Observations of the Weddell Sea Anomaly in the ground-based and space-borne TEC measurements

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Contents lists available at ScienceDirect

Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp

Observations of the Weddell Sea Anomaly in the ground-based and space-borne TEC measurements Irina Zakharenkova a, b, *, Iurii Cherniak a, c, Irk Shagimuratov a a b c

Kaliningrad Department of IZMIRAN, 41 Pobeda Av., 236010, Kaliningrad, Russia Institut de Physique du Globe de Paris, 75013, Paris, France COSMIC Program Office, University Corporation for Atmospheric Research, Boulder, Colorado, USA

A R T I C L E I N F O

A B S T R A C T

Keywords: Weddell Sea Anomaly Swarm TEC Topside ionosphere

The Weddell Sea Anomaly (WSA) is a summer ionospheric anomaly, which is characterized by a greater nighttime ionospheric density than that in daytime in the region near the Weddell Sea. We investigate the WSA signatures in the ground-based TEC (vertical total electron content) by using GPS and GLONASS measurements of the dense regional GNSS networks in South America. We constructed the high-resolution regional TEC maps for December 2014–January 2015. The WSA effects of the TEC exceed the noontime values are registered starting from 17 LT, it reaches its maximum at 01–05 LT and starts to disappear after 09 LT. Maximal TEC enhancements were as large as a factor of 2.5–3.5 and were registered at 03–04 LT. This effect was mainly localized in the geographical region of 55
S-75
S latitude and 80
W-30
W longitude, close to the Antarctic Peninsula. Further, we examined the WSA occurrence in the topside ionosphere by using GPS measurements from a zenith-looking GPS antenna on board three Swarm satellites to determine topside TEC (above ~500 km altitude) at the topside ionosphereplasmasphere system. Global maps of the topside TEC indicated that the zone with significant WSA effect in the topside TEC (TEC increase ~2–4 times the noontime level) had a large spatial extent over southern Pacific and Atlantic Ocean. It was observed around 150
W-20
W and between 40
S and 70
S during 23 LT - 06 LT. For the first time, the WSA signatures were shown in the topside TEC data derived from the GPS measurements onboard the Swarm constellation. Independently, two other instruments - FORMOSAT-3/COSMIC radio occultation electron density profiles and in situ measurements by the Langmuir Probe instrument onboard Swarm satellites – were able to confirm: (1) the same location of the WSA zone as revealed in Swarm TEC; (2) the most-pronounced WSA effect, as a maximal electron density exceed over the noontime values, corresponds to altitudes above 400–500 km.

1. Introduction The Weddell Sea Anomaly (WSA) effect was found first in ionosonde measurements in the 1950s after the International Geophysical Year in 1957 (Bellchambers and Piggott, 1958; Penndorf, 1965) from the Falkland Islands (52
S, 60
W, dip 50.4) to the southern shore of the Weddell Sea (around 75
S, 30
W, dip 64.0). According to the ionosonde observations, during the local summer the ionospheric F2 peak density (NmF2) varied in such a way that the daily peak occurs at local night instead of during daytime hours. The WSA feature only appeared during the southern hemisphere summer, in the months around the December solstice. The observations of the NmF2 maximum displacement to the post-midnight sector during local summer (November–February) for the Argentine Islands ionosonde (65.3
S, 64.3
W) were reported by Wrenn

et al. (1987) and then they were compared with model simulations of the Utah State University Time-Dependent Ionospheric Model (TDIM) by Sojka et al. (1988). However, the main constrain of the ground-based facilities (in particular, ionosondes) is that they were able to provide only limited coverage of this area at the Antarctic Peninsula and a few islands nearby. Also a total number of the deployed ionosondes in this region was significantly reduced since International Geophysical Year in 1957. That is why real spatial extent of the WSA was not discovered for several decades. Such limitations of the sparse ground-based instruments could be overcome only by satellite observations. One decade ago, new LEO (Low Earth Orbit) satellite missions aroused a spate of interest to the WSA phenomenon. Horvath and Essex (2003) and Horvath (2006) used data from the TOPEX/Poseidon satellite to study the WSA appearance in 1998/1999 and 1996/1997 respectively,

* Corresponding author. Kaliningrad Department of IZMIRAN, 41 Pobeda Av., 236010, Kaliningrad, Russia. E-mail addresses: [email protected] (I. Zakharenkova), [email protected] (I. Cherniak), [email protected] (I. Shagimuratov). http://dx.doi.org/10.1016/j.jastp.2017.06.014 Received 29 July 2016; Received in revised form 12 May 2017; Accepted 25 June 2017 Available online 27 June 2017 1364-6826/© 2017 Elsevier Ltd. All rights reserved.

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

and they were first to found that it actually extended over the South Pacific as far as 160
W at the time of its formation near dusk in local solar time. The TOPEX mission provided altimeter measurements, which could be converted into the ionospheric TEC (Total Electron Content) for the altitudinal range of 0–1336 km, but only over the ocean surface. Jee et al. (2009) have analyzed more than 13 years of TOPEX TEC data to study seasonal and solar activity variations of the WSA. Next very successful mission was FORMOSAT-3/COSMIC launched in 2006. This mission consists of six identical satellites at six different orbits of 700–800 km altitude and provides unprecedented number of electron density profiles retrieved from the GPS Radio Occultation (RO) experiment. Burns et al. (2008) reported one of the first results of the WSA features observed in the COSMIC data. They extracted and analyzed the ionospheric F2 peak parameters (density, NmF2, and height, hmF2) from RO profiles for April 2006–August 2007 period. They suggested that the WSA was a continuation of the southern, summer equatorial anomaly that had been displaced southward. They also discussed modelling efforts to simulate the WSA using the Coupled Magnetosphere Ionosphere Thermosphere (CMIT) model. These runs did not produce the WSA. He et al. (2009) analyzed the WSA appearance in the ionospheric F2 peak using the global NmF2 and hmF2 maps derived from COSMIC RO profiles for more than two years of observations. They report that the NmF2 change is associated with the hmF2 change, while the latter is correlated closely with the components of the geomagnetic field. Lin et al. (2009) showed first threedimensional density structure of the WSA in December 2007 using the COSMIC RO profiles. The electron density slices in the altitudinal range of 200–500 km were constructed by data averaging during one-month period. The WSA and the nearby density enhancement region in the southern hemisphere occur over a large area between 30
S and 90
S latitudes and 150
W-30
E longitudes. The WSA feature is most significantly seen at an altitude of 300 km. Furthermore, it was found that a WSA-like feature with similar electron density enhancement occurs also in the Northern Hemisphere (Horvath and Lovell, 2009), namely at northeast Asia, close to Russian Yakutsk and the Okhotsk Sea. As the midlatitude nighttime (evening) density enhancements were found to exist in both hemispheres (Luan et al., 2008; Thampi et al., 2009; Lin et al., 2010), it was concluded that it is a general midlatitude ionospheric phenomenon and it was also named as “Mid-latitude Summer Nighttime Anomaly” (MSNA). Thampi et al. (2009) highlighted two major features of the MSNA: (1) the electron density is higher during night than daytime and (2) at night, the electron density at mid-latitudes remains higher than lower latitudes. Liu et al. (2010) regarded the MSNA as a phase reversal of the diurnal cycle and found three such regions on the globe, namely at East Asian, Northern Atlantic, and South Pacific (which included the WSA region), by making use of 6 years of in situ electron density measurements at 400 km from the CHAMP satellite. Using the Thermosphere, Ionosphere, Mesosphere Energetics and Dynamics/Global Ultraviolet Image (TIMED/GUVI) observations. Zhang et al. (2013) investigated the so-called midlatitude arcs (MLA) as an example of the nightside enhancements of ionospheric electron density at 20
–45
magnetic latitudes in both hemispheres and concluded that MSNA may be a subset of MLA. From both ionosonde and satellite observations the WSA is considered mainly as the F2 peak phenomenon. However, Horvath and Lovell (2009) analyzed the in situ observations onboard the DMSP (Defense Meteorological Satellite Program) F12 satellite to study the evening/ nighttime topside ionosphere during the 1996/1997 southern summer. The constructed maps of the topside ionosphere's plasma density parameters at an altitudinal range of about 850–870 km were able to track a complete nighttime WSA structure. More recently, Slominska et al. (2014) reported about detection of the WSA and MSNA signatures in the in situ data onboard the DEMETER satellite during the years 2005–2010. Though due to a sun-synchronous orbit these observations were limited to 10.5 LT and 22.5 LT time interval, but an altitude of measurements was about 700 km (later decreased to 620 km), i.e. the region greatly above the F2 peak. Further, Liu et al. (2014) revealed the anticorrelation

between electron temperature (Te) and electron density in the topside ionosphere along the WSA latitudes, which were observed concurrently by COSMIC, DEMETER and Tatiana-2 satellites at an altitudinal range of 660–830 km. However, a number of papers on the WSA observations in the topside ionosphere is very limited. Here, we should also note the ground-based GPS TEC measurements. Today ground-based GPS TEC and global ionospheric maps (GIMs) of TEC are one of the most widely used observations for the Earth's ionosphere monitoring, climatology and space weather applications (Jakowski, 1996; Garner et al., 2008). A number of the ground-based receivers within the global and regional networks explosively grows from several hundreds worldwide in the 1990s to more than 6000 stations today. However, there are a few publications on the WSA observations in the ground-based GPS TEC data (Mosert et al., 2011) and there are only few attempts to use GIMs TEC for this purpose as a supportive material (He et al., 2009; Zhang et al., 2013) or as a major data source (Meza et al., 2015). There are several possible explanations of this fact. First, TEC represents an integral electron content along the related ray path between a ground-based station and GPS satellite, i.e. electron content within an altitudinal range of 0–20,200 km. In case of weak WSA effects in the F2 region peak density, integral nature of the TEC observations could diminish the WSA effect in TEC. Second, the GIMs TEC, produced by several IGS (International GNSS Service) centres, are also based on the ground-based receivers, which are relatively sparse in the Southern Hemisphere, dominated by oceans (e.g., Mannucci et al., 1998). Another important issue of the GIMs TEC is that they are constructed based on a rather limited set of stations (each IGS center specifies their own set of input stations and processes ~150–250 stations worldwide to generate own GIMs product), so only a few IGS stations from the South America continent (most close to the WSA region) are usually processed during generation of the GIMs TEC, for other neighbour grid points the data are interpolated. So, manifestation of the WSA feature in the GIMs TEC depends on the strength of the WSA effects and number of GPS stations used for this GIM generation. Deployment of denser regional GNSS networks close to the WSA region can provide a new opportunity to test the WSA manifestation in the ground-based TEC observations. On the other side, a lot of modern LEO satellites are equipped with a dual-frequency GPS receiver and a zenithlooking antenna that can be used not only for precise orbit determination (POD) purposes but can also provide new data on the plasma density distribution at the topside ionosphere and plasmasphere (Heise et al., 2002; Zakharenkova and Cherniak, 2015). The aim of this paper is to gain new knowledge about physical mechanisms of the WSA development in a framework of its spatial and altitudinal extent by involving the new space-borne GPS TEC observations and its combination with dense networks of the ground-based GPS measurements. Our study is focused only on the WSA phenomenon, i.e. the Southern Hemisphere manifestation of the MSNA. 2. Database 2.1. Ground-based GPS data In this study we consider the longitudinal sector of South America. Currently several networks of permanent ground-based stations provide GNSS measurements with an open access. A lot of stations contribute measurements to the International GNSS Service (IGS) and the University NAVSTAR Consortium (UNAVCO) networks and these services have long-term databases since the years 2000s. Rather recently the national networks for continuous GNSS monitoring were deployed in Brazil and Argentina - the Brazilian Network for Continuous Monitoring (RBMC – Rede Brasileira de Monitoramento Continuo dos Sistemas GNSS) and Red Argentina de Monitoreo Satelital Continuo (RAMSAC CORS). It allows us to increase significantly the GNSS data coverage over the southern part of this continent. Fig. 1 presents the geographical location of the available 106

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Fig. 1. Location of the GNSS stations belonging to the IGS, UNAVCO, RBMC and RAMSAC networks in the South America region. Signs 1,2,3 indicate location of LPGS, FALK and PAL2 stations correspondingly. Geomagnetic equator is shown by the black solid line.

2.2. Space-borne GPS data

GNSS stations over this region. We made use of raw GPS and GLONASS measurements provided by the ground-based networks of GNSS receivers. A significant advantage of the GLONASS, as compared to the GPS, is that the GLONASS has an orbit inclination of ~65
, that is ten degree higher than the GPS orbit inclination. This feature is important for high latitude regions, where a multisystem GNSS receiver can track the GLONASS navigation signals for much longer time and with higher elevation angles than GPS ones. Sampling interval of the raw data was 30 s. The well-known algorithms of TEC estimation from the frequency-differenced GPS measurements were used from Blewitt (1990); Hofmann-Wellenhof (2001). Firstly, the slant TEC (sTEC), defined as the line integral of the electron density from a GPS satellite to a receiver, is derived from the differential P-code pseudorange and the carrier phase. Further, there are applied algorithms for detection and correction of cycle slips and loss-of-lock, remove of outliers, and resolving of the phase ambiguity. The sTEC is converted to the vertical TEC (hereafter TEC) using a mapping function and an approximation of the ionosphere as a spherical thin shell at 350 km altitude. The detailed description of the TEC retrieval technique can be found in Zakharenkova et al. (2016a). To obtain diurnal TEC variation over a ground-based station from the sTEC dataset, we use an algorithm of trigonometrical polynomial expansion of degree 6. Unknown satellites and receiver differential code biases (DCBs) are derived during solution of the overdetermined system of polynomial expansion equations with use of least-squares fitting technique and singular value decomposition. For the ground-based GNSS data we select the elevation angle cut-off as 30
to avoid the multi-path effect. As a result we derived TEC values along all visible satellites with 30 s sampling rate and diurnal TEC variation averaged with 10 min interval. TEC values are presented in TEC units, 1 TECU ¼ 1016 el/m2.

To study the WSA signatures in the topside ionosphere we make use of the GPS measurements onboard Swarm satellites. The Swarm mission was launched on 22 November 2013 and it consists of three identical satellites – Swarm Alpha (A), Bravo (B) and Charlie (C) (named hereafter SWA, SWB and SWC respectively) - two of them (SWA&SWC) fly in a tandem separated by 1
–1.4
in longitude at an orbit altitude of 460 km while the third satellite has an orbit altitude of 510 km. The nominal orbit inclination is 87.5
. Each satellite has a 8-channel dual frequency GPS receiver onboard. Data of the upward-looking antenna of a spaceborne GPS receiver can be used to estimate the topside TEC between a LEO satellite and a GPS satellite. It is necessary to mention that satellites in the Swarm constellation are in a non-sun-synchronous orbit, so orbital LT is continuously changes. It is an important issue for the WSA studies, when we require satellite measurements perfectly matched with: (1) geographical region, (2) month of the WSA maximal manifestation, and (3) specific LT coverage. Aside from that, the upper satellite (SWB) orbit slowly separates in local time from the tandem satellites (SWA&SWC). For season of the September 2014 equinox, the LT separation between satellites was about 40–60 min, whereas it reached 3 h at the December 2015 solstice (Zakharenkova et al., 2016b). Analysis of the Swarm database for 2.5 years reveals that the most fortunate satellite configuration for the WSA observation was period of the December 2014 solstice (December 2014–January 2015), when satellites provided data for adjacent LT sectors (LT separation was 1.2–1.4 h) with proper coverage of day-time and nighttime LT sectors. For December 2015 solstice the Swarm configuration covered early morning-evening LT sectors. So, GPS measurements onboard the Swarm constellation for the 107

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

stations located close to the WSA region. As many of the southern GNSS stations in the Argentina RAMSAC network were deployed and/or provide measurements since the years 2013–2014, we select three separated stations with better data coverage during the years 2011–2015 from the UNAVCO service. These three GPS stations located at different latitudes with ~15
separation (see Fig. 1) have following geographic/geomagnetic coordinates: LPGS (34.91
S; 57.93
W; 24.97
S GLAT; 9.92
GLON), FALK (51.69
S; 57.87
W; 39.29
S GLAT, 10.79
GLON) and PAL2 (64.77
S; 64.05
W; 50.47
GLAT, 9.49
GLON). For each station we derive diurnal TEC values for each day of five years from December 2010 till January 2016. Further, from the initial TEC dataset with 10 min resolution we calculated the 27-day sliding window medians that are the median value for the sets {TEC(LT, DOY13),…., TEC(LT, DOYþ13)}, where LT is Local Time, DOY is day of year (e.g. Ratovsky et al., 2014; Ratovskiy et al., 2015). We assume that the 27-day sliding window medians represent variability associated with diurnal, seasonal and long-term solar activity variations since the influence of short-term variations (shorter than 27 days) are suppressed. We analyzed both absolute and noon-normalized diurnal-seasonal variations of TEC. The noon-normalized variations were calculated as TEC(LT, DOY)/TEC(Noon, DOY), where the noontime TEC was considered as an average TEC value for 12.0–14.0 LT period. These noon-normalized TEC variations can demonstrate precisely a level of the diurnal anomaly development, i.e. how TEC at a given LT is larger than that of local noon. This simple method was previously used by Mamrukov (1971) to reveal the Yakutsk Anomaly occurrence. Recently Klimenko et al. (2015) used

December 2014 solstice were processed and analyzed in this study. We have used the algorithms of calculation and calibration of the GPS TEC modified for a case of a moving space-based GPS receiver. The differential code bias for a LEO receiver was estimated by an algorithm generally used in the UCAR/CDAAC for LEO GPS data processing and described in details by Yue et al. (2011). Slant TEC values were scaled to estimate vertical TEC using a geometric factor derived by assuming the plasma occupies a spherical thin shell, the shell altitude was selected as 550 km above the Earth's surface for Swarm satellites. To avoid the multipath effect the elevation angle cut-off was selected as 40
. The more detailed description of the vertical TEC determination from LEO GPS measurements was described previously in Yue et al. (2011); Zakharenkova and Astafyeva (2015); Zakharenkova and Cherniak (2015). 2.3. Satellite-based electron density measurements Complementary to the ground-based and space-borne GPS measurements we use the in situ electron density provided by the Langmuir Probe instrument onboard Swarm satellites and COSMIC RO electron density profiles accumulated for a period of December 2014–January 2015. 3. Results 3.1. WSA signatures in the long-term series of the ground-based TEC Firstly, we consider variability of the ground-based TEC for GNSS

Fig. 2. Diurnal-seasonal variations of the median TEC for LPGS, FALK and PAL2 stations for the years 2011–2015. The black line presents sunrise and sunset terminator line at 100 km altitude. White area indicates total absence of data. 108

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

TEC enhancement varied in the years 2011–2015 and correlated with solar activity level of the current 24th solar cycle. Fig. 3 shows the noon-normalized TEC values for considered stations. The color coding depicts an intensity of the diurnal TEC anomaly: 1) yellow color (ratio below 0.6) indicates TEC values quite smaller than the noontime ones, which should be a normal behaviour of the diurnal TEC variation for a most part of a day; 2) red color (ratio close to 1.0) indicates TEC values that are equal or close to the noontime TEC values; 3) dark red and blue colours (ratio above 1.5 and 2.0 correspondingly) depict that diurnal TEC variability has an anomalous behaviour with TEC values much larger than that of local noon. Results with the noon-normalized TEC values for the low-latitude LPGS station (Fig. 3a) reveal clearly signatures of a normal behaviour in the diurnal TEC variability: for all considered years and all seasons the peak TEC values were observed during sunlit time and close to local noon (10–16 LT). For the FALK station (Fig. 3b) normal TEC behaviour with a single peak around noon was observed only during local winter season (May–September), whereas during local summer, especially on December–January, maximal values were moved to the pre-noon sector (06–12 LT). Diurnal-seasonal variations of the noon-normalized TEC values for the Antarctic PAL2 station show the pronounced WSA effects in December–January, where amplitude of the nighttime (20-08 LT) TEC enhancement can be as much as two-three times the noontime TEC values (dark blue color on Fig. 3c). The maximal WSA effects coincide in time with the polar day, when ionization should be produced throughout all 24 h of a day (see terminator line on Fig. 3c). These specific WSA signatures were observed for all

the similar approach and positive difference between nighttime and noontime NmF2 values to determine WSA, Yakutsk Anomaly and MSNA. Fig. 2 presents variability of the median TEC values at these three stations during the years 2011–2015. Position of the solar terminator at 100 km altitude is shown by black line. For all station the maximal TEC values are clearly observed from October to April, i.e. during local summer. As expected, the highest TEC values (above 50 TECU) are found at LPGS station, located closely to the southern crest of the equatorial ionization anomaly. Median TEC values over the low-latitude station LPGS (Fig. 2a) demonstrate normal behaviour in the diurnal TEC variability with a clear TEC peak close to local noon (12–15 LT). Results for the FALK station (Fig. 2b) show that the TEC peak in diurnal variation was observed around noon in local spring and autumn seasons, but during local summer the diurnal peak TEC values were moved to the prenoon sector (8–12 LT). Results for the most southern station PAL2, located in Antarctica, revealed quite different features. Here, normal TEC behaviour with daily maximum near noon was observed in local winter and equinoxes (Fig. 2c). During local summer an inverse situation to a normal one was observed: minimal TEC values were registered around local noon, whereas maximal TEC values were found in local nighttime (18 LT - 08 LT). For instance, in December 2013 the daily minimum of 19 TECU was observed at 12–15 LT and daily maximum of 50 TECU was registered at 03–05 LT. It is important to note, that these signatures of abnormal diurnal behaviour corresponded to that part of local summer, when in fact the ionosphere over this station was sunlit for a whole day, i.e. 00–24 LT (see terminator line on Fig. 2c). Magnitude of the nighttime

Fig. 3. The same as Fig. 2 but for the noon-normalized TEC variations (ratio of TEC to the noontime TEC). 109

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

no data were marked as blank ones. The resulting global TEC maps correspond to the topside electron content in an altitudinal range of ~500–20,000 km, as satellite orbit altitude was ~460 km and ~510 km for SWA and SWB correspondingly. Fig. 5 presents global maps of the Swarm-derived topside TEC and noon-normalized TEC for different moments of LT at the December 2014 solstice. For conditions of 13 LT the global distribution of the topside TEC derived from SWA GPS measurements (left panel of Fig. 5a) depicts the maximal TEC values (above 30–35 TECU) in equatorial and low-latitude region with some extension of high TEC values to the Southern (summer) hemisphere relative to the geomagnetic equator. We can note also another interesting feature of the topside TEC distribution at 13 LT. There is clearly seen a decrease of electron density within the midlatitude region (40
S–70
S) of South America-Antarctica with an extension to the southeastern Pacific and southwestern Atlantic regions, mainly within the longitude range of 170
W–0
W. It is a signature of socalled “daytime WSA” which is characterized as an unusually large electron density depletion that reaches its best-developed form at around 14 LT (e.g. Horvath, 2006). The noon-normalized results for 13 LT (right panel of Fig. 5a) obviously demonstrate that all ratio values are close to 1, i.e. to the noontime TEC values which are averages of 12.0–14.0 LT period. Comparison with the subsequent LT maps of the topside TEC distribution during 16–18 LT (Fig. 5b–d) indicates a gradual TEC decrease in the equatorial region. Aside of that, a noticeable TEC enhancement is registered in the whole southern Pacific and a well-defined zone with the TEC increase of 1.5–2.0 times larger values than that of the noontime starts to be recognized in the southern Pacific region with a longitudinal extent of 160
W–60
W (right panel of Fig. 5c and d). Next set of the global maps (Fig. 5e–l) corresponds to the nighttime LT conditions. We can note here a disappearance of the equatorial ionization anomaly around the geomagnetic equator and development of the distinguished signature of the nightitme WSA enhancements in the topside TEC at the background of low nighttime electron density. It is most important to note that these WSA signatures are registered in the topside TEC values, above 500 km altitude. After 23 LT the WSA enhancement in the topside TEC forms near the Antarctic Peninsula and extends further across the South Pacific, albeit at lower latitudes towards New Zealand and Indonesia. This enhancement appears within a longitudinal range of 120
W–60
W (Fig. 5e), then it gradually increases in size and magnitude with time after midnight. We observe the maximal values of the WSA enhancements in the topside TEC at 03 LT (Fig. 5i) and they were about 20 TECU, which contributed up to 50–60% to the co-located ground-based TEC. WSA signatures in the topside ionosphere start to fade away in the morning after 05–06 LT (Fig. 5k and l). Analysis of the noon-normalized TEC values further indicates an occurrence of the clearly distinguished WSA zone in the southern Pacific and Atlantic region. Development of the zone with the TEC enhancement by a factor of 2 and more above the noontime level (dark blue color on graphs) can be clearly seen in post-midnight hours. The maximal TEC enhancement by a factor of 3–4 above the noontime level was registered close to 60
W longitude during 03–04 LT. We want to emphasize, that this obvious signature of the WSA was observed in the topside TEC by both Swarm satellites separated in altitude and longitude. This effect was consistent in time and space and it was detected in a relatively great dataset of the raw GPS measurements, as the Swarm GPS receiver can track simultaneously up to 8 different GPS satellites in a great spatial volume and it is able to observe the WSA effect ahead/behind/aside LEO position and for much longer time than in situ cross-section by Langmuir Probe instrument or GPS RO event or altimeter-based TEC measurements. The zone with the significant WSA effect in the topside TEC (TEC exceed by a factor of 2 and more above the noontime level) had a rather large spatial extent and it covered a wide area of the southeastern Pacific and southwestern Atlantic Oceans. It was located around 150
W-20
W and between 40
S and 70
S from late evening till early morning time. So,

considered years (solar activity levels) with slightly smaller ratio values during low solar activity periods, in particular on January 2011 and December 2015 where F10.7 ≈ 100 s.f.u. Analysis of the ground-based TEC variability over three separated GNSS stations along ~60
W longitude revealed a strong north-south difference in the diurnal TEC behaviour with an appearance of the most pronounced signatures of the WSA at the southern station PAL2, located at the Antarctic Peninsula. In order to estimate spatial extent and magnitude of the WSA signature seen in the ground-based TEC data we investigated in detail data provided by all available GNSS stations in South America for the December 2014 solstice. 3.2. WSA signatures observed in the high-resolution regional TEC maps GPS and GLONASS observations of ~280 permanent stations were used to construct the regional TEC maps with high spatial resolution. Regional GNSS networks in Brazil and Argentina provide a major contribution to this dataset (see Fig. 1). We process GPS&GLONASS data for December 2014–January 2015 to obtain the monthly averaged TEC maps over the South America region. TEC values along all visible satellite passes were binned and averaged in cells of 1
 1
resolution in geographical latitude and longitude. No interpolation were used here, empty cells with no data were marked as blank ones. This approach is rather close to the MIT TEC maps generation (Rideout and Coster, 2006), but we used both GPS and GLONASS observations (instead of GPS only in the MIT processing) and higher cut-off elevation angles (30
instead of 7
). In the same way as for the diurnal TEC variation over a particular GNSS station, the regional TEC maps were constructed in the LT domain (hourly maps) and they were also normalized to the noontime TEC values, which were calculated by averaging of all TEC values for 12.0–14.0 LT. Fig. 4 presents regional maps of the absolute and noon-normalized TEC values over the South America region for different moments of LT. The hourly maps of the absolute TEC values (left graph of each panel) illustrate spatial plasma density distribution with time. It is clearly seen the development of the equatorial ionization anomaly close to the geomagnetic equator with the pronounced maximum in the afternoon hours (13–17 LT). For the low latitude region the minimal TEC values were observed in the post-midnight hours. The contrast behaviour of the diurnal TEC variation is found over the southern tip of South America and the Antarctic Peninsula – minimal TEC values were registered after the noon hours and maximal ones - at the early morning LT hours. Right graphs of each panel on Fig. 4 with the noon-normalized TEC values had revealed this inverse signature of the diurnal TEC behaviour over the Weddell Sea region in more detail. The WSA effects of the TEC exceed the noontime values were registered starting from 17 to 19 LT, they reached its maximum at 01–05 LT and started to disappear after 09 LT. Maximal ratio values of 2.5–3.5 were registered at 03–04 LT. This effect was mainly localized within the geographical region of 55
S-70
S latitude and 80
W-40
W longitude. However, lack of the ground-based measurements does not allow us to recognize a spatial extent of the WSA. 3.3. WSA signatures observed in global topside TEC maps Further, to get a global overview on the WSA development we constructed the global maps of the topside TEC values derived from the GPS measurements onboard Swarm satellites. We processed 50 days of observations (±25 days from the December 2014 solstice) to obtain hourly maps of the monthly averaged TEC values. The Swarm orbit configuration during this period was very fortunate to cover afternoon and postmidnight LT sectors. The TEC values along all satellite links (SwarmGPS) were binned and averaged in cells of 5
 15
resolution in geographical latitude and longitude. The global TEC maps were constructed as hourly maps in the LT domain for available LT hours and they were also normalized to the noontime TEC values (data average for 12.0–14.0 LT period). No interpolation were used here, empty cells with 110

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Fig. 4. Regional maps of TEC and noon-normalized TEC (ratio to the noon values) for different moments of LT in December 2014–January 2015. Geomagnetic equator is shown by the black solid line.

111

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Fig. 5. Global maps of the Swarm-derived topside TEC and noon-normalized TEC (ratio to the noon values) for different moments of LT in December 2014–January 2015. The WSA is clearly identified around 17–06 LT. The solid black lines indicate the geomagnetic equator. SWA and SWB means Swarm A and Swarm B satellites correspondingly.

extent of the WSA effect covered mainly the region of 50
S–80
S. The prominent WSA effect with ratio of 2–3 is clearly seen at longitude of 100
W during the whole period 23-05 LT (Fig. 6b). Further displacement in longitude to 140
W (Fig. 6a) depicts a decrease of the WSA effect to ratio values of ~1.5 within a slightly shifted latitudinal range of 40
S–70
S. Displacement in longitude in the east direction to 5
E (Fig. 6d) shows also rather weak WSA effect with the ratio values of ~1.6–1.8 observed at 03 LT and 05 LT, whereas in previous LT the ratio values were close or even lesser than the noontime TEC values. Which means that extension of the pronounced WSA effects to the African longitudinal sector of southern Atlantic occurred at post-midnight LT hours and it coincided with a maximal development of the WSA.

the WSA is a large-scale ionospheric feature that occurs over oceans and there is no ground-based instruments or networks of stations which can be able to observe its whole structure from the ground level. In order to demonstrate in more detail the WSA signatures in the topside TEC with geographical latitude we constructed slices of the considered global maps. Fig. 6 presents the meridional slices of the noonnormalized TEC variability at 140
W, 100
W, 60
W and 5
E geographic longitudes. The main feature that can be found here is the significant difference in the WSA effect manifestation depending on longitude. The most pronounced WSA effects were found at longitude of 60
W (Fig. 6c), where at 23 LT the WSA enhancement already reached value of 2 (TEC exceed by 2 times (or by 100%) the noon-time values), then it further increased and reached the highest values of 4 at 03 LT. The latitudinal

112

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Fig. 6. Appearance of the WSA in the noon-normalized Swarm TEC at meridional slices evaluated along longitudes (a) 140
W, (b) 100
W, (c) 60
W, and (d) 5
E. Color lines correspond to different LT hours. Dotted line shows level of equality to the noontime values (ratio is equal to 1). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

LT, 02–04 LT and 04–06 LT bands. For these intervals both SWA and SWB in situ measurements are available (last two graphs from the bottom). The COSMIC-derived global maps of Ne at an altitudinal range of 300–500 km demonstrate a further decrease of Ne at equatorial region (compare Figs. 7i, j and 8c, d, j and k) and more precise separation of the Ne enhancement zone in the southern Pacific. This zone extends across the South Pacific towards New Zealand and Indonesia. This extension is clearly observed in both COSMIC and Swarm in situ data in the topside ionosphere. It is important to note, that the absolute values of electron density inside and outside of the WSA are gradually decreased with altitude, but the noon-normalized values, which illustrated the strength of this daily anomaly, remain rather high, which is an evidence of large extension of the WSA anomaly to altitudes much above the F2 layer peak. Results presented in Figs. 7 and 8 indicate that for conditions of the December 2014 solstice the WSA effects were observed for a whole night from 22 LT till 06 LT.

3.4. Altitudinal extent of the WSA signatures observed in global maps In addition to the ground-based and space-borne GPS TEC observations, we also analyzed the ionospheric electron density (Ne) measurements at the different altitudinal regions. For this aim we processed more than 33,000 COSMIC RO profiles for December 2014–January 2015 period. All distorted RO profiles detected during the special quality control procedures were removed from our analysis. Also we should note that number of COSMIC RO profiles is significantly smaller in 2014–2015 as comparing to the beginning of the mission time, so we should not expect a very good data coverage. We constructed global maps of the Ne distribution by averaging of the COSMIC-derived Ne data to several altitudinal ranges: 100–200 km, 200–300 km, 300–400 km, 400–500 km and 500–600 km. These maps were generated in the LT domain with 2 h resolution, actual data were binned in cells of 5
 15
resolution in geographical latitude and longitude. In the similar way we constructed hourly global LT maps of the Swarm-derived Ne distribution at two fixed altitudes of 460 and 510 km. Figs. 7 and 8 present the global maps of absolute Ne and noonnormalized Ne data for the different altitudinal ranges during 22 LT 06 LT. Analysis of the global maps of the COSMIC-derived Ne for 22-00 LT band (left graphs of Fig. 7a–e) reveals that distinct enhancements of electron density over the Pacific region are registered at 300–600 km altitudes, with the highest Ne values at two altitudinal ranges: 300–400 km and 400–500 km. These enhancements seem to be connected with the high Ne values at equatorial latitudes, normally observed at late evening time. Independent in situ measurements provided by the SWA satellite at ~460 km altitude (Fig. 7f) revealed a rather good agreement with the COSMIC Ne distribution for 400–500 km range (Fig. 7d) in the occurrence of the similar zone of the Ne enhancement over the southern Pacific region and connected to the equatorial Ne enhancement. The noon-normalized Ne values detect the presence of the WSA signature at different altitudinal ranges. It is worth to note that location of this zone and its magnitude (factor of 2 and more in its center) coincided perfectly with the WSA signatures observed in the Swarmderived topside TEC (Fig. 6). The similar structure is clearly seen in the SWA data (right panel of Fig. 7f). So, additionally two independent instruments – COSMIC RO and Swarm Langmuir Probe – confirmed the same location of the most pronounced WSA effect over the Southern Pacific. We can note that COSMIC-derived noon-normalized Ne values demonstrate the occurrence of the zone of the maximal enhancements starting from 300 km, but the strongest anomaly values (Ne exceed over the noontime values by a factor of 2 and more) are observed in the topside ionosphere at 400–500 km and 500–600 km. Results of the SWA satellite also confirms that the WSA effect, as a more pronounced day/ night difference, had a higher magnitude at 460 km than at 300–400 km (right panels of Fig. 7f and c). Next figures illustrate an evolution of the WSA effects during 00–02

4. Discussion and conclusions In the given paper we reported an occurrence of the WSA signatures in the ground-based and space-borne GPS TEC observations. Firstly, we try to use new opportunities offered by the recently deployed and rather dense GNSS networks in South America. Nowadays, from all groundbased instruments only GPS/GNSS can provide permanent and continuous measurements of the ionospheric density with high temporal resolution and over a large surface. We analyzed the long-term series of the TEC observations derived from the GNSS stations separated in latitude along ~60
W longitude: the FALK station at the Falkland Islands, where the only one in the WSA region ionosonde (Port Stanley) operates now, and the PAL2 station at the Antarctic Peninsula. Results indicated a strong north-south difference in the diurnal TEC behaviour with the appearance of the much more pronounced signatures of the WSA at the most southern PAL2 station. Here, the nighttime TEC enhancement exceeded by a factor of 2–3 the noontime values and such maximal WSA effects coincided in time with the polar day, when ionization should be produced all 24 h. The WSA effects at the FALK station were much more weaker in magnitude. Between these two stations there is the Weddell Sea with no more ground-based instruments. Further, we analyze all available GNSS stations in the South America region. We process not only standard GPS measurements, but also take an opportunity provided by new GNSS receivers which are able to measure signals of several GNSS. Due to its higher orbit inclination, the GLONASS system provides a significant avantage at high latitudes, where a multisystem GNSS receiver can track the GLONASS navigation signals for much longer time and with higher elevation angles than GPS ones. The regional TEC maps with the high spatio-temporal resolution revealed that the WSA effects of the TEC exceed the noontime values are registered starting from 17 to 19 LT, it reached its maximum at 01–05 LT and started 113

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Fig. 7. Global maps of absolute Ne and noon-normalized Ne (ratio to the noon values) data derived for different altitudes from COSMIC RO and Swarm in situ measurements for 22-00 LT (left panel) and 00–02 LT (right panel) in December 2014–January 2015. The WSA is clearly identified in the noon-normalized Ne values for all altitudes above 300 km. Empty cells are marked by while color.

114

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

Fig. 8. The same as Fig. 7 but for 02–04 LT and 04–06 LT.

towards the oceans, but actual ground-based instruments’ distribution does not able to cover this region neither today or in the near future. Analysis of the ground-based TEC data for both long-term TEC data series and for December 2014–January 2015 period reveals that: 1) the WSA effects are clearly recognized in the ground-based TEC even during

to disappear after 09 LT. The maximal ratio values of 2.5–3.5 were registered at 03–04 LT. This effect was mainly localized within the geographical region of 55
S-70
S latitude and 80
W-40
W longitude between the southern tip of the continent and the Atlantic Peninsula. Obviously, this region with the WSA enhancement should extent further 115

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

radiation absence. The obtained results of the nighttime enhancements in the topside TEC can support the plasmaspheric contribution to the WSA formation. Lin et al. (2009) used the COSMIC RO data and demonstrated the WSA connection with the southern crest of the equatorial anomaly, noted the downward plasma diffusion from the topside ionosphere or plasmasphere that may also contribute to its maintenance. Our results derived from Swarm measurements also confirmed that the WSA effects in the topside ionosphere could be connected with the electron density enhancements in the equatorial anomaly region, especially during premidnight hours. Horvath (2006) indicates that the Weddell Sea Anomaly itself exhibits both a longitudinal and a latitudinal variation. Its development over a wider area of the Weddell Sea is due to the photoionization taking place throughout the summer nights as the ionosphere remains illuminated between November and February because of the tilted dipole of the Earth (Sojka et al., 1985). Our results derived from the ground-based GPS/ GNSS data are in agreement with such conclusions as the most pronounced effect in TEC was observed at the GPS/GNSS stations inside the polar circle region. In terms of thermospheric tidal effects as a possible source of the neutral wind driven WSNA, both the WSA region and WSNA in the Northern Hemisphere were intensively studied by satellite observations (Lin et al., 2010; Xiong and Lühr, 2014) and first principle model simulations (Chen et al., 2013; Chang et al., 2015). The long-term series of in situ measurements onboard the CHAMP mission (altitude of 300–400 km) and the unique K Band Ranging (KBR) measurements between two GRACE A and B satellites (altitude of 400–500 km) were actively involved into investigation of the global signatures of the MSNA manifestation (Lin et al., 2010; Xiong and Lühr, 2014). It is also important to note, that Liu et al. (2010) explained the MSNA formation by a combination of the neutral wind effect in the geomagnetic frame, the photoionization, and the thermal contraction of the ionosphere at sunset and they concluded that the WSA/MSNA signatures should be more prominent at altitudes near and above hmF2. So, our results demonstrate a general agreement with previous observational research based on COSMIC and TOPEX/Jason measurements and simulation research, particularly of Chen et al. (2011), but using new multi-instrumental data sources onboard the Swarm mission we revealed the occurrence of the strong WSA effects in the topside ionosphere (above ~500 km altitude). In conclusion, we note that the number of the current LEO missions applicable to ionosphere research is rather small, besides of that: (1) the C/NOFS satellite had re-entry on November 28, 2015; (2) degrading performance of the COSMIC mission (~150 profiles daily as of April 2017); (3) launch of the follow-on mission COSMIC-2 is expected in September 2017 but it will be a low-inclination segment of satellites. So, ones of the most promising missions to study the WSA signatures from space are still altimeter missions Jason-2/Jason-3 and the Swarm constellation, while only the latter provides currently the in situ measurements of the ionospheric plasma density for altitudes below 550 km. Due to the orbit inclination of 66
and specifics of the radar altimeter measurements missions like TOPEX or Jason have several limitations: (1) the ionospheric TEC data can be derived over oceans only; (2) geographic latitude coverage is between 66
S and 66
N; (3) significant change of LT along a single pass. From this point of view, the multi-satellite Swarm mission with a polar circular orbit can provide measurements with real global coverage and with a fixed LT along a pass (the orbital LT change is about 6–7 min per day). Apart of that, the Swarm GPS receiver can track simultaneously up to 8 different GPS satellites with 1 Hz rate in a great spatial volume and it is able to observe the WSA effect ahead/behind/ aside LEO position and for much longer time than in situ cross-section by the Langmuir Probe instrument or COSMIC GPS RO event or altimeterbased TEC measurements, so global data coverage by the Swarm GPS measurements would be much better after comparing these techniques within the same period of data accumulation. In the present paper, we

moderate solar activity; 2) the WSA effects are more pronounced for the GPS/GNSS stations located inside a polar circle where 24 h ionization could maintain the higher density values; 3) the WSA effects as the pronounced TEC enhancements can be as large as a factor of 2 or more over the noontime TEC values, they are observed during 20-08 LT in December–January; 4) relative strength of the WSA effects (ratio to the noontime TEC) at the Antarctic GPS station has a tendency to increase with an increase of the solar activity level. Further, we examined the WSA occurrence in the topside ionosphere by using the GPS measurements onboard three Swarm satellites. The global maps of the topside TEC indicated that the zone with the significant WSA effect in the topside TEC (TEC increase ~2–4 times the noontime level) had a large spatial extent over southern Pacific and Atlantic Ocean. It was observed around 150
W-20
W and between 40
S and 70
S during 23 LT - 06 LT. For a first time, the WSA signatures were shown in the topside TEC data derived from the GPS measurements onboard the Swarm constellation. Additionally, we present results derived independently by two other instruments - COSMIC radio occultation electron density profiles and in situ measurements by Langmuir Probe instrument onboard Swarm satellites (Figs. 7–8). These results confirm: (1) the same location of the WSA zone as revealed in the Swarm TEC; (2) the most-pronounced WSA effect, as a maximal electron density exceed over the noontime values, corresponded to altitudes above 400–500 km. It is important to note, that these WSA signatures were found in the TEC, which is an integral of electron density through the bottomside, topside ionosphere and plasmasphere, but we could not split TEC into these parts. Bilitza (2009) reported that the bottomside ionosphere contributes only about 20% of the ground-based TEC and the dominant contribution comes from the topside (>80%). Estimations made with the topside GPS TEC derived from GOCE satellite and ground-based GIMs TEC demonstrate that for conditions of the December 2013 solstice and 05–06 LT the region above 250 km contributes about 80% to the groundbased TEC in the WSA zone of the Southern Pacific (Zakharenkova and Cherniak, 2015). That is why if the WSA is well developed at the F2 peak and above, it should be recognized in the ground-based TEC observations too. By taking advantages of a GPS receiver onboard LEO satellite we can separate altitudinal region and extract TEC above a fixed LEO orbit altitude. The Swarm-derived results presented in this paper indicate that the distinct WSA enhancements were observed both in the in situ electron density at ~460 km and ~510 km altitude, as well as in the topside TEC above these altitudes during all nighttime hours (23 LT – 06 LT). This is a clear evidence that the WSA was well developed in the topside ionosphere, which should be coincided with an uplift and/or extension of the whole F2 layer to higher altitudes. Today, there is still no commonly accepted view on the mechanisms of the WSA formation. The major physical mechanisms that have been suggested are the neutral winds (Dudeney and Piggott, 1978), an electric field (Burns et al., 2011), the photoionization (Horvath, 2006), and the downward diffusion from the plasmasphere (Burns et al., 2008). Chen et al. (2011) performed SAMI2 model simulations of the WSA and concluded that: (1) the equatorward neutral wind is the critical driver for the WSA formation, since it sustains the ionospheric layer at higher altitude and results in plasma accumulation; (2) the plasmaspheric downward flux provides plasma source after 22 LT which maintains the intensity of the WSA density structure. With a high degree of similarity with simulation results of Chen et al. (2011, Fig. 1, p.2) our results (shown in Figs. 7–8) demonstrate an enhancement of ionospheric electron density at 300–400 km altitudes during 20-04 LT. Moreover, we should emphasize that noticeable enhancement of ionospheric electron density occurred also at higher ionospheric altitudes (400–600 km) during all nighttime hours, which was found in both in situ and topside TEC data. It is important to note that the topside TEC (above 500 km) increased gradually during nighttime hours and reached its maximal values at ~03 LT in the conditions of solar 116

I. Zakharenkova et al.

Journal of Atmospheric and Solar-Terrestrial Physics 161 (2017) 105–117

demonstrated that the space-borne GPS TEC measurements can be effectively used for investigation of the WSA signatures in the topside ionosphere and current non-ionospheric LEO satellite missions with a GPS receiver onboard like GRACE, TerraSAR-X, MetOP, SAC-C, Sentinel, Swarm, placed at different orbit altitudes, can provide new data and new clue to investigation of the large-scale ionospheric plasma density anomalies at the topside ionosphere and plasmasphere.

measurements. J. Geophys. Res. 114, A04307. http://dx.doi.org/10.1029/ 2008JA013801. Klimenko, M.V., Klimenko, V.V., Karpachev, A.T., Ratovsky, K.G., Stepanov, A.E., 2015. Spatial features of Weddell Sea and Yakutsk Anomalies in foF 2 diurnal variations during high solar activity periods: interkosmos-19 satellite and ground-based ionosonde observations, IRI reproduction and GSM TIP model simulation. Adv. Sp. Res. 55, 2020–2032. http://dx.doi.org/10.1016/j.asr.2014.12.032. Lin, C.H., Liu, J.Y., Cheng, C.Z., Chen, C.H., Liu, C.H., Wang, W., Burns, A.G., Lei, J., 2009. Three-dimensional ionospheric electron density structure of the Weddell Sea Anomaly. J. Geophys. Res. 114, A02312. http://dx.doi.org/10.1029/2008JA013455. Lin, C.H., Liu, C.H., Liu, J.Y., Chen, C.H., Burns, A.G., Wang, W., 2010. Midlatitude summer nighttime anomaly of the ionospheric electron density observed by FORMOSAT-3/COSMIC. J. Geophys. Res. 115, A03308. http://dx.doi.org/10.1029/ 2009JA014084. Liu, H., Thampi, S.V., Yamamoto, M., 2010. Phase reversal of the diurnal cycle in the midlatitude ionosphere. J. Geophys. Res. 115, A01305. http://dx.doi.org/10.1029/ 2009ja014689. Liu, J.Y., Chang, F.Y., Oyama, K.-I., Kakinami, Y., Yeh, H.C., Yeh, T.L., Jiang, S.B., Parrot, M., 2014. Topside ionospheric electron temperature and density along the Weddell Sea latitude. J. Geophys. Res. Sp. Phys. 120, 609–614. http://dx.doi.org/ 10.1002/2014JA020227. Luan, X., Wang, W., Burns, A., Solomon, S.C., Lei, J., 2008. Midlatitude nighttime enhancement in F region electron density from global COSMIC measurements under solar minimum winter condition. J. Geophys. Res. Sp. Phys. 113, 1–13. http:// dx.doi.org/10.1029/2008JA013063. Mamrukov, A.P., 1971. Evening anomalous enhancement of ionization in F region. Geomagn. Aeron. 21 (6), 984–988. Mannucci, A.J., Wilson, B.D., Yuan, D.N., Ho, C.H., Lindqwister, U.J., Runge, T.F., 1998. A global mapping technique for GPS-dericed ionospheric total electron content measurements. Radio Sci. 33, 565–582. http://dx.doi.org/10.1029/97RS02707. Meza, A., Natali, M.P., Fernandez, L.I., 2015. PCA analysis of the nighttime anomaly in far-from-geomagnetic pole regions from VTEC GNSS data. Earth Planets Sp. 67, 106. http://dx.doi.org/10.1186/s40623-015-0281-4. Mosert, M., McKinnell, L.A., Gende, M., Brunini, C., Araujo, J., Ezquer, R.G., Cabrera, M., 2011. Variations of foF2 and GPS total electron content over the Antarctic sector. Earth Planets Sp. 63, 327–333. http://dx.doi.org/10.5047/eps.2011.01.006. Penndorf, R., 1965. The average ionospheric conditions over the Antarctic. In: Waynick, A.H. (Ed.), Geomagnetism and Aeronomy, Antarct. Res. Ser., vol. 4. AGU, Washington, D. C, pp. 1–45. Ratovskiy, K., Medvedev, A., Oynats, A., 2015. Similarities and differences between regular variations of F2-layer parameters of polar and midlatitude ionosphere in East Siberian sector. Solar Terr. Phys. 1, 70–79. http://dx.doi.org/10.12737/7832. Ratovsky, K.G., Shi, J.K., Oinats, A.V., Romanova, E.B., 2014. Comparative study of highlatitude, mid-latitude and low-latitude ionosphere on basis of local empirical models. Adv. Sp. Res. 54, 509–516. http://dx.doi.org/10.1016/j.asr.2014.02.019. Rideout, W., Coster, A., 2006. Automated GPS processing for global total electron content data. GPS Solut. 10, 219–228. http://dx.doi.org/10.1007/s10291-006-0029-5. Slominska, E., Blecki, J., Lebreton, J.-P., Parrot, M., Slominski, J., 2014. Seasonal trends of nighttime plasma density enhancements in the topside ionosphere. J. Geophys. Res. Sp. Phys. 119, 6902–6912. http://dx.doi.org/10.1002/2014JA020181. Sojka, J.J., Raitt, W.J., Schunk, R.W., Parish, J.L., 1985. Diurnal variations of the dayside, ionospheric, mid-latitude trough in the southern hemisphere at 800 km: model and measurement comparison. Planet. Space Sci. 33, 1375–1382. Sojka, J.J., Schunk, R.W., Wrenn, G.L., 1988. A comparison of foF2 obtained from a timedependent ionospheric model with Argentine Islands' data for quiet conditions. J. Atmos. Terr. Phys. 50, 1027–1039. http://dx.doi.org/10.1016/0021-9169(88) 90092-X. Thampi, S., Lin, C.H., Liu, H., Yamamoto, M., 2009. First tomographic observations of the midlatitude summer nighttime anomaly (MSNA) over Japan. J. Geophys. Res. 114, A10318. http://dx.doi.org/10.1029/2009JA014439. Wrenn, G.L., Rodger, A.S., Rishbeth, H., 1987. Geomagnetic storms in the Antarctic Fregion. I. Diurnal and seasonal patterns for main phase effects. J. Atmos. Terr. Phys. 49, 901–913. Xiong, C., Lühr, H., 2014. The midlatitude summer night anomaly as observed by CHAMP and GRACE: interpreted as tidal features. J. Geophys. Res. Sp. Phys. 119 http:// dx.doi.org/10.1002/2014JA019959. Yue, X., Schreiner, W.S., Hunt, D.C., Rocken, C., Kuo, Y.-H., 2011. Quantitative evaluation of the low Earth orbit satellite based slant total electron content determination. Sp. Weather 9, S09001. http://dx.doi.org/10.1029/2011SW000687. Zakharenkova, I., Astafyeva, E., 2015. Topside ionospheric irregularities as seen from multi-satellite observations. J. Geophys. Res. Space Phys. 120 (1), 807–824. http:// dx.doi.org/10.1002/2014JA020330. Zakharenkova, I., Cherniak, Iu, 2015. How can GOCE and TerraSAR-X contribute to the topside ionosphere and plasmasphere research? Sp. Weather 13, 271–285. http:// dx.doi.org/10.1002/2015SW001162. Zakharenkova, I., Astafyeva, E., Cherniak, I., 2016a. GPS and GLONASS observations of large-scale traveling ionospheric disturbances during the 2015 St. Patrick's Day storm. J. Geophys. Res. Space Phys. 121 http://dx.doi.org/10.1002/2016JA023332. Zakharenkova, I., Astafyeva, E., Cherniak, Iu, 2016b. GPS and in situ Swarm observations of the equatorial plasma density irregularities in the topside ionosphere. Earth Planets Sp. 68 (1), 1–11. http://dx.doi.org/10.1186/s40623-016-0490-5. Zhang, Y., Paxton, L.J., Kil, H., 2013. Nightside midlatitude ionospheric arcs: TIMED/ GUVI observations. J. Geophys. Res. Sp. Phys. 118, 3584–3591. http://dx.doi.org/ 10.1002/jgra.50327.

Acknowledgements We thank the European Space Agency (ESA) for providing the Swarm data (http://earth.esa.int/swarm) and the UCAR/CDAAC for COSMIC RO electron density profiles (http://cdaac-www.cosmic.ucar.edu/cdaac/ products.html). We acknowledge the use of the raw GPS&GLONASS data provided by IGS (ftp://cddis.gsfc.nasa.gov), UNAVCO (ftp://data-out. unavco.org), RAMSAC CORS of National Geographic Institute of Argentina (www.igm.gov.ar/NuestrasActividades/Geodesia/Ramsac/), and Brazilian network for continuous monitoring (ftp://geoftp.ibge.gov. br/RBMC/). We also thank IGS for providing orbits products. This work was funded by RFBR according to the research project N 16-05-01077a. References Bellchambers, W.H., Piggott, W.R., 1958. Ionospheric measurements made at Halley Bay. Nature 182, 1596–1597. http://dx.doi.org/10.1038/1821596a0. Bilitza, D., 2009. Evaluation of the IRI-2007 model options for the topside electron density. Adv. Space Res. 44 (6), 701–706. http://dx.doi.org/10.1016/ j.asr.2009.04.036. Blewitt, G., 1990. An automatic editing algorithm for GPS data. Geophys. Res. Lett. 17, 199–202. Burns, A.G., Zeng, Z., Wang, W., Lei, J., Solomon, S.C., Richmond, A.D., Killeen, T.L., Kuo, Y., 2008. Behavior of the F2 peak ionosphere over the South Pacific at dusk during quiet summer conditions from COSMIC data. J. Geophys. Res. 113, A12305. http://dx.doi.org/10.1029/2008JA013308. Burns, A.G., Solomon, S.C., Wang, W., Jee, G., Lin, C.H., Rocken, C., Kuo, Y.H., 2011. The summer evening anomaly and conjugate effects. J. Geophys. Res. 116, A01311. http://dx.doi.org/10.1029/2010JA015648. Chen, C.H., Huba, J.D., Saito, A., Lin, C.H., Liu, J.Y., 2011. Theoretical study of the ionospheric Weddell sea anomaly using SAMI2. J. Geophys. Res. 116, A04305. http://dx.doi.org/10.1029/2010JA015573. Chen, C.H., Lin, C.H., Chang, L.C., Huba, J.D., Lin, J.T., Saito, A., Liu, J.Y., 2013. Thermospheric tidal effects on the ionospheric midlatitude summer nighttime anomaly using SAMI3 and TIEGCM. J. Geophys. Res. Space Phys. 118, 3836–3845. http://dx.doi.org/10.1002/jgra.50340. Chang, L.C., Liu, H., Miyoshi, Y., Chen, C.H., Chang, F.Y., Lin, C.H., Liu, J.Y., Sun, Y.Y., 2015. Structure and origins of the Weddell Sea Anomaly from tidal and planetary wave signatures in FORMOSAT-3/COSMIC observations and GAIA GCM simulations. J. Geophys. Res. A Sp. Phys. 120, 1325–1340. http://dx.doi.org/10.1002/ 2014JA020752. Dudeney, J.R., Piggott, W.R., 1978. Antarctic ionospheric research. In: Lanzerotti, L.J., Park, C.G. (Eds.), Upper Atmosphere Research in Antarctica, Ant. Res. Ser, vol. 29. AGU, Washington, D. C, pp. 200–235. Garner, T.W., Gaussiran II, T.L., Tolman, B.W., Harris, R.B., Calfas, R.S., Gallagher, H., 2008. Total electron content measurements in ionospheric physics. Adv. Space Res. 42 (4), 720–726. http://dx.doi.org/10.1016/j.asr.2008.02.025. He, M., Liu, L., Wan, W., Ning, B., Zhao, B., Wen, J., Yue, X., Le, H., 2009. A study of the Weddell sea anomaly observed by FORMOSAT-3/COSMIC. J. Geophys. Res. Sp. Phys. 114, 1–10. http://dx.doi.org/10.1029/2009JA014175. Heise, S., Jakowski, N., Wehrenpfennig, A., Reigber, C., Lu, H., 2002. Sounding of the topside ionosphere/plasmasphere based on GPS measurements from CHAMP: initial results. Geophys. Res. Lett. 29 (14), 1699. http://dx.doi.org/10.1029/ 2002GL014738. Hofmann-Wellenhof, B., 2001. Global Positioning System: Theory and Practice. Springer, New-York. Horvath, I., 2006. A total electron content space weather study of the nighttime Weddell Sea Anomaly of 1996/1997 southern summer with TOPEX/Poseidon radar altimetry. J. Geophys. Res. 111, A12317. http://dx.doi.org/10.1029/2006JA011679. Horvath, I., Essex, E.A., 2003. The Weddell sea anomaly observed with the Topex satellite data. J. Atmos. Solar-Terrestrial Phys. 65, 693–706. http://dx.doi.org/10.1016/ S1364-6826(03)00083-X. Horvath, I., Lovell, B.C., 2009. Investigating the relationships among the South Atlantic Magnetic Anomaly, southern nighttime midlatitude trough, and nighttime Weddell Sea Anomaly during southern summer. J. Geophys. Res. Sp. Phys. 114, 1–18. http:// dx.doi.org/10.1029/2008JA013719. Jakowski, N., 1996. TEC monitoring by using satellite positioning systems. In: Kohl, H., Ruester, R., Schlegel, K. (Eds.), Modern Ionospheric Science. EGS, Berlin, pp. 371–390. Jee, G., Burns, A.G., Kim, Y.H., Wang, W., 2009. Seasonal and solar activity variations of the Weddell Sea Anomaly observed in the TOPEX total electron content

117