Winds and waves in the middle atmosphere at Saskatoon (52°N, 107°W), Collm (52°N, 15°E) and Badary (52°N, 105°E)

Winds data from three radar systems in the U.S.S.R. G.D.R. and Canada, which are well-spaced along the 52°N latitude circle, are used to illustrate longitudinal/regional variations in the dynamics of the upper middle atmosphere 80–97 km. Responses to the stratospheric warming of 1982/3 are noted at all three locations, but the zonal wind does not reverse at Badary, consistent with the flow there being eastward during all months of the year. Planetary wave period oscillations (2–30 days) are observed at all locations, and highly significant cross-spectral coherences exist between the three stations.     

Winds oscillations (∼ 6h–6days) in the upper middle atmosphere at Monpazier (France, 45°N, 1°E) and Saskatoon (Canada, 52°N, 107°W) in 1979–1980

The medium frequency radar (∼ 2.2 MHz) at Saskatoon has been run continuously since 1978 and the Meteor Radar at Monpazier ran continuously for ∼ 10 day intervals in most months of 1979/1980. The radars are separated by ∼ 8000 km. Because of the excellent quality of the data, spectral and harmonic analyses have been completed from ∼ 70 to 100km and oscillations with periods of ∼ 6h–6days studied.There are substantial similarities in the mean zonal winds, both with regard to magnitudes and times of seasonal reversals; also in annual and semi-annual oscillations. In general, the semi-diurnal tide has similar amplitudes, phases and vertical wavelengths : there are consistent summer (long λ) and winter (short λ) characteristics, with rapid transitions between them. Differences between the timing of these transitions and in some of the mid-season tides are discussed. The diurnal tide is less regular and of smaller amplitude at both locations, often being too small to reliably characterize at Monpazier. However, seasonal variations between summer and winter months may be discerned.In addition to the 24 and 12 h tidal oscillations, which traditionally are studied in most detail, there is clear evidence for additional osculations near 6,8, ∼ 10 and ∼ 16 h and longer periods of ∼ 2 and ∼ 5 days. The amplitudes of these are often comparable or larger than the ‘dominant’ 24 and 12 h tides. The monthly and seasonal variations of these additional oscillations are studied, as a function of height, at the two locations. There is evidence for large scale (global) and small scale (local) disturbances in the wind field.     

Combination of Primrose Lake (54°N, 110°W) ROCOB winds (20–60 km) and Saskatoon (52°N, 107°W) M.F. radar winds (60–110 km): 1978–1983

Comparisons are made between data from the unique Saskatoon medium frequency radar set, which is continuous from mid 1978–1983 and the ROCOB data from Primrose Lake, which is only 340 km northwest of Saskatoon. Until 1981 there were 2–3 firings per week and now there are 4–5 per month. While the final agreement is satisfactory, special care was required when matching the two regions: particular problems are the low rocket sampling rate and the unexpectedly large amplitude of the diurnal tide. Combination of the two data sets is made.The Canadian zonal winds are quite similar to CIRA 72, especially in the summer months, however, the winter winds show much more systematic variability due to 12- and 6-month periods of oscillation and stratwarms. Canadian meridional winds differ significantly from existing data models.     

Quasi-biennial oscillation in ionospheric parameters measured at Juliusruh (55°N, 13°E)

When seasonal variations were eliminated by evaluating 12-month running means, the ionospheric parameters foE, foF2 and hmF2 at Juliusruh (54.6°N, 13.4°E) showed large solar cycle variations. However, when further 3-yr running averages were evaluated and subtracted, QBO (Quasi-biennial oscillations) were noticed in all these parameters. Sunspot series did not reveal a QBO, but geomagnetic Ap did show a QBO. The peaks of the ionospheric QBO and QBO of Ap could be roughly compared, with lags or leads of a few months. Also, these compared roughly with the well-known QBO peaks of tropical stratospheric (50 mb) zonal winds. Similar analyses at other locations are warranted.     

Tidal winds in the radio-meteor region over Trivandrum (8.5°N, 77°E)

Wind measurements using a meteor trail radar which performed during two June solstice and three December solstice months were analysed to study the tidal winds over Trivandrum (8.5°N). The ter-diurnal (8 h) component of wind was found to be as significant as the diurnal and semi-diurnal components. The modulations in the altitude profiles of amplitude and phase of the tidal wind components are interpreted to be due to higher order tidal modes. The amplitudes and vertical wavelengths of the tidal modes observed in the meteor zone are discussed.     

Dynamics of the upper middle atmosphere (80–110 km) at Tromsø (70°N) and Saskatoon (52°N), June–December 1987, using the Tromsø and Saskatoon MF radars

A real-time-winds (RTW) system from Saskatoon operated with the Tromsø MF (partial reflection) radar system on a continuous basis in the period June–December 1987. This interval includes MAC/SINE and EPSILON. Profiles with 3-km resolution were obtained every 5 min—weak ionization and few geomagnetic disturbances limited the observations normally to 80–110 km. However, daily mean winds, diurnal and semidiurnal tidal characteristics (amplitudes, phases and wavelengths) and gravity wave characteristics (intensities, mean directions) are available throughout this interval. This is particularly valuable in defining the background state for some experiments, e.g. rockets, and for comparison with related parameters from the lidar and other radars (EISCAT, SOUSY-VHF). Comparisons with these dynamical parameters from Saskatoon (52 N) are made : the zonal circulation was weaker at Troms0, tidal amplitudes smaller, and summer 12-h tidal wavelengths shorter ( ~ 80 km vs ~ 100 km). The fall transition for this tide occurred in September at Troms0, earlier than observed elsewhere.     

Long period wind oscillations in the meteor region over Trivandrum (8°N, 77°E)

The spectra of long period wind oscillations in the meteor zone over Trivandrum are presented. The spectral amplitudes were found to be much larger during June 1984 when the QBO in the stratospheric zonal wind was in a strong easterly phase compared with June 1987 when the zonal winds at the altitude of maximum QBO were weak westerlies. Zonal wind amplitudes for periods of 15 and 5 days were found to be most significant during these two June months. The amplitudes of these two oscillations in meridional wind were found to be as large as the amplitudes in the zonal wind. The vertical wavelength in both zonal wind and meridional winds of the 15-day oscillation is very large whereas for the 5-day oscillation the vertical wavelengths were 80 and 65 km during June 1984 and June 1987, respectively. The results are discussed.     

Variations of mean winds and tides in the upper middle atmosphere over a solar cycle,Saskatoon, Canada, 52°N, 107°W

Saskatoon (52 N, 107 W) medium frequency (MF) radar data from 1979 to 1990 have been analyzed to investigate the solar activity effects on upper middle atmospheric winds and tidal amplitudes. The period of study covers two solar maxima and a solar minimum; the continuous data allow a systematic analysis of solar cycle dependence on mean winds and tides. The height region of 79–97 km sampled in the study shows an apparent but very weak dependence of mean winds and tidal amplitudes on solar activity variation. The observed features are fairly consistent with the early results reported by Sprenger and Schmindkr [(1969) J. atmos. terr. Phy. 31, 217). The mean zonal wind and the semidiurnal tidal amplitudes appear to exhibit positive and negative correlations with the solar activity, respectively; the statistical significances of these correlations are generally low. There is a biennial periodicity evident in the zonal wind oscillations but this docs not have a consistent phase relationship with the equatorial stratospheric wind oscillations (QBO). The meridional winds and the tidal amplitudes are characterized with different and quite irregular periods of oscillations (2–5 yr). The diurnal tidal variations over the solar cycle are small and irregular, although amplitudes are slightly larger during the solar minimum years.     

Simultaneous meteor radar observations at Monpazier (France, 44°N) and Punta Borinquen (Puerto Rico, 18°N). II—Mean zonal wind and long period waves

Meteor radar wind observations have been simultaneously conducted by the CNET/CRPE meteor radar group at Monpazier (France, 44°N) and at Punta Borinquen (Puerto Rico, 18°N) in 1977–1978. The Puerto Rico results yielded some characteristics of the mean zonal wind and long period motions at low latitudes which have been compared with the characteristics of the same components observed at middle latitudes. Some differences (both time and altitude shift) are seen for the winter and spring circulation between low and mid-latitudes and even high latitudes (Kiruna, 68°N) where a CNET radar was established in 1974–1975.More energy was observed at low latitudes in the long period wave range and with the systematic existence of peaks at periods 2.5 days and between 3 and 10 days. The long period motions can be generally considered as ‘evanescent’ waves at low and middle latitudes, except during the July–August campaign in Puerto Rico.The existence of two components of 6-day and 60 h periods at both stations in March 1978 allowed us to assume that these waves were planetary waves and to study their longitudinal and latitudinal structure.     

Optical interferometer measurements of thermospheric temperature at Kavalur (12.5°N, 78.5°E), India

First results on the behaviour of thermospheric temperature over Kavalur (12.5°N, 78.5°E geographic; 2.8°N geomagnetic latitude) located close to the geomagnetic equator in the Indian zone are presented. The results are based on measurements of the Doppler width of O(¹D) night airglow emission at 630 nm made with a pressure-scanned Fabry-Perot interferometer (FPI) on 16 nights during March April 1992. The average nighttime (2130-0430 IST) thermospheric temperature is found to be consistently higher than the MSIS-86 predictions on all but one of the nights. The mean difference between the observed nightly temperatures and model values is 269 K with a standard error of 91 K. On one of the nights (9/10 April 1992, Ap = 6) the temperature is found to increase by ~250 K around 2330 IST and is accompanied by a ‘midnight collapse’ of the F-region over Ahmedabad (23°N, 72°E, dip 26.3°N). This relationship between the temperature increase at Kavalur and F-region height decrease at Ahmedabad is also seen in the average behaviour of the two parameters. The temperature enhancement at Kavalur is interpreted as the signature of the equatorial midnight temperature maximum (MTM) and the descent of the F-region over Ahmedabad as the effect of the poleward neutral winds associated with the MTM.     

Tidal winds at mesopause altitudes over Arecibo (18°N, 67°W), 5–11 April 1989 (AIDA '89)

An imaging Doppler interferometer (IDI) radar was operated during the three AIDA '89 campaigns in Puerto Rico over the period March–May of 1989. The output of the IDI analysis characterizes radar scattering in terms of a number of discrete ‘scattering points,’ also referred to as ‘multiple scattering centers,’ IDI/MSC for short. For each of these points the three-dimensional location, radial velocity and amplitude and phase are determined, similar to the output of meteor radars. We have applied the conventional Groves [(1959) J. atmos. terr. Phys. 16, 344–356] meteor wind radar analysis to the scattering points to produce the mean apparent motions over the height range from 70 to 110 km which are presented here. The mean apparent motion of the scattering centers is the quantity that would correspond to the neutral atmosphere wind or bulk motion if the scattering points are physical entities (such as turbulent eddies) whose motions are determined solely by advection. This is the quantity which is treated as the ‘wind’ in the analysis which follows and which should be compared to the wind measurements as deduced from the other methods employed during this campaign. There is, however, a caveat which supports the contention of Hineset al. [(1993) J. atmos. terr. Phys. 55, 241–287] that extreme care must be used in interpreting the velocities measured by partial reflection radars as winds. The current application of the Groves method of analysis has revealed motions from which one would infer a typical equatorial easterly circulation, with mean meridional circulation becoming significant only above 96 km. A periodogram analysis of the complete data interval (5–11 April) has shown the diurnal tide to be the most significant feature of the wind field at these altitudes, with zonal amplitudes up to some 50 m/s and meridional amplitudes approximately half this value. The 12 and 6 h tides become as significant as the diurnal above 100 km. The two day (48 ± 5 h) wave is the next most significant feature, with zonal amplitude increasing with height up to 30 m/s at 110km. The semidiurnal tide is not at all well developed below 100 km. However, analysis on a day by day basis reveals a significant semidiurnal component which is not phase coherent over the total interval. Mean vertical velocities are of the order of tens of centimeters per second and are considered to be more realistic than the meters per second velocities usually inferred from analyses of meteor trail drifts.     

Mesopause temperatures and integrated band brightnesses calculated from airglow OH emissions recorded at Maynooth (53.2°N, 6.4°W) during 1993

Spectra of the hydroxyl emissions in the wavelength range 1.0–1.6 μm, which originate at mesopause altitudes, have been obtained, using a Fourier transform spectrometer at Maynooth (53.2°N, 6.4°W), on all suitable nights during the period January–December 1993. Rotational temperatures and integrated band brightnesses have been calculated from the spectra of the OH(3, 1) and (4, 2) vibration-rotation bands. The mean annual temperatures calculated over all measurements were T(3, 1)=200±19 K and T(4, 2)=206±19 K, where the uncertainty represents the standard deviation on the measurements. Harmonic analysis of the nightly averaged temperature values revealed an amplitude of 27 ± 1 K and a phase of 95 ± 2 days in the annual variation of the (3, l) band at our latitude. The semiannual component was found to have an amplitude of 7 ± 1 K and a phase of −51 ± 9 days for this band. Results for the (4, 2) band were identical in both amplitude and phase for the annual component, while the semiannual component gave an amplitude of 8 ± 1 K and a phase of − 43 ± 7 days. These results are compared with data recorded by the SME satellite, and with the predictions of the MSISE-90 model for a station at 53° latitude. Temperatures predicted by the MSISE-90 model for Maynooth are consistently below the values obtained in this study by 15–20 K. Excellent agreement is observed between the absolute value of temperature, in the case of the SME satellite, and in the amplitude and phase of the annual variation predicted by MSISE-90. The phase of the semiannual component observed in our data deviates somewhat from the −99 ± 1 days predicted by MSISE-90.The annual mean brightness of the OH (3, 1) band was found to be 75 ± 18 kR, while that of the (4, 2) band was 106±26 kR. Diurnal variations generally showed a steady decrease from dusk to dawn, apart from a brief period in June and July. Monthly average values of band brightness have been calculated for each band and are compared with the predictions of a recent photochemical model (Le Texier et al., 1987). The model shows some elements of agreement with our observations, particularly a pair of maxima near the equinoxes, but it does not predict the broad winter maximum observed in both bands at this latitude.     

Some regularities of the lower thermosphre wind regime in summer and autumn from meteor radar measurements in obninsk (55°N, 38°E)

Meteor wind measurements made at Obninsk during MAC/SINE and MAC/EPSILON, and in the summer and autumn of other years since 1973, are reported. The zonal wind, which is presumed to be in the 93–95 km height region, is generally westerly, and the meridional wind is mainly northerly. Quasi-two-day oscillations are studied, as are semidiurnal tides. There is some evidence for a 22-yr periodicity in the amplitude of the semidiurnal tide     

Upper atmosphere semidiurnal tides at Christchurch (44°S) and Scott Base (78°S)

Lunar and solar semidiurnal tides have been determined from winds measurements in the 82 to 100 km height range at Christchurch and Scott Base made during 1983–1984. At Christchurch, the solar tide has maximum amplitudes in April and December, while at Scott Base, only the December maximum is present at all heights. Phases at Scott base mostly agree with those measured earlier at Mawson, but vertical wavelengths are always long. The lunar tide was difficult to isolate at Christchurch, but the winter to summer phase reversal was clear. This was also seen at Scott Base.     

The quasi 2-day wave at Durham (43°N): solar and magnetic effects

A quasi 2-day oscillation has been observed in the meteor winds at Durham since 1970. On a four-day basis the oscillation occurs throughout the year with amplitudes of 10 m s⁻¹ and standing wave or evanescent (λ_z > 150 km) behavior with height. During late summer the oscillation increases in amplitude to ~30 m s⁻¹ with increased phase coherence. When analyzed as a 48 h component the time of maximum of the North-South oscillation prefers the value 15 h LST implying some interaction with the solar tides. The amplitude of the 2-day component is correlated with the daily magnetic index A_p indicating magnetic activity as a possible forcing for this oscillation.     

Mean winds at 60–90 km observed with the MU radar (35°N)

Mean winds at 60–90 km altitudes observed with the MU radar (35°N, 136°E) in 1985–1989 are presented in this paper. The zonal wind at 70 km became westward and eastward in summer and winter, respectively, with a maximum amplitude of 45 m s⁻¹ westward in early July and 80 m s⁻¹ eastward at the end of November. The meridional wind below 85 km was generally northward with the amplitudes less than 10 m s⁻¹. In September to November, the meridional wind at 75–80 km becomes as large as 20–30 m s⁻¹. Those zonal wind profiles below 90 km show good coincidence with the CIRA 1986 model, except for the latter half of winter, from January to March, when the observational result showed a much weaker eastward wind than the CIRA model. The height of the reversal of the summer wind from westward to eastward was determined as being 83–84 km, which is close to the CIRA 1986 model of 85 km. The difference between the previous meteor radar results at 35–40°N, which showed the reversal height below 80 km, could be due to interannual variations or the difference in wind measurement technique. In order to clarify that point, careful comparative observations would be necessary. These mean winds were compared with Adelaide MF radar observations, and showed good symmetry between the hemispheres, including the summer reversal height, except for the short period of eastward winds above Kyoto and the long period over Adelaide.     

The characteristics of the mesospheric two-day wave as observed at Grahamstown (33.3°S, 26.5°E)

The characteristics of the quasi two-day wave, as observed in meteor wind data recorded at Grahamstown (33°19′S, 26°30′E) between April 1986 and April 1989, are described and discussed in the context of existing knowledge. As is typical, the wave amplitude has been largest during the summer months December–February with a maximum in late January, but this amplitude has differed markedly from year-to-year. The period and phase are found to be variable and, where possible, have been obtained as functions of date. January 1987 and 1989 were both characterised by clear drifts toward longer periods, with corresponding drifts in phase. At each of the summer maxima observed the period was found to be close to 48 h, with the phase of the meridional component within 3 h of local midnight. Direct comparison of local data for January 1987 with published data obtained simultaneously in Australia and Antarctica confirms the well-established westward propagation of the wave, but with an apparent zonal wavenumber somewhat smaller than the expected value of 3. It is shown that the discrepancy could arise from a combination of observational effects and a northward tilt to the wave velocity. The best result of an attempt to detect a horizontal phase gradient from local data alone is, however, more consistent with a southward tilt. There is theoretical support for both conclusions, but the matter cannot be resolved with the data presently available. It is also concluded that the background circulation at meteor heights has little influence on the wave parameters, and that indications of wave activity outside the summer season are of doubtful significance.     

Saturated gravity wave spectra derived from foil chaff experiments performed during campaign ‘DYANA’ at 44°N (Biscarrosse,Southwest France)

Two series of measurements of horizontal wind velocity, one performed in winter 1990 at 44°N latitude, and the other in summer 1988 at 69°N latitude, were subjected to a spectral analysis. The two series covered a height interval of 11–15 km centred at about the same height (92.5km in winter, 91.5km in summer). In both cases the slopes of the spectra were close to −3. A dominant wavelength (if it existed) must have been larger than the height interval covered.     

Total ozone changes in the northern hemisphere mid-latitudinal belt (30–60°N) derived from Dobson spectrophotometer measurements, 1964–1988

A statistical analysis is presented of the daily values of total column ozone obtained during 1964–1988 from a network of 24 Dobson instrument stations located within the mid-latitudinal belt 30–60°N. The year-round-, as well as the summer- and the winter data have been analyzed to detect year-to-year changes in the annual pattern of the probability density function of total ozone (PDFTO) in the belt. Long-term variations in the PDFTO have been deduced from trend analyses of the following statistical characteristics of the annual pattern of the PDFTO: mean, standard deviation, median, 10 and 90 percentile, and the asymmetry coefficient. The trends have been calculated from a multiple regression model adjusted for serial correlations, the 11-yr solar cycle, and long-term stratospheric circulation fluctuation effects. We have found that the pattern of the PDFTO has been drifting uniformly (without changes in its shape) towards lower ozone values, with some seasonal peculiarities in the drift (the annual rate of total column ozone loss). The winter pattern of the PDFTO moves (with a drift of 2.3% per decade) towards lower ozone values faster than the year round pattern (drift equals 1.8% per decade) and the summer pattern of the PDFTO (drift equals 1.5% per decade).