First observations of quasi 120 day oscillation in Mesospheric winds and temperature: Observations inferred from Meteor Radar

Authors


Corresponding authors: S. S. Das, Space Physics Laboratory, Vikram Sarabhai Space Centre ISRO-Department of Space, Trivandrum-695022, India. (dassiddhu@yahoo.com)

Abstract

[1] For the first time, the presence of quasi 120 day oscillation in the neutral winds and temperature at mesosphere and lower thermosphere (MLT) region, measured using meteor wind radar over an equatorial station Thumba (8.5°N, 76.5°E) is discussed. The amplitudes of quasi 120 day oscillation as a function of height have been established over Thumba. The origin of the quasi 120 day oscillation in geophysical parameters at MLT heights are discussed in terms of the atmospheric angular momentum, F10.7, and Lyman-alpha solar flux and nonlinear interactions between annual and semiannual oscillations. The significance of the present study lies in providing the observational evidence for quasi 120 day oscillation in the MLT region and in discussing the possible source mechanisms for the observed oscillation.

1 Introduction

[2] The middle atmospheric region is considered as the coupling region between the troposphere (where all weather phenomena take place) and the upper atmosphere (where the dynamical processes are directly influenced by solar UV radiation). It is well-established that the middle atmospheric thermal and dynamical structures are dominated by tides, planetary and gravity waves in the neutral winds, and temperature. Owing to the importance these atmospheric waves, various ground (e.g., radar and lidar), in situ (rocket), and satellite (e.g., High Resolution Doppler Lidar (HRDL) and Thermosphere-Inosphere-Mesosphere Energetics and Dynamics (TIMED) Doppler Interferometer (TIDI)) based measurements have been carried out to improve our understanding of these dynamical processes. Extensive studies were carried out in the past to explore the planetary waves in the middle and lower thermosphere (MLT) region, which includes quasi 2 day [e.g., Craig et al., 1981, 1983; Salby, 1981; Chshyolkova et al., 2005; Li et al., 2008; Babu et al., 2011], quasi 7 day [e.g., Hocking and Kumar, 2011], and quasi 16 day waves [e.g., Kingsley et al., 1978; Reddi and Ramkumar, 1995; Mitchell et al., 1999; Pancheva et al., 2004; Lima et al., 2006; Das et al., 2010]. In the MLT region, there were studies focusing on 27 day [e.g., Luo et al., 2001] and 50–70 days oscillations [e.g., Kumar et al., 2007a]. Apart from these planetary waves and intraseasonal oscillations, there were extensive studies on the semiannual oscillation (SAO), annual oscillations (AO), and quasi-biennial oscillation in the MLT region [e.g., Kumar et al., 2011]. However, the MLT region community did not pay much attention to wind and temperature oscillations with a period of quasi 120 day although the magnitude is as large as intraseasonal oscillations and SAO. It was proposed that the origin of this oscillation is due to the exchange of angular momentum between the solid Earth and the atmosphere [Djurovic and Pâquet, 1989; Djurovic, 1983]. It is also to be noted that the atmosphere excitation is related to solar activity, which oscillates with the same periods. Quasi 120 day oscillations were observed in solar activity, atmospheric angular momentum, the geomagnetic index, and interplanetary magnetic field [Djurovic and Pâquet, 1989; Pâquet et al., 1997]. The main intent of this paper is to present and discuss the signature of quasi 120 day oscillation in winds and temperature in the MLT region over Thumba. For the first time, the amplitude and phase of quasi 120 day oscillation as a function of height are established using 5 years of daily mean winds obtained from All sky Interferometric meteor radar (commercially known as SKiYMET). It is envisaged that the present results will contribute to better understanding of the dynamics of the MLT region.

2 Observational Techniques and Data Analysis

[3] The meteor radar (SKiYMET) located at an equatorial station Thumba (8.5°N, 76.5°E) employs a multichannel coherent receiver, operating at 35.25 MHz with a peak power of 40 kW and duty cycle up to 15%. This meteor radar has been operating continuously since June 2004 for the measurement of winds and temperature in the MLT region. For the present study, daily mean wind components were obtained at six height levels: 82, 85, 88, 91, 94, and 98 km, and temperature averaged over 86–94 km from January 2005 to December 2009 (5 years). The width of the height bins used for the wind averaging is 2 km [Hocking et al., 2001]. The winds derived from Thumba meteor radar are well comparable with MF radar [Kumar et al., 2007b] and TIDI observations [John et al., 2011]. The temperature-pressure parameter and thereafter daily mean absolute temperature between 86–90 km can be determined from lifetime of meteor trials [Hocking et al., 1997]. Details of system description, operating mode, and meteor detection algorithm can be found in Hocking et al. [1997, 2001] and Kumar et al. [2007b].

[4] Daily mean temperature derived from Thumba meteor radar corresponding to 90 km height region is in good agreement with the space borne measurement by the Sounding of the Atmosphere using Broadband Emission (SABER) instrument on board the TIMED satellite [Kumar, 2007; Das et al., 2012] and colocated multiwavelength day glow photometer [Vineeth et al., 2005]. In the present study, data gaps in both the wind and temperature were interpolated with the linear interpolation method. However, the observed data gaps are very few and thus do not affect the present analysis.

3 Results and Discussion

[5] Most of the atmospheric parameters show the seasonal patterns like AO and SAO. Thus, before subjecting the time series of wind and temperature to harmonics analysis, one must remove these existing seasonal components. Because our main aim is to detect the signal corresponding to 120 day oscillation and the third harmonic of AO is 120 days, which may interfere with the real signal, we need to first remove the AO component from the original time series data. A time series (signal) corresponding to the period of 365 ± 10 day is generated using least squares fitting to the original time series data of winds (at different height levels) and mesopause temperature, and then it is removed from the original time series to get the residual signals. The residual signals of zonal and meridional winds are then subjected to harmonic analysis to obtain the spectra at different height levels (82, 85, 88, 91, 94, and 98 km). In harmonic analysis, sinusoidal functions with periods from 1 to 1825 days were successively fitted using least-squares method to the time series of the zonal and meridional wind independently to obtain the periodogram at various height levels. All of the analysis carried out in the present study is based on the daily mean data. Similarly, harmonic analysis was applied to the time series of temperature obtained from meteor radar observations. Figures 1a–1c show the spectra of zonal and meridional winds at 88 km and temperatures corresponding to 90 km obtained from meteor radar. These figures reveal the presence of well-known planetary waves and oscillations of various periods. With the exceptions of the quasi 120 day oscillation, which is the secondary maximum in the spectra, other oscillations in the wind components in the MLT region were well reported and discussed in the past as mentioned earlier.

Figure 1.

Harmonic spectra for (a) zonal, (b) meridional wind at 88 km, (c) mesopause temperature, (d) atmospheric angular momentum, (e) solar flux (F10.7), and (f) Lyman-alpha after removing the seasonal component of annual oscillation from 2005 to 2009.

[6] The amplitude of the SAO is about ~12 m/s in zonal wind component, and it is interesting to note that the amplitude of quasi 120 day oscillation is as high as ~7 m/s. It is well known that SAO amplitudes are negligible in the meridional component, which can be confirmed from Figure 1b. The amplitude of quasi 120 day oscillation in the meridional component is about 4–5 m/s. The spectrum of temperature shown in Figure 1c also reveals that the secondary peak is at around 120 day with amplitude around ~2 K. The amplitude of SAO in temperature is ~ 3.5 K. From these periodograms, it is evident that the quasi 120 day oscillation is a prominent oscillation in the MLT region.

[7] As mentioned in the introduction section, Djurovic and Pâquet [1989] reported the presence of 120 day oscillation in the atmospheric angular momentum, solar activity, and other geophysical phenomena. We have also processed the recent data set of 6 hourly atmospheric angular momentum obtained from National Centers for Environmental Prediction (NCEP) and National Center for Atmospheric Research (NCAR) reanalysis data (http://www.esrl.noaa.gov/psd/map/clim/aam.rean.shtml). Figure 1d shows the amplitude spectrum of atmospheric angular momentum, which clearly shows a secondary prominent peak (after SAO) at 120 days. Similarly, we have considered solar 10.7 cm (F10.7) flux, which is used as the proxy for slowly varying component of the solar UV flux in the 160–400 nm region for our further analysis. The solar flux measurements have been taken at Penticton, Canada at local noon hours (20:00 GMT) and the data are available at www.ngdc.noaa.gov/stp/solar/flux.html. Figure 1e shows the periodogram of F10.7 flux, which clearly shows the presence of quasi 120 day oscillation. Apart from F10.7 flux, we have also considered Lyman-alpha irradiance, which is the brightest solar vacuum ultraviolet (λ < 200 nm) emission. The solar Lyman-alpha photons penetrate into the mesosphere and deposit their energy mainly by molecular oxygen dissociation at 70–110 km [Woods et al., 2000]. As well, the Lyman-alpha radiation is the dominant component in the solar spectrum to derive the atmospheric changes in the mesospheric height as it plays a major role in mesospheric chemistry. The data are available at http://lasp.colorado.edu/lisird/tss. Figure 1f shows the periodogram of Lyman-alpha, which also shows the presence of quasi 120 day oscillation. Even though, the presence of 120 day oscillation in the solar flux and Lyman-alpha provides evidence for the solar origin for this oscillation in wind and temperature in the MLT region, more information and scientific justification is needed to ascertain this claim.

[8] To study the annual variability of quasi 120 day oscillation, a band pass Butterworth filter of 105–135 days was applied to the time series of winds, temperature, and solar flux. Figure 2 shows the filtered time series plots of (a) zonal (left panels) and meridional (right panels) winds at various height levels, (b) mesopause temperature obtained from meteor radar, and (c) solar flux. These plots clearly illustrate the seasonal dependence of quasi 120 day amplitudes. In lower heights, winds show maximum amplitudes between July 2005 and January 2007, which gradually amplifies toward January 2008 (onwards) with increasing heights. Temperature shows maximum amplitude between January 2007 and January 2009. One can note that the amplitude of solar flux is maximum between July 2005 and January 2007, which is well correlated with the zonal wind maxima in lower heights. However, the modulation of 120 day oscillation in winds and temperature by AO, SAO, and quasi-biennial oscillation is not observed.

Figure 2.

Filtered time series of 105–135 day (a) zonal and meridional winds at various heights, (b) temperature at 88 km, and (c) solar flux.

[9] To extract the height-time information of amplitude and phase of the quasi 120 day oscillation, the periodograms at each height level of zonal and meridional winds for individual years are separately obtained. Figure 3 shows the height-time intensity plot of amplitude (left panels) and phase (right panels) for zonal (top panels) and meridional (bottom panels) winds corresponding to quasi 120 day oscillation from 2005 to 2009. The maximum amplitude observed in zonal wind is below 86 km, whereas for the meridional wind it is between 88 and 92 km. For both zonal and meridional wind, the phase remains almost constant. However, interannual variabilities are also observed in the amplitude and phase profiles. The entire time series of the wind data is again subjected to harmonic analysis and the averaged height profiles of amplitude and phase corresponding to quasi 120 day oscillation are extracted. Figures 4a and 4b show the height profiles of amplitude (Figure 4a) and phase (Figure 4b) corresponding to quasi 120 day oscillations in zonal and meridional winds, respectively. The height profile of amplitude shows a peak at ~85 km in zonal winds and at ~91 km in meridional winds, which is in consistence with the analysis carried out for individual years. The phase profile of zonal wind shows an increasing tendency from lower heights to the height of maximum amplitude, whereas for meridional wind, it shows a decreasing phase with increasing height from the height of maximum amplitude to upper height region.

Figure 3.

Height-time intensity plot of zonal (top) and meridional wind (bottom) amplitude (left) and phase (right) corresponding to quasi 120 day oscillation from 2005 to 2009.

Figure 4.

Height profiles of amplitude for (a) zonal wind, (b) meridional wind, and (c) cross-correlation coefficient of zonal wind with meridional wind, temperature, and solar flux.

[10] To establish the phase relation between winds (zonal and meridional), temperature, and solar flux, the quasi 120 day component of these parameters are extracted using a filter and correlated. Figure 4c shows the cross-correlation of zonal wind with meridional wind, temperature at 88 km, and solar flux. The figure clearly shows that zonal wind leads meridional wind and temperature by ~ 50 and ~20 days, respectively. However, the zonal wind and solar flux are in the same phase. The amplitudes of quasi 120 day oscillations in winds and temperature observed in the present study are quite significant and thus it is important that the studies on MLT dynamics especially over the equatorial region, should quantify this oscillation with long-term databases.

4 Concluding Remarks

[11] For the first time, the presence of quasi 120 day oscillation in neutral winds and temperature in the MLT region over an equatorial station has been discussed. The analysis revealed that the amplitude of quasi 120 day oscillations peaks at ~ 85 km in zonal winds and at ~ 91 km in meridional winds. The observed amplitude was ~ 7 m/s in zonal winds and ~4.5 m/s in meridional winds. In temperature, the amplitude is found to be around ~ 2 K at 90 km. The presence of the quasi 120 day oscillation in the solar flux of F10.7, Earth angular momentum, and Lyman alpha indicate that the observed oscillation in winds and temperature in the MLT region may be of solar origin. However, further studies are needed to substantiate this assertion because Earth's angular momentum also exhibits 120 day oscillations. Thus, present observations provided the evidence for the presence of 120 day oscillation in the MLT region for the first time, and emphasized its importance by estimating the vertical structure of amplitudes in wind and temperature, which may have implication in better understanding the MLT dynamics.

Acknowledgment

[12] The Thumba meteor radar belongs to Space Physics Laboratory, Vikram Sarabhai Space Centre (VSSC) ISRO-Department of Space, Government of India. Authors would like to thank, Director SPL, W.K. Hocking, and all the technical staffs for their support in the operation of the radar.

Ancillary