from more abundance of atomic oxygen
above 3 hPa to more abundance of ozone below 3 hPa.
suggested that the dominant cause of annual ozone amplitude is the annual
variation in the production rate of odd oxygen. In the middle and lower
stratosphere, the production rate of odd oxygen is roughly equal to the
production rate of ozone because of Reaction (R2). This explains the
summer ozone maximum when the production rate of odd oxygen is maximal.
Figure (magenta line, left panel) and Fig. (blue line)
show that the annual ozone maximum at 10 hPa occurs during early summer. The
red line in Fig. represents the phase of the ozone-AO given as the
month of maximum, and in the ozone-AO maximum at 10 hPa we can clearly see
the phase (red line) around 5–6 months, i.e. May–June (early summer). This
summer ozone maximum is also influenced by the meridional transport of
ozone-rich air from the tropics .
Semi-annual oscillation (SAO)
The profile of the amplitude of ozone-SAO in ppm (blue line) and its
phase given as the month of the maximum (red line) are shown in the left and
middle panel. The amplitude of the ozone-SAO depending on the months of the
year is represented in % in the right panel.
The semi-annual oscillation (SAO) has a period of 6 months. The ozone-SAO
amplitude maximum is around 3 hPa (40 km) and exhibits its maximum of
slightly over 0.4 ppm (blue line in Fig. , left panel) or around
7 % (cyan line in Figure , right panel).
obtained the same order of magnitude for the amplitude of ozone-SAO
calculated through an iterative spectral analysis with data from GROMOS for
the period 1994–2004. By using the same regression method and data
(July 2002–April 2012) from MIPAS got an ozone-SAO
amplitude for our latitude around 0.3 ppmv at 40 km.
showed an amplitude of ozone-SAO of the order of 0.3 ppmv around 40 km and
at 40∘ N latitude with measurements from SABER (Sounding of the
Atmosphere using Broadband Emission Radiometry) for the period 2002–2005.
On the other hand, found an SAO amplitude of about
0.5 ppmv around 3 hPa with 9 years (October 1978–September 1987) of
SBUV (Solar Backscatter Ultraviolet)
ozone-mixing ratio data. The middle panel of Fig. shows the phase
of ozone-SAO, the month of the maximum is between 3.5 and 4 months at 3 hPa,
within the period of the SAO. This can be confirmed in the right panel of
Fig. , where the ozone-SAO amplitude is represented against the
months of the year. These peak amplitudes coincide in time with the
equinoxes, whereas the minima coincide with the solstices. The most
interesting point about the ozone-SAO is that its maximal amplitude is
located in the same altitude region (upper stratosphere) as the maximal
anti-correlation between ozone and temperature .
Amplitude of the temperature-, zonal wind-, meridional wind-,
vertical wind-, ozone-SAO throughout the year at 3 hPa. The ozone-SAO is
multiplied by 10.
In order to understand the ozone-SAO, we have also performed a regression
analysis for temperature, zonal wind, meridional wind and vertical wind, all
of them from the ECMWF operational analysis data over Bern. Since the purpose
of this study is to explain the natural variations of ozone only, we are not
going to provide explanations for the ECMWF products used but we take them as
given. Figure shows the semi-annual cycle at 3 hPa along the
year. The amplitude of the temperature (red line), zonal wind (orange line),
meridional wind (cyan line), vertical wind (green line) and ozone (blue
line)-SAO are plotted together in order to have an overview of which
parameter or parameters play a major role in the ozone-SAO. In
Fig. the ozone-SAO amplitude is multiplied by 10. The ozone-SAO is
1.5 month out of phase with that of temperature in the upper stratosphere.
This is consistent with ozone photochemistry . At 3 hPa, the
first positive maximum of ozone occurs at the beginning of March, 1 month
after the time of the first SAO-temperature maximum. The stratopause zonal
wind-SAO is characterised by the occurrence of easterly winds during the
solstice seasons and westerly winds during the equinoxes. The warm phase
descends with the easterly shear zone in the solstice season. The cold phase
descends with the westerly shear zone around the equinox. The zonal wind
reversal of the SAO from westward to eastward wind is likely driven by
gravity waves. Otherwise, zonal wind reversal from SAO eastward to westward
wind is assumed to be mainly driven by horizontal advection and meridional
momentum transport of extratropical planetary waves .
High-speed Kelvin wave, gravity wave propagation into the upper
stratosphere and the resultant deposition of eddy zonal momentum are thought
to be the most significant forcing of the westerly acceleration
. The out-of-phase relationship between temperature and
vertical wind (upwelling (w>0), green line in Fig. ) is expected
due to adiabatic cooling associated with the ascent of ozone-rich air.
Therefore, the ozone maximum is found during the equinoxes at 3 hPa, as
shown in Fig. .
Ozone VMR from GROMOS (blue line), temperature from TEMPERA (black
line) and ECMWF (red line) at 3 hPa for the period from January 2012 to
January 2015.
Easterly accelerations, on the other hand, appear to be forced meridionally
by planetary waves and meridional advection . During
wintertime planetary wave breaking in the upper stratosphere is observed,
producing poleward transport and a downward flow (w<0, green line in
Fig. ). The downwelling, through air compression, yields an
increase in the temperature. The anti-correlation of ozone and temperature is
maximal around 3 hPa (upper stratosphere) during winter over Bern
. In Fig. , we observe that an increase in
temperature of around 20–30 K causes a decrease in ozone volume-mixing
ratios in winter of the order of 1–1.5 ppmv. Figure shows the
ozone VMR (volume mixing ratio) from GROMOS (blue line), the temperature from TEMPERA radiometer
(black line) and the temperature from ECMWF operational data (red line) near
3 hPa for the period between January 2012 and January 2015. The good
agreement between the temperature from TEMPERA and from ECMWF at this
altitude is clearly seen in Fig. . We observe the two-peaked ozone
curve in each calendar year (ozone-SAO), the minima in ozone during
wintertime and summertime and the maxima during spring and autumn. Regarding
the temperature, we notice the fluctuations during wintertime, the increase
in spring- and summertime and the decrease after the summer solstice. The
aforementioned anti-correlation of ozone and temperature is clearly visible,
displayed by the upper stratospheric warming events and the associated ozone
decrease in winter. These stratospheric warming events in winter are related
to planetary wave breaking (PWB). Planetary waves can break and cause
disruptions to the polar vortex and rapid warmings of the stratosphere. The
occurrence of warming events are even more frequent in the upper stratosphere
compared to the mid-stratosphere as the climatology study by
showed. Thus, the upper stratospheric warmings contribute
to the temperature-SAO and the latter contributes to the ozone-SAO. A
possible role of stratospheric warmings for the generation of the ozone-SAO
at midlatitudes was previously mentioned by and
. and showed that the
strongest oscillations of ozone above Bern are because of displacements of
the polar vortex by planetary wave breaking. The temperature-induced ozone
change is mainly dominated by the NOx catalytic ozone destruction cycle,
Reaction (R5) and in lesser extent by Reaction (R2), which are
sensitive to temperature variations . The rate
of NOx cycle and Reaction (R5) increases when the temperature rises
whereas the rate of Reaction (R2) slows down as the temperature
increases. Furthermore, the observed summertime amplitude dip in ozone (blue
line in Fig. ) is also due to warming; in this case to the increase
of temperature in summer. The ozone destruction rates are accelerated during
spring as the temperature rises, causing the observed amplitude drop in ozone
after March.
Quasi-biennial oscillation (QBO)
Amplitude of the ozone-QBO determined by multilinear regression for
the time interval from January 1997 to January 2015.
The quasi-biennial oscillation (QBO) dominates the variability of the
equatorial stratosphere and is easily identified as an alternation of
descending westerly and easterly wind regimes, with a variable average period
length of approximately 28 months. Even though the QBO is a tropical
phenomenon, it affects the stratospheric flow from pole to pole and
consequently some chemical constituents such as ozone are affected through
modulation of extratropical wave propagation induced by the QBO
. The QBO was implemented in the multilinear parametric
trend model using the Singapore zonal winds at 30 and
50 hPa as a proxy, which are approximately phase-shifted by a quarter period
so that they are sine and cosine functions of the same period (28.8 months)
. Their combination can emulate any QBO phase shift
. were the first to use the QBO proxies
in this multilinear parametric trend model. In Fig. and in
Fig. (orange line) we can observe that the amplitude of the
ozone-QBO maximum is located around 0.15 ppmv (not exceeding 3 %) near
30 hPa (24 km). Our results are in relatively good agreement with those of
, with ozone-QBO amplitudes of about 5 % at 25 km for
the period 1994–2004. These findings also agree with those by
, who obtained an ozone-QBO amplitude of slightly
over 0.2 ppmv in our latitude region at 25 km between 2002 and 2012. During
the easterly (westerly) shear zone of the Singapore zonal winds the amplitude
of ozone-QBO observed in Fig. is positive (negative). This effect
is related to the QBO-induced meridional circulation and confirms the
midlatitude ozone-QBO out of phase relationship with the equatorial
ozone-QBO. The amplitude of the equatorial ozone-QBO is positive (negative)
during the westerly (easterly) shear zone .
The meridional circulation affects chemical tracers such as ozone and gives
rise to strong ozone-QBO signals in such tracers at all latitudes
. Some models and observational studies
found that the downward
(upward) vertical motion in the equatorial westerly (easterly) shear zone
induces an increase (decrease) in ozone whereas in the midlatitudes it
induces a decrease (increase) in ozone, which we observe in our study.
Solar activity cycle
The solar radiation between 200 and 240 nm is primarily responsible for the
formation of ozone in the stratosphere. Changes in solar UV spectral
irradiance directly modify the production rate of ozone in the upper
stratosphere. The solar activity cycle is dominated by an 11-year solar
cycle. The solar activity cycle effect on upper stratospheric ozone is a
direct consequence of heating and photochemistry (Reactions R1 and
R2 and NOx cycle). The lower stratospheric response in ozone occurs
mainly by means of a dynamical response to solar UV variations
. Further, the 11-year solar cycle influence on planetary
wave propagation will influence the strength of the polar vortex as well as
the strength of the mean meridional Brewer–Dobson (BD) circulation since this
is forced by planetary wave transfer of momentum. Therefore, a solar cycle
variation in ozone amount is to be expected . In order to
take into account the effect of solar activity cycle in the stratospheric
ozone, a time series of the normalised solar radio flux at 10.7 cm, which is
a proxy of the solar activity, was fitted to the ozone monthly means during
the regression analysis. Stratospheric ozone has a delayed response to the
solar activity cycle e.g.. Therefore, in
order to estimate this time lag we checked the time series of total ozone
above Arosa (Switzerland) for the four solar activity cycles from 1953 to 1996.
Generally, the ozone maxima seem to have lagged around 1 year after the solar
maximum regarding the time series of solar radio flux at 10.7 cm. Further,
mentioned a lag of about 1 year behind the maxima of
the 10.7 cm solar radio flux time series for ozone profiles measured at
Hohenpeissenberg (Germany). Accordingly, we have performed the regression
analysis by shifting the solar activity cycle proxy 1 year.
Amplitude of the ozone-solar F10.7 flux fitted by multilinear
regression for the time interval from January 1997 to January 2015. The green
line represents the time series of the monthly mean 10.7 cm solar radio flux
as a measure of the solar activity.
In Fig. the normalised time series of the monthly
mean 10.7 cm solar radio flux (green line) is represented along with the amplitude of the
response in ozone to this proxy between January 1997 and January 2015. We can
observe the increase in ozone volume-mixing ratio in the stratosphere during
periods of increased solar activity, and the opposite during the solar
minimum. The amplitude of the 11-year response in ozone is of the order of
5 % or 0.25 ppmv in the lower stratosphere and around 3 % or smaller
than 0.1 ppmv in the stratopause (blue line in Fig. ). The result
in the lower stratosphere is in agreement with the study by
, in which they studied the natural variability of
stratospheric ozone from GROMOS for the time interval between 1994 and 2004
with an iterative spectral analysis. Additionally,
found a peak-to-peak amplitude up to 7 % of ozone variation from DIAL
data at Hohenpeissenberg related to the 11-year solar cycle over the time
period 1979 to early 2003.
Observational studies indicate
a large solar influence in the lower stratosphere as a dynamic result of the
interaction in the upper stratosphere between ozone and UV radiation. With
our data set we obtain a smaller solar influence at the stratopause compared
with the lower stratosphere. This makes sense since the increased absorption
of solar irradiance during the solar maximum increases the temperature but
also increases the production of ozone, which is the main absorber of
radiation in this altitude region. In turn the temperature-dependent ozone
photochemistry (anti-correlation between temperature and ozone) may partly
compensate the solar irradiance influence at the stratopause.
suggested that the impact the solar radiation has on the
upper atmosphere and stratopause region influences the lower stratosphere
through modulation of the internal mode of variation in the polar night jet
and a change in the Brewer–Dobson circulation.
Various mechanisms have been proposed for the 11-year solar cycle influence
on ozone variability. For instance, a change in mean meridional lower
stratospheric dynamics between solar maximum and solar minimum may be the
main factor for the ozone variability due to the solar activity signal. This
change in ozone transport is supported by model and observational results
and references therein. Model simulations
show that zonal asymmetries in ozone, water vapour and
temperature fields are modulated by the 11-year solar cycle, providing a link
between the solar cycle, zonal asymmetries, planetary waves and the
Brewer–Dobson circulation, all of them affecting the stratospheric ozone
distribution .
The surprisingly strong ozone amplitude
observed by GROMOS (5 % in the lower stratosphere) cannot be explained
solely by the solar activity cycle. We hypothesise that the solar variability
is in competition with other factors and together they emphasise the solar
cycle response in ozone at this station. For example, strong ozone anomalies
resulting from other factors may enhance the ozone anomaly related to the
solar activity cycle. Our study only considers a time interval of about
1.5 cycles of the solar activity. Thus, the derived 11-year amplitude could
be influenced by ozone anomalies of other reasons which may accidentally
occur during the solar maximum or the solar minimum phase.
stated that the understanding of observed local changes
may be improved if the change in regional wave patterns and associated mixing
processes induced by the solar cycle are taken into account. Moreover, there
is a lack of long-term observational studies on the response of ozone to the
solar cycle in the midlatitudes.
However the exact mechanism of the dynamical response to solar cycle
variations is not fully understood and cannot be reproduced fully by
chemistry-climate models . On the other hand,
have proposed self-sustained oscillators in the
atmosphere for the first time to support the theory that the oscillations they studied
are excited internally in the atmosphere. They stated that the periods and
phases observed might be interpreted as synchronisation effects that are
typical of non-linear oscillators. The connection with our study relies on
the fact that both analyses are performed in Europe, in the middle atmosphere
and at the same periods of oscillation. A self-excited oscillator might
explain our strong amplitude in the response of ozone related to the
11-year period. Nevertheless, this is a novel interpretation which needs
further analysis and with our experimental data we cannot prove the theory.
El Niño–Southern Oscillation (ENSO)
El Niño–Southern Oscillation (ENSO) is the globally dominating mode of
interannual climate variability and an important driver for modulating the
stratospheric climate in the Northern Hemisphere (NH). ENSO-forced variations
in tropical upwelling lead to temperature and water vapour variations in the
tropical lower stratosphere and impact the chemistry and transport of ozone.
Atmospheric teleconnections lead to ENSO-related effects on the strength of
planetary waves and the Brewer–Dobson circulation, both affecting the
distribution of stratospheric ozone at middle and high latitudes
. Particularly at midlatitudes, the ENSO effects have a
longitudinal variability . Observational studies show that
warm ENSO events are associated with a weak and warm polar vortex over the
Arctic during wintertime . During the warm ENSO winters,
enhanced planetary wave activity propagates from the troposphere to the
stratosphere where it decelerates the zonal mean flow and accelerates the
meridional (poleward) transport .
We used the Multivariate ENSO index (MEI) as a proxy of the El
Niño–Southern Oscillation (ENSO) signal. The MEI index monitors the ENSO
phenomenon with six variables (sea-level pressure, zonal and meridional
components of the surface wind, sea surface temperature, surface air
temperature and total cloudiness fraction of the sky). Due to the delayed
response of the ENSO signal in the northern extratropical atmosphere we have
performed the regression analysis with a time lag of 1 season in the MEI
index. The northern ENSO effects manifest from late winter to early spring,
i.e. 1 season delayed regarding the tropical ENSO, and this is supported by
observational and modelling studies e.g..
Amplitude of the ozone-ENSO fitted during multilinear regression for
the time interval from January 1997 to January 2015. The green line
represents the MEI index.
In Fig. the amplitude of the ozone-ENSO for the time interval
since January 1997 to January 2015 is represented. Even though the amplitude
of the ozone-ENSO is quite small, for most of the altitude range around 1 %
(red line in Fig. ) we can clearly observe the response in ozone at
northern midlatitudes due to ENSO, especially the strong ENSO event in
1997–1998 (Fig. ). The main feature is the opposite phase between
the amplitudes of ozone-ENSO in the lower and middle stratosphere. The green
line in Fig. represents the MEI index, which is positive during
the warm El Niño phase (wENSO) and negative during the cold La Niña
phase (cENSO). The amplitude of ozone-ENSO decreases (increases) during wENSO
(cENSO) in the lower stratosphere while it increases (decreases) in the
middle stratosphere. This feature is also observed over Europe in the study
by by the SOCOL chemistry-climate model. The mechanism
which explains this pattern is a strengthening of the residual mean
circulation caused by anomalously vertically propagating waves and
references therein.