Topside equatorial zonal ion velocities measured by C / NOFS during rising solar activity

The Ion Velocity Meter (IVM), a part of the Coupled Ion Neutral Dynamic Investigation (CINDI) instrument package on the Communication/Navigation Outage Forecast System (C/NOFS) spacecraft, has made over 5 yr of in situ measurements of plasma temperatures, composition, densities, and velocities in the 400–850 km altitude range of the equatorial ionosphere. These measured ion velocities are then transformed into a coordinate system with components parallel and perpendicular to the geomagnetic field allowing us to examine the zonal (horizontal and perpendicular to the geomagnetic field) component of plasma motion over the 2009–2012 interval. The general pattern of local time variation of the equatorial zonal ion velocity is well established as westward during the day and eastward during the night, with the larger nighttime velocities leading to a net ionospheric superrotation. Since the C/NOFS launch in April 2008, F10.7 cm radio fluxes have gradually increased from around 70 sfu to levels in the 130–150 sfu range. The comprehensive coverage of C/NOFS over the low-latitude ionosphere allows us to examine variations of the topside zonal ion velocity over a wide level of solar activity as well as the dependence of the zonal velocity on apex altitude (magnetic latitude), longitude, and solar local time. It was found that the zonal ion drifts show longitude dependence with the largest net eastward values in the American sector. The pre-midnight zonal drifts show definite solar activity (F10.7) dependence. The daytime drifts have a lower dependence on F10.7. The apex altitude (magnetic latitude) variations indicate a more westerly flow at higher altitudes. There is often a net topside subrotation at low F10.7 levels, perhaps indicative of a suppressed F region dynamo due to low field line-integrated conductivity and a low F region altitude at solar minimum.


Introduction
The study of F region plasma drifts and neutral winds is important to an understanding of the dynamics and thermal balance of the thermosphere and ionosphere.Early observations of zonal and vertical plasma flow at low latitudes were largely limited to those provided by ground-based techniques such as the Jicamarca incoherent scatter radar in Peru (e.g., Woodman, 1972;Fejer et al., 1979aFejer et al., , b, 1981Fejer et al., , 1985Fejer et al., , 1991)).Such measurements provide comprehensive coverage with respect to local time, seasonal, and solar activity dependences at one geographic location.In situ satellite-based observations from spacecraft such as Atmosphere Explorer, Dynamics Explorer 2, San Marco, DMSP, ROCSAT-1, and C/NOFS (e.g., Maynard et al., 1988Maynard et al., , 1995;;Coley and Heelis, 1989;Coley et al., 1994;Hartman and Heelis, 2007;Huang et al., 2010;Fejer et al., 2013) have been used to examine ionospheric drifts (or equivalently electric fields) over spatially more extensive regions, with the limitation of mixing local time, latitude, and longitude.The electric fields measured, in turn, are the result of a complicated electrodynamic interaction between the E and F regions that varies greatly from day to night.Neutral tidal winds create the E region dynamo field that maps along equipotential field lines to the F region.The F region dynamo field is largely shorted through the E region during the day but becomes more important in the evening as the E region conductivity declines (e.g., Heelis et al., 1974;Heelis, 2004).The basic characteristics of zonal plasma drift at the equator have been presented by Fejer et al. (1981Fejer et al. ( , 1985) ) and Fejer (2011) using Jicamarca data.A typical diurnal cycle consists of westward drifts of about 50 m s −1 during the day that are solar cycle independent and solar-cycle-dependent nighttime eastward drifts that peak pre-midnight with an average value of around 130 m s −1 .Near the F peak when the zonal flow is averaged over all local times, a net superrotation is normally observed.The daytime east-west F region drifts are representative of the E region neutral winds that generate the vertical electric fields in the equatorial F region (Woodman, 1972).At night the coupling between the E and F regions decreases, and the nighttime zonal drifts become more nearly equal to the F region neutral winds that generate these fields (Woodman, 1972;Heelis et al., 1974).The local time variation of the east-west drifts has been examined by Maynard et al. (1988), Coley andHeelis (1989), andColey et al. (1994) using DE 2 satellite measurements of the electric field during a high solar flux period.Near solar maximum Pacheco et al. (2011) looked at magnetic latitude variations of zonal drifts using measurements from the ROCSAT-1 spacecraft.Recently, Pfaff et al. (2010) have shown the first results for a limited period of 2008 of F region zonal and vertical drift from the Vector Electric Field Instrument (VEFI) instrument on board the Communication/Navigation Outage Forecast System (C/NOFS) spacecraft.Fejer et al. (2013) also present VEFI results but for the more extended period of 2008-2011 observing longitudinal structure and wave-4 signature in the electric field.
The Coupled Ion Neutral Dynamic Investigation (CINDI) is an experiment on board the C/NOFS spacecraft that was launched into orbit on 16 April 2008.C/NOFS is in an ideal position to study the general behavior of the equatorial topside ionosphere as a function of solar activity.Early C/NOFS climatology results using CINDI indicate that the ionosphere is extremely cold (∼ 600 K) at night, and the O + to H + transition height was extremely low during the 2008 solar minimum and has latitudinal and LT variations (Heelis et al., 2009;Coley et al., 2010).The launch of C/NOFS coincided with the deepest solar minimum since the space age began with extended periods of no sunspots and F10.7 cm radio fluxes in the 60-70 sfu range.This low solar activity continued until near the end of 2009 when there was a recurrence of low-level sunspot activity that gradually increased through 2012.In this paper we present the observed variations of the topside zonal plasma velocity as measured by CINDI for the 2009-2012 time period as a function of solar activity, longitude and solar local time (SLT).

Spacecraft instrumentation, orbit, and data
The Ion Velocity Meter (IVM) is the part of the CINDI instrument designed to measure the thermal plasma parameters.It consists of a retarding potential analyzer (RPA) used to determine total ion density (N i ), ion temperature (T i ), composition, and the ram ion velocity component and an ion drift meter (DM) to measure the cross-track ion velocity.The RPA can resolve the fractional composition of H + , He + , and O + in the ambient plasma and determines a single mean ion temperature for all species present.A detailed description of the functioning of the RPA and DM may be found in Heelis and Hanson (1998).Depending on instrument mode, the CINDI RPA measurements are made at either a 1 or a 2 Hz cadence, and the DM operates at 24 Hz.This allows determination of the vector plasma motion at the RPA cadence.In order to ensure that only good quality ram velocity determinations were used for this study, a minimum ion density of 10 3 cm −3 was required.In addition, since the least-squares analysis algorithm that produces the ram velocity produces the most accurate results in an environment that is a mixture of light ions (H + and He + ) and O + , we required that the fractional composition of the ionosphere be between 40 % and 90 % O + .Under these conditions the estimated accuracy of the RPA ram velocity is less than 10 m s −1 (decreasing with increasing plasma density), and the estimated accuracy of the cross-track measurement is 2 m s −1 .For the purposes of this paper the plasma velocity measurements have been converted into a magnetic coordinate system.If R is the vector from the center of the Earth to the spacecraft, then the components are parallel to the geomagnetic field (B g ), the zonal component, and the meridional component where the zonal component is in the R × B g direction (positive eastward) and the meridional component is in the zonal direction × B g (positive upward).
The initial 402 × 851 km altitude 13 • inclination orbit of C/NOFS undergoes precession in such a fashion that the spacecraft perigee samples all local times over the course of approximately 67 days.For the purposes of this paper we have averaged the data into bins according to date (1 month bin size), apex altitude (10 km bin size), longitude (10 • bin size), and SLT (1 h bin size).Along with the nominal 100 % duty cycle of the spacecraft, this allows combining the bins in order to obtain statistical time and longitudinal pictures of all the measured parameters.

Observations
As previously stated, the first part of the C/NOFS mission covered a period of very low to moderate solar activity.We first examine the overall climatology of the lowlatitude topside zonal ion drifts averaged over 1-year periods beginning with the year 2012, a period of moderate solar activity that more closely resembles the conditions under which previous studies (e.g., Herrero and Mayr, 1986) have been conducted than the earlier years of the C/NOFS mission.Figure 2 shows a contour plot of the zonal ion velocity measured by CINDI as a function of longitude and solar local time using data from 2012 in the 400-1000 km apex altitude range.For this and all subsequent plots in this paper, a positive zonal drift value represents eastward flow and a negative value is westward.The monthly averaged F10.7 flux for this year was in the 100-140 sfu range.As might be expected from the previous work mentioned above, we see a pattern of westward drift during the day and eastward drift during the night.The morning change from eastward to westward occurs in the 04:00-05:00 SLT interval.This is consistent with the reversal times seen by Herrero and Mayr (1986) under similar levels of solar activity.The afternoon reversal back to eastward occurs at around 18:00-19:00 SLT, somewhat later than the 16:30-17:00 SLT reported by Herrero and Mayr (1986).We also see significant longitude variations in the afternoon and evening local times.The early evening exhibits an eastward peak that weakens in the 300-360 • longitude range, while around 16:00 SLT there are westward extrema near 30 • , 130 • , and 310 • .Indeed, the 300-360 • region has weaker eastward flow at night and stronger daytime westward flow than any other longitude sector.The standard deviations of the data are about 40 m s −1 during the day and 80 m s −1 at night reflecting the large geophysical variability of the drifts during this period of moderate solar activity.
The longitudinal structure can be examined in more detail in Fig. 3 the daytime westward drift displays a much weaker dependence on solar activity as shown in Fig. 9.This plot uses data covering the 10:00-16:00 SLT interval.Almost all the data are westward, and the slope is only 0.50 m s −1 sfu −1 .
Another useful view of the data is obtained by looking at the SLT variation of the zonal drift over a limited longitude range for each of the four complete years of data used in this study.Figure 10 gives this plot for the 260-270 • longitude, a location where we see the highest value of the premidnight eastward flow in the moderate solar activity conditions of 2012.On the dayside we see that while westward flow is always present, there is less variation for the extremely low F10.7 conditions of 2009 with an average drift of around −50 m s −1 .For the solar active period of 2012, the daytime variation is also small with an average value of about −30 m s −1 .In the nighttime we see a consistent pattern with Figure 11 presents the SLT variation of the zonal drift in the 170-180 • longitude range for 2012 for four different overlapping ranges of apex altitude (magnetic latitude).Each range covers 400 km of apex altitude, and the altitude displayed on the plot is the base altitude of the associated range (i.e., the trace labeled 400 km covers 400-800 km altitude).Here we see that the net zonal drift is organized by apex altitude with the strongest eastward drift associated with the lowest apex altitude (lowest magnetic latitude) range.The average drifts get smaller and then reverse to become negative (westward) as the apex altitude increases.

Discussion
The ground-based measurements that are most easily compared to the CINDI data set come from the Jicamarca incoherent scatter radar (located at 11.9 • N, 283.2 • E; 2 • N magnetic dip latitude).An earlier comparison by Coley and Heelis (1989) between in situ eastward velocities from DE 2 and the Jicamarca radar found generally good agreement during the high solar activity period of 1981-1983.In that study the DE 2 measurements were made in the 200-600 km altitude range while the Jicamarca measurements were taken in the 300-400 km altitude range.The most comparable Jicamarca measurements to the CINDI data in this study were reported in Fejer et al. (2013)  ratios during this time.As Fig. 11 demonstrates for the 2012 data, the magnitude of the nighttime peak depends on the apex altitude of the measurement, with lower apex altitudes (lower magnetic latitudes) having a more eastward flow.This would imply that Jicamarca might see more easterly zonal flow at night than CINDI.
VEFI data from C/NOFS are presented by Pfaff et al. (2010) and Fejer et al. (2013).These papers show data that have both similarities and differences to the results given in this paper.The zonal drifts derived from a limited amount of 2008 data in Pfaff et al. (2010) show a similar pattern to other studies but with a more limited range of 25 m s −1 westward in daytime to 50 m s −1 eastward at night.The more extensive VEFI data set used in Fejer et al. (2013) is binned so as to demonstrate seasonal variations of the zonal drift, but nevertheless shows great similarity to the data of Fig. 2, particularly in the general longitude variations.The drifts in the Jicamarca longitude sector are similar in magnitude and phase to those seen by CINDI.
An important factor here is that most previous measurements of topside zonal plasma drift have been made under moderate to active solar conditions as determined by standard EUV proxies.Since the level of solar activity under which C/NOFS has operated has varied from unusually low to moderate, this mission has provided an extension to the climatological database on the topside ionospheric region during conditions of an extremely contracted ionosphere (Heelis et al., 2009;Coley et al., 2010).This means that C/NOFS was effectively operating further above the primary altitudes of E and F region dynamo activity than other equatorial spacecraft and that the neutral winds responsible for the F region dynamo are generally well below the spacecraft altitude.The nighttime zonal flow as measured would then probably be driven by neutral winds at off-equatorial latitudes whose electric fields would map along magnetic field lines up to the region of C/NOFS.During the solar maximum period of 1999-2003, Pacheco et al. (2011) )  latitude variations of zonal drifts using measurements from the ROCSAT-1 spacecraft.They find that the eastward peak decreases in magnitude with increasing magnetic latitude, while the daytime westward flow is relatively stable suggesting that the decrease of superrotation with latitude is mainly produced by neutral winds that also decrease with latitude.
Given the weaker nature of the nighttime zonal plasma drifts in the 2008-2009 period, it is reasonable to assume that the F region dynamo is also weaker during this period in the off-equatorial latitudes.This view is supported by Huang et al. (2012), who find a correlation between VEFI zonal drifts and GRACE neutral densities and suggest that the zonal ion drifts may be used as a proxy for neutral wind measurements.

Summary
Measurements of the low-latitude topside zonal ion drift were made by the CINDI instrument on board the C/NOFS spacecraft during the 2009-2012 time interval, a period of very low to moderate solar activity as determined from F10.7 cm radio fluxes.The drifts observed under moderate solar activity in the 280 • longitude region are organized into a diurnal pattern that closely resembles the drifts seen previously by the Jicamarca incoherent scatter radar.There are clear longitude variations with smaller nighttime eastward drifts at longitudes away from the American sector.The pre-midnight drifts show a definite F10.7 dependence, while the daytime drifts show relatively little variation with solar activity.The apex altitude (magnetic latitude) variation of the drifts indicates a more westerly flow at higher altitudes.The decrease in the nighttime eastward ion flow with lower solar activity and altitude seems to indicate reduced F region zonal neutral winds that are in turn producing a considerably weakened F region dynamo during the 2009-2010 period.

Fig. 6 .
Fig. 6.Contour plot of CINDI zonal velocities versus SLT (hours) and longitude for the year 2009.All available data in the 400-1000 km apex altitude range were used.

Fig. 7 .
Fig. 7. Monthly averaged zonal drifts (diamonds) versus monthly averaged F10.7 flux for August 2008-December 2012.CINDI IVM data in the 19:00-24:00 SLT range, all longitudes, and the 400-1200 km apex altitude range were used.The least-squares fit straight line is shown, and the fit coefficients are indicated at the bottom right.See text for details.

Fig. 10 .
Fig. 10.CINDI zonal ion drift versus SLT in the 260-270 • longitude range and the 400-100 km apex altitude range for each of the 4 yr from 2009 to 2012.The 24 h averaged zonal velocity is indicated for each year.