We report derivations of short-term periodic features observed in time series of the radiative cooling of the Earth’s thermosphere. In particular, we diagnose observations of the infrared emission from nitric oxide (NO) at 5.3 μm to reveal periodicities equal to the solar rotation period (27 days) and its next three harmonics. From 2002 to 2009 we observe 27 day, 13.5 day, 9 day and (occasionally) 6.75 day periods in the thermospheric NO cooling, the solar wind speed and the Kp geomagnetic index. Periodic features shorter than 27 days are absent in the time series of the 10.7 cm radio flux (F10.7) over this same time period. The periodic features in the NO cooling are found to occur throughout the depth of the thermosphere and are strongest at high latitudes. These results confirm the persistent coupling between the solar corona, the solar wind and the energy budget of the thermosphere.
In the last few years, short-term periodic features of 9 days have been observed in a number of parameters in the thermosphere, including density (Lei et al. 2008), composition (Crowley et al. 2008) and energy budget (Mlynczak et al. 2008). The physical cause of these periodic features has been attributed to the recurrence of high-speed solar wind streams emanating from coronal holes on the Sun. The 9 day periodicity is the third harmonic of the solar rotation period of 27 days. In this paper, we further examine the occurrence of short-term periodicities (27 days and shorter) in the infrared radiative cooling of the thermosphere over a timespan of 8 years. The radiative cooling data are derived from observations made by the sounding of the atmosphere using the broadband emission radiometry (SABER) instrument launched on the NASA thermosphere–ionosphere–mesosphere energetics and dynamics (TIMED) satellite in December 2001.
We examine the entire SABER dataset from January 2002 to November 2009 to determine the occurrence of the 27 day periodicity and its higher harmonics in the radiative cooling rates owing to emission from nitric oxide (NO) at 5.3 μm. Our prior studies have focused on periodic features in the globally averaged power (watts) radiated by NO. In this present study, we not only examine the global power on an annual basis but we also search for periodic features as a function of altitude and latitude. As will be shown later, we find that, while the periodic features exist throughout the thermosphere, there exists a definite latitude dependence in the periodic features. In addition, periodic features as short as 6.75 days (27/4) are shown to occur as well.
In the next section, we briefly review the processes by which the cooling owing to NO emission is derived. We follow this with an examination of periodic features in the solar wind speed, the 10.7 cm radio flux index (F10.7) and the Kp geomagnetic index. These are compared, on an annual basis, with the periodic features found in the NO radiative cooling. The occurrence of periodic features as a function of latitude and altitude is then given. A summary section concludes the paper.
2. Data preparation
The SABER instrument is a limb-scanning radiometer that measures spectrally integrated radiance (W cm−2 sr) in 10 discrete spectral channels ranging from 1.27 μm to approximately 16 μm (Russell et al. 1999). A fundamental scientific objective of the instrument is to make measurements to quantify the radiative energy budget of the thermosphere and mesosphere (Mlynczak 1996, 1997). The instrument was launched in December 2001 on the TIMED satellite into an orbit inclined 74° to the equator. Routine scientific operations began in January 2002. The instrument has operated nearly continuously since commencing science operations, recording in excess of 97 per cent of available data.
SABER views the atmosphere at 90° to the spacecraft velocity vector. This fact, combined with the satellite inclination, results in latitude coverage from 55° north (south) to 83° south (north). The TIMED spacecraft undergoes a yaw manoeuvre every 60 days to prevent the Sun from illuminating the SABER telescope radiator. This manoeuvre results in a shifting of the latitude coverage from north (south) to south (north). Thus, the atmosphere poleward of 55° is observed in only one hemisphere at a time and for a 60 day period.
The SABER instrument continuously scans the Earth’s limb recording profiles of infrared radiance. A single limb scan, from approximately 400 km to the hard Earth surface, requires approximately 50 s. One of the SABER channels records NO infrared emission at 5.3 μm. NO is the dominant infrared emitter responsible for cooling the atmosphere above about 120 km (Kockarts 1980).
In this paper, we examine the power (watts) radiated by the NO molecule from the thermosphere (100–250 km). The process by which the SABER radiance measurements are converted to profiles of radiative cooling, radiative fluxes and radiated power is described in Mlynczak et al. (2008). Briefly, the measured limb radiance profile is inverted assuming the emission is in the weak-line radiative transfer limit. This yields, after a correction for the limited SABER bandpass, a vertical profile of radiative cooling rate in watts per cubic metre. These cooling rates are then integrated vertically to yield values of fluxes of infrared radiation that exit the thermosphere (W m−2). Integration around latitude circles with respect to area yields the power (watts) radiated by the thermosphere. A final integration from pole-to-pole yields the total global power radiated by NO.
In these studies, we also use time series of the solar wind speed, the F10.7 index and the Kp geomagnetic index. The Kp and F10.7 data are obtained from the Space Physics Interactive Data Resource at the National Geophysical Data Center, NOAA Satellite and Information Service. The solar wind speed data are obtained from measurements made by the Solar Wind Electron Proton and Alpha Monitor (SWEPAM) instrument on the Advanced Composition Explorer (ACE) satellite (McComas et al. 1998).
Shown in figure 1 are four time series comprising, from top to bottom, the daily global power (watts) radiated by NO from January 2002 to October 2009, as derived from the SABER instrument; the daily solar wind speed obtained from the SWEPAM instrument on the ACE satellite; the daily average Kp geomagnetic index; and the daily average F10.7 solar radio flux. As discussed in Mlynczak et al. (2010), the NO time series shows the long-term decline in the NO cooling owing to the 11 year solar cycle. The effects of the prolonged solar minimum are manifest in the lower levels of NO power observed from 2007 onward. The ‘spikes’ observed in the data are due to the ‘thermostat’ effect (Mlynczak et al. 2003) caused by the impact of coronal mass ejections (CMEs) and other geomagnetic phenomena. There is substantial high-frequency content visually evident in the data.
The solar wind speed and Kp indexes do not exhibit a dramatic decline with time as does the NO power. They do, however, exhibit substantial persistent short-term variability. In particular, both Kp and the solar wind speed appear more uniform with similar short-term structure from mid-2005 onward. In contrast, the F10.7 index in the bottom panel in figure 1 exhibits a clear short-term variability that will be shown to be associated with the 27 day solar rotation period. F10.7 exhibits a decrease with time since 2002 that is associated with the declining phase of the current solar cycle. From 2008 onwards, there is an absence of short-term periodic features in the F10.7, which is associated with the absence of sunspots during this time.
Figure 2 shows the lomb normalized periodogram (LNP) for the solar wind speed, for each year 2002–2009 to date. The 27 day periodicity appears in all years but 2005. The first harmonic of the solar rotation period (13.5 days) is evident, with different strengths, in all years. The second harmonic (9 days) is strongest in 2005, 2006, 2007 and 2008. The fourth harmonic (6.75 days or 27/4) is evident in 2006, 2007 and 2009. Occasionally, other higher harmonics appear, such as the fifth harmonic (27/5 or 5.4 days) in 2004 and 2007. These features are typically not statistically significant owing to their small magnitude.
Short-term periodic features in the Kp index are shown in figure 3. The 27 day and 13.5 day periods are evident in most years. However, the 13.5 day period is strongest in the years when the 27 day period is weakest. The 9 day period is prominent in 2004–2008. The 6.75 day period is strongest in 2006. In figure 4, the LNP for the F10.7 cm solar radio flux exhibits periodic features of approximately 27 days. No higher harmonics of the solar rotation period are evident in the LNP for this parameter in any year from 2002 to 2009.
In summary, the solar wind and Kp indexes exhibit, depending on year, periodicities equal to the 27 day solar rotation period. The second, third and fourth harmonics are also evident at times. The most prominent of all the harmonics is the third harmonic at 9 days followed by the second harmonic at 13.5 days. Both Kp and the solar wind speed exhibit a period of 6.75 days (fourth harmonic) in 2003, 2006 and 2007. In contrast, the F10.7 exhibits only the fundamental 27 day first harmonic.
The LNP for the NO cooling is provided for each year 2002–2009 (to date) in figure 5. The 27 day solar rotation period is evident in most years, but not all. The occurrence of the 27 day period in the NO power is correlated most strongly in the years in which it is also strong in the solar wind speed and the Kp index. The second harmonic 13.5 day period is evident in the NO cooling most strongly in 2008 although it appears present, but weaker, in other years. It appears well correlated with the occurrence of the 13.5 day period in Kp and the solar wind speed.
The most prominent harmonic in the NO power is the 9 day period, occurring strongly in 2005, 2006, 2007 and 2008. The fourth harmonic of the solar rotation period at 6.75 days is most evident in 2006 and 2007. The occurrence of the 9 day periodicity was tied to the occurrence of coronal holes on the Sun that were spaced approximately 120° apart in solar longitude (e.g. Temmer et al. 2005). New work by Lei et al. (submitted) links similar behaviour in the distribution of coronal holes in 2008 with the occurrence of thermospheric density variations with periods of 13.5 and 9 days. This result strongly implies that the periodic features observed in NO in 2008 are due to the distribution of coronal holes, as in 2005. We suggest that, given the prolonged solar minimum from 2005 to 2008, the 13.5, 9 and 6.75 day periods observed in the infrared cooling by NO are associated with the occurrence of high-speed streams associated with coronal holes on the Sun in these years.
An advantage of the Lomb technique is that it affords a test of the significance of the various periodic features in the data. Shown in table 1, as indicated by an ‘x’, are the occurrences of the various periodicities in solar wind speed (Vsw), the Kp index and the NO infrared power in the years 2002–2009, at the 90 per cent confidence level. The 9 day periodicity in the NO emission is significant at this level in 2005–2008, while the 6.75 day periodicity is significant only in 2006.
The LNP results shown in figure 5 are for the daily, global, vertical integrated (from 100 to 250 km) power radiated by NO. To examine whether there is any variation in the occurrence or strength of the periodic features as a function of altitude or latitude we compute time series of the global power, but integrating only over 10 km bins in the vertical. We then compute the LNP for the global time series in each of these bins. This approach will elucidate whether or not there is any altitude variation in the occurrence of the periodicities. Shown in figure 6 is the LNP for the daily global time series of NO power in five different 10 km deep altitude bins. The length of the record assessed here is nearly 8 years in each LNP. The results in figure 6 show that the 27 day, 13.5 day and 9 day features occur throughout the depth of the thermosphere—that is, they are not limited to a certain altitude range. This result is fundamental as it demonstrates that the geomagnetic coupling responsible for the periodicities in radiative cooling exists throughout the entire depth of the thermosphere.
Finally, we examine the latitude dependence of the periodicity in the radiated NO power. To do so, we compute the NO power in eight latitude bins spanning the equator to the pole in each hemisphere. Seven of the bins are 11° wide, covering the equator to 77° latitude, and the eighth bin covers 77° to the pole. Although SABER data exist only to approximately ±83° latitude, we assume the SABER-derived flux (W m−2) is the same between 83° and 90° latitudes as between 77° and 83° latitudes. To obtain the power between 77° and the pole we multiply the flux measured between 77° and 83° by the area between 77° and 90° latitudes. Furthermore, as discussed above, the polar regions are observed in alternate periods of 60 days owing to the SABER viewing geometry and the spacecraft orbital inclination. Thus, we examine time series of 60 day length, which is sufficient to see the short-term periodicities of 27 days or less.
Shown in figures 7 and 8 are the LNP for the NO power in several latitude bins for two successive 60 day ‘yaw periods’ in 2005–2006. Period 23 (figure 7) is from September to November 2005 and period 24 (figure 8) is from November 2005 to January 2006. The prominent 9 day periodicity is apparent most strongly in the polar regions and mid-latitudes while it is very weak but still visually apparent in the tropics. This latitude dependence is consistent with the theory that the periodicity is geomagnetic in origin (Mlynczak et al. 2008).
We have presented studies of the occurrence of short-term periodic features in the radiative cooling of the thermosphere associated with infrared emission by the NO molecule. The infrared emission measurements are taken by the SABER instrument on the TIMED spacecraft and now span 8 years. We specifically look for periodic features associated with the solar rotation period (27 days) and its higher harmonics. Over the 8 years, we observe periods of 27 days, 13.5 days, 9 days and, in 2006 and 2007, 6.75 days. These periods correspond to the solar rotation period and its subsequent three harmonics.
The origin of the periodic features in the radiative cooling data is confirmed to be from coupling of the solar wind to the thermosphere. In particular, the 9 day and 6.75 day periods, when they occur in the NO cooling, are simultaneously present in the solar wind speed and the Kp geomagnetic index. Periods shorter than 27 days are absent in the time series of the 10.7 cm radio flux, thus excluding solar ultraviolet radiation as the cause of the periodic features.
The occurrence of the short-term periodic features was also examined as a function of altitude and latitude. We observe the periodic features throughout the depth of the thermosphere (defined herein as 100–250 km) over the duration of the dataset to date. We also find a distinct latitude dependence of the periodicities, with the periodic features being strongest at the poles and weakest in the tropics. These data and related analyses provide insight into the persistent and pervasive coupling between processes in the solar corona, the solar wind and the terrestrial thermosphere.
One contribution of 8 to a Special feature ‘Geospace effects of high-speed solar wind streams’.
- Received February 9, 2010.
- Accepted April 23, 2010.
- © 2010 The Royal Society