a Long Term

8/6/2019 a Long Term

http://slidepdf.com/reader/full/a-long-term 1/20

ANTARCTICA’S ROLE IN UNDERSTANDING LONG-TERM CHANGE

IN THE UPPER ATMOSPHERE

MARTIN J. JARVIS British Antarctic Survey, Madingley Road, Cambridge, CB3 0ET, U.K.

E-mail: [email protected]

(Received 8 December 2000; accepted 8 August 2001)

Abstract. Within the global context, Antarctica has a key role to play in understanding long-term

change in the upper atmosphere, both because of its isolation from the rest of the world and because

of its unique geophysical attributes. Antarctic upper atmosphere data can provide global change ob-

servations regarding the mesosphere, thermosphere, ionosphere, plasmasphere and magnetosphere.

It will not only provide trend estimates but, just as importantly, it will define the background variab-

ility which exists in the upper atmosphere and against which these trends must be resolved. Upperatmospheric change can be driven both from within the Earth’s near environment primarily through

changing atmospheric composition, dynamics or geomagnetic field, or it can be driven externally,

predominantly by the Sun. Recent observations are discussed in the light of increasing interest in

global change issues and sun-weather relationships.

Keywords: Antarctica, global change, mesosphere, thermosphere, ionosphere, plasmasphere

1. Introduction

The Earth’s upper atmosphere, taken here to be the region upward of 60 km alti-

tude extending outwards more than ten Earth radii to the magnetopause, includesthe mesosphere, thermosphere, ionosphere, plasmasphere and magnetosphere. It

is observed to undergo changes on time scales from less than a second through

to one hundred years and beyond. The latter limit merely reflects the longest

period over which we have been recording any absolute measurements at these

altitudes but time scales for change can be expected to extend through millions of

years. Both short-term variability and long-term change can be driven both from

sources within the Earth’s environmental envelope, delineated by the bounds of

the geomagnetic field, and from external sources primarily within the solar system

and predominantly by the Sun. The upper atmosphere can respond more strongly

and relatively more simply than the troposphere to complex changes occurring in

the troposphere itself and consequently can act as a ‘litmus test’ of underlying

long-term changes, possibly of anthropogenic cause. For instance, a temperature

increase near the Earth’s surface resulting from increased ‘greenhouse gases’ is

expected to be accompanied by a decrease in temperature twenty times greater in

the thermosphere at 250 km. However, in order to extract these trends against a

background of shorter-term variability and non-anthropogenic secular change, it

Surveys in Geophysics 22: 155–174, 2001.

© 2001 Kluwer Academic Publishers. Printed in the Netherlands.



156 MARTIN J. JARVIS

is essential that we first understand and quantify the naturally-varying influences

of the Sun, geomagnetic field, lower atmosphere and all other external influences

upon upper atmospheric processes.

Antarctica has an influential role to play in research into long-term change

processes. It provides important opportunities for testing critical hypotheses re-

lating observations to current theories, both because of its advantageously unique

geophysical attributes and because of its remote location with respect to localised

anthropogenic sources. This paper will discuss ways in which Antarctic research

can help us to understand how and why our planet’s upper atmosphere may be

slowly changing.

2. The Thermosphere

Increased greenhouse gas concentrations in the Earth’s atmosphere lead to in-

creased temperatures in the troposphere. The increase that will be invoked by

doubling concentrations of ‘greenhouse gases’ (primarily carbon dioxide and meth-

ane) compared to pre-industrialised times has been predicted to be of the order of

2 K (Kattenberg et al., 1996) and is forecast to occur by the year 2100. However,

numerical models show that higher up in the atmosphere this doubling of green-

house gases should result in a decrease in temperature several times greater. Roble

and Dickinson (1989), for instance, demonstrated that the thermosphere should

cool by approximately 50 K and the mesosphere by approximately 10 K. A dra-

matic illustration of such a cooling effect is evident on the upper atmosphere of

Venus which has a very high concentration (96%) of carbon dioxide in its atmo-

sphere and thus is an extreme case of a planet experiencing the greenhouse effect.

This raises the temperature of the Venusian lower atmosphere to 700 K – over twicethat on Earth. At the same time, however, its effect is to cool the Venusian daytime

thermosphere to just 300 K (Keating and Bougher, 1992) typically less than one

third of the corresponding temperature on Earth. This cooling is due primarily to

very strong radiative infrared cooling associated with CO2 at 15 µm wavelength

and the consequent enhancement in emissivity of the planet’s upper atmosphere.

One current goal within upper atmospheric research is to confirm this predicted

temperature decrease by detecting its signature in geophysical data series. The re-

liability of temperature trend estimates is dependent on our ability to retrieve those

trends against naturally occurring variability in the parameter that we are trying

to measure. The magnitude of both the predicted trend and the natural variability

vary with altitude. The predicted values for a scenario where greenhouse gases

are doubled are shown in Figure 1. Near the ground we might expect a rise in

temperature of just 2 K and this will be seen against a diurnal variability of the

order of 20 K.

Lübken and Von Zahn (1991) grouped winter and summer polar mesospheric

temperature profiles measured in situ in the northern hemisphere by tracking



ANTARCTICA’S ROLE IN UNDERSTANDING CHANGE IN THE UPPER ATMOSPHERE 157

Figure 1. The expected change in temperature for a doubled greenhouse gas scenario compared todiurnal variability at different levels in the atmosphere.

the rate of descent of inflated spheres ejected from meteorological rockets. They

showed that the temperature variation in the high latitude mesosphere throughout

the summer covers a range of <20 K at 82 km; during winter the temperature is

more variable and this figure rises to approximately 60 K. Thus the polar summer

mesosphere provides a relatively stable background against which to detect secu-

lar change. Against this background the mesosphere should experience cooling of

approximately 10 K (Roble and Dickinson, 1989). Higher up, in the thermosphere,

the doubled greenhouse gas scenario leads to a predicted decrease in temperat-ure of approximately 50 K (Roble and Dickinson, 1989) against a background

temperature that can vary by >400 K from day to day.

The upper atmosphere is consequently potentially the most sensitive region of

the atmosphere in which to detect a ‘signal’ of a greenhouse gas-induced long-

term trend seen against the ‘noise’ of the diurnal variability. In the summer polar

mesosphere this signal-to-noise ratio is five times larger than near the Earth’s

surface. In the thermosphere it is similar to that near the surface. However, the

thermosphere has dual advantages over the lower atmosphere: first, the trend is

an order of magnitude greater than that near the surface and second, much of the

diurnal thermospheric temperature variability can be removed because of its known

dependence upon the strength of solar radiation.

There are no direct long-term measurement series of thermospheric temper-

ature. However, a number of workers have used proxies to indirectly investigate

temperature trends. Rishbeth (1990) demonstrated that the temperature decreases

predicted by Roble and Dickinson (1989) for a doubled greenhouse gas scenario

would result in a lowering in the height of the ionospheric F2-layer peak height




(hmF2) by around 20 km. This is due primarily to the change in altitude of a

fixed pressure level as a direct consequence of cooling. Bremer (1992) utilised

this thermospheric temperature proxy by deriving hmF2 from ionograms recorded

at Juliusruh (54 ◦N, 13 ◦E) over 38 years. He found a downward trend in hmF2

averaged over the whole year of 0.24 km yr−1. Ulich and Turunen (1997) made a

similar analysis of hmF2 over 39 years at Sodankyla (67 ◦N, 27 ◦E) with similar

results.

These European observations were brought into a global context by an ana-

lysis of 38 years of ionospheric sounding data from Argentine Islands (65 ◦S,

64 ◦W) on the Antarctic Peninsula and Port Stanley (52 ◦S, 58 ◦W) in the Falk-

land Islands (Jarvis et al., 1998). Not only were these measurements from the

opposite hemisphere to previous results, but the remoteness of Antarctica away

from localised anthropogenic inputs means that its atmosphere tends to represent

a global average. Furthermore, the fact that in the southern hemisphere the offset

between the geomagnetic and geographic poles is significantly greater than that in

the northern hemisphere, coupled with the location of Argentine Islands and PortStanley at the opposite geographic longitude to that of the geomagnetic pole, means

that these two ionosonde stations have the highest possible geographic latitudes

for their geomagnetic latitude (see Figure 2). Thus they are able to sample the

high latitude atmosphere with minimal introduction of variability into the data

through geomagnetic activity and are therefore extremely valuable sites for such

measurements. Jarvis et al. (1998) showed that hmF2 at these two sites was also

decreasing. Taking into account any residual effects of geomagnetic activity or

changing thermospheric winds they determined the trend for each month of the

year. The results were in general agreement with those of Bremer (1992) and Ulich

and Turunen (1997) implying that this phenomenon was indeed a global effect.

There have since been a number of similar analyses which have begun to con-fuse the picture; at some stations downward trends are detected and at some sites

upward trends are detected. One reason for confusion is the different methodology

used by different authors. During the analysis it is important to minimise the effects

of changing solar radiation, geomagnetic activity and any diurnal variation in trend

induced by secular variations in thermospheric wind. It is also important to care-

fully consider the detrimental consequences both of poorly calibrated or disrupted

data series and of inadequate time series duration [Bremer (1992) demonstrated

that a dataset of at very least two solar cycles in duration is necessary to get reliable

results for this work]. Those published results that fully meet these criteria (e.g.,

Bremer (1992), Bremer (1998), Jarvis et al. (1998)) indicate that there is generally

a negative trend except over part of Eastern Europe (east of 30 ◦E) where there is

a positive trend. Figure 3 gives a summary of these results; the European resultsshown are the median values of all individual trends in the region west or east of

30 ◦E (from Bremer (1998) Figure 2).

The minimisation of solar cycle effects in hmF2 trend analyses has generally

been achieved through the generation of an empirical relationship between sunspot




Figure 2. Antarctica showing the offset of the geomagnetic invariant and geographic poles and the

locations of research stations mentioned in this paper. Geographic (solid) and geomagnetic (dashed)

latitude lines are drawn at 10◦ intervals. The invariant geomagnetic pole is denoted by M.P.

number or 10.7 cm flux and the hmF2 data itself. This has then been used to remove

as much of the solar cycle influence from the data analysis as possible. However,

this is a relatively simplistic approach and does not account for the possibility

that the thermosphere has long-term conditioning. Such conditioning was reported

by Field and Rishbeth (1997) who found F2-layer data for different solar cycles

showed a markedly different response to the same level of geomagnetic activity.

Thus Field and Rishbeth (1997) suggested that the response to a given level of

geomagnetic activity depends not just on the current level of solar flux but on

its past ‘history’. This emphasises the need for data series covering several solar

cycles.

Mikhailov and Marin (2000) interpret trends in the peak electron concentration

of the F-region (foF2) in terms of the long-term variation in geomagnetic activity

and consequent changes in the occurrence characteristics of geomagnetic storms.

Danilov and Mikhailov (2001) extend this concept to consider whether hmF2




Figure 3. The derived trend in F-region altitude from ionosonde and satellite observations. Argentine

Is. and Port Stanley ionosonde data (by month) from Jarvis et al. (1998); European ionosonde data

(mean of all months) from Bremer (1998); satellite drag data from Keating et al. (2000).

trends may be linked to long-term geomagnetic activity changes. In a reanalysesof the Argentine Islands and Port Stanley data presented by Jarvis et al. (1998)

they conclude that the local time variation of the correlation between geomag-

netic activity and hmF2 trend principally agrees with the current understanding of

ionospheric storms. Conversely, in spite of finding a positive correlation between

hmF2 trend and geomagnetic activity, they reproduce the negative hmF2 trend

found by Jarvis et al. (1998) even though it occurred during a long-term increase in

geomagnetic activity (Clilverd et al., 1998) (see Section 5). There is clearly more

research required before the implications of the observed hmF2 trends can be fully

understood.

The ionospheric results from Antarctica and western Europe have recently been

corroborated by a completely independent method. Keating et al. (2000) have con-

firmed that the thermosphere is cooling by measuring the atmospheric drag on five

satellites to show that the density of the thermosphere at 350 km has decreased by

approximately 10% over two solar cycles. This is equivalent to a decrease in hmF2

of 0.25 km yr−1 (also shown in Figure 3). Comparison with the calculations of

Rishbeth (1990) indicates that this is equivalent to a 0.6 K yr−1 drop in temperature.




3. The Mesosphere

Lower down, in the mesosphere, the consensus of observations also show long-

term cooling. At around 60 kin altitude cooling rates between 0.25 K yr−1

and0.4 K yr−1 are found. Measurement techniques at this altitude include lidar (Aikin

et al., 1991; Hauchecorne et al., 1991) and sounding rockets (Keckhut et al., 1999;

Golitsyn et al., 1996). Golitsyn et al. (1996) present the results from over 7000

rocket launches between 1964 and 1995 at 5 sites covering a latitudinal range from

Molodeshneya (68 ◦S, 46 ◦E), Antarctica, to Heiss Island (81◦N, 58 ◦E). Temper-

ature measurements were made between approximately 22 km and 75 km altitude

using a resistance thermometer and showed cooling of the order of 0.6 K yr−1

at heights between 60 and 70 km. Using an entirely different technique, Tauben-

heim et al. (1997) monitored the reflection height of 164 kHz radiowaves in the

lower D region at 50 ◦N over 30 years and demonstrated that the column-mean

temperature between the stratopause and 82 km altitude has systematically fallen

at a linear rate of 0.6 K yr−1. At about 82 km altitude one of the most striking

indications of long-term change in the upper atmosphere is visible from the ground

unaided. Noctilucent clouds (e.g., Fogle and Haurwitz, 1966), which are formed of

ice particles which appear in the extremely low temperatures near the polar summer

mesopause, have doubled in occurrence rate over northern Europe in thirty years

(Gadsden, 1990). This may be an indication of steadily declining temperatures at

that altitude, hence providing sufficiently low temperatures for ice particle form-

ation more frequently than in the past. If this is so, then it is estimated that the

increase in occurrence rate reflects a temperature drop of 0.3 K yr−1 (Gadsden,

1990). An alternative possibility is that methane increases in the lower atmosphere

have led to increased water vapour presence at the mesopause promulgating ice

particle formation (Thomas, 1996). However, Lübken et al. (1996) do not findevidence of any long term change in temperature below the mesopause from a

survey of over 30 years of measurements at high latitude (66 ◦N – 71 ◦N). In the

early years (1960s) these measurements were carried out using rocket grenades; in

the later years (1990s) the mesospheric temperature profiles were derived from the

rate of descent of rocket-launched falling spheres. A more detailed presentation of

this result is given by Lübken (2000).

In order to investigate the implications of this result, which does not sit eas-

ily with observations above and below this altitude, we need to fully understand

how mesospheric temperatures are controlled or what the characteristics of the

background variability are. One way to try to resolve this problem is to compare

mesospheric temperatures and drivers in the Arctic with those in the Antarctic.

Until recently mesopause temperature in the Antarctic had been believed to be

several degrees warmer than that in the Arctic, primarily because the upward-

propagating gravity wave activity, which generates the cold summer mesopause

by forcing adiabatic cooling, has been argued to be weaker in the Antarctic than in

the Arctic. There are three ways to test this.




The first is to make temperature measurements in the Antarctic mesosphere for

direct comparison with those in the Arctic. While Arctic mesospheric temperatures

have been well documented for different seasons and for over 30 years, the first

in situ measurements of the Antarctic mesospheric temperature profile were not

carried out until very recently (Lübken et al., 1999a). A series of rocket-released

falling spheres were launched and tracked from Rothera (68 ◦S, 68 ◦W) research

station on the Antarctic Peninsula during the austral summer. Rothera is ideally loc-

ated for such a north-south comparison because it is almost exactly co-latitudinal to

Andoya (69 ◦N, 16 ◦E), Norway, from where the majority of northern hemisphere

in situ measurements have been made. The logistical constraints of a rocket cam-

paign from the Antarctic meant that the first launch of the two-month campaign

took place three weeks after summer solstice. The first temperature measurement

on 5 January 1998 showed the mesopause temperature to be as cold as 129 K at 87

km surprisingly close to the northern hemisphere July mean value of 130 K (July

being the equivalent northern month with respect to the solstice). Twenty-four tem-

perature profiles where taken from Rothera approximately two days apart and thetemperature through January at 82 km was found to be close to that in the Arctic,

but by February it was a few degrees warmer than its Arctic equivalent. This close

symmetry of the Antarctic and Arctic upper atmosphere thermal structure during

January/July implies that the physical processes dominating the energy budget are

similar. This is contrary to expectation given the clear interhemispheric difference

in surface topography and land/ocean distribution which might be expected to

cause different gravity wave activity and hence different dynamical forcing in the

two polar mesospheres.

These first in situ mesospheric temperature profiles over Antarctica provide an

important benchmark for future long-term change studies. Remotely sensed tem-

perature comparisons between the Antarctic and Arctic mesospheres have recentlybeen published by Huaman and Balsley (1999). Limb-scanned satellite observa-

tions from the High Resolution Doppler Imager (HRDI) were averaged across

seven years and sampled between 64◦ and 68◦ latitude over an altitude range of

84–87 km. These data indicate that the mesosphere in the Antarctic is some 15 K

warmer than that in the Arctic for a period extending from about 10 days before to

about 15 days after the summer solstice. This difference then reduces rapidly and

levels off to 3 ± 2 K by three weeks after solstice – just after the Antarctic rocket

campaign began. Temperatures at 82 km altitude from the falling spheres over

Rothera show close agreement with those in the Arctic for January/July but indicate

that the Antarctic mesosphere became 4 ± 6 K warmer by mid-February. Figure

4 shows a superposition of data taken from Figure 4, of Lübken et al. (1999a)

and Figure 1 of Huaman and Balsley (1999). The Antarctic in situ and remotely-sensed measurements at 82 km tend towards disagreement in the data closest to

solstice (i.e., 15–20 days after solstice) and further in situ Antarctic observations,

this time taken within a day or two of the solstice, are plainly needed in order to

provide a more definitive result. It will also be necessary to consider whether the




Figure 4. Comparison of the Antarctic-Arctic temperature difference in the summer mesosphere at

82 km altitude from in situ (Lübken et al., 1999a) and limb-scanned (Huaman and Balsley, 1999)

observations. Individual in situ measurements are noted by filled squares and the best-fit through

these by a dashed line. Limb-scanned observations are denoted by a solid line with error bars.

interhemispheric temperature comparison has the same characteristics at 82 km as

it does at the mesopause.

The second way to test whether forced adiabatic cooling is weaker in the Ant-

arctic than in the Arctic, is to compare the occurrence of temperature-dependent

phenomena. As noted earlier, noctilucent cloud formation is highly temperaturedependent, providing a benchmark of when mesopause temperatures drop below

a certain value. Unfortunately, however, there have been few noctilucent cloud

observations in Antarctica (cf. Thomas, 1996), not because of their absence, but

because land in the ideal latitude range for observations (55–65◦) is very limited

and sparsely inhabited and tropospheric cloud hampers observation because of the

presence of the Antarctic Convergence. That said, there have surprisingly been

‘out-of-season noctilucent cloud observations in the Antarctic implying, again,

that we do not properly understand the energy budget of the mesosphere – a

necessary prerequisite for interpreting trend estimates. For instance, Griffiths and

Shanklin (1987) and Shanklin (1988) documented two occurrences of noctilucent

cloud at Faraday research station (65 ◦S) on the Antarctic peninsula in June 1985.

In other words, this summer phenomenon was observed near the winter solstice.

Similarly, Warren et al. (1997) observed noctilucent clouds in April (four months

after summer solstice) at South Pole Station (90 ◦S) and also provided supporting

evidence that unusually cold temperatures can occur in the Antarctic mesosphere

at unexpected times of year. They presented examples of OH temperature meas-




urements from 88 km altitude over South Pole which show transitory reductions

over a period of a few hours to values as low as 130 K at the end of May and

August – five and four months respectively from summer solstice. While noctilu-

cent clouds are rarely observable in Antarctica, two other manifestations of similar

phenomena, Polar Mesospheric Summer Echoes (PMSE) and Polar Mesospheric

Clouds (PMC), make for easier measurement. PMC are optically thin layers of

cloud typically observed during summer at polar latitudes >70◦ by satellite via

limb-scanning at ultra-violet wavelengths. Olivero and Thomas (1986) showed

that, statistically, northern PMC are inherently brighter than southern PMC pre-

sumably because of larger particles or higher particle concentrations. This leads to

the presumption that if their formation is primarily controlled by temperature, the

southern polar mesosphere is warmer than that in the north. PMSE, which are radar

echoes from thin layers at similar height and with similar characteristics as noctilu-

cent clouds, provide similar indications. Few measurements have yet been made

in the Antarctic, but those that have (Woodman et al., 1999) suggest that PMSE

in the southern hemisphere are weaker than their northern counterparts. Howeverdue to the sparcity of the occurrence statistics, further southern hemisphere PMSE

observations are necessary to confirm this.

The third method of investigating whether upward-propagating gravity wave

activity is weaker in the Antarctic than in the Arctic is to directly observe how

the gravity wave field in the Antarctic mesosphere compares with that used in

numerical computer models of the atmosphere. To this end airglow imagers have

recently been installed at South Pole (90 ◦S) (Ejiri et al., 1999) and Halley (76 ◦S,

27 ◦W) (Clilverd et al., 2000b). Yoshiki and Sato (2000) have recently published

a statistical analysis of gravity waves observed in the polar stratosphere over a

period of 10 years from radiosonde measurements. They use data from 33 sites,

15 of which are poleward of 60

◦

S, but particularly concentrate on data from theAntarctic stations of Syowa (69 ◦S, 40 ◦E) and Casey (66 ◦S, 110 ◦E). They suggest

that gravity waves observed in the Arctic are forced by topography whereas in the

Antarctic some sources may exist in the stratosphere. They also find evidence that,

while most gravity waves transfer energy upward in the Arctic, there is a relatively

high percentage of downward energy propagation in the Antarctic in winter and

spring.

Of the three methods described above, none are without difficulty. In situ

temperature measurements by rocket and remote measurement of temperature by

limb-scanning satellites detect the effect (temperature change) rather than the cause

(gravity waves). Statistical data on the occurrence rate and strength of PMSE and

NLC from the ground or PMC from satellites provide only an indirect proxy to

the physical processes. Observation of the gravity wave field suffers from its de-pendence on darkness and clear skies and thus south of the Antarctic Circle is not

possible for the scientifically critical summer solstice period. It is only through a

combination of these techniques and additional numerical modelling work that a

better understanding of the physics is likely to be forthcoming.




4. The Plasmasphere

So far we have discussed only signatures of change to the neutral atmosphere, but

Antarctica also offers the opportunity to assess what change might be occurringeven higher in altitude where the atmosphere is completely ionised. Clilverd et al.

(2000a) monitored VLF radio waves received at Faraday Research Station, Argen-

tine Islands, on the Antarctic Peninsula and at Dunedin, New Zealand, which were

transmitted from the opposite hemisphere by two U.S. Navy transmitters in the

U.S.A. Whistler-mode electromagnetic signals propagated along the geomagnetic

field lines from the northern to the southern hemisphere, their group delay being

determined by the field line length and the plasma density along it. The group delay

time was measured by direct comparison with the travel time of the subionospheric

signal and provided an estimation of tubular plasmaspheric electron concentration

along the field line. Exactly which field line the signals travelled along (i.e., the

L-shell of the magnetospheric duct) was determined from the difference in group

delay time for the two different transmitting frequencies. Using this method, Clil-verd et al. (2000a) studied the depletion of the plasmasphere in the aftermath of

magnetic storms and found that depletions during the solar minimum of 1995 were

significantly deeper than during the minimum of 1986, independent of the time

of year. The depletion factor was approximately 2 in 1986 but was 3–4 in 1995.

In addition these depletions seen at both Faraday and Dunedin were even deeper

than those observed in 1958 and 1961 Carpenter (1962) using naturally occurring

whistlers.

These results suggest some long-term change in the plasmasphere taking place

over a timescale of at least 38 years, but what this change might be is difficult to

determine until the exact cause of the plasmaspheric depletions is understood. Ana-

lytical and numerical model comparisons with the depletion data were carried outby Clilverd et al. (2000a) but these did not clearly identify any mechanism. These

comparisons did demonstrate that it was likely that the depletions were neither

solely the result of modifications to the chemical balance or electron concentration

in the thermosphere, nor the result of outward E × B drift of plasma.

5. The Effect of the Geomagnetic Field

In addition to long-term change triggered by changes in the atmosphere itself, we

must consider what long-term changes are driven by geomagnetic and extrater-

restrial processes. For instance, there may be some as yet unexplained link between

the greater storm-time plasmaspheric depletions and an increase in geomagnetic

activity. Clilverd et al. (1998) demonstrated that all phases of the solar cycle have

shown an enhancement in geomagnetic activity, indicated by changes to the aa

index, since solar cycle 14 (i.e., 1910). They showed that changes due to drifting

magnetic latitude of the sites used to derive the aa index would be barely signi-




ficant and that the prime cause of the aa index increase was an increase in solar

activity. That such an increase in solar activity has occurred has been indicated by

Lockwood et al. (1999).

For both short-term atmospheric variability and long-term atmospheric change,

the Antarctic provides the best location for separating the signatures of geomagnet-

ically controlled processes from those dependent on the Earth’s orbital geometry

(e.g., solar radiation and meteoric bombardment). The reason for this is the large

offset between the geographic south pole and the geomagnetic south pole; the

geographic and geomagnetic invariant poles are separated by 17◦ of latitude in

the Antarctic (see Figure 2), compared to only 11◦ in the Arctic. This eccentricity

of the southern geomagnetic lines of latitude enables comparison of records from

geophysical observatories with a similar geomagnetic latitude but with a variety of

geographic latitudes (or vice versa) and hence the isolation of phenomena driven

by changes in the internally-generated geomagnetic field or by externally-driven

changes to the balance of the magnetosphere from those driven by solar zenith

angle or geographic symmetry.It is in the mesosphere and lower thermosphere which together form the trans-

ition region between the fully-ionised collisionless plasma of geospace and the

virtually unionised collision-dominated air of the troposphere and stratosphere,

that this geomagnetic-geographic offset becomes most dominant in its role on the

physics and chemistry of the atmosphere. An example of this is demonstrated by

Crowley et al. (1999) who highlight the implications of presenting nitric oxide ob-

servations in the geographic coordinate system, which prevails as standard practice

amongst chemists. In geographic coordinates, limb-scanned satellite measurements

of nitric oxide at around 105 km altitude at equinox show a 25% hemispheric

asymmetry: smaller values of NO concentration are found in the southern hemi-

sphere than in the northern hemisphere. However, when viewed in geomagneticcoordinates the asymmetry disappears belying the importance of charged particle

precipitation, which is magnetospherically configured, on NO production. As

Crowley et al. (1999) note, this result is a clear demonstration of the importance

of the geographic-geomagnetic polar offset for understanding magnetospheric and

solar influences on various minor species in the mesosphere and lower thermo-

sphere. It is readily apparent that minor species can have a dramatic long-term

effect on the upper atmosphere and this means of separating causal mechanisms is

more effective in the Antarctic than in the Arctic because of the greater offset of

the poles there.

Secular changes in solar magnetic activity produce statistical changes in the

position of the auroral oval (Feynman and Ruzmaikin, 1999) and hence changes in

charged particle precipitation into the upper atmosphere which in turn can changenitric oxide concentration. The changes over 90 years in the location of the equat-

orward edge of the auroral oval for specified ‘quiet’ and ‘disturbed’ geomagnetic

activity levels which were presented by Feynman and Ruzmaikin (1999) (their Fig-

ure 4) for the northern hemisphere. In this paper those changes have been mapped




Figure 5. The equatorward movement (indicated by shading) over 90 years starting in 1904 of the

position of the equatorward edge of the auroral oval under both quiet and disturbed geomagnetic

conditions. This has been mapped here as the geomagnetic conjugate to the northern hemisphere

result of Feynman and Ruzmaikin (1999). Also marked are the positions of the research stations

named in Figure 2.

to their magnetic conjugate positions in the southern hemisphere and these are

shown in Figure 5.

6. Heliomagnetic Influence

Lockwood et al. (1999) inferred from near-Earth interplanetary magnetic field

measurements that the total magnetic flux leaving the Sun has increased by a factor

of 1.4 between 1964 and 1995. They also showed that the aa index appears to

make a good proxy for these magnetic field measurements and hence, by inference

from earlier aa records, they suggest that the magnetic flux leaving the Sun has

increased by a factor of 2.3 since 1901. Longer-term variations than these can only

be modelled using some other permanently preserved proxy to extrapolate back

in time. Solar radiance and solar magnetic activity both directly and indirectly

affect the balance of the upper atmosphere. Solar radiance directly heats the upper

atmosphere but also changes the climatology in the lower atmosphere which then




feeds back to the upper atmosphere through changes in upward-propagating wave

fields and hence momentum forcing of the mesosphere and lower thermosphere.

There is evidence that solar irradiance and solar magnetic activity are posit-

ively correlated (Willson et al., 1986), and consequently Beer et al. (1990) have

suggested that 10Be can be used as the permanently preserved proxy needed to

study variations in solar activity pre-dating accurate solar records. The magnetic

field frozen into the solar wind is dependent upon solar activity. It deflects cos-

mic rays, particularly those with the lowest energies, from penetrating the Earth’s

atmosphere. The clearest demonstration of this effect is the Forbush decrease ob-

served before geomagnetic storms (Forbush, 1937). This reduction in the cosmic

ray flux leads to a reduction in the production of cosmogenic radionuclides in

the atmosphere. 10Be has a short lifetime in the atmosphere (compared to 14C,

for example) and is thus best suited to monitoring changes in solar activity on

timescales as short as the 11-year solar cycle. It is also relatively unaffected by

anthropogenic influences (Beer et al., 1990). Beer et al. (1990) compared 10Be in

polar ice-core records with the aa index over a period>100 years and demonstratedthat the 10Be provides a possible proxy for heliomagnetic activity. The greatest

uncertainty in interpretation were short-term fluctuations presumed due to changes

in precipitation rate, atmospheric mixing and scavenging efficiency but Beer et al.

(1990) suggested that a possible way to reduce these uncertainties is to combine

records from different sites such as the Antarctic and Arctic with different climate

conditions but similar production rates.

Dreschhoff and Zeller (1998) have demonstrated that Antarctic ice cores can

also be used to provide a historical record of major solar proton events. Shea et al.

(1999) identified large transient concentrations of nitrates in the cores as signatures

of these events. Solar proton events can have a significant impact on the upper

atmosphere and any long-term changes in their occurrence characteristics couldchange both the ambient ionised and neutral characteristics. For instance, Jackman

et al. (2000) have shown that SPEs can produce HOx and NOy constituents in the

mesosphere and stratosphere and that these NOy constituents in the stratosphere

have the capability of affecting ozone for years past the events.

Cosmic rays have been suggested as a missing link between solar output and

terrestrial weather, with maximum effect at subauroral latitudes. Feynman and

Ruzmaikin (1999) have noted that there are two ways that the long-term increase

in solar wind noted by Lockwood et al. (1999) might affect tropospheric cloud

cover and hence climate change. First, it is well documented that the solar wind

suppresses the cosmic ray flux reaching the earth (e.g., the Forbush decrease) and

thus an increased solar wind will lead to a decrease in cosmic ray flux. A long-

term decrease in cosmic ray flux has been observed by Stozhkov et al. (2000)and argued by Ahuwalia (2000) to be of solar influence. Tinsley and Deen (1991)

correlated cyclone intensity in the northern hemisphere with cosmic ray flux on a

decadal timescale. On shorter time scales, Veretenenko and Pudovkin (1994) obser-

vationally correlated a 10% change of cloud cover at 50–60◦ geomagnetic latitude




with Forbush decreases in the flux of galactic cosmic rays with energies between

several hundred MeV and 1 GeV. At lower latitudes no effect was apparent; this

is consistent with a cosmic ray causal mechanism because at low geomagnetic

latitudes the geomagnetic field shields the Earth from cosmic rays so that particles

with energies <1 GeV cannot penetrate to the troposphere (Smart and Shea, 1985).

In the polar cap the effect was more complex and less clearly determinable; here

the field lines are open to the solar wind and particles of all rigidities and incident

from all directions can find their way to the troposphere. Changes in cosmic ray

flux may then affect surface temperature because of its sensitivity to cloud cover.

Svensmark and FriisChristensen (1997) suggested that cosmic rays form a miss-

ing link in solar-climate relationships and they relate changes in cosmic ray flux

to long-term trends in surface temperature, although there is considerable debate

about this suggestion (e.g., Jorgansen and Hansen (2000)). Second, the increase

in solar wind will trigger more magnetospheric storms and substorms which will

move the subauroral region further equatorward (see Figure 5) and hence move

the latitude of the cloud-effective cosmic ray activity further equatorward also. Ina simple comparison, Cliver et al. (1998), though ignoring the influence of other

known drivers, found a strong correlation between minimum aa geomagnetic index

values and the Earth’s surface temperature.

7. Vertical Dynamic Coupling

Clearly there are poorly understood complex inter-relationships here – ones which

could be crucial for understanding the natural variability of our atmospheric envir-

onment and how past and future climate changes are influenced by extraterrestrialprocesses. To study these relationships further there is a need both for long-term

data series of a global nature and for collocated measurements from a host of

instrumentation looking at all aspects of the atmospheric and solar-terrestrial envir-

onment on shorter timescales. Halley Research Station (76 ◦S, 27 ◦W), Antarctica,

provides a unique opportunity in this respect. First, it is located at a subauroral

latitude and yet because of the large offset of the geographic and geomagnetic poles

in the southern hemisphere, it lies at a relatively high geographic latitude and thus

experiences 4 months of continuous darkness in winter. Goldberg (1979) noted that

many Sun-weather relationships show their best correlation at high latitudes during

local winter, when direct solar insolation is least competitive, thereby qualifying

the high (geographic) latitude atmosphere as a primary region for causal mechan-

ism research. This winter dependence is demonstrated in the Wilcox effect (Wilcox

et al., 1973) relating solar sector boundary crossings to the atmospheric vorticity

index – a then controversial finding which Hines and Halevy (1977) examined

critically and extensively and were obliged to concede to the physical reality of the

correlation. Second, Halley has an impressive array of collocated instrumentation




routinely measuring a multitude of atmospheric parameters from the troposphere

right through to the magnetosphere.

The lower and upper atmospheres cannot be considered in isolation. Climate

change in the troposphere will produce changes in the upper atmosphere. The

converse is also true. For instance, Arnold and Robinson (1998) computationally

explored the ways in which the propagation of planetary-scale waves between

the troposphere and thermosphere responds to solar-cycle-induced changes. They

showed that the presence of planetary waves coupled the upper and lower atmo-

spheres through a feedback mechanism where planetary waves actively modify

the atmospheric winds which in turn affect the propagation of the waves. Con-

sequently the wave field between the ground and the lower thermosphere (e.g.,

105 km) is sensitive to changes in that upper boundary which, in turn, is sens-

itive to solar cycle input. Planetary waves are important in the bulk transport of

chemical constituents in the middle atmosphere and there will consequently be

other, additional feedbacks. The Antarctic is an ideal location for studying such

effects because of the pole-centred topology of the continent and the extremesof sunlight encountered even as high as the lower thermosphere. A circumpolar

network of mesospheric wind radars, situated at a growing number of sites around

the Antarctic (e.g., Rothera, Davis, McMurdo, Halley), will enable planetary wave

modal and propagation characteristics to be determined and compared to prevailing

lower thermospheric conditions. To this end the growing network of SuperDARN

radars (Greenwald et al., 1995) may also prove highly beneficial due to their ability

to monitor planetary waves in meteor winds (Jenkins et al., 1998).

8. Summary

In all the research areas discussed in this paper there is a need for a multiplicity

of observations and a growing requirement to understand the atmospheric pro-

cesses in the context of the whole ‘Earth system’. This will require observations

at all levels of the atmosphere using a combination of both ground-based and

satellite observing techniques. Ground-based observations can provide temporal

continuity at a fixed geographic location, often going back to the pre-satellite

era, and the ability to observe small scale phenomena. Satellites have the obvi-

ous advantage that they can provide much greater geographic coverage but at the

expense of often having poorer spatial resolution and usually sampling in a fixed or

slowly drifting local time sector. Many examples of the complementary nature of

ground-based and satellite data have been shown here. For example, satellite drag

observations and ionosonde observations are able to corroborate thermospheric

temperature changes through completely independent techniques; Antarctic-Arctic

differences in satellite observations of PMC complement Antarctic-Arctic differ-

ences in ground–based observations of PMSE strength; rocket-launched falling

spheres and satellite-based limb-scanning provide complementary mesospheric




temperature data; historical ground-based geomagnetic data and recent satellite

data can be used to study different aspects of the effect of geomagnetic storms on

the upper atmosphere; satellite measurements of the interplanetary magnetic field

provide a benchmark for the use of the aa index as a proxy for the Sun’s magnetic

flux going back 100 years. In all these cases the scientific value of having both

ground-based and satellite data far outweighs the scientific value of the sum of the

parts.

Long–term change tends by its very nature to be a global phenomena. Therefore

no one area of the world can be isolated for carrying out such research. However,

studies in the Antarctic are essential in order to prove that any upper atmospheric

changes observed elsewhere in the world are, indeed, global in nature and not

the result of some localised phenomenon, either natural or anthropogenic. In this

role Antarctica stands out from all other continents. It is the most isolated from

human influence having low levels of direct pollution from chemicals (including

greenhouse gases), particulates, electromagnetic radiation and direct heating. Not

only does this provide a ‘clean’ environment for carrying out trend measurements,but any lasting residual influence of these pollutants detected in Antarctica will

tend to represent a global mean. Even more importantly Antarctica has a unique

role to play in several specific areas of long-term change study regarding the upper

atmosphere. Ionospheric soundings from Antarctica will help unravel the problem

of why different regions of the world observe different trends in the height of F-

region peak electron density. It is particularly important in this work because high

geographic latitude stations there have relatively low geomagnetic latitudes. Com-

parison between the Antarctic and Arctic will enable us to understand the balance

of mechanisms behind mesospheric energetics and hence the reason behind trends

in mesospheric temperature and noctilucent cloud occurrence. Research stations on

the Antarctic peninsula have a role to play in investigating trends in plasmasphericstorm response because of their geomagnetic position in the Southern Hemisphere

with respect to northern hemisphere VLF transmitters. The Antarctic also offers

the potential of permanently preserved records of solar activity and solar proton

events in ice-cores, a subauroral location with months of continuous darkness

for investigating sun-weather relationships and their long-term consequences, and

a pole-centred land-base for studying feedbacks between the upper and lower

atmospheres through planetary wave activity and the subsequent signatures of

tropospheric trends in the upper atmosphere. The issues are complex and much

research is needed. This will be partly based on the excellent long-term data series

that have been taken in Antarctica by scientists of many nations since the Inter-

national Geophysical Year in 1957 and partly based upon new observations from

the expanding network of upper atmospheric instrumentation around the Antarcticcombined with that from upcoming satellite missions such as TIMED and Odin.




References

Ahuwalia, H.S.: (2000), On galactic cosmic ray flux decrease near solar minima and IMF intensity,

Geophys. Res. Lett. 27, 1603–1606.Aikin A.C., Chanin, M.-L., Nash, J. and Kendig, D.J.: (1991), Temperature trends in the lower

mesosphere, Geophys. Res. Lett. 18, 416–419.

Arnold, N.F. and Robinson, T.R.: (1998), Solar cycle changes to planetary wave propagation and

their influence on the middle atmosphere circulation, Ann. Geophysicae 16, 69–76.

Beer J., Blinov, A., Bonani, G., Finkel, R.C., Hofmann, H.J., Lehmann, B., Oeschger, H., Sigg, A.,

Schwander, J., Staffelbach, T., Stauffer, B., Suter, M. and Wolfli, W.: (1990), Use of 10Be in polar

ice to trace the 11-year cycle of solar activity, Nature 347, 164–347.

Bremer, J.: (1992), Ionospheric trends in mid-latitudes as a possible indicator of the atmospheric

greenhouse effect, J. Atmos. Terr. Phys. 54, 1505–1511.

Bremer, J.: (1998), Trends in the ionospheric E and F regions over Europe. Ann. Geophysicae 16,

986–996.

Carpenter, D.L.: (1962), New experimental evidence of the effect of magnetic storms on the

magnetosphere, J. Geophys. Res. 76, 135–145.

Clilverd M.A., Clark, T.D.G., Clarke, E., Rishbeth, H.: (1998), Increased magnetic storm activity

from 1868 to 1995, J. Atmos. Sol.-Terr. Phys. 60, 1047–1056.

Clilverd M.A., Jenkins, B. and Thomson, N.B.: (2000a), Plasmaspheric storm time erosion, J.

Geophys. Res. 105, 2997–13008.

Clilverd M.A., Taylor, M.J., Jarvis, M.J., Espy, P. and Rose, M.C.: (2000b), Studying mesospheric

dynamics in the Antarctic using the upgraded instrument cluster at Halley. In First S-RAMP

Conference, Sapporo, Japan, p328.

Cliver, E.W., Boriakoff, V. and Feynman, J.: (1998), Solar variability and climate change: Geomag-

netic aa index and global surface temperature. Geophys. Res. Lett. 25, 1035–1038.

Crowley, G., Ridley, A., Winningham, D., Frahm, R., Scharber, J. and Russell III, J.: (1999), On the

hemispheric symmetry in thermsopheric nitric oxide, Geophys. Res. Lett. 26, 1545–1548.

Danilov, A.D. and Mikhailov, A.V.: (2001), F2-layer parameters long-term trends at the Argentine

Islands and Port Stanley stations, Ann. Geophysicae 19, 341–349.

Dreschhoff, G. and Zeller, E.J.: (1998), Ultra-high resolution nitrate in polar ice as indicator of pastsolar activity, Solar Physics 177, 365–374.

Ejiri M., Aso, T., Okada, M., Tsutsumi, M., Taguchi, M., Sato, N. and Okano, S.: (1999), Japanese

research project on Arctic and Antarctic observations of the middle atmosphere, Adv. Space Res.

24(12), 1689–1692.

Feynman, J. and Ruzmaikin, A.: (1999), Modulation of cosmic ray precipitation related to climate,

Geophys. Res. Lett. 26, 2057–2060.

Field, P.R. and Rishbeth, H.: (1997), The response of the ionospheric F2-layer to geomagnetic

activity: an analysis of worldwide data, J. Atmos. Sol.-Terr. Phys. 59, 163–180.

Fogle, B. and Haurwitz, B.: (1966), Noctilucent clouds, Space Sci. Rev. 6, 279–340.

Forbush S.E.: (1937), On the effects in cosmic ray intensity observed during the recent magnetic

storm, Phys. Rev. 51, 1108–1109.

Gadsden, M.: (1990), A secular change in noctilucent cloud occurrence, J. Atmos. Terr. Phys. 52,

247–251.

Goldberg, R.A.: (1979), An experimental search for causal mechanisms in Sun/weather-climaticrelationships, in B.M. McCorinac and T.A. Seliga (eds.), Solar-terrestrial Influences on Weather

and Climate, D. Reidel, The Netherlands, pp. 161–174.

Golitsyn, G.S., Seminov, A.I., Shefov, N.N., Fishkova, L.M., Lysenko, E.V. and Perov, S.P.: (1996),

Long-term temperature trends in the middle and upper atmosphere, Geophys. Res. Lett. 23, 1741–

1744.




Greenwald, R.A., Baker, K.B., Dudeney, J.R., Pinnock, M., Jones, T.B., Thomas, E.C., Villain, J.P.,

Cerisier, J.C., Senior, C., Hanuise, C., Hunsucker, R.D., Sofko, G., Koehler, J., Nielson, E. Pel-

linen, R., Walker, A.D.M., Sato, N. and Yamagishi, H.: (1995), DARN/SuperDARN: a global

view of the dynamics of high-latitude convection, Space Sci. Rev. 71, 761–793.

Griffiths, L.A. and Shanklin, J.D.: (1987), An observation of noctilucent cloud in Antarctic winter,Weather 42, 391.

Hauchecorne, A., Chanin, M.-L. and Keckhut, P.: (1991), Climatology and trends in the middle

atmosphere temperature (33–87 km) as seen by Rayleigh lidar over France, J. Geophys. Res. 96,

15297–15309.

Hines, C.O. and Halevy, I.: (1977), On the reality and nature of a certain Sun-weather correlation, J.

Atmos. Sci. 34, 382–404.

Huaman, M.M. and Balsley, B.B.: (1999), Differences in near-mesopause summer winds, tem-

peratures, and water vapor at northern and southern latitudes as possible causal factors for

inter-hemispheric PMSE differences, Geophys. Res. Lett. 26, 1529–1532.

Jackman C.H., Fleming, E.L. and Vitt, F.M.: (2000), Influence of extremely large solar proton events

in a changing stratosphere, J. Geophys. Res. 105, 11659–11670.

Jarvis M.J., Jenkins, B. and Rodgers, G.A.: (1998), Southern hemisphere observations of a long-

term decrease in F-region altitude and thermospheric wind providing possible evidence for global

thermospheric cooling, J. Geophys. Res. 103, 20774–20787.

Jenkins, B., Jarvis, M.J. and Forbes, D.M.: (1998), Mesospheric wind observations derived from

Super Dual Auroral Radar Network (SuperDARN) HF radar meteor echoes at Halley, Antarctica:

Preliminary results, Radio Sci. 33, 957–965.

Jorgansen, T.S. and Hansen, A.W.: (2000), Comments on ‘Variation of cosmic ray flux and global

cloud coverage – a missing link in solar terrestrial relationships’, J. Atmos. Sol.-Terr. Phys. 62,

73–77.

Kattenberg, A., Giorgi, F., Grassi, H., Meehl, G.A., Mitchell, J.F.B., Stouffer, R.J., Tokioka, T.,

Weaver, A.J. and Wigley, T.M.L.: (1996), Climate models – projections of future climate, in

J.T. Houghton, L.C.M. Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell (eds.),

Climate Change 1995: The Science of Climate Change, Cambridge University Press, Cambridge,

U.K., pp. 285–357.

Keating, G. and Bougher, S.W.: (1992), Isolation of major Venus cooling mechanism and implica-

tions for Earth and Mars, J. Geophys. Res. 97, 4189–4197.Keating, G.M., Tolson, R.H. and Bradford, M.S.: (2000), Evidence of long term global decline in the

Earth’s thermospheric densities apparently related to anthropogenic effects, Geophys. Res. Lett.

27, 1523–1526.

Keckhut P., Schmidlin, F.J., Hauchecorne, A. and Chanin, M.-L.: (1999), Stratospheric and meso-

spheric cooling trend estimates from US rocketsondes at low latitude stations (8◦S–34◦N), taking

into account instrumental changes and natural variability, J. Atmos. Sol.-Terr. Phys. 61, 447–459.

Lockwood M., Stamper, R. and Wild, M.N.: (1999), A doubling of the Sun’s coronal magnetic field

during the past 100 years, Nature 399, 437–439.

Lübken, F.-J.: (1999), Thermal structure of the Arctic summer mesosphere, J. Geophys. Res. 104,

9135–9149.

Lübken, F.-J.: (2000), Nearly zero temperature trend in the polar summer mesosphere, Geophys. Res.

Lett. 27, 3603–3606.

Lübken F.-J. and Von Zahn, U.: (1991), Thermal structure of the mesopause region at polar latitudes,

J. Geophys. Res. 96, 20841–20857.Lübken, F.-J., Fricke, K.H., Langer, M.: (1996), Noctilucent clouds and the thermal structure near

the Arctic mesopause in summer, J. Geophys. Res. 101, 9489–9508.

Lübken, F.-J., Jarvis, M.J. and Jones, G.O.L.: (1999), First in situ temperature measurements at the

Antarctic summer mesopause, Geophys. Res. Lett. 26, 3581–3584.




Mikhailov, A.V. and Marin, D.: (2000), Geomagnetic control of foF2 trends, Ann. Geophysicae 18,

653–665.

Olivero J.J. and Thomas, G.E.: (1986), Climatology of polar mesospheric clouds, J. Atmos. Sci. 43,

1263–1274.

Rishbeth H.: (1990), A greenhouse effect in the ionosphere? Planet. Space Sci. 38, 945–948.Roble, R.G. and Dickinson, R.E.: (1989), How will changes in carbon dioxide and methane modify

the mean structure of the mesosphere and thermosphere? Geophys. Res. Lett. 16, 1441–1444.

Shanklin, J.D.: (1988), Reply, Weather 43, 368; 43, 380.

Shea, M.A., Smart, D.F. and Dreschhoff, G.A.M.: (1999), Identification of major proton fluence

events from nitrates in polar ice cores, Radiation Measurements 30, 309–316.

Smart D.F. and Shea, M.A.: (1985), Galactic cosmic radiation and solar energetic particles, in A.

Jursa (ed.), Handbook of Geophysics and Space Environment , Air Force Geophys. Lab., Bedford,

MA.

Stozhkov, Y.I., Pokrevsky, P.E. and Okhlopkov, V.P.: (2000), Long term negative trend in cosmic ray

flux, J. Geophys. Res. 105, 9–17.

Svensmark H. and FriisChristensen, E.: (1997), Variation of cosmic ray flux and global cloud

coverage – a missing link in solar-climate relationships, J. Atmos. Terr. Phys. 59, 1225–1232.

Taubenheim, J., Entzian, G. and Berendorf, K.: (1997), Long-term decrease of mesospheric tem-

perature, 1963–1995, inferred from radiowave reflection heights, Adv. Space Res. 20(11),

2059–2063.

Thomas, G.: (1996), Global change in the mesosphere-lower thermosphere region: has it already

arrived? J. Atmos. Terr. Phys. 58(14), 1629–1656.

Tinsley, B.A. and Deen, G.W.: (1991), Apparent response to MeV–GeV particle-flux variations – a

connection via electrofreezing of supercooled water in high-level clouds, J. Geophys. Res. 96,

22283–22296.

Ulich, T. and Turunen, E.: (1997), Evidence for long-term cooling of the upper atmosphere in

ionosonde data, Geophys. Res. Lett. 24, 1103–1106.

Veretenenko, S.V. and Pudovkin, M.I.: (1994), Effects of Forbush decrease of galactic cosmic rays

in variation of general cloudiness, Geomagnetizm I Aeronomiya 34, 38–44.

Warren, S.G., Thomas, G.E., Hernandez, G. and Smith, R.W.: (1997), Noctilucent cloud observed in

late April at South Pole Station: temperature anomaly or meteoritic dust? J. Geophys. Res. 102,

1991–2000.Willson R.C., Hudson, H.S., Frohlich, C. and Brusa, R.W.: (1986), Long-term downward trend in

total solar irradiance, Science 234, 1114–1117.

Wilcox, J.M., Scherrer, P.H., Svalgaard, L., Roberts, W.O. and Olson, R.H.: (1973), Solar magnetic

structure: influence on stratospheric circulation, Science 180, 185–186.

Woodman, R.F., Balsley, B.B., Aquino, F., Flores, L., Vazqez, E., Sarango, M., Huaman. M.M.

and Soldi, H.: (1999), First observations of polar mesospheric summer echoes in Antarctica,

J. Geophys. Res. 104, 22577–22590.

Yoshiki M. and Sato, K.: (2000), A statistical study of gravity waves in the polar regions based on

radiosonde data, J. Geophys. Res. 105, 17995–18011.

a Long Term

Documents