sign in sign up
<<
Home , Ionosphere

GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L19105, doi:10.1029/2008GL035208, 2008

Ionosphere response to solar high-speed streams Jiuhou Lei,1 Jeffrey P. Thayer,1 Jeffrey M. Forbes,1 Qian Wu,2 Chengli She,3 Weixing Wan,3 and Wenbin Wang2 Received 1 July 2008; revised 27 August 2008; accepted 28 August 2008; published 8 October 2008.

[1] We present 9- and 7- periodic oscillations in the [3] In this Letter, we investigate recurrent geomagnetic global mean Total Content (TEC) from 1 January activity with periods less than 13 days caused by periodic 2005 to 31 December 2006. Spectral analysis indicates that high speed streams and relate these observations the pronounced periodicities of 9 and 7 days observed in to periodicities observed in the global mean total electron TEC are associated with variations in solar wind high-speed content (TEC) of the ionosphere. In addition, we analyze streams and geomagnetic activity. Neutral and neutral temperature and winds measured by a Fabry-Perot winds near 250 km, measured by a Fabry-Perot Interferometer (FPI) at Resolute Bay (74.73°N, 94.89°E) Interferometer at Resolute Bay, also exhibit 9- and 7-day from the 630.0 nm emission in the F region. The ionosphere periodicities. These pronounced periodicities support and observations help to illustrate the perva- simultaneous observations of 9- and 7-day periodicities in siveness of the periodicities produced by geomagnetic thermosphere neutral density (Lei et al., 2008a; Thayer et activity. The investigation is motivated by a new discovery al., 2008). It is anticipated that the ionospheric response at 9 of thermosphere mass density oscillations at 400 km altitude and 7 days represents some combination of effects due to that occur at periods near 5, 7, and 9 days. These thermo- chemical loss, neutral winds, and disturbance dynamo- sphere density oscillations, with perturbations of 20–40%, driven electric fields. Citation: Lei, J., J. P. Thayer, J. M. have been correlated with similar periodicities in Kp and Forbes, Q. Wu, C. She, W. Wan, and W. Wang (2008), solar wind speed, whose origins are related to coronal hole Ionosphere response to solar wind high-speed streams, Geophys. distribution on the ’s chromosphere and the production Res. Lett., 35, L19105, doi:10.1029/2008GL035208. of solar wind high speed streams [Lei et al., 2008a, 2008b; Thayer et al., 2008]. The persistence of coronal holes over 1. Introduction many solar rotations and the related solar wind disturbances lead to recurrent geomagnetic activity, especially during the [2] Variability in the ionosphere F-region can be caused declining phase of the and in solar minimum. by solar activity, seasonal and diurnal variations, geomag- Recurrent geomagnetic activity leads to periodic heating of netic activity and meteorological influences [e.g., Rishbeth the and a subsequent response in global and Mendillo, 2001]. By studying repeatable ionospheric thermosphere density. This new finding introduces a new variability on time intervals of days, processes such as solar-terrestrial connection that produces periodicities in meteorological and geomagnetic influences on the F region the thermosphere at subharmonics of the 27-day solar can be studied in general isolation from the other processes rotation. An interesting question is whether these perio- listed above. Of particular attention in recent years is the dicities are observed and correlated with ionospheric influence of the lower atmosphere on the F-region [e.g., variability and other thermosphere properties such as wind Altadill and Apostolov, 2003; Mendillo et al., 2002; and temperature. Forbes et al., 2000]. Tantalizing periodicities observed in the F-region ionosphere of 2, 5, 10, and 16 days have prompted studies invoking lower atmosphere influences due 2. Data Sets to similar periodicities observed in planetary wave activity [4] We use the global mean vertical TEC time series from [e.g., Altadill and Apostolov, 2003; Lasˇtovie`ka, 2006]. In all 1 January 2005 to 31 December 2006. This time series is these studies, it has been recognized that geomagnetic calculated from global ionospheric TEC maps (GIMs) and activity can be a significant contributor to F region vari- corresponds to a time interval where pronounced periodici- ability on time scales of 2–30 days, but little work has been ties of 9 and 7 days were observed in thermosphere neutral pursued to investigate the solar-terrestrial connection that density, solar wind velocity and geomagnetic activity [Lei et can cause such variability in the ionosphere and whether the al., 2008a; Thayer et al., 2008]. The GIMs are provided by observed multi-day periodicities in the ionosphere may be Jet Propulsion Laboratory based on the GPS TEC observa- due to periodicities in solar wind forcing. tions from around 200 GPS receivers. The details of the retrieval are given by Mannucci et al. [1998]. The GIMs used here have a longitude and latitude resolution of 5°

1 2.5° and temporal resolution of 2 hours. Therefore, there are Department of Aerospace Engineering Sciences, University of Color- 72 cells along both the longitude and latitude directions ado, Boulder, Colorado, USA. 2High Altitude Observatory, National Center for Atmospheric Research, with 5184 total cells in each snapshot of the TEC map. The Boulder, Colorado, USA. 3Institute of Geology and , Chinese Academy of Sciences, Beijing, China.

Copyright 2008 by the American Geophysical Union. 0094-8276/08/2008GL035208$05.00

L19105 1of5 L19105 LEI ET AL.: IONOSPHERE RESPONSE TO SOLAR WIND L19105

advantage of this method is to mitigate local features in the ionospheric density [Afraimovich et al., 2008]. [5] In addition to the TEC data, the neutral temperature and winds near 250 km from the FPI at Resolute Bay are used to illustrate the pervasive nature of these periodicities in the ionosphere and thermosphere. The information about the FPI at Resolute Bay and the error analysis of the measurements are provided by Wu et al. [2004]. The neutral temperature from days 263 to 321 in 2005 and from days 265 to 366 in 2006, and geographic meridional winds from days 301 to 366 in 2006 are used. The FPI winds in 2005 are not used in this study because of data gaps caused by cloud weather. [6] The hourly averaged solar wind speed at the L1 point observed by the ACE and the 3-hourly planetary magnetic activity index, Kp, will also be used in this study. The solar wind data are obtained from the GSFC/SPDF OMNIWeb interface and Kp data are provided by NGDC database.

3. Results

[7] Figure 1a shows the variations of the global mean TEC as a function of day number in 2005 and 2006. The annual, semi-annual, and 27-day variations of the global mean TEC are clearly seen in Figure 1a; these variations have been studied by Afraimovich et al. [2008] and Hocke [2008]. Imbedded in the large time scale variations are shorter oscillations with multi-day periods (less than the 27-day solar rotation period). These multi-day oscillations can cause a few units of TEC (TECu) change in the global mean TEC. More significantly, these oscillations were persistent for the entire two year period under study. [8] To illustrate the characteristics of the periodic oscil- lations, Lomb-Scargle (LS) periodograms [Lomb,1976; Figure 1. Variations of (a) global mean TEC as a function Scargle, 1982] were calculated on the global mean TEC of day number and (b) the corresponding Lomb-Scargle in 2005 and 2006 (Figure 1b). Dominant spectral compo- periodogram for years (left) 2005 and (right) 2006. The unit nents above the 95% significance level in the 2005 data set, of TEC is TECu (1 TECu = 1016 /m2). Lomb- in order of highest to lowest amplitudes, are found at Scargle periodograms of percent residuals from 11-day periods of 27, 9 and 13.5 days with amplitudes of 1.3, 0.4 running mean TEC at (c) high (60°N–90°N), (d) middle and 0.3 TECu. In the 2006 data set, periods of 27, 7 and 9 (30°N–60°N) and (e) low (0°N–30°N) latitude bands, days are predominant with the amplitudes of 0.7, 0.23 and (f) solar wind velocity (Sw), and (g) Kp in (left) 2005 and 0.2 TECu (we refer to oscillations with periods from 22 to (right) 2006. The horizontal dashed lines represent the 95% 32 days as the 27-day oscillation). From the LS periodo- significance level. grams, we can see that the amplitudes of the 7-, 9-, and 13.5-day oscillations are around one third of the 27-day oscillation. Hocke [2008] also found the 7-, 9-, and 27-day global mean TEC is calculated from the GIMs by the oscillations from a spectral analysis using global mean TEC following equation [see Afraimovich et al., 2008]: from January 1995 to March 2007, and showed a perfect correlation in the 27-day oscillation between global mean X TEC and solar EUV. However, Hocke suggested other X1 TEC ¼ Ii;j Á Si;j ð1Þ mechanisms rather than solar irradiation should contribute Si;j i;j to the 7-, and 9-day oscillations in global mean TEC i;j because the 9-day periodicity is absent in the Mg II and 7-day oscillation of solar EUV is negligible. where i, j are the indices along the longitude and latitude of Note that the spectral analysis for the solar EUV radiation the GIM cell; I is the vertical TEC values in each GIM i,j flux from the SEM/SOHO during 2005–2006 also indicates cell; and S denotes the corresponding GIM cell area. In the i,j that solar radiation variations are too small [Lei et al., calculation of the global mean TEC, i =1,2,..., 72, and j = 2008b] to explain the observed 7- and 9-day variations in 1, 2, ..., 72. In addition to the global mean TEC, the mean TEC. TEC values at high (60°N–90°N), middle (30°N–60°N) [9] Now we focus on the investigation of whether the 7- and low (0°N–30°N) latitudinal bands in the Northern and 9-day periodicities in TEC correspond well with recur- Hemisphere are also calculated from equation (1). The main

2of5 L19105 LEI ET AL.: IONOSPHERE RESPONSE TO SOLAR WIND L19105

Figure 2. Evolution of the wavelet power spectra of percent residual from (top) 11-day running global mean TEC, (middle) solar wind velocity, and (bottom) Kp in (left) 2005 and (right) 2006. The white solid contours denote the regions of the wavelet spectrum above 95% confidence level, and the horizontal white dashed lines indicate the positions of the periods of 5, 7, 9, and 13 days. Highest intensities are in red and lowest in dark blue.

rent geomagnetic activity and periodic high speed solar observe the time evolution of the periodicities over the two wind streams. Note that the 13.5-day periodicity will not be year period. The Morlet wavelet analysis was applied to the studied here, although this component is present in the TEC percent residuals in the global mean TEC from an 11-day spectrum in 2005. This is because the 13.5-day periodicity running mean (Figure 2, top). The spectral power of TEC in can be present in both solar EUV radiation and geomagnetic 2005 shows a distinct enhancement around the band at a activity [e.g., Mursula and Zieger, 1996]. A comparison of 9 day period, around days 20–100, days 120–150 and days LS periodograms of percent residual TEC from 11-day 240–360. It can also be seen that the spectral power evolves running mean (obtained from N5.5 to N+5.5 days, where toward a 7-day period from days 120 to 160. For year 2006, N is the current day) with those of solar wind velocity and the 7-day oscillation is predominant, and it occurs from geomagnetic activity index Kp in 2005 and 2006 is also days 90 to 140 and days 250 to 360; the oscillations at shown in Figures 1c–1g. Here we use the mean TEC at periods of 9 day or 4–5 day also appear in a few short time high, middle and low latitude bands instead of the global intervals. The time evolution of the power in TEC pertur- mean TEC in order to examine the latitudinal dependence of bations is similar to those in both solar wind speed and Kp the 7- and 9-day periodicities in TEC. It is evident that the in both 2005 and 2006 (Figure 2), indicating the observed 7- 7- and 9-day periodicities in TEC are similar to the and 9-day oscillations in TEC are related to the variations of oscillations in solar wind and Kp. All variables (i.e., TEC, solar wind and geomagnetic activity. solar wind and Kp) have a dominant peak at 9 days in 2005, and pronounced peaks at 7- and 9-day periods in 2006. The 4. Discussion and Conclusions relative amplitudes in TEC, with 3–4% for the periodicities at the periods of 9 days in 2005 and 7 days in 2006, can be [11] We have presented the 9- and 7-day periodic oscil- observed at all latitudes, with some indication of greater lations in the global mean TEC during 2005–2006. The change at high latitudes than at low latitudes. The relative relative fluctuations in the 9- and 7-day periodicities in TEC fluctuations are about ±6% if a band-pass filter with a width are about ±6% and introduce significant recurring variations of 2 days, centered on 9 days or 7 days, is applied due to the in the ionospheric electron density. Our spectral analysis nonsinusoidal nature of theses periodicities. Hocke [2008] reveals that the pronounced periodicities of 9 and 7 days showed that the relative amplitude for the 27-day periodic- observed in TEC are associated with the variations of solar ity is about 8–14% during 2005–2006 [see Hocke, 2008, wind stream and geomagnetic activity. The remaining Figure 6]. Although the 7- and 9-day periodicities in TEC question is why the periodicities of solar wind and geo- are smaller than 27-day periodicity, they still can introduce magnetic activity cause periodic ionospheric oscillations at significant day-to-day variations in the ionospheric density 9 days or 7 days. because of their much higher occurrence , which [12] Lei et al. [2008a] and Thayer et al. [2008] reported will be discussed later. the pronounced periodicities of 9 and 7 days observed in the [10] The correlative periodicities of the 9- and 7-day CHAMP neutral densities in 2005, and 2006, respectively. periodicities in TEC with the solar wind and Kp index The neutral mass density measured by the CHAMP satellite supports similar relations observed in the thermosphere is primarily responding to temperature changes in the density by Lei et al. [2008a] and Thayer et al. [2008]. In thermosphere that lead to a redistribution of mass density. a similar approach to that described by Thayer et al. [2008], Although this type of mass redistribution is not necessary to we apply a Morlet wavelet analysis to the TEC data to affect the relative concentrations of atomic and molecular

3of5 L19105 LEI ET AL.: IONOSPHERE RESPONSE TO SOLAR WIND L19105

important role in the periodicities of electron density in F2 region or TEC. [15] In addition, the F region electron density or TEC can be significantly changed by transport through neutral winds and electric fields [e.g., Pro¨lss, 1995]. The variations of Joule and particle heating at high latitudes due to the recurrent geomagnetic activity may force neutral winds to oscillate at the periods of 7 and 9 days. As shown in Figure 3, the 6–7 day oscillation above 95% confidence level is clearly seen in the Resolute Bay FPI meridional winds in 2006, although the 6–7 day oscillation is not as pronounced as in the neutral temperature and there is also a broad spectrum of the wind periodicities in the range of 2– 7 days. This may be related to less reliable wind data due to cloud weather. It must be recognized that observations at this location may not be representative of global or even regional conditions. On the other hand, electric fields are generated by neutral winds in the E region in daytime and in the F region at night, or they penetrate from to ionosphere. The penetration electric fields from magne- tosphere can not last multi-days because of the shielding of the region 2 current, whereas electric fields may oscillate at the same as neutral winds through the distur- bance wind dynamo process [Blanc and Richmond, 1980]. Thus, the 7-, and 9-day periodicities in electric fields or neutral winds modulated by the solar wind and geomagnetic activity are also potential causes for the corresponding periodicities in electron density and TEC. [16] Mlynczak et al. [2008] reported a 9-day periodicity Figure 3. Lomb-Scargle periodograms of neutral tempera- in the infrared cooling of the , and also sug- ture in (top) 2005 and (middle) 2006 and (bottom) gested a direct coupling between the Sun’s upper atmo- meridional winds in 2006 measured by the Resolute Bay sphere and the infrared energy budget of the lower FPI in the winter seasons. The dashed lines represent the thermosphere. It is interesting that the periodic fast solar 95% significance level. wind streams and recurrent geomagnetic activity even affect the lower thermosphere. The excited lower thermosphere by species, there remains an impact on the ionosphere, through the recurrent geomagnetic activity may play a role in changes in its peak height influenced by neutral temperature producing the periodicities in electron density and TEC in for example [e.g., Zhang et al., 1999]. complex and coupled pathways. Due to the impact on the [13] Figure 3 shows the neutral temperature spectra in the lower thermosphere, it is interesting to note that studies winter seasons (when the northern cap was in night- using mesosphere/lower thermosphere winds to extract time condition) of 2005 (Figure 3, top) and 2006 (Figure 3, planetary wave periodicities, such as by Altadill and middle) measured by the Resolute Bay FPI instrument. In Apostolov [2003], may be contaminated by the recurrent 2005, the 9-day periodicity is clearly seen in the neutral periodicities in geomagnetic activity. This can further con- temperature with an amplitude as large as 60 K. In 2006, the fuse the relation of planetary wave periodicities to correla- spectral amplitude at 6–7 days is most distinct, with less tions in F-region variability. pronounced peaks at 9, 4, and 13 days. It is evident that the [17] An extensive and impressive review of TEC contri- spectra of temperature are also similar to those of solar wind butions to ionospheric storms is given by Mendillo [2006]. and Kp in Figure 1. The temperature periodicities are Our findings further enhance the utility of TEC measure- expected to affect the electron density in the F2 region or ments for ionosphere study. However, the processes at play TEC by changing the recombination rate and the F2 layer in modifying the ionosphere require additional data sources height of the ionosphere. and model simulations to understand the relative contribu- [14] The O/N2 ratio may exhibit similar oscillatory be- tion of the periodicities of the neutral density, neutral havior linked to recurrent geomagnetic activity as mass temperature, neutral winds, neutral composition and electric density and neutral temperature. Daytime O+ densities up fields associated with periodic high speed solar wind to the F2 peak are strongly influenced by the of O streams and recurrent geomagnetic activity in driving the and its chemical loss [e.g., Pro¨lss, 1995]. It is lost by ionospheric periodic oscillations. We expect these multi- chemical reactions, primarily with N2 and O2. As a result, day periodicities to be observed in other thermosphere/ + the O density is often related to changes in the O/N2 ratio ionosphere/magnetosphere data sets in solar minimum during the day and time integrated changes in 1/N2 density thus enabling a much more extensive analysis of the at night. Thus the periodicities in neutral composition entire geospace response to solar wind high speed stream caused by the recurrent geomagnetic activity may play an disturbances.

4of5 L19105 LEI ET AL.: IONOSPHERE RESPONSE TO SOLAR WIND L19105

[18] Acknowledgments. We thank the Jet Propulsion Laboratory, Mendillo, M., H. Rishbeth, R. G. Roble, and J. Wroten (2002), Modelling California Institute of Technology for the development and operation of F2-layer seasonal trends and day-to-day variability driven by coupling GIM. This work was supported by the AFOSR MURI Award FA9550-07-1- with the lower atmosphere, J. Atmos. Sol. Terr. Phys., 64, 1911–1931. 0565 and by Grant ATM-0719480 from the National Science Foundation as Mlynczak, M. G., F. J. Martin-Torres, C. J. Mertens, B. T. Marshall, R. E. part of the Program. The Resolute FPI observation is Thompson, J. U. Kozyra, E. E. Remsberg, L. L. Gordley, J. M. Russell supported by a NSF award ATM0404790 to NCAR. NCAR is supported III, and T. Woods (2008), Solar-terrestrial coupling evidenced by periodic by the National Science Foundation. behavior in geomagnetic indexes and the infrared energy budget of the thermosphere, Geophys. Res. Lett., 35, L05808, doi:10.1029/ References 2007GL032620. Mursula, K., and B. Zieger (1996), The 13.5-day periodicity in the Sun, Afraimovich, E. L., E. I. Astafyeva, A. V. Oinats, Y. V. Yasukevich, and solar wind, and geomagnetic activity: The last three solar cycles, I. V. Zhivetiev (2008), Global electron content: A new conception to J. Geophys. Res., 101(A12), 27,077–27,090. track solar activity, Ann. Geophys., 26, 335–344. Pro¨lss, G. W. (1995), Ionospheric F-region storms, in Handbook of Atmo- Altadill, D., and E. M. Apostolov (2003), Time and scale size of planetary spheric Electrodynamics, vol. 2, edited by H. Volland, pp. 195 –248, wave signatures in the ionospheric F region: Role of the geomagnetic CRC Press, Boca Raton, Fla. activity and mesosphere/lower thermosphere winds, J. Geophys. Res., Rishbeth, H., and M. Mendillo (2001), Patterns of ionospheric variability, 108(A11), 1403, doi:10.1029/2003JA010015. J. Atmos. Sol. Terr. Phys., 63, 1661–1680. Blanc, M., and A. D. Richmond (1980), The ionospheric disturbance Scargle, J. D. (1982), Studies in astronomical time series analysis. II. Sta- dynamo, J. Geophys. Res., 85, 1669–1686. tistical aspects of spectral analysis of unevenly spaced data, Astrophys. J., Forbes, J. M., S. Palo, and X. Zhang (2000), Variability of the ionosphere, 263, 835–853. J. Atmos. Sol. Terr. Phys., 62, 685–693. Thayer, J. P., J. Lei, J. M. Forbes, E. K. Sutton, and R. S. Nerem (2008), Hocke, K. (2008), Oscillations of global mean TEC, J. Geophys. Res., 113, Thermospheric density oscillations due to periodic solar wind high-speed A04302, doi:10.1029/2007JA012798. streams, J. Geophys. Res., 113, A06307, doi:10.1029/2008JA013190. Lasˇtovie`ka, J. (2006), Forcing of the ionosphere by waves from below, Wu, Q., R. D. Gablehouse, S. C. Solomon, T. L. Killeen, and C.-Y. She J. Atmos. Sol. Terr. Phys., 68, 479–497. (2004), A new Fabry-Perot interferometer for upper atmospheric re- Lei, J., J. P. Thayer, J. M. Forbes, E. K. Sutton, and R. S. Nerem (2008a), search, Proc. SPIE Int. Soc. Opt. Eng., 5660, 218–227. Rotating solar coronal holes and periodic modulation of the upper atmo- Zhang, S.-R., S. Fukao, W. L. Oliver, and Y. Otsuka (1999), The height of sphere, Geophys. Res. Lett., 35, L10109, doi:10.1029/2008GL033875. the maximum ionospheric electron density over the MU , J. Atmos. Lei, J., J. P. Thayer, J. M. Forbes, E. K. Sutton, R. S. Nerem, M. Temmer, Sol. Terr. Phys., 61, 1367–1383. and A. Veronig (2008b), Global thermospheric density variations caused by high-speed solar wind streams during the declining phase of solar cycle 23, J. Geophys. Res., doi:10.1029/2008JA013433, in press. J. M. Forbes, J. Lei, and J. P. Thayer, Department of Aerospace Lomb, N. R. (1976), Least-squares frequency analysis of unequally spaced Engineering Sciences, University of Colorado, 1000 Engineering, Boulder, data, Astrophys. Space Sci., 39, 447–462. CO 80309, USA. ([email protected]; [email protected]) Mannucci, A. J., et al. (1998), A global mapping technique for GPS-derived C. She and W. Wan, Institute of Geology and Geophysics, Chinese ionospheric total electron content measurements, Sci., 33, 565– Academy of Sciences, Beijing 100029, China. 582. W. Wang and Q. Wu, High Altitude Observatory, National Center for Mendillo, M. (2006), Storms in the ionosphere: Patterns and processes for Atmospheric Research, 3080 Center Green, Boulder, CO 80301, USA. total electron content, Rev. Geophys., 44, RG4001, doi:10.1029/ 2005RG000193.

5of5