- Email: [email protected]
TEC variation during high and low solar activities over South American sector
Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Contents lists available at ScienceDirect
Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp
TEC variation during high and low solar activities over South American sector O.F. Jonah a, E.R. de Paula a, M.T.A.H. Muella b, S.L.G. Dutra a, E.A. Kherani a, P.M.S. Negreti a, Y. Otsuka c a
Instituto Nacional de Pesquisas Espaciais (INPE), Av. de Astronautas, 1758, Jd da Granja, São José dos Campos, São Paulo 12227-010, Brazil Instituto de Pesquisa e Desenvolvimento (IP&D), Universidade do Vale do Paraíba (UNIVAP), Av. Shishima Hifumi, 2911, São José dos Campos, São Paulo 12444-000, Brazil c Nagoya University, 3-13 Honohara, Toyokawa, Aichi 442-8507, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 29 January 2015 Received in revised form 1 October 2015 Accepted 4 October 2015 Available online 9 October 2015
Using dual frequency GPS receivers in the South American sector, the measurement of absolute ionospheric Total Electron Content (TEC) has been estimated applying the Nagoya ionospheric model for both the years of 2009 and 2001, which represent low and high solar activities, respectively. The diurnal, dayto-day, monthly, seasonal, latitudinal and longitudinal variations of TEC were studied for equatorial and low latitude regions of South America. The strength and characteristics of the Equatorial Ionization Anomaly (EIA) were equally analyzed. The analyses reveal the diurnal, seasonal and semidiurnal TEC variation, as well as the nighttime variability during the low and high solar activities. Wavelet power spectra analysis was employed to check the periodicities of the TEC data, F10.7 and zonal and meridional wind velocities measured by Meteor radar at ∼100 km altitude. Periods such as 27, 16, 8–10, 1–5 days were found to be dominant in the zonal and meridional wind velocity corresponding with those of TEC periodicities. Hence, besides the solar radiation, we suggest that there are contributions of tides and planetary waves in spatial and temporal TEC enhancement and variations during the geomagnetic quiet periods of both solar activities. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Total Electron Content (TEC) Diurnal, seasonal, latitudinal and longitudinal variations of TEC TEC periodicity during high and low solar activities Equatorial Ionization anomaly at different seasons
1. Introduction Exploring the ionosphere is of utmost interest due to the numerous complexities associated with this region (Rabiu et al., 2007). Although, over the last century humanity has learned to use the properties of the ionosphere in a tremendous way, there are more to understand about the chemical and physical changes of this region of the Earth’s atmosphere. One of the parameters that can be used to study the ionosphere is the Total Electron Content (TEC). Study of TEC variability over the South American continent is of great interest due to the possibility to investigate the processes responsible for the ionospheric behavior over this region. TEC is significant in helping us to understand the short and long term changes of our upper atmosphere during major phenomena caused by factors like solar activities, geomagnetic storms and meteorological influences (e.g. Forbes et al., 2000; Kane, 2003; Rishbeth and Mendillo, 2001). These changes in the ionosphere affect navigation systems, surveillance systems and modern technologies such as communication systems, since the signal from the satellite to the receiver must pass through the ionized layer (Bagiya et al., 2009), which causes a delay at such signal. A good description of the ionosphere is also needed in order to http://dx.doi.org/10.1016/j.jastp.2015.10.005 1364-6826/& 2015 Elsevier Ltd. All rights reserved.
improve the performance of the ionospheric models (Bilitza, 2000). Many researchers have studied the morphological features of ionospheric electron density and TEC at low and equatorial latitudes (Dabas et al., 1993; Kane, 2003; Batista and Abdu, 2004; Costa et al., 2004; Rama Rao et al., 2006; Bhuyan and Borah, 2007; Abdu et al., 2007; Bagiya et al., 2009, Jonah et al., 2014, DuarteSilva et al., 2015). Dabas et al. (1993) studied the variations in TEC with different solar indices, i.e. EUV, F10.7 solar flux and smoothed sunspot number (SSN) for summer, winter and equinoxes over the Indian sector. They showed that TEC exhibited nonlinear relationship with SSN in general and linear variations with EUV and F10.7 solar flux. Kane (2003) studied day-to-day variability of quiet-time ionosphere using F2-peak electron density data. He observed oscillations of day-to-day variation with peak spacing of 7 days at several locations and indicated that in the absence of solar or geomagnetic effects, planetary waves dominate the dayto-day variability. Batista and Abdu (2004) also studied the ionospheric variability from observations of the F2 layer peak density over the Brazilian low and equatorial latitudes. Costa et al. (2004) investigated diurnal and seasonal variations of TEC during a year of low solar activity (1997) at Presidente Prudente, a station located
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
near the equatorial ionization anomaly crest in Brazil. It was found that TEC was maximum during summer months (from November to February) and minimum during winter months (from May to August), with intermediate values during the equinoctial months, but the springer equinox (September–October) presenting larger TEC values than the autumn equinox (March–April). Bhuyan and Borah (2007) reported that the ionospheric diurnal variability is in general less significant at the magnetic equator, and tend to increase progressively towards the crest regions of the equatorial ionization anomaly. Abdu et al. (2007) studied the solar flux effects on equatorial ionization anomaly and total electron content over Brazil, using maximum frequency of F2 layer (foF2) data set between 1996 and 2003 and TEC data set between 2002 and 2003. They found similar solar flux dependence of seasonal variation in both foF2 and TEC, which showed maximum during equinox and minimum during June solstice. Solar activity dependence of TEC has also been studied by a large number of researchers (e.g. Balan et al., 1993 and references cited there in). We also highlight the climatologic studies derived from TOPEX/POSEIDON measurements (e.g. Codrescu et al., 1999, 2001; Jee et al., 2004; Scherliess et al., 2008) which established many interesting facts about longitudinal, seasonal, wave number four, and geomagnetic TEC variations. The South American sector is associated with highly variable electrodynamics processes which are particularly prominent at the equatorial and low-latitude regions. The declination of the magnetic field lines is the highest at the Brazilian region and gradient between the trough and the crest is very sharp, which results in large temporal and spatial variation of the ionospheric electron content (Dasgupta et al., 2007; Muella et al., 2010). The largest fraction of the solar radiant energy is also centered mainly at the equatorial and low latitudes, hence many interesting phenomena are presented at these regions. However, other effects associated to the strength of the equatorial electrojet (EEJ), changes of the Earth’s magnetic field at the South American Magnetic Anomaly region (SAMA), added to the effects of tides, waves and the thermospheric neutral winds contribute to the variability in TEC. The ionospheric electron density variability during geomagnetic disturbed times have been observed by numerous authors (e.g. Fedrizzi et al., 2001; Tsurutani et al., 2008; de Siqueira et al., 2011), and only few studies (e.g. Jee et al., 2004; Scherliess et al., 2008) have been reported for quiet time period. Yet the motions in the upper atmosphere are of two kinds, those whose immediate sources of energy are confined in the upper atmosphere itself, and those whose energy are transmitted from the lower atmosphere (Charney and Drazin, 1961). Hence, this study aims at generally understanding the TEC variations during geomagnetic quiet time at the ionospheric region of the South America sector. This present paper is organized in the following way: Firstly describe the data and methodology used to estimate the absolute GPS total electron content by using the Nagoya ionospheric model. Next, in Section 3, we present the results and the discussions on the diurnal, monthly and seasonal variations of TEC during a year of high solar activity (2001) and a year of low solar activity (2009), and also on the periodicities of TEC. Finally, in Section 4, we present the conclusions.
2. GPS-TEC measurement methodology and data The Absolute TEC (ATEC) used in the work was estimated from the Nagoya model by making use of the following principal equations (Otsuka et al., 2002). The slant TEC at a piercing point is mapped to the vertical TEC (VTEC) at that point and it is given by
T i (t ) = S (εi (t )) V i (t ),
(1)
23
where εi = εi (t ) is the elevation angle of the GPS satellite, Vi(t) is the vertical TEC, S (ϵi (t )) is the slant factor (mapping function) given by τi/τo , and τi is the oblique length of ray path between 300 and 550 km altitude while τo is equal to the ionospheric thickness of 250 km for zenith path. In order to eliminate errors, which increase relative to the slant factor, the cutoff elevation angle was fixed at 30°. The absolute TEC I i (t ), measured by satellite-i at epoch t, can therefore be obtained by equation:
I i (t ) = S (εi (t )) V i (t ) + Bi
(2)
i
where B is the instrumental biases of both the receiver and the satellite. The Bi is calculated by using the least square fitting method as shown by the following residual equation: 2 ⎡ ¯ ⎛ ⎛ 1 ⎞ ⎞⎤ Ns Ni I ⎟⎟ Bi⎟ ⎥ , E = Σ Σ W ki ⎢ k − ⎜⎜ Vk + ⎜⎜ ¯ i ) ⎠ ⎟⎥ ⎢ S (ϵi ) ⎝ i k ⎝ S (ϵ ⎠⎦ ⎣ k k
W ki =
1 S (ϵik )
(3)
where Wki is the weighting function, Vk, k ¼ 1,2,…,Nt and Bi, i ¼1,2, …,Ns where Nt is the number of hourly TEC average, and Ns is the number of satellites which are observed by a receiver (note that variables with overline denote averaged values). Taking the partial derivatives of E with respect to Vk and Bi, and setting them to zero, yield equations which can be solved for the desired parameters (Vk and Bi). To reduce the estimation errors of hourly TEC average caused by the assumptions of both the shell model and the spatial uniformity of the hourly TEC average, Wki is selected as an inverse of the slant factor (second part of Eq. (3)) and it becomes smaller with decreasing elevation angle. For further explanation readers are referred to Otsuka et al. (2002). In order to study the variation of TEC due to local time, season and solar activity over South America sector, we have obtained data from the following data bases: 1. SOPAC: Scripts and Permanent Array Centre Garner GPS archive (known as SOPAC GARNER) contains files of observation and navigation from the GPS global network. The data base belongs to the International GNSS Service (IGS). It is available at 〈ftp:// garner.ucsd.edu/pub/rinex/〉. 2. RBMC/IBGE: Brazilian Network for Continuous Monitoring of the Institute of Brazilian Geography and Statistics. The data of this database can be accessed at 〈ftp://geoftp.ibge.gov.br/RBMC/ dados/〉. The distribution of the ground-based dual frequency GPS receivers used in this work are given in Fig. 1. The F10.7 cm solar flux and Kp data were obtained from the National Geophysical Data Center (NGDC) at National Oceanic and Atmospheric Administration (NOAA) data base (http://omniweb.gsfc.nasa.gov/form/dx1. html). In the top panel of Fig. 2 is shown the F10.7 cm solar flux in sfu (1 sfu ¼10 22 W m 2 Hz 1) from 2000 to 2009. In the middle and bottom panels of Fig. 2 are shown the daily ∑Kp index throughout 2001 and 2009, respectively. Only data from daily ∑Kpr 24 were considered in this work. The MLT wind data measured at Santa Maria (29.4°S, 53.3°W, dip latitude 17.8°S) and Cachoeira Paulista (22°S, 45°W, dip latitude 15°S) are used to infer the meridional and zonal wind. The meteor radar system is an All-Sky Interferometric Meteor Radar (SKiYMET) type with operating frequency of 35.24 MHz. It utilizes 13 pulse width, with a peak power of 12 kW and pulse repetition frequency of 2 kHz. Due to practical limitations, the received data are obtained within the zenith angle range 17–70°, since at lower zenith angle (o 15°) radial velocity errors are excessive and at higher zenith angle (4 70°) the signal gets contaminated with ground clutter and airplane echoes. The radar detects about 5000 meteor echoes per day for which it estimates angular position,
24
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Fig. 1. The South America map showing the location of the observatories with dual-frequency GPS receivers (marked with yellow circles). The dashed line in red indicates the magnetic equator. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
range, and radial velocity (Jonah et al., 2014). Details about the radar can be found in Hocking et al. (2001). The received horizontal wind data are analyzed at a range resolution of 3 km and time resolution of 1 h in the altitude range 81–100 km.
3. Results and discussion To observe the diurnal, monthly, seasonal, solar activity, longitudinal, and latitudinal variation of TEC from the magnetic equator to the Equatorial Ionization Anomaly ( EIA), GPS data of the following locations were processed by the Nagoya Model: Arequipa (16.5°S, 71.6°W, dip latitude 5°S) and São Luís (2°S, 44°W, dip latitude 2°S) located close to the dip equator and at both longitudinal ends of the South America region, Cachoeira Paulista (23°S, 45°W, dip latitude 16°S) and Porto Alegre (30°S, 51°W, dip latitude 19°S) located at low latitude region close to the southern crest of the EIA. The seasonal variation of TEC for both solar minimum (2009) and maximum (2001) are studied by grouping data into three seasons and taking their hourly average and standard deviation. All the analyses were done over the southern hemisphere therefore the months of November, December, January, February are considered as summer solstice, the months of May, June, July, August as winter solstice, while the months of March, April, September and October as equinox. Both low and high solar activities data obtained throughout the years of 2009 and 2001, respectively, were analyzed in this study. 3.1. Diurnal and monthly variation in TEC for solar minimum of 2009 Fig. 3 shows mass plots of monthly variation of TEC for typical
geomagnetic quiet days (∑Kpo24) during a period of 12 months from January to December 2009. The plots with blue and green lines represent, respectively, the stations close to dip equator (Arequipa and São Luís) and the low latitude stations close to the crest of the EIA (Cachoeira Paulista and Porto Alegre). The monthly F10.7 index is also indicated to the corresponding months at the right side of the plots. Almost similar TEC patterns are observed for all months of different seasons. In general, the diurnal variation of TEC shows a pre-dawn minimum, a steady early morning increase, followed by an afternoon maximum and gradual fall after sunset. It is clear from the figure that day minimum in TEC occurs between about 0500 LT and 0600 LT, while day maximum occurs between 1300 LT and 1700 LT. Larger variations of TEC are observed in daytime for different stations while nighttime variations are observed only for some stations. The TEC enhancement particularly at the low latitude stations during daytime can be associated with the upward drift of plasma caused by the fountain →
→
effect, as a result of E × B drift and the consequent diffusion along magnetic field lines due to gravity force and pressure gradient, which form two peaks at low latitude, known as the Equatorial Ionization Anomaly (EIA) (Abdu et al., 2007). The low latitude stations exhibit lower value of TEC during May, June, July and August compared to the same months of the equatorial stations. This implies that the formation of EIA is weaker during these months, which represents winter solstices. The EIA strength and characteristics can be observed between the Cachoeira Paulista and Porto Alegre stations, as it is possible to observe that in almost all the months, TEC enhancement is larger at the Cachoeira Paulista station than at Porto Alegre station that is little further away. This is expected because 2009 is a low solar activity period and anomaly peak is restricted to much lower latitudes due to less
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
25
Fig. 2. The top panel shows the daily variability of F10.7 cm solar flux index from year 2000 to 2009. The dashed boxes in red indicate the two periods of solar flux data used in this study. The middle and bottom panels show, respectively, the daily sum of Kp for both 2009 and 2001. The horizontal red lines are used here to mark the threshold level for quiet geomagnetic period. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
ionization and weaker EIA compared to high solar activity. Longitudinal variation can be observed between São Luís and Arequipa. The Arequipa station shows slightly larger TEC enhancement in some months compared to the São Luís station. This is expected because São Luís is closer to the magnetic equator than the Arequipa station and there is always a trough at the equator due to fountain effect as explained above. To further observe the variability of TEC in Fig. 3, the standard deviation (STD) was obtained. The STD was found to vary according to the diurnal variation, exhibiting low and steady STD between 1 and 2 TECU during predawn and morning hours, a buildup of up to 4 TECU in daytime and reaching 6 TECU after sunset. It is interesting to see that during winter solstice the diurnal TEC variations are lower at low latitude than at the equator, while during summer and equinox the diurnal TEC variations at the low latitude are larger than during winter. This is expected because the fountain effect is more developed at the equator during summer and equinox than during winter (Jee et al., 2004). In general, the day-to-day variability in 2009 low solar activity is low (highest F10.7 ¼76.8 for December month) compared with the high solar activity period of 2001. 3.2. Diurnal and monthly variation in TEC for solar maximum of 2001 Fig. 4 shows the mass plots of monthly variation of TEC from January to December 2001. Similar diurnal patterns are observed for all months as compared to 2009, except that the rate of post sunset and nighttime changes are different due to the prereversal enhancement in the ionospheric vertical drift velocities that are prominently associated to solar maximum period. It is important to note that the scale has increased from 40 to 150 TECU depending on local time, with the values of TEC increasing drastically by more than 400% on average from solar minimum to solar maximum for all stations. Table 1 shows the percentage change from 2009 to 2001 for Arequipa and São Luís. This increase is mainly due to the corresponding increase in the solar flux represented by F10.7 (shown in Fig. 2), which increased from
76.8 sfu (highest in 2009) to 235.6 sfu (highest in 2001). It is clearly seen from Fig. 4 that the weakened EIA at sunset around 1800–1900LT becomes intensified again during the postsunset hours. Anderson and Klobuchar (1983) affirmed that the primary source of the enhancement at equatorial anomaly latitudes after sunset is the evening increase of the equatorial fountain. The strength of the EIA at sunset that is intensified by an enhancement in the zonal (eastward) electric field, also known as prereversal enhancement (PRE), is dependent on the solar flux values. A corresponding increase in the zonal wind and of the ratio between the average Perdersen conductivity along the magnetic field line in the F and E regions, leads to an increase in the prereversal zonal electric field (Fejer et al., 1991). Furthermore, some monthly differences were observed between Arequipa and São Luís equatorial stations in 2001 and 2009 high and low solar activities. For example, in 2001, Arequipa showed a slightly larger TEC value throughout the year except in the months of April, August and September. Similarly, in 2009 solar minimum, Arequipa station exhibits higher percentages of TEC than São Luís, except for winter months of June and July. The monthly percentage changes are depicted in Table 2. A plausible explanation for this longitudinal difference can be attributed to the larger magnetic declination at São Luís station ( 20° west) compared to the Arequipa station ( 0.5° west). Using the same analytical method the STD of TEC variation of 2001 solar maximum were calculated and were found to vary according to the diurnal patterns in some cases (i.e. they exhibit low STD late night to early morning hour, highest during afternoon and evening time), from June to August. However, the months from March to April and September to October exhibit high post sunset variability in TEC as observed from the STD. We noted that STD increased from 6 TECU (highest value observed in the 2009 solar minimum) to 30 TECU in 2001 (highest value observed in the high solar period) and it is dominant after sunset during months with relatively high F10.7. This behavior is found mostly at stations located in low latitude region. This indicates that day-to-day variability for monthly basis of TEC is larger during solar maximum. A similar variation has been reported by Abdu et al. (2007)
26
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Fig. 3. Monthly diurnal TEC variation during 2009. Blue and green lines represent the stations close to dip equator (Arequipa and São Luís) and the low latitude stations within the EIA crest (Cachoeira Paulista and Porto Alegre) respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
27
Fig. 4. Monthly diurnal TEC variation during 2001. Blue and green lines represent the stations close to dip equator (Arequipa and São Luís) and the low latitude within the EIA crest (Cachoeira Paulista and Porto Alegre) respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
28
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Table 1 TEC percentage change from solar minimum of 2009 to solar maximum of 2001 for Arequipa (Areq) and São Luís (Salu). Months
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Areq (%) Salu (%)
373.6 406.7
358.7 403.8
349.5 411.6
372.6 445.6
306.3 320.6
339.3 294.1
308.5 269.1
368.1 380.7
382.1 403.2
400.3 448.5
334.3 370.8
341.1 426.6
Table 2 Monthly TEC percentage change between Arequipa and São Luís for low and high solar activities of 2009 and 2001 respectively. Months
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Monthly percentage change between Areq and Salu in 2009 Areq(%) 51.2 52.4 53.8 51.2 Salu(%) 48.8 47.6 46.2 48.8
50.6 49.4
49.4 50.6
48.7 51.3
49.3 50.7
50.0 50.0
52.1 47.9
52.3 47.7
54,4 45.6
Monthly percentage change between Areq and Salu in 2001 Areq(%) 50.5 50.1 50.6 47.6 Salu(%) 50.5 49.9 49.4 52.4
50.7 50.3
52.1 47.9
51.2 48.8
48.6 51.4
49.0 51.0
50.8 50.2
50.3 49.7
50.0 50.0
Fig. 5. Seasonal variability during 2001 and 2009. The panel from left to right represents Arequipa, São Luís, Cachoeira Paulista and Porto Alegre. The blue, green and red error bars represent winter solstices, equinox and summer solstices, while small circles represent 2001 data. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
using foF2 data measured by digital ionosondes in the equatorial and low latitude locations in Brazil from 1996 to 2003. The TEC and STD during day time and nighttime at both Cachoeira Paulista and Porto Alegre exhibit different levels of TEC due to the EIA amplitude. This is because there is higher photoionization effect and more prereversal enhancement effect that could make the fountain effect to be well developed during this period. As reported by Santos et al. (2013) the equatorial E B drift is also strongly dependent on EUV solar flux and the ambient ionosphere, and as a result during solar maximum period the EIA may reach much higher latitudes (as seen from the plots of Porto Alegre station for 2001).
High variabilities in TEC, as observed from the STDs, were observed between 1900 LT and 2300 LT. The months of March–May, September and October exhibited the largest nighttime variation in STD during post sunset followed by February and November. The months of June, July and August indicate lowest nighttime variability at the low latitude stations, as observed from Cachoeira Paulista and Porto Alegre. Generally, the day-to-day analysis revealed that the month of January had very few nighttime variations, which increased through February and had maximum values during the months of March, September and October. Besides the clear evidence from Fejer et al. (1991), Abdu et al. (2007) and Santos et al. (2013) that solar flux plays a principal role
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
29
Fig. 6. The monthly TEC average contour plots for low (left) and high (right) solar activity as a function of month and local time. The TECU has been changed to decibel unit in both cases.
in the vertical prereversal drift enhancements and late high peaks in TEC, another important factor that could be responsible for the evening prereversal enhancement and the consequent nighttime variations in TEC could be associated to the faster decrease of the E region ionospheric conductivity after sunset relative to the F region ionospheric conductivity. 3.3. Seasonal variability in TEC The mean and the standard error bar for hourly seasonal variation in TEC for both 2009 and 2001 are plotted together for better comparison as shown in Fig. 5 to represent the seasonal hourly dispersion of TEC. The blue, green and red colors represent winter solstice, equinox and summer solstice, respectively. Each season is represented by groups of four months GPS-TEC data as
explained earlier in Section 3. This is to demonstrate that there are important seasonal and long term variations in the ionosphere, as explained below. It is possible to observe the following:
The TEC scale for solar maximum is about four times larger than for solar minimum and the highest TEC value stands around 14:00LT. During winter solstice the TEC values are larger at the equator than at the off equatorial stations. A low value of TEC is observed over all the stations for both minimum and maximum solar activities during winter compared with the other seasons.
It can be observed from Fig. 5 that TEC values of summer solstices is higher than equinox during 2009 solar minimum activity at low latitude stations but the variability is larger at equinox. This could result from the more production of electron
30
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Table 3 The comparisons of TEC behavior between 2009 low solar activity and 2001 high solar activity. S/N 2001 1
2 3 4 5
6 7 8
2009
The nighttime resurgence of the anomaly (occasioning a secondary peak in TEC) There is no nighttime resurgence at all stations during low solar flux of 2009. associated to the evening prereversal enhancement is observed at Cachoeira Paulista and Porto Alegre stations, particularly during high solar activity (2001), mainly in the months of January–May and October–December, but not observed at São Luís because it is very close to the dip equator, and very little peaks are shown in Arequipa station because it is relatively further away from the equator compared to São Luís station. The São Luís station shows slightly larger TEC enhancement in some months The Arequipa station shows slightly larger TEC enhancement in almost all the compared to the Arequipa station. months compared to the São Luís station. The TEC and STD in Fig. 6 during daytime and nighttime at both Cachoeira The TEC and STD are closely of the same value for both the low latitude stations Paulista and Porto Alegre exhibit different levels of TEC due to the EIA amplitude. during 2009 periods. Our results suggest that prereversal peaks are dominant during equinoxes for There are no prereversal peaks for all seasons during 2009 solar minimum. 2001 solar phase, which signifies a semidiurnal variation. Fig. 5 shows that there is no much longitudinal difference between Arequipa and Fig. 5 shows larger TEC peaks for summer months of November and December of São Luís for 2001, except that São Luís exhibits slightly larger TEC in 2001 period 2009 at the low latitude stations of Cachoeira Paulista and Porto Alegre than during summer months of January and February of the same year than Arequipa for same period and Arequipa shows smaller peaks at sunset compared to São Luís during equinox of 2001 period. STD shows up to 30 TECU in 2001 the highest value observed in the high solar The highest STD value observed in the 2009 solar minimum is 6 TECU period. Generally, in both 2009 and 2001 solar phases we can observe the seasonal peaks between summer and winter (annual variation) and the equinox maximum compared to the solstices (semiannual variation). The solar flux variations between the two extreme solar activities are clearly reflected by the value of TEC between the two periods.
by photoionization at the summer solstices hemisphere enhancing the background electron density at the region. The seasonal effect can be clearly observed particularly between summer and winter solstices (e.g. December solstice and June solstice), and the equinox and solstices of our results during both solar activities. The TEC variability (obtained from the STD) is reduced from by about 30% for stations at the equator and 200% for stations at the crest of the EIA from December to June solstice conditions for both solar activities. The Sun–Earth distance between June and December period and the Earth’s inclination could be a causative factor (Jee et al., 2004).
Another obvious TEC-EIA characteristic observed during 2001 solar maximum activity is that TEC values during equinox are higher than the TEC values during the summer solstices of the same year and for all stations. This is mainly because the sun overhead is around the equator during equinoxes, which could lead to more intense ionospheric current in E region associated with eastward electric field and, consequently causing a welldeveloped EIA than during the solstices (Wu et al., 2004; Abdu et al., 2007). This equinox–solstices variation is referred to as semiannual variation. The semiannual variation of the EIA could also be due to the combined effect of the solar zenith angle and
Fig. 7. Distribution of TEC variation as function of season and location. The vertical axis represents number of occurrence and the horizontal axis stands for residual TEC. The blue color represents low solar activity and the red represents the high solar activity. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
31
Fig. 8. The Power spectra analysis distribution of 2009 hourly average TEC contour plots for Cachoeira Paulista, Porto Alegre, Arequipa and São Luís from top to bottom respectively. The vertical axis represents the daily period and the horizontal axis represent the day of the year.
magnetic field geometry (Wu et al., 2004). On the other hand, during the solstices, photoionization at the equator decrease because the Sun overhead moves to higher latitudes and fountain effect is expected to be low (Kumar et al., 2014). Lastly, the TEC daily peak show semiannual variation with highest peak during the equinox in 2001 solar maximum (Fig. 5), while the daily highest peak of TEC variation is observed during the summer of 2009 solar minimum (Fig. 5), and lowest daily peaks (troughs) are observed during winter solstice for both solar activities. This implies that equinoctial semiannual variation depend largely on solar flux.
Generally, the seasonal and semiannual variations vary from one station to another. For example, during 2009 low solar activity,
the low-latitude stations present higher TEC daily peaks during summer and equinox, while the stations close to the equator (São Luís and Arequipa) exhibit low anomaly peaks. On the other hand, during winter solstice TEC daily peaks are larger at equatorial stations than at low latitude stations. The reason for this has been explained above. However taken a close look at Fig. 5, it can be observed that during summer solstice the equatorial stations presents slightly larger TEC daily peaks than the low latitude station, while during equinox period larger TEC daily peaks are observed at the lower stations than at the equator. This could be due to wind that is more effective during solar maximum and transequatorial winds blowing from the summer to the winter hemisphere which act to transport plasma towards the equator. Thus wind-driven migration of large density atomic oxygen [O]
32
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
may result in increased O/N2 density ratio during summer (hence comparative low TEC value is observed at low latitudes) and towards the low latitudes during equinox (hence lower TEC values are observed at equator than at lower latitudes). It is important to mention that unlike 2009 low solar activity that exhibited low F10.7 seasonal variation, 2001 high solar activity was associated with high F10.7 seasonal variation. Therefore further observation is needed to clarify the wind contribution for the 2001 TEC seasonal variation. 3.4. Summary of TEC variations Fig. 6 shows color contour plots of monthly average of TEC for the low solar activity (left panel) and high solar activity (right panel) as a function of month and local time, longitude (Arequipa and São Luís) and latitude (São Luís, Cachoeira Paulista and Porto Alegre). In order to use the same scaling system for both solar
activity plots, we used:
⎛ TEC ⎞ ⎟ 20 log10 ⎜ ⎝ TECref ⎠
(5)
where TECref ¼ 1TECU. Solar cycle 23 is characterized by both extreme minimum (2009) and extreme maximum (2001) activity period that is unique in many ways and it is therefore important to observe the extent of the effect on TEC variation. The interesting facts in Fig. 6 are that it is possible to observe the daily, monthly, seasonal, longitudinal and latitudinal variation as function of time. The main TEC features for both low and high solar activities of 2009 and 2001 respectively are outlined in Table 3. 3.5. General TEC distribution Fig. 7 presents the residual distribution of average seasonal
Fig. 9. (a) The Power spectra analysis distribution of 2009 solar minimum daily average TEC contour plots of Cachoeira Paulista, Porto Alegre, Arequipa and São Luís from top to bottom respectively not considering the daily and semiannual periodicities. The vertical axis represents the daily period and the horizontal axis represent the day of the year. (b) Power spectra analysis distribution of 2009 solar minimum daily average zonal wind, meridional winds (both at ∼100 km) and solar flux from top to bottom respectively. The vertical axis represents the daily period and the horizontal axis represent the day of the year. (c) The Power spectra analysis distribution of 2001 solar maximum daily average TEC contour plots of Cachoeira Paulista, Porto Alegre, Arequipa and São Luís from top to bottom respectively not considering the daily and semiannual periodicities. The vertical axis represents the daily period and the horizontal axis represent the day of the year. (d) Power spectra analysis distribution of 2001 solar maximum daily average zonal wind, meridional winds and solar flux from top to bottom respectively. The vertical axis represents the daily period and the horizontal axis represent the day of the year.
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
33
Fig. 9. (continued)
change in TEC pattern (ΔTEC ) determined by: ΔTEC = (TEC − TECs )), where TECs is the difference between the average values of each seasonal TEC from each data set of the season in particular (TEC). The vertical axis represents number of occurrence and the horizontal axis stands for residual TEC. The blue color represents low solar activity and the red represents the high solar activity. The TEC distribution ratio of 1:4 TECU during equinox and summer solstice and a ratio of 1:3 during winter solstice are observed. Generally low solar activity of 2009 exhibits high kurtosis around zero, which confirms the low variability observed at earlier results during this period. Also during low solar activity period TEC distribution pattern exhibit a form of Gaussian distribution around zero mean, indicating that on average TEC experience as many increases as decreases. This characteristic is observed for all the stations. However, during high solar activity of 2001 the opposite is the case. The kurtosis is flat confirming the high standard deviation that was observed during the period. The TEC patterns skew left but spread out to the right during winter indicating more decreases than increases in TEC. Otherwise, during summer solstice TEC pattern skew right and also spread out to the right in most of the plots indicating more increases than decreases in TEC. Fig. 7 also reveals that during the equinox data form a close normal distribution in all stations showing as much increase as decreases
in TEC, except in Arequipa station. 3.6. Wavelet analysis In this section we show other mechanisms that could be responsible for TEC variation during magnetic quiet periods. It is well known that the nonlinear interaction of quasi-stationary planetary waves that are trapped below the mesopause can give origin to large changes in both migrating and non-migrating tides (Goncharenko et al., 2010). These tides are known to exhibit an amplitude maximum at low-latitudes and can propagate into the lower thermospheric region. Moreover, the propagating tides can modulate electric fields through the ionospheric wind dynamo at E region and consequently they map along magnetic field lines to high altitudes of F region height and then influence ionospheric variability (Sripathi and Bhattacharyya, 2012). Jee et al. (2005) used a one dimensional middle-latitude ionospheric model to study the sensitivity of quiet-time TEC to the atmospheric and ionospheric parameters including the neutral wind and found out that during both day and night, the magnetic meridional component of the neutral wind significantly affect TEC, and the geographic zonal wind can cause noticeable longitudinal variation in TEC due to the longitudinal variation in the declination
34
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
angle, which is more prominent in the South American sector (the region in focus). On these evidences and since we are working with only geomagnetic quiet period, it is important to observe the effect of upward propagating waves on the E layer of which may have consequent effect on TEC as explained above. To carry out this investigation we applied wavelet spectral analysis distribution of the daily means of meridional and zonal winds at the altitude of ∼100 km, and compared their periodicities with the daily means of TEC data over Cachoeira Paulista, Porto Alegre, São Luís and Arequipa stations. Apart from the well-known daily peaks and annual and semiannual periodicities such as 27, 16, 1–5 days are found to be dominant and consistent. Fig. 8 is the wavelet spectral analysis plots of hourly mean TEC for the solar minimum year of 2009. The one day peak period along the year and the half day peak period with a semiannual anomaly, can be clearly seen in the figure particularly over Cachoeira Paulista and Porto Alegre. However the semiannual period is not so obvious over São Luís and Arequipa stations. In order to see clearly the other periodicities, we eliminated the daily and the semiannual periodicities by using only the daily averages in the following analysis. The remaining periodicities are clearly shown in Fig. 9(a) which shows the wavelet analysis distribution of daily average of TEC for 2009. Periodicities of 1–5 days are stronger at Arequipa and São Luís stations, 16 days and 27 days are also observed at Cachoeira Paulista and Porto Alegre. Likewise, to explain the TEC periodicities plotted in Fig. 9(a), we made the wave analysis for the zonal and meridional wind at ∼100 km height. From top to bottom in Fig. 9(b), is depicted the daily average zonal and meridional winds, and in the last panel the daily mean of F10.7 for 2009. The periods within the black line contour have 95% confidence level and we used the Morlet function (wave number of 6) as a mother wavelet. It can be clearly seen the 27 days solar periodicity indicated by the F10.7 coinciding well with the TEC. Other periodicities with 16 days, which corresponds to the same period in both zonal and meridional winds around the same days, are also prominent. The 16 days periodicities can be associated with lunar tide effects, which causes strong enhancement on the semidiurnal variations. Moreover, strong 1–5 and 7day periods are also observed at all stations, but not so significant in the low latitude stations of Cachoeira Paulista and Porto Alegre. The same wavelet power spectra analysis applied for 2009 were also analyzed for the TEC measurements of solar maximum year 2001. The results are shown in Fig. 9(c) and (d). The intensity of the F10.7 indicated by the 27-day periodicity is stronger as expected for all the stations in solar maximum. The 16 days periodicity is observed only in the meridional wind (Fig. 9(d)) and in TEC for all stations (Fig. 9(c)), while 1–2, 5–7 are observed in TEC, and 8–10 days oscillations are observed in both the zonal and meridional winds. The present results are also in accordance with the features observed from the decomposition analysis presented by Abdu et al. (2006), that reported dominance of 4–5 days and 7 days periods on observations of ionospheric vertical drift velocities and on meridional and zonal winds at ∼100 km height. Generally, these results demonstrate the existence of a strong vertical coupling through upward propagating waves leading to day-to-day oscillation in TEC. A possible electrodynamic coupling mechanism connecting these oscillations could be included in future studies. Oscillations in TEC as observed in the panels of Fig. 9(a) and (c) with different periodicities from 3 to 16 days have been attributed to planetary waves that propagate horizontally and vertically up to the lower thermosphere, affecting the ionosphere in an integrated sense. The planetary waves at equatorial and lowlatitudes are generally Kelvin waves with a zonal wave number from 1 to 3 and periods of 3–4 days (ultra-fast waves), 6–10 days (fast waves) and 15–20 days (slow waves). The quasi- 16-days
periodicities have also been identified as Rossby waves with zonal wave number 1 (Pancheva et al., 2008). The oscillations observed here in the spectral analysis of TEC with periods from 1 to 10 days were more significant during the equinoxes and the December solstice months of the year 2001 for all the stations, corresponding with some of those oscillations also observed in the wind data, which suggests upward propagation of planetary waves from the lower atmosphere. The fast and mainly ultra-fast Kelvin waves are more capable to propagate upward, and are crucial at tropical latitudes for the vertical coupling between the neutral and ionized atmosphere (Phanikumar et al., 2014). However, throughout the year 2009 the results depicted in Fig. 9 revealed that the presence of waves with shorter periods is more significant in the wind data than in the TEC, which suggests that during the solar minimum period the oscillations in the ionosphere might not be potentially coupled with the lower atmosphere through ultra-fast planetary waves propagating upward.
4. Conclusions In this study we have used up to 730 days of dual frequency GPS data to calculate the TEC over four stations located in the South American sector. The analysis were used to describe the TEC variation in the ionosphere in terms of diurnal, day-to-day, seasonal, latitudinal and longitudinal variations during quiet time for low (year 2009) and high (year 2001) solar flux. The results were also used to investigate the strength and characteristics of the ionosphere at the EIA. In addition, using the power wavelet analysis spectra, periodic TEC variations were analyzed and some physical mechanisms responsible for these variations were suggested. We made mass hourly plots of each month, and grouped month into seasons for both phases of the solar cycle. The standard deviation (STD) was also used to describe the local and spatial TEC variability at different locations on the South American sector. Our results indicate that the local TEC variability, as observed in the STD, is largest around the locations of low latitude stations with maximum around afternoon and late in the evening. The largest STD values were between 5 and 6 TECU during solar minimum and between 25 and 30 TECU during solar maximum. TEC variability was larger in equinox followed by summer solstice and low in winter solstice for solar maximum, while for solar minimum larger TEC variability was observed during summer solstice followed by equinox and lowest during winter solstice. The TEC value and TEC variability increased from 2009 solar minimum to 2001 solar maximum by ∼400% and ∼200% respectively. It should be mentioned that during winter solstices for both phases of the solar cycle, TEC diurnal values at the crest of the anomaly region exhibits lower values than TEC at the equatorial regions. This could be because sun radiation during winter solstice is not as intense as during equinox and summer. Therefore fountain effect is not well developed and plasma is not effectively transported off the magnetic equator during this season. An annual seasonal variation is observed in the solstices and a semiannual variation is observed in the equinoxes. In order to investigate mechanisms which are responsible for TEC periodicities as observed in our data, we applied wavelet spectral analysis to daily means of TEC data over Cachoeira Paulista, Porto Alegre, São Luís and Arequipa stations, as well as to the daily means of meridional and zonal winds at the altitude of ∼100 km along with F10.7 to compare their periodicities. Apart from the strong a 27-day period and a 16-day period observed in F10.7 and wind parameter respectively, we also observed that a strong 1–5 and 7-day periods were common both to all the stations but not so strong at both Cachoeira Paulista and
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Porto Alegre stations and the two parameters (zonal and medional winds). A day-to-day variation in TEC is reported to contain a component driven by planetary waves enhanced by tides as they propagate upward. A strong vertical coupling through the upward propagating waves can also give rise to day-to-day oscillation in TEC. Our study also indicates that apart from solar radiation effect, changes in the meridional wind or zonal winds can also lead to corresponding changes in TEC.
Acknowledgments Jonah O.F. world like to acknowledge the supports from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) under process 133429/2011-3 and Dr. Paulo Prado Batista for the meteor radar data. E.R. de Paula is grateful to AFOSRFA9550-10-10564 and CNPq under Grant No. 3056842010-8. Marcio Muella would like to thank the support from CNPq under Grant No. 304674/2014-1. The RBMC/IBGE GPS data and magnetic and solar flux data used in this work were obtained from IBGE website http:// www.ibge.gov.br/home/geociencias/download/tela_ inicial.php? tipo¼8) and National Geophysical Data Center (NGDC) at NOAA website (http://www.ngdc.noaa.gov/ngdcinfo/onlineaccess.html) respectively.
References Abdu, M.A., Batista, P.P., Batista, I.S., M. Brum, C.G.M., Carrasco, A.J., 2006. Equatorial evening prereversal electric field and spread F. J. Geophys. Res. . http://dx.doi. org/10.1029/1005GL024837 Abdu, M.A., Brum, C.G.M., Batista, I.S., Sobral, J.H.A., de Paula, E.R., Souza, J.R., 2007. Solar flux effects on equatorial ionization anomaly and total electron content over Brazil: observational results versus IRI representations. J. Adv. Space Res. v. 42, 617–625. Anderson, D.N., Klobuchar, J.A., 1983. Modeling the total electron content observations above Ascension Island. J. Geophys. Res. 88, 8020–8024. Bagiya, M.S., Joshi, H.P., Iyer, K.N., Aggarwal, M., Ravindran, S., Pathan, B.M., 2009. TEC variations during low solar activity periods (2005-2007) near the equatorial ionospheric anomaly crest region in India. Ann. Geophys. (Germany) 27, 1047–1057. Balan, N., Bailey, G.J., Jayachandran, B., 1993. Ionospheric evidence for a non-linear relationship between the solar EUV and 10.7-cm fluxes during an intense solar cycle. Planet. Space Sci. 41, 141–145. Batista, I.S., Abdu, M.A., 2004. Ionospheric variability at Brazilian low and equatorial latitudes comparison between observations and IRI model. J. Adv. Space Res. 34 1894–1894. Bhuyan, P.K., Borah, R.R., 2007. TEC derived from GPS network in India and comparison with the IRI. Adv. Space Res. 39, 830–840. Bilitza, D., 2000. Report from 33rd COSPAR Scientific Assembly, Warsaw, Poland, 16–23 July 2000. IRI News Lett. 7, 1–4. Charney, J.G., Drazin, P.G., 1961. Propagation of planetary-scale disturbances from the lower into upper atmosphere. J. Geophys. Res. 66, 83–110. Codrescu, M.V., Beierle, K.L., Fuller-Rowell, T.J., Palo, S.E., Zhang, X., 2001. More total electron content climatology from TOPEX/Poseidon measurements. Radio Sci. . http://dx.doi.org/10.1029/1999RS002407 Codrescu, M.V., Palo, S.E., Zhang, X., Fuller-Rowell, T.J., Poppe, C., 1999. TEC climatology derived from TOPEX/POSEIDON measurements. J. Atmos. Sol. Terr. Phys. 61, 281–298. http://dx.doi.org/10.1016/S1364- 6826(98)00132-1. Costa, A.M., Boas, J.W.V., Fonseca, J.R., 2004. GPS total electron content measurements at low latitudes in Brazil for low solar activity. Geofís. Int. 43, 129–137. Dabas, R.S., Lakshmi, D.R., Reddy, B.M., 1993. Solar activity dependence of ionospheric electron content and slab thickness using different solar indices. Pure Appl. Geophys. (France) 140, 721–728. Dasgupta, A., Paul, A., Das, A., 2007. Ionospheric Total Electron Content (TEC) studies with GPS in the equatorial region. Indian J. Radio Space Phys. 36, 278–292. de Siqueira, P.M., de Paula, E.R., Muella, M.T.A.H., Rezende, L.F.C., Abdu, M.A.,
35
Gonzalez, W.D., 2011. Storm-time total electron content and its response to penetration electric fields over South America. Ann. Geophys. . http://dx.doi. org/10.5194/angeo-29-1765-2011 Duarte-Silva, M.H., Muella, M.T.A.H., Silva, L.C.C., de Abreu, A.J., Fagundes, P.R., 2015. Ionospheric slab thickness at the equatorial anomaly region after the deep solar minimum of cycle 23/24. Adv. Space Res. . http://dx.doi.org/10.1016/j. asr.2015.05.010 Fejer, B.G., de Paula, E.R., Gonzalez, S.A., Woodman, R.F., 1991. Average vertical and zonal F region plasma drifts over Jicamarca. J. Geophys. Res. . http://dx.doi.org/ 10.1029/91JA01171 Fedrizzi, M., de Paula, E.R., KantorI, J., 2001. Ionospheric effects of the August 26, 1998 magnetic storm over South America, In: International Beacon Satellite Symposium Proceeding, Boston. Forbes, J.M., Palo, S.E., Zhang, X., 2000. Variability of the ionosphere. J. Atmos. Sol. Terr. Phys. 62, 685–693. Goncharenko, L.P., Coster, A.J., Chau, J.L., Valladares, C.E., 2010. Impact of sudden stratospheric warmings on equatorial ionization anomaly. J. Geophys. Res. . http://dx.doi.org/10.1029/2010JA015400 Hocking, W.K., Fuller, B., Vandepeer, B., 2001. Real time determination of meteor related parameters utilizing modern digital technology. J. Atmos. Sol. Terr. Phys. 63, 155–169. Jee, G., Schunk, R.W., Scherliess, L., 2004. Analysis of TEC data from the TOPEX/ Poseidon mission. J. Geophys. Res. . http://dx.doi.org/10.1029/2003JA010058 Jee, G., Schunk, R.W., Scherliess, L., 2005. On the sensitivity of total electron content (TEC) to upper atmospheric/ionospheric parameters. J. Atmos. Sol. Terr. Phys. 67, 1040–1052. Jonah, O.F., de Paula, E.R., Kherani, E.A., Dutra, S.L.G., Paes, R.R., 2014. Atmospheric and ionospheric response to stratospheric sudden warming of January 2013. J. Geophys. Res. 119. http://dx.doi.org/10.1002/2013JA019491. Kane, R.P., 2003. Day-to-day variability of quiet-time ionospheric foF2. Indian J. Radio Space Phys. 32, 344–348. Kumar, S., Singh, A.K., Lee, J., 2014. Equatorial ionospheric anomaly and comparison with IRI model during descending phase of solar activity (2005–2009). Adv. Space Res. 53, 724–733. Muella, M.T.A.H., Kherani, E.A., de Paula, E.R., Cerruti, A.P., Kintner, P.M., Kantor, I.J., MitchelL, C.N., Batista, I.S., Abdu, M.A., 2010. Scintillation-producing Fresnelscale irregularities associated with the regions of steepest TEC gradients adjacent to the equatorial ionization anomaly. J. Geophys. Res. 115. http://dx.doi. org/10.1029/2009JA014788. Otsuka, Y., Ogawa, T., Saito, A., Tsugawa, T., Fukao, S., Miyazaki, S., 2002. A new technique for mapping of total electron content using GPS network in Japan. J. Earth Planets Space 54, 63–70. Pancheva, D.V., Mukhtarov, P.J., MitchelL, N.J., Fritts, D.C., Riggin, D.M., Takahashi, H., Batista, P.P., Clemesha, B.R., Gurubaran, S., Ramkumar, G., 2008. planetary wave coupling (5–6-day WAVES) in the low-latitude atmosphere–ionosphere system. J. Atmos. Sol. Terr. Phys. 70, 101–122. Phanikumar, D.V., Niranjan, K.K., Sanjay, K., 2014. Signatures of ultra fast kelvin WAVES in low latitude ionospheric TEC during January 2009 stratospheric warming event. J. Atmos. Sol. Terr. Phys. 117, 48–53. Rabiu, A.B., Mamukuyomi, A.I., Joshua, E.O., 2007. Variability of equatorial ionosphere inferred from geomagnetic field measurements. Bull. Astron. Soc. India 35, 607–618. Rama Rao, P.V.S., Gopi, K.S., Niranjan, K., Prasad, D.S.V.V.D., 2006. Temporal and spatial variations in TEC using simultaneous measurements from the Indian GPS network of receivers during the solar activity period of 2004–2005. Ann. Geophys. 24, 3279–3292. Rishbeth, H., Mendillo, M., 2001. Patterns of F2‐layer variability. J. Atmos. Sol. Terr. Phys. 63, 1661–1680. Santos, A.M., Abdu, M.A., Sobral, J.H.A., Mascarenhas, M., Nogueira, P.A.B., 2013. Equatorial evening prereversal vertical drift dependence on solar EUV flux and F10.7 index during quiet and disturbed periods over Brazil. J. Geophys. Res. 118, 4662–4671. http://dx.doi.org/10.1002/jgra.50438. Scherliess, L., Thompson, D.C., Schunk, R.W., 2008. Longitudinal variability of lowlatitude total electron content: Tidal influences. J. Geophys. Res. . http://dx.doi. org/10.1029/2007JA012480 Sripathi, S., Bhattacharyya, A., 2012. Quiet time variability of the GPS TEC and EEJ strength over Indian region associated with major sudden stratospheric warming events during 2005/2006. J. Geophys. Res. 117. http://dx.doi.org/ 10.1029/2011JA017103, A05305. Tsurutani, B.T., Verkhoglyadova, O.P., Mannucci, A.J., Saito, A., Araki, T., Yumoto, K., Tsuda, T., Abdu, M.A., Sobral, J.H.A., Gonzalez, W.D., Mccreadie, H., Lakhina, G.S., Vasyliünas, V.M., 2008. Prompt penetration electric fields (PPEFs) and their ionospheric effects during the great magnetic storm of 30–31 October 2003. J. Geophys. Res. . http://dx.doi.org/10.1029/2007JA012879 Wu, C.C., fry, C.D., Liu, J.–Y., Liou, K., Tseng, C.L., 2004. annual TEC variation in the equatorial anomaly region during the solar minimum: September 1996–August 1997. J. Atmos. Terr. Phys. 66, 199–207.
Contents lists available at ScienceDirect
Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp
TEC variation during high and low solar activities over South American sector O.F. Jonah a, E.R. de Paula a, M.T.A.H. Muella b, S.L.G. Dutra a, E.A. Kherani a, P.M.S. Negreti a, Y. Otsuka c a
Instituto Nacional de Pesquisas Espaciais (INPE), Av. de Astronautas, 1758, Jd da Granja, São José dos Campos, São Paulo 12227-010, Brazil Instituto de Pesquisa e Desenvolvimento (IP&D), Universidade do Vale do Paraíba (UNIVAP), Av. Shishima Hifumi, 2911, São José dos Campos, São Paulo 12444-000, Brazil c Nagoya University, 3-13 Honohara, Toyokawa, Aichi 442-8507, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 29 January 2015 Received in revised form 1 October 2015 Accepted 4 October 2015 Available online 9 October 2015
Using dual frequency GPS receivers in the South American sector, the measurement of absolute ionospheric Total Electron Content (TEC) has been estimated applying the Nagoya ionospheric model for both the years of 2009 and 2001, which represent low and high solar activities, respectively. The diurnal, dayto-day, monthly, seasonal, latitudinal and longitudinal variations of TEC were studied for equatorial and low latitude regions of South America. The strength and characteristics of the Equatorial Ionization Anomaly (EIA) were equally analyzed. The analyses reveal the diurnal, seasonal and semidiurnal TEC variation, as well as the nighttime variability during the low and high solar activities. Wavelet power spectra analysis was employed to check the periodicities of the TEC data, F10.7 and zonal and meridional wind velocities measured by Meteor radar at ∼100 km altitude. Periods such as 27, 16, 8–10, 1–5 days were found to be dominant in the zonal and meridional wind velocity corresponding with those of TEC periodicities. Hence, besides the solar radiation, we suggest that there are contributions of tides and planetary waves in spatial and temporal TEC enhancement and variations during the geomagnetic quiet periods of both solar activities. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Total Electron Content (TEC) Diurnal, seasonal, latitudinal and longitudinal variations of TEC TEC periodicity during high and low solar activities Equatorial Ionization anomaly at different seasons
1. Introduction Exploring the ionosphere is of utmost interest due to the numerous complexities associated with this region (Rabiu et al., 2007). Although, over the last century humanity has learned to use the properties of the ionosphere in a tremendous way, there are more to understand about the chemical and physical changes of this region of the Earth’s atmosphere. One of the parameters that can be used to study the ionosphere is the Total Electron Content (TEC). Study of TEC variability over the South American continent is of great interest due to the possibility to investigate the processes responsible for the ionospheric behavior over this region. TEC is significant in helping us to understand the short and long term changes of our upper atmosphere during major phenomena caused by factors like solar activities, geomagnetic storms and meteorological influences (e.g. Forbes et al., 2000; Kane, 2003; Rishbeth and Mendillo, 2001). These changes in the ionosphere affect navigation systems, surveillance systems and modern technologies such as communication systems, since the signal from the satellite to the receiver must pass through the ionized layer (Bagiya et al., 2009), which causes a delay at such signal. A good description of the ionosphere is also needed in order to http://dx.doi.org/10.1016/j.jastp.2015.10.005 1364-6826/& 2015 Elsevier Ltd. All rights reserved.
improve the performance of the ionospheric models (Bilitza, 2000). Many researchers have studied the morphological features of ionospheric electron density and TEC at low and equatorial latitudes (Dabas et al., 1993; Kane, 2003; Batista and Abdu, 2004; Costa et al., 2004; Rama Rao et al., 2006; Bhuyan and Borah, 2007; Abdu et al., 2007; Bagiya et al., 2009, Jonah et al., 2014, DuarteSilva et al., 2015). Dabas et al. (1993) studied the variations in TEC with different solar indices, i.e. EUV, F10.7 solar flux and smoothed sunspot number (SSN) for summer, winter and equinoxes over the Indian sector. They showed that TEC exhibited nonlinear relationship with SSN in general and linear variations with EUV and F10.7 solar flux. Kane (2003) studied day-to-day variability of quiet-time ionosphere using F2-peak electron density data. He observed oscillations of day-to-day variation with peak spacing of 7 days at several locations and indicated that in the absence of solar or geomagnetic effects, planetary waves dominate the dayto-day variability. Batista and Abdu (2004) also studied the ionospheric variability from observations of the F2 layer peak density over the Brazilian low and equatorial latitudes. Costa et al. (2004) investigated diurnal and seasonal variations of TEC during a year of low solar activity (1997) at Presidente Prudente, a station located
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
near the equatorial ionization anomaly crest in Brazil. It was found that TEC was maximum during summer months (from November to February) and minimum during winter months (from May to August), with intermediate values during the equinoctial months, but the springer equinox (September–October) presenting larger TEC values than the autumn equinox (March–April). Bhuyan and Borah (2007) reported that the ionospheric diurnal variability is in general less significant at the magnetic equator, and tend to increase progressively towards the crest regions of the equatorial ionization anomaly. Abdu et al. (2007) studied the solar flux effects on equatorial ionization anomaly and total electron content over Brazil, using maximum frequency of F2 layer (foF2) data set between 1996 and 2003 and TEC data set between 2002 and 2003. They found similar solar flux dependence of seasonal variation in both foF2 and TEC, which showed maximum during equinox and minimum during June solstice. Solar activity dependence of TEC has also been studied by a large number of researchers (e.g. Balan et al., 1993 and references cited there in). We also highlight the climatologic studies derived from TOPEX/POSEIDON measurements (e.g. Codrescu et al., 1999, 2001; Jee et al., 2004; Scherliess et al., 2008) which established many interesting facts about longitudinal, seasonal, wave number four, and geomagnetic TEC variations. The South American sector is associated with highly variable electrodynamics processes which are particularly prominent at the equatorial and low-latitude regions. The declination of the magnetic field lines is the highest at the Brazilian region and gradient between the trough and the crest is very sharp, which results in large temporal and spatial variation of the ionospheric electron content (Dasgupta et al., 2007; Muella et al., 2010). The largest fraction of the solar radiant energy is also centered mainly at the equatorial and low latitudes, hence many interesting phenomena are presented at these regions. However, other effects associated to the strength of the equatorial electrojet (EEJ), changes of the Earth’s magnetic field at the South American Magnetic Anomaly region (SAMA), added to the effects of tides, waves and the thermospheric neutral winds contribute to the variability in TEC. The ionospheric electron density variability during geomagnetic disturbed times have been observed by numerous authors (e.g. Fedrizzi et al., 2001; Tsurutani et al., 2008; de Siqueira et al., 2011), and only few studies (e.g. Jee et al., 2004; Scherliess et al., 2008) have been reported for quiet time period. Yet the motions in the upper atmosphere are of two kinds, those whose immediate sources of energy are confined in the upper atmosphere itself, and those whose energy are transmitted from the lower atmosphere (Charney and Drazin, 1961). Hence, this study aims at generally understanding the TEC variations during geomagnetic quiet time at the ionospheric region of the South America sector. This present paper is organized in the following way: Firstly describe the data and methodology used to estimate the absolute GPS total electron content by using the Nagoya ionospheric model. Next, in Section 3, we present the results and the discussions on the diurnal, monthly and seasonal variations of TEC during a year of high solar activity (2001) and a year of low solar activity (2009), and also on the periodicities of TEC. Finally, in Section 4, we present the conclusions.
2. GPS-TEC measurement methodology and data The Absolute TEC (ATEC) used in the work was estimated from the Nagoya model by making use of the following principal equations (Otsuka et al., 2002). The slant TEC at a piercing point is mapped to the vertical TEC (VTEC) at that point and it is given by
T i (t ) = S (εi (t )) V i (t ),
(1)
23
where εi = εi (t ) is the elevation angle of the GPS satellite, Vi(t) is the vertical TEC, S (ϵi (t )) is the slant factor (mapping function) given by τi/τo , and τi is the oblique length of ray path between 300 and 550 km altitude while τo is equal to the ionospheric thickness of 250 km for zenith path. In order to eliminate errors, which increase relative to the slant factor, the cutoff elevation angle was fixed at 30°. The absolute TEC I i (t ), measured by satellite-i at epoch t, can therefore be obtained by equation:
I i (t ) = S (εi (t )) V i (t ) + Bi
(2)
i
where B is the instrumental biases of both the receiver and the satellite. The Bi is calculated by using the least square fitting method as shown by the following residual equation: 2 ⎡ ¯ ⎛ ⎛ 1 ⎞ ⎞⎤ Ns Ni I ⎟⎟ Bi⎟ ⎥ , E = Σ Σ W ki ⎢ k − ⎜⎜ Vk + ⎜⎜ ¯ i ) ⎠ ⎟⎥ ⎢ S (ϵi ) ⎝ i k ⎝ S (ϵ ⎠⎦ ⎣ k k
W ki =
1 S (ϵik )
(3)
where Wki is the weighting function, Vk, k ¼ 1,2,…,Nt and Bi, i ¼1,2, …,Ns where Nt is the number of hourly TEC average, and Ns is the number of satellites which are observed by a receiver (note that variables with overline denote averaged values). Taking the partial derivatives of E with respect to Vk and Bi, and setting them to zero, yield equations which can be solved for the desired parameters (Vk and Bi). To reduce the estimation errors of hourly TEC average caused by the assumptions of both the shell model and the spatial uniformity of the hourly TEC average, Wki is selected as an inverse of the slant factor (second part of Eq. (3)) and it becomes smaller with decreasing elevation angle. For further explanation readers are referred to Otsuka et al. (2002). In order to study the variation of TEC due to local time, season and solar activity over South America sector, we have obtained data from the following data bases: 1. SOPAC: Scripts and Permanent Array Centre Garner GPS archive (known as SOPAC GARNER) contains files of observation and navigation from the GPS global network. The data base belongs to the International GNSS Service (IGS). It is available at 〈ftp:// garner.ucsd.edu/pub/rinex/〉. 2. RBMC/IBGE: Brazilian Network for Continuous Monitoring of the Institute of Brazilian Geography and Statistics. The data of this database can be accessed at 〈ftp://geoftp.ibge.gov.br/RBMC/ dados/〉. The distribution of the ground-based dual frequency GPS receivers used in this work are given in Fig. 1. The F10.7 cm solar flux and Kp data were obtained from the National Geophysical Data Center (NGDC) at National Oceanic and Atmospheric Administration (NOAA) data base (http://omniweb.gsfc.nasa.gov/form/dx1. html). In the top panel of Fig. 2 is shown the F10.7 cm solar flux in sfu (1 sfu ¼10 22 W m 2 Hz 1) from 2000 to 2009. In the middle and bottom panels of Fig. 2 are shown the daily ∑Kp index throughout 2001 and 2009, respectively. Only data from daily ∑Kpr 24 were considered in this work. The MLT wind data measured at Santa Maria (29.4°S, 53.3°W, dip latitude 17.8°S) and Cachoeira Paulista (22°S, 45°W, dip latitude 15°S) are used to infer the meridional and zonal wind. The meteor radar system is an All-Sky Interferometric Meteor Radar (SKiYMET) type with operating frequency of 35.24 MHz. It utilizes 13 pulse width, with a peak power of 12 kW and pulse repetition frequency of 2 kHz. Due to practical limitations, the received data are obtained within the zenith angle range 17–70°, since at lower zenith angle (o 15°) radial velocity errors are excessive and at higher zenith angle (4 70°) the signal gets contaminated with ground clutter and airplane echoes. The radar detects about 5000 meteor echoes per day for which it estimates angular position,
24
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Fig. 1. The South America map showing the location of the observatories with dual-frequency GPS receivers (marked with yellow circles). The dashed line in red indicates the magnetic equator. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
range, and radial velocity (Jonah et al., 2014). Details about the radar can be found in Hocking et al. (2001). The received horizontal wind data are analyzed at a range resolution of 3 km and time resolution of 1 h in the altitude range 81–100 km.
3. Results and discussion To observe the diurnal, monthly, seasonal, solar activity, longitudinal, and latitudinal variation of TEC from the magnetic equator to the Equatorial Ionization Anomaly ( EIA), GPS data of the following locations were processed by the Nagoya Model: Arequipa (16.5°S, 71.6°W, dip latitude 5°S) and São Luís (2°S, 44°W, dip latitude 2°S) located close to the dip equator and at both longitudinal ends of the South America region, Cachoeira Paulista (23°S, 45°W, dip latitude 16°S) and Porto Alegre (30°S, 51°W, dip latitude 19°S) located at low latitude region close to the southern crest of the EIA. The seasonal variation of TEC for both solar minimum (2009) and maximum (2001) are studied by grouping data into three seasons and taking their hourly average and standard deviation. All the analyses were done over the southern hemisphere therefore the months of November, December, January, February are considered as summer solstice, the months of May, June, July, August as winter solstice, while the months of March, April, September and October as equinox. Both low and high solar activities data obtained throughout the years of 2009 and 2001, respectively, were analyzed in this study. 3.1. Diurnal and monthly variation in TEC for solar minimum of 2009 Fig. 3 shows mass plots of monthly variation of TEC for typical
geomagnetic quiet days (∑Kpo24) during a period of 12 months from January to December 2009. The plots with blue and green lines represent, respectively, the stations close to dip equator (Arequipa and São Luís) and the low latitude stations close to the crest of the EIA (Cachoeira Paulista and Porto Alegre). The monthly F10.7 index is also indicated to the corresponding months at the right side of the plots. Almost similar TEC patterns are observed for all months of different seasons. In general, the diurnal variation of TEC shows a pre-dawn minimum, a steady early morning increase, followed by an afternoon maximum and gradual fall after sunset. It is clear from the figure that day minimum in TEC occurs between about 0500 LT and 0600 LT, while day maximum occurs between 1300 LT and 1700 LT. Larger variations of TEC are observed in daytime for different stations while nighttime variations are observed only for some stations. The TEC enhancement particularly at the low latitude stations during daytime can be associated with the upward drift of plasma caused by the fountain →
→
effect, as a result of E × B drift and the consequent diffusion along magnetic field lines due to gravity force and pressure gradient, which form two peaks at low latitude, known as the Equatorial Ionization Anomaly (EIA) (Abdu et al., 2007). The low latitude stations exhibit lower value of TEC during May, June, July and August compared to the same months of the equatorial stations. This implies that the formation of EIA is weaker during these months, which represents winter solstices. The EIA strength and characteristics can be observed between the Cachoeira Paulista and Porto Alegre stations, as it is possible to observe that in almost all the months, TEC enhancement is larger at the Cachoeira Paulista station than at Porto Alegre station that is little further away. This is expected because 2009 is a low solar activity period and anomaly peak is restricted to much lower latitudes due to less
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
25
Fig. 2. The top panel shows the daily variability of F10.7 cm solar flux index from year 2000 to 2009. The dashed boxes in red indicate the two periods of solar flux data used in this study. The middle and bottom panels show, respectively, the daily sum of Kp for both 2009 and 2001. The horizontal red lines are used here to mark the threshold level for quiet geomagnetic period. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
ionization and weaker EIA compared to high solar activity. Longitudinal variation can be observed between São Luís and Arequipa. The Arequipa station shows slightly larger TEC enhancement in some months compared to the São Luís station. This is expected because São Luís is closer to the magnetic equator than the Arequipa station and there is always a trough at the equator due to fountain effect as explained above. To further observe the variability of TEC in Fig. 3, the standard deviation (STD) was obtained. The STD was found to vary according to the diurnal variation, exhibiting low and steady STD between 1 and 2 TECU during predawn and morning hours, a buildup of up to 4 TECU in daytime and reaching 6 TECU after sunset. It is interesting to see that during winter solstice the diurnal TEC variations are lower at low latitude than at the equator, while during summer and equinox the diurnal TEC variations at the low latitude are larger than during winter. This is expected because the fountain effect is more developed at the equator during summer and equinox than during winter (Jee et al., 2004). In general, the day-to-day variability in 2009 low solar activity is low (highest F10.7 ¼76.8 for December month) compared with the high solar activity period of 2001. 3.2. Diurnal and monthly variation in TEC for solar maximum of 2001 Fig. 4 shows the mass plots of monthly variation of TEC from January to December 2001. Similar diurnal patterns are observed for all months as compared to 2009, except that the rate of post sunset and nighttime changes are different due to the prereversal enhancement in the ionospheric vertical drift velocities that are prominently associated to solar maximum period. It is important to note that the scale has increased from 40 to 150 TECU depending on local time, with the values of TEC increasing drastically by more than 400% on average from solar minimum to solar maximum for all stations. Table 1 shows the percentage change from 2009 to 2001 for Arequipa and São Luís. This increase is mainly due to the corresponding increase in the solar flux represented by F10.7 (shown in Fig. 2), which increased from
76.8 sfu (highest in 2009) to 235.6 sfu (highest in 2001). It is clearly seen from Fig. 4 that the weakened EIA at sunset around 1800–1900LT becomes intensified again during the postsunset hours. Anderson and Klobuchar (1983) affirmed that the primary source of the enhancement at equatorial anomaly latitudes after sunset is the evening increase of the equatorial fountain. The strength of the EIA at sunset that is intensified by an enhancement in the zonal (eastward) electric field, also known as prereversal enhancement (PRE), is dependent on the solar flux values. A corresponding increase in the zonal wind and of the ratio between the average Perdersen conductivity along the magnetic field line in the F and E regions, leads to an increase in the prereversal zonal electric field (Fejer et al., 1991). Furthermore, some monthly differences were observed between Arequipa and São Luís equatorial stations in 2001 and 2009 high and low solar activities. For example, in 2001, Arequipa showed a slightly larger TEC value throughout the year except in the months of April, August and September. Similarly, in 2009 solar minimum, Arequipa station exhibits higher percentages of TEC than São Luís, except for winter months of June and July. The monthly percentage changes are depicted in Table 2. A plausible explanation for this longitudinal difference can be attributed to the larger magnetic declination at São Luís station ( 20° west) compared to the Arequipa station ( 0.5° west). Using the same analytical method the STD of TEC variation of 2001 solar maximum were calculated and were found to vary according to the diurnal patterns in some cases (i.e. they exhibit low STD late night to early morning hour, highest during afternoon and evening time), from June to August. However, the months from March to April and September to October exhibit high post sunset variability in TEC as observed from the STD. We noted that STD increased from 6 TECU (highest value observed in the 2009 solar minimum) to 30 TECU in 2001 (highest value observed in the high solar period) and it is dominant after sunset during months with relatively high F10.7. This behavior is found mostly at stations located in low latitude region. This indicates that day-to-day variability for monthly basis of TEC is larger during solar maximum. A similar variation has been reported by Abdu et al. (2007)
26
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Fig. 3. Monthly diurnal TEC variation during 2009. Blue and green lines represent the stations close to dip equator (Arequipa and São Luís) and the low latitude stations within the EIA crest (Cachoeira Paulista and Porto Alegre) respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
27
Fig. 4. Monthly diurnal TEC variation during 2001. Blue and green lines represent the stations close to dip equator (Arequipa and São Luís) and the low latitude within the EIA crest (Cachoeira Paulista and Porto Alegre) respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
28
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Table 1 TEC percentage change from solar minimum of 2009 to solar maximum of 2001 for Arequipa (Areq) and São Luís (Salu). Months
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Areq (%) Salu (%)
373.6 406.7
358.7 403.8
349.5 411.6
372.6 445.6
306.3 320.6
339.3 294.1
308.5 269.1
368.1 380.7
382.1 403.2
400.3 448.5
334.3 370.8
341.1 426.6
Table 2 Monthly TEC percentage change between Arequipa and São Luís for low and high solar activities of 2009 and 2001 respectively. Months
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Monthly percentage change between Areq and Salu in 2009 Areq(%) 51.2 52.4 53.8 51.2 Salu(%) 48.8 47.6 46.2 48.8
50.6 49.4
49.4 50.6
48.7 51.3
49.3 50.7
50.0 50.0
52.1 47.9
52.3 47.7
54,4 45.6
Monthly percentage change between Areq and Salu in 2001 Areq(%) 50.5 50.1 50.6 47.6 Salu(%) 50.5 49.9 49.4 52.4
50.7 50.3
52.1 47.9
51.2 48.8
48.6 51.4
49.0 51.0
50.8 50.2
50.3 49.7
50.0 50.0
Fig. 5. Seasonal variability during 2001 and 2009. The panel from left to right represents Arequipa, São Luís, Cachoeira Paulista and Porto Alegre. The blue, green and red error bars represent winter solstices, equinox and summer solstices, while small circles represent 2001 data. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
using foF2 data measured by digital ionosondes in the equatorial and low latitude locations in Brazil from 1996 to 2003. The TEC and STD during day time and nighttime at both Cachoeira Paulista and Porto Alegre exhibit different levels of TEC due to the EIA amplitude. This is because there is higher photoionization effect and more prereversal enhancement effect that could make the fountain effect to be well developed during this period. As reported by Santos et al. (2013) the equatorial E B drift is also strongly dependent on EUV solar flux and the ambient ionosphere, and as a result during solar maximum period the EIA may reach much higher latitudes (as seen from the plots of Porto Alegre station for 2001).
High variabilities in TEC, as observed from the STDs, were observed between 1900 LT and 2300 LT. The months of March–May, September and October exhibited the largest nighttime variation in STD during post sunset followed by February and November. The months of June, July and August indicate lowest nighttime variability at the low latitude stations, as observed from Cachoeira Paulista and Porto Alegre. Generally, the day-to-day analysis revealed that the month of January had very few nighttime variations, which increased through February and had maximum values during the months of March, September and October. Besides the clear evidence from Fejer et al. (1991), Abdu et al. (2007) and Santos et al. (2013) that solar flux plays a principal role
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
29
Fig. 6. The monthly TEC average contour plots for low (left) and high (right) solar activity as a function of month and local time. The TECU has been changed to decibel unit in both cases.
in the vertical prereversal drift enhancements and late high peaks in TEC, another important factor that could be responsible for the evening prereversal enhancement and the consequent nighttime variations in TEC could be associated to the faster decrease of the E region ionospheric conductivity after sunset relative to the F region ionospheric conductivity. 3.3. Seasonal variability in TEC The mean and the standard error bar for hourly seasonal variation in TEC for both 2009 and 2001 are plotted together for better comparison as shown in Fig. 5 to represent the seasonal hourly dispersion of TEC. The blue, green and red colors represent winter solstice, equinox and summer solstice, respectively. Each season is represented by groups of four months GPS-TEC data as
explained earlier in Section 3. This is to demonstrate that there are important seasonal and long term variations in the ionosphere, as explained below. It is possible to observe the following:
The TEC scale for solar maximum is about four times larger than for solar minimum and the highest TEC value stands around 14:00LT. During winter solstice the TEC values are larger at the equator than at the off equatorial stations. A low value of TEC is observed over all the stations for both minimum and maximum solar activities during winter compared with the other seasons.
It can be observed from Fig. 5 that TEC values of summer solstices is higher than equinox during 2009 solar minimum activity at low latitude stations but the variability is larger at equinox. This could result from the more production of electron
30
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Table 3 The comparisons of TEC behavior between 2009 low solar activity and 2001 high solar activity. S/N 2001 1
2 3 4 5
6 7 8
2009
The nighttime resurgence of the anomaly (occasioning a secondary peak in TEC) There is no nighttime resurgence at all stations during low solar flux of 2009. associated to the evening prereversal enhancement is observed at Cachoeira Paulista and Porto Alegre stations, particularly during high solar activity (2001), mainly in the months of January–May and October–December, but not observed at São Luís because it is very close to the dip equator, and very little peaks are shown in Arequipa station because it is relatively further away from the equator compared to São Luís station. The São Luís station shows slightly larger TEC enhancement in some months The Arequipa station shows slightly larger TEC enhancement in almost all the compared to the Arequipa station. months compared to the São Luís station. The TEC and STD in Fig. 6 during daytime and nighttime at both Cachoeira The TEC and STD are closely of the same value for both the low latitude stations Paulista and Porto Alegre exhibit different levels of TEC due to the EIA amplitude. during 2009 periods. Our results suggest that prereversal peaks are dominant during equinoxes for There are no prereversal peaks for all seasons during 2009 solar minimum. 2001 solar phase, which signifies a semidiurnal variation. Fig. 5 shows that there is no much longitudinal difference between Arequipa and Fig. 5 shows larger TEC peaks for summer months of November and December of São Luís for 2001, except that São Luís exhibits slightly larger TEC in 2001 period 2009 at the low latitude stations of Cachoeira Paulista and Porto Alegre than during summer months of January and February of the same year than Arequipa for same period and Arequipa shows smaller peaks at sunset compared to São Luís during equinox of 2001 period. STD shows up to 30 TECU in 2001 the highest value observed in the high solar The highest STD value observed in the 2009 solar minimum is 6 TECU period. Generally, in both 2009 and 2001 solar phases we can observe the seasonal peaks between summer and winter (annual variation) and the equinox maximum compared to the solstices (semiannual variation). The solar flux variations between the two extreme solar activities are clearly reflected by the value of TEC between the two periods.
by photoionization at the summer solstices hemisphere enhancing the background electron density at the region. The seasonal effect can be clearly observed particularly between summer and winter solstices (e.g. December solstice and June solstice), and the equinox and solstices of our results during both solar activities. The TEC variability (obtained from the STD) is reduced from by about 30% for stations at the equator and 200% for stations at the crest of the EIA from December to June solstice conditions for both solar activities. The Sun–Earth distance between June and December period and the Earth’s inclination could be a causative factor (Jee et al., 2004).
Another obvious TEC-EIA characteristic observed during 2001 solar maximum activity is that TEC values during equinox are higher than the TEC values during the summer solstices of the same year and for all stations. This is mainly because the sun overhead is around the equator during equinoxes, which could lead to more intense ionospheric current in E region associated with eastward electric field and, consequently causing a welldeveloped EIA than during the solstices (Wu et al., 2004; Abdu et al., 2007). This equinox–solstices variation is referred to as semiannual variation. The semiannual variation of the EIA could also be due to the combined effect of the solar zenith angle and
Fig. 7. Distribution of TEC variation as function of season and location. The vertical axis represents number of occurrence and the horizontal axis stands for residual TEC. The blue color represents low solar activity and the red represents the high solar activity. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
31
Fig. 8. The Power spectra analysis distribution of 2009 hourly average TEC contour plots for Cachoeira Paulista, Porto Alegre, Arequipa and São Luís from top to bottom respectively. The vertical axis represents the daily period and the horizontal axis represent the day of the year.
magnetic field geometry (Wu et al., 2004). On the other hand, during the solstices, photoionization at the equator decrease because the Sun overhead moves to higher latitudes and fountain effect is expected to be low (Kumar et al., 2014). Lastly, the TEC daily peak show semiannual variation with highest peak during the equinox in 2001 solar maximum (Fig. 5), while the daily highest peak of TEC variation is observed during the summer of 2009 solar minimum (Fig. 5), and lowest daily peaks (troughs) are observed during winter solstice for both solar activities. This implies that equinoctial semiannual variation depend largely on solar flux.
Generally, the seasonal and semiannual variations vary from one station to another. For example, during 2009 low solar activity,
the low-latitude stations present higher TEC daily peaks during summer and equinox, while the stations close to the equator (São Luís and Arequipa) exhibit low anomaly peaks. On the other hand, during winter solstice TEC daily peaks are larger at equatorial stations than at low latitude stations. The reason for this has been explained above. However taken a close look at Fig. 5, it can be observed that during summer solstice the equatorial stations presents slightly larger TEC daily peaks than the low latitude station, while during equinox period larger TEC daily peaks are observed at the lower stations than at the equator. This could be due to wind that is more effective during solar maximum and transequatorial winds blowing from the summer to the winter hemisphere which act to transport plasma towards the equator. Thus wind-driven migration of large density atomic oxygen [O]
32
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
may result in increased O/N2 density ratio during summer (hence comparative low TEC value is observed at low latitudes) and towards the low latitudes during equinox (hence lower TEC values are observed at equator than at lower latitudes). It is important to mention that unlike 2009 low solar activity that exhibited low F10.7 seasonal variation, 2001 high solar activity was associated with high F10.7 seasonal variation. Therefore further observation is needed to clarify the wind contribution for the 2001 TEC seasonal variation. 3.4. Summary of TEC variations Fig. 6 shows color contour plots of monthly average of TEC for the low solar activity (left panel) and high solar activity (right panel) as a function of month and local time, longitude (Arequipa and São Luís) and latitude (São Luís, Cachoeira Paulista and Porto Alegre). In order to use the same scaling system for both solar
activity plots, we used:
⎛ TEC ⎞ ⎟ 20 log10 ⎜ ⎝ TECref ⎠
(5)
where TECref ¼ 1TECU. Solar cycle 23 is characterized by both extreme minimum (2009) and extreme maximum (2001) activity period that is unique in many ways and it is therefore important to observe the extent of the effect on TEC variation. The interesting facts in Fig. 6 are that it is possible to observe the daily, monthly, seasonal, longitudinal and latitudinal variation as function of time. The main TEC features for both low and high solar activities of 2009 and 2001 respectively are outlined in Table 3. 3.5. General TEC distribution Fig. 7 presents the residual distribution of average seasonal
Fig. 9. (a) The Power spectra analysis distribution of 2009 solar minimum daily average TEC contour plots of Cachoeira Paulista, Porto Alegre, Arequipa and São Luís from top to bottom respectively not considering the daily and semiannual periodicities. The vertical axis represents the daily period and the horizontal axis represent the day of the year. (b) Power spectra analysis distribution of 2009 solar minimum daily average zonal wind, meridional winds (both at ∼100 km) and solar flux from top to bottom respectively. The vertical axis represents the daily period and the horizontal axis represent the day of the year. (c) The Power spectra analysis distribution of 2001 solar maximum daily average TEC contour plots of Cachoeira Paulista, Porto Alegre, Arequipa and São Luís from top to bottom respectively not considering the daily and semiannual periodicities. The vertical axis represents the daily period and the horizontal axis represent the day of the year. (d) Power spectra analysis distribution of 2001 solar maximum daily average zonal wind, meridional winds and solar flux from top to bottom respectively. The vertical axis represents the daily period and the horizontal axis represent the day of the year.
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
33
Fig. 9. (continued)
change in TEC pattern (ΔTEC ) determined by: ΔTEC = (TEC − TECs )), where TECs is the difference between the average values of each seasonal TEC from each data set of the season in particular (TEC). The vertical axis represents number of occurrence and the horizontal axis stands for residual TEC. The blue color represents low solar activity and the red represents the high solar activity. The TEC distribution ratio of 1:4 TECU during equinox and summer solstice and a ratio of 1:3 during winter solstice are observed. Generally low solar activity of 2009 exhibits high kurtosis around zero, which confirms the low variability observed at earlier results during this period. Also during low solar activity period TEC distribution pattern exhibit a form of Gaussian distribution around zero mean, indicating that on average TEC experience as many increases as decreases. This characteristic is observed for all the stations. However, during high solar activity of 2001 the opposite is the case. The kurtosis is flat confirming the high standard deviation that was observed during the period. The TEC patterns skew left but spread out to the right during winter indicating more decreases than increases in TEC. Otherwise, during summer solstice TEC pattern skew right and also spread out to the right in most of the plots indicating more increases than decreases in TEC. Fig. 7 also reveals that during the equinox data form a close normal distribution in all stations showing as much increase as decreases
in TEC, except in Arequipa station. 3.6. Wavelet analysis In this section we show other mechanisms that could be responsible for TEC variation during magnetic quiet periods. It is well known that the nonlinear interaction of quasi-stationary planetary waves that are trapped below the mesopause can give origin to large changes in both migrating and non-migrating tides (Goncharenko et al., 2010). These tides are known to exhibit an amplitude maximum at low-latitudes and can propagate into the lower thermospheric region. Moreover, the propagating tides can modulate electric fields through the ionospheric wind dynamo at E region and consequently they map along magnetic field lines to high altitudes of F region height and then influence ionospheric variability (Sripathi and Bhattacharyya, 2012). Jee et al. (2005) used a one dimensional middle-latitude ionospheric model to study the sensitivity of quiet-time TEC to the atmospheric and ionospheric parameters including the neutral wind and found out that during both day and night, the magnetic meridional component of the neutral wind significantly affect TEC, and the geographic zonal wind can cause noticeable longitudinal variation in TEC due to the longitudinal variation in the declination
34
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
angle, which is more prominent in the South American sector (the region in focus). On these evidences and since we are working with only geomagnetic quiet period, it is important to observe the effect of upward propagating waves on the E layer of which may have consequent effect on TEC as explained above. To carry out this investigation we applied wavelet spectral analysis distribution of the daily means of meridional and zonal winds at the altitude of ∼100 km, and compared their periodicities with the daily means of TEC data over Cachoeira Paulista, Porto Alegre, São Luís and Arequipa stations. Apart from the well-known daily peaks and annual and semiannual periodicities such as 27, 16, 1–5 days are found to be dominant and consistent. Fig. 8 is the wavelet spectral analysis plots of hourly mean TEC for the solar minimum year of 2009. The one day peak period along the year and the half day peak period with a semiannual anomaly, can be clearly seen in the figure particularly over Cachoeira Paulista and Porto Alegre. However the semiannual period is not so obvious over São Luís and Arequipa stations. In order to see clearly the other periodicities, we eliminated the daily and the semiannual periodicities by using only the daily averages in the following analysis. The remaining periodicities are clearly shown in Fig. 9(a) which shows the wavelet analysis distribution of daily average of TEC for 2009. Periodicities of 1–5 days are stronger at Arequipa and São Luís stations, 16 days and 27 days are also observed at Cachoeira Paulista and Porto Alegre. Likewise, to explain the TEC periodicities plotted in Fig. 9(a), we made the wave analysis for the zonal and meridional wind at ∼100 km height. From top to bottom in Fig. 9(b), is depicted the daily average zonal and meridional winds, and in the last panel the daily mean of F10.7 for 2009. The periods within the black line contour have 95% confidence level and we used the Morlet function (wave number of 6) as a mother wavelet. It can be clearly seen the 27 days solar periodicity indicated by the F10.7 coinciding well with the TEC. Other periodicities with 16 days, which corresponds to the same period in both zonal and meridional winds around the same days, are also prominent. The 16 days periodicities can be associated with lunar tide effects, which causes strong enhancement on the semidiurnal variations. Moreover, strong 1–5 and 7day periods are also observed at all stations, but not so significant in the low latitude stations of Cachoeira Paulista and Porto Alegre. The same wavelet power spectra analysis applied for 2009 were also analyzed for the TEC measurements of solar maximum year 2001. The results are shown in Fig. 9(c) and (d). The intensity of the F10.7 indicated by the 27-day periodicity is stronger as expected for all the stations in solar maximum. The 16 days periodicity is observed only in the meridional wind (Fig. 9(d)) and in TEC for all stations (Fig. 9(c)), while 1–2, 5–7 are observed in TEC, and 8–10 days oscillations are observed in both the zonal and meridional winds. The present results are also in accordance with the features observed from the decomposition analysis presented by Abdu et al. (2006), that reported dominance of 4–5 days and 7 days periods on observations of ionospheric vertical drift velocities and on meridional and zonal winds at ∼100 km height. Generally, these results demonstrate the existence of a strong vertical coupling through upward propagating waves leading to day-to-day oscillation in TEC. A possible electrodynamic coupling mechanism connecting these oscillations could be included in future studies. Oscillations in TEC as observed in the panels of Fig. 9(a) and (c) with different periodicities from 3 to 16 days have been attributed to planetary waves that propagate horizontally and vertically up to the lower thermosphere, affecting the ionosphere in an integrated sense. The planetary waves at equatorial and lowlatitudes are generally Kelvin waves with a zonal wave number from 1 to 3 and periods of 3–4 days (ultra-fast waves), 6–10 days (fast waves) and 15–20 days (slow waves). The quasi- 16-days
periodicities have also been identified as Rossby waves with zonal wave number 1 (Pancheva et al., 2008). The oscillations observed here in the spectral analysis of TEC with periods from 1 to 10 days were more significant during the equinoxes and the December solstice months of the year 2001 for all the stations, corresponding with some of those oscillations also observed in the wind data, which suggests upward propagation of planetary waves from the lower atmosphere. The fast and mainly ultra-fast Kelvin waves are more capable to propagate upward, and are crucial at tropical latitudes for the vertical coupling between the neutral and ionized atmosphere (Phanikumar et al., 2014). However, throughout the year 2009 the results depicted in Fig. 9 revealed that the presence of waves with shorter periods is more significant in the wind data than in the TEC, which suggests that during the solar minimum period the oscillations in the ionosphere might not be potentially coupled with the lower atmosphere through ultra-fast planetary waves propagating upward.
4. Conclusions In this study we have used up to 730 days of dual frequency GPS data to calculate the TEC over four stations located in the South American sector. The analysis were used to describe the TEC variation in the ionosphere in terms of diurnal, day-to-day, seasonal, latitudinal and longitudinal variations during quiet time for low (year 2009) and high (year 2001) solar flux. The results were also used to investigate the strength and characteristics of the ionosphere at the EIA. In addition, using the power wavelet analysis spectra, periodic TEC variations were analyzed and some physical mechanisms responsible for these variations were suggested. We made mass hourly plots of each month, and grouped month into seasons for both phases of the solar cycle. The standard deviation (STD) was also used to describe the local and spatial TEC variability at different locations on the South American sector. Our results indicate that the local TEC variability, as observed in the STD, is largest around the locations of low latitude stations with maximum around afternoon and late in the evening. The largest STD values were between 5 and 6 TECU during solar minimum and between 25 and 30 TECU during solar maximum. TEC variability was larger in equinox followed by summer solstice and low in winter solstice for solar maximum, while for solar minimum larger TEC variability was observed during summer solstice followed by equinox and lowest during winter solstice. The TEC value and TEC variability increased from 2009 solar minimum to 2001 solar maximum by ∼400% and ∼200% respectively. It should be mentioned that during winter solstices for both phases of the solar cycle, TEC diurnal values at the crest of the anomaly region exhibits lower values than TEC at the equatorial regions. This could be because sun radiation during winter solstice is not as intense as during equinox and summer. Therefore fountain effect is not well developed and plasma is not effectively transported off the magnetic equator during this season. An annual seasonal variation is observed in the solstices and a semiannual variation is observed in the equinoxes. In order to investigate mechanisms which are responsible for TEC periodicities as observed in our data, we applied wavelet spectral analysis to daily means of TEC data over Cachoeira Paulista, Porto Alegre, São Luís and Arequipa stations, as well as to the daily means of meridional and zonal winds at the altitude of ∼100 km along with F10.7 to compare their periodicities. Apart from the strong a 27-day period and a 16-day period observed in F10.7 and wind parameter respectively, we also observed that a strong 1–5 and 7-day periods were common both to all the stations but not so strong at both Cachoeira Paulista and
O.F. Jonah et al. / Journal of Atmospheric and Solar-Terrestrial Physics 135 (2015) 22–35
Porto Alegre stations and the two parameters (zonal and medional winds). A day-to-day variation in TEC is reported to contain a component driven by planetary waves enhanced by tides as they propagate upward. A strong vertical coupling through the upward propagating waves can also give rise to day-to-day oscillation in TEC. Our study also indicates that apart from solar radiation effect, changes in the meridional wind or zonal winds can also lead to corresponding changes in TEC.
Acknowledgments Jonah O.F. world like to acknowledge the supports from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) under process 133429/2011-3 and Dr. Paulo Prado Batista for the meteor radar data. E.R. de Paula is grateful to AFOSRFA9550-10-10564 and CNPq under Grant No. 3056842010-8. Marcio Muella would like to thank the support from CNPq under Grant No. 304674/2014-1. The RBMC/IBGE GPS data and magnetic and solar flux data used in this work were obtained from IBGE website http:// www.ibge.gov.br/home/geociencias/download/tela_ inicial.php? tipo¼8) and National Geophysical Data Center (NGDC) at NOAA website (http://www.ngdc.noaa.gov/ngdcinfo/onlineaccess.html) respectively.
References Abdu, M.A., Batista, P.P., Batista, I.S., M. Brum, C.G.M., Carrasco, A.J., 2006. Equatorial evening prereversal electric field and spread F. J. Geophys. Res. . http://dx.doi. org/10.1029/1005GL024837 Abdu, M.A., Brum, C.G.M., Batista, I.S., Sobral, J.H.A., de Paula, E.R., Souza, J.R., 2007. Solar flux effects on equatorial ionization anomaly and total electron content over Brazil: observational results versus IRI representations. J. Adv. Space Res. v. 42, 617–625. Anderson, D.N., Klobuchar, J.A., 1983. Modeling the total electron content observations above Ascension Island. J. Geophys. Res. 88, 8020–8024. Bagiya, M.S., Joshi, H.P., Iyer, K.N., Aggarwal, M., Ravindran, S., Pathan, B.M., 2009. TEC variations during low solar activity periods (2005-2007) near the equatorial ionospheric anomaly crest region in India. Ann. Geophys. (Germany) 27, 1047–1057. Balan, N., Bailey, G.J., Jayachandran, B., 1993. Ionospheric evidence for a non-linear relationship between the solar EUV and 10.7-cm fluxes during an intense solar cycle. Planet. Space Sci. 41, 141–145. Batista, I.S., Abdu, M.A., 2004. Ionospheric variability at Brazilian low and equatorial latitudes comparison between observations and IRI model. J. Adv. Space Res. 34 1894–1894. Bhuyan, P.K., Borah, R.R., 2007. TEC derived from GPS network in India and comparison with the IRI. Adv. Space Res. 39, 830–840. Bilitza, D., 2000. Report from 33rd COSPAR Scientific Assembly, Warsaw, Poland, 16–23 July 2000. IRI News Lett. 7, 1–4. Charney, J.G., Drazin, P.G., 1961. Propagation of planetary-scale disturbances from the lower into upper atmosphere. J. Geophys. Res. 66, 83–110. Codrescu, M.V., Beierle, K.L., Fuller-Rowell, T.J., Palo, S.E., Zhang, X., 2001. More total electron content climatology from TOPEX/Poseidon measurements. Radio Sci. . http://dx.doi.org/10.1029/1999RS002407 Codrescu, M.V., Palo, S.E., Zhang, X., Fuller-Rowell, T.J., Poppe, C., 1999. TEC climatology derived from TOPEX/POSEIDON measurements. J. Atmos. Sol. Terr. Phys. 61, 281–298. http://dx.doi.org/10.1016/S1364- 6826(98)00132-1. Costa, A.M., Boas, J.W.V., Fonseca, J.R., 2004. GPS total electron content measurements at low latitudes in Brazil for low solar activity. Geofís. Int. 43, 129–137. Dabas, R.S., Lakshmi, D.R., Reddy, B.M., 1993. Solar activity dependence of ionospheric electron content and slab thickness using different solar indices. Pure Appl. Geophys. (France) 140, 721–728. Dasgupta, A., Paul, A., Das, A., 2007. Ionospheric Total Electron Content (TEC) studies with GPS in the equatorial region. Indian J. Radio Space Phys. 36, 278–292. de Siqueira, P.M., de Paula, E.R., Muella, M.T.A.H., Rezende, L.F.C., Abdu, M.A.,
35
Gonzalez, W.D., 2011. Storm-time total electron content and its response to penetration electric fields over South America. Ann. Geophys. . http://dx.doi. org/10.5194/angeo-29-1765-2011 Duarte-Silva, M.H., Muella, M.T.A.H., Silva, L.C.C., de Abreu, A.J., Fagundes, P.R., 2015. Ionospheric slab thickness at the equatorial anomaly region after the deep solar minimum of cycle 23/24. Adv. Space Res. . http://dx.doi.org/10.1016/j. asr.2015.05.010 Fejer, B.G., de Paula, E.R., Gonzalez, S.A., Woodman, R.F., 1991. Average vertical and zonal F region plasma drifts over Jicamarca. J. Geophys. Res. . http://dx.doi.org/ 10.1029/91JA01171 Fedrizzi, M., de Paula, E.R., KantorI, J., 2001. Ionospheric effects of the August 26, 1998 magnetic storm over South America, In: International Beacon Satellite Symposium Proceeding, Boston. Forbes, J.M., Palo, S.E., Zhang, X., 2000. Variability of the ionosphere. J. Atmos. Sol. Terr. Phys. 62, 685–693. Goncharenko, L.P., Coster, A.J., Chau, J.L., Valladares, C.E., 2010. Impact of sudden stratospheric warmings on equatorial ionization anomaly. J. Geophys. Res. . http://dx.doi.org/10.1029/2010JA015400 Hocking, W.K., Fuller, B., Vandepeer, B., 2001. Real time determination of meteor related parameters utilizing modern digital technology. J. Atmos. Sol. Terr. Phys. 63, 155–169. Jee, G., Schunk, R.W., Scherliess, L., 2004. Analysis of TEC data from the TOPEX/ Poseidon mission. J. Geophys. Res. . http://dx.doi.org/10.1029/2003JA010058 Jee, G., Schunk, R.W., Scherliess, L., 2005. On the sensitivity of total electron content (TEC) to upper atmospheric/ionospheric parameters. J. Atmos. Sol. Terr. Phys. 67, 1040–1052. Jonah, O.F., de Paula, E.R., Kherani, E.A., Dutra, S.L.G., Paes, R.R., 2014. Atmospheric and ionospheric response to stratospheric sudden warming of January 2013. J. Geophys. Res. 119. http://dx.doi.org/10.1002/2013JA019491. Kane, R.P., 2003. Day-to-day variability of quiet-time ionospheric foF2. Indian J. Radio Space Phys. 32, 344–348. Kumar, S., Singh, A.K., Lee, J., 2014. Equatorial ionospheric anomaly and comparison with IRI model during descending phase of solar activity (2005–2009). Adv. Space Res. 53, 724–733. Muella, M.T.A.H., Kherani, E.A., de Paula, E.R., Cerruti, A.P., Kintner, P.M., Kantor, I.J., MitchelL, C.N., Batista, I.S., Abdu, M.A., 2010. Scintillation-producing Fresnelscale irregularities associated with the regions of steepest TEC gradients adjacent to the equatorial ionization anomaly. J. Geophys. Res. 115. http://dx.doi. org/10.1029/2009JA014788. Otsuka, Y., Ogawa, T., Saito, A., Tsugawa, T., Fukao, S., Miyazaki, S., 2002. A new technique for mapping of total electron content using GPS network in Japan. J. Earth Planets Space 54, 63–70. Pancheva, D.V., Mukhtarov, P.J., MitchelL, N.J., Fritts, D.C., Riggin, D.M., Takahashi, H., Batista, P.P., Clemesha, B.R., Gurubaran, S., Ramkumar, G., 2008. planetary wave coupling (5–6-day WAVES) in the low-latitude atmosphere–ionosphere system. J. Atmos. Sol. Terr. Phys. 70, 101–122. Phanikumar, D.V., Niranjan, K.K., Sanjay, K., 2014. Signatures of ultra fast kelvin WAVES in low latitude ionospheric TEC during January 2009 stratospheric warming event. J. Atmos. Sol. Terr. Phys. 117, 48–53. Rabiu, A.B., Mamukuyomi, A.I., Joshua, E.O., 2007. Variability of equatorial ionosphere inferred from geomagnetic field measurements. Bull. Astron. Soc. India 35, 607–618. Rama Rao, P.V.S., Gopi, K.S., Niranjan, K., Prasad, D.S.V.V.D., 2006. Temporal and spatial variations in TEC using simultaneous measurements from the Indian GPS network of receivers during the solar activity period of 2004–2005. Ann. Geophys. 24, 3279–3292. Rishbeth, H., Mendillo, M., 2001. Patterns of F2‐layer variability. J. Atmos. Sol. Terr. Phys. 63, 1661–1680. Santos, A.M., Abdu, M.A., Sobral, J.H.A., Mascarenhas, M., Nogueira, P.A.B., 2013. Equatorial evening prereversal vertical drift dependence on solar EUV flux and F10.7 index during quiet and disturbed periods over Brazil. J. Geophys. Res. 118, 4662–4671. http://dx.doi.org/10.1002/jgra.50438. Scherliess, L., Thompson, D.C., Schunk, R.W., 2008. Longitudinal variability of lowlatitude total electron content: Tidal influences. J. Geophys. Res. . http://dx.doi. org/10.1029/2007JA012480 Sripathi, S., Bhattacharyya, A., 2012. Quiet time variability of the GPS TEC and EEJ strength over Indian region associated with major sudden stratospheric warming events during 2005/2006. J. Geophys. Res. 117. http://dx.doi.org/ 10.1029/2011JA017103, A05305. Tsurutani, B.T., Verkhoglyadova, O.P., Mannucci, A.J., Saito, A., Araki, T., Yumoto, K., Tsuda, T., Abdu, M.A., Sobral, J.H.A., Gonzalez, W.D., Mccreadie, H., Lakhina, G.S., Vasyliünas, V.M., 2008. Prompt penetration electric fields (PPEFs) and their ionospheric effects during the great magnetic storm of 30–31 October 2003. J. Geophys. Res. . http://dx.doi.org/10.1029/2007JA012879 Wu, C.C., fry, C.D., Liu, J.–Y., Liou, K., Tseng, C.L., 2004. annual TEC variation in the equatorial anomaly region during the solar minimum: September 1996–August 1997. J. Atmos. Terr. Phys. 66, 199–207.