On the response of the ionospheric F region over Indian low-latitude station Gadanki to the annular solar eclipse of 15 January 2010



[1] The response of the ionospheric F region over Indian low-latitude regions to the annular solar eclipse of 15 January 2010 is investigated. The foF2 corresponding to an electron density increase of ∼21% at the F2 peak is seen over Gadanki (13.5°N, 79°E) during the course of the eclipse in comparison with the control day behavior. After the peak phase the foF2 shows a large decrease (∼19%) compared to the mean control day pattern. The total electron content (TEC) at Bangalore (13°N, 78°E) which is located very close to Gadanki is expected to follow a similar pattern of temporal evolution. This TEC shows reduction with respect to control day both at the peak phase (17%) and in the postpeak phase (30%). The enhanced foF2 from the start to the peak phase of the eclipse is attributed to the effect of the weakened equatorial ionization anomaly (EIA). At altitude regions below 270 km, the eclipse induced cutoff of solar insolation results in chemical recombination becoming dominant and thus contributes to the decrease in columnar content in spite of foF2 increase. The post peak phase steep decrease of both foF2 and TEC is attributed to the substantial increase in the poleward meridional winds, the inhibition of the EIA, and persistent depletion in the lower-altitude electron densities. In summary, this study demonstrates the modifications in electrodynamics, recombination, and neutral dynamics acting in concert to produce the observed effects at low latitudes during an eclipse.

1. Introduction

[2] Solar eclipses are celestial events which have evoked human curiosity since time immemorial. The impact of solar eclipses on Earth's atmosphere has been the subject matter of many studies. Rishbeth [1968] described the eclipse effects in ionospheric E and F regions taking theoretical factors into consideration. Studies involving the depletion of E and F1 electron densities [Van Zandt et al., 1960], induced variations of electron and ion temperature, changes in drift velocity [Baron and Hunsucker, 1973], generation of gravity waves [Chimonas and Hines, 1970] and occurrence of counter electrojet [St.-Maurice et al., 2010] have been presented by different workers. Variation in F2 parameters in response to solar eclipses have been reported at various geomagnetic latitudes. Evans [1965] reported an increase in foF2at temperate latitudes and attributed the phenomenon to variations in electron and ion temperatures resulting in downward diffusion of ionization from heights above F2 layer. He also pointed out that this effect is mainly relevant at midlatitude to high-latitude regions. Rashid et al. [2006] have reported the response of high-latitude ionosphere to a solar eclipse which occurred during the recovery phase of a magnetic storm. Latitudinal dependence of the ionospheric response was discussed in detail by Le et al. [2009], and they reported smaller NmF2 changes during eclipses at lower latitudes compared to mid latitudes. A reduced thickness of F2 layer was observed by Chandra et al. [2007] during a total solar eclipse at middle latitudes of 11 August 1999. Modeling studies related to solar eclipse also suggest an enhancement in foF2 due to contraction and subsequent reduction in [O]/[N2] ratio post eclipse [Müller-Wodarg et al., 1998] at mid latitudes. Tomás et al. [2007] have studied the effects of a solar eclipse in 2005 at an equatorial station using the data from CHAMP satellite wherein the modulation of electrojet and an asymmetry in the EIA due to the effect of solar eclipse is discussed. Momani et al. [2010] have reported multi-instrument analysis of a solar eclipse at high-latitude regions. They have shown clear depletion in electron density, electron temperature, ion velocity and plasma cutoff frequency during the solar eclipse passage at different stations. Their radar measurements show clear differences in the ionospheric response between the E and F layers of ionosphere and between ion and electron temperature in the F layer.

[3] On Friday, 15 January 2010, an annular eclipse of the Sun was witnessed from within a 300 km wide track that traversed half of Earth. The track of the Moon's shadow began in westernmost Central African Republic at 05:14 UT. The eclipse path is 371 km wide at its start as the antumbra quickly travels east-southeast. The antumbral shadow passed through Chad, Central African Republic, Democratic Republic of the Congo, Uganda, Kenya, and Somalia. After leaving Africa, the path crosses the Indian Ocean and continues into Asia through India, Bangladesh, Burma (Myanmar), and China. The first contact at the city of Trivandrum, a magnetic equatorial station (8.5°N, 77°E) was at 11:04 India Standard Time (IST) and persisted up to 15:05 IST, the time of peak of obscuration ∼13:15 IST. At the low-latitude station Gadanki, the time of first contact was ∼11:16 IST, peak of obscuration ∼13:30 IST and the time of fourth contact ∼15:11 IST. Choudhary et al. [2011] examined the response of equatorial and low-latitude regions over India to this solar eclipse using total electron content (TEC) (different locations), magnetic field (different locations) and ionosonde at magnetic equator. They have attributed the eclipse and post eclipse variations of TEC at low-latitude stations to inhibition of the EIA. In the present study we are examining the same event, with the objective of explaining the unique response of the ionosphere over low-latitude station Gadanki and nearby stations. Our analysis highlights the new and interesting aspect that, the responses of foF2 and TEC near low-latitude region near Gadanki are divergent. The detailed analysis and interpretation are given in the subsequent sections.

[4] The noontime occurrence of the solar eclipse with the path of maximum obscuration crossing over the magnetic equator on 15 January 2010 proved to be ideal to investigate the equatorial and low-latitude ionospheric response. The maximum obscuration of the eclipse at the low-latitude station Gadanki (13.5°N, 79°E) was 82% at ∼13:30 IST.

2. Data and Method of Analysis

[5] The details of the data used and the method of estimating the different parameters are explained in the following subsections.

2.1. Magnetic Field

[6] The magnetic field (H) data is acquired from the magnetometer at the magnetic equatorial location Tirunelveli (TIRU) (8.73°N, 77.73°E). ΔH values are deduced by subtracting the mean nighttime level from the magnetic field H.

2.2. Ionosonde NmF2 and Electron Density

[7] Ionosonde data from Gadanki (13.5°N, 79°E) is utilized to study the temporal variations in electron density. The electron density versus real height profiles are derived from the manually scaled ionogram data using POLAN software. The ionosonde data from Trivandrum (TRV) is used in conjunction with that at Gadanki to study the evolution of the EIA crest.

[8] The electron density is obtained from the foF2 at Gadanki using the relation,

display math

where NmF2 is the peak electron density of F2 region and foF2 is the maximum frequency in MHz reflected from the F2 region [Rishbeth and Garriott, 1969].

2.3. Meridional Wind

[9] The meridional wind estimation is made using ionosonde data from TRV and Gadanki. The method given by Krishna Murthy et al. [1990] is used for estimating the meridional winds. Here, the vertical drift of the F layer is estimated for the two stations (TRV and Gadanki) after accounting for the effects due to recombination and diffusion. The F region vertical drift at the magnetic equatorial region over Trivandrum is purely electro dynamical in nature while that at a low-latitude station like Gadanki has contribution from the meridional winds (U) in addition to diffusion along the magnetic field lines. The observed vertical drift velocities (Vo) are initially derived from the rate of change of h'F (d(h'F)/dt). The true vertical drift is obtained from Vd = Vo-βH, where β is the effective recombination coefficient and H is the scale height given by H = (1/N dN/dh)−1. N is the electron density and βH is the correction due to recombination. The actual meridional wind for the period of study is estimated from the equations for the vertical drift at the two stations as

display math

where, VD and V are the vertical drifts over Trivandrum and Gadanki, respectively. I is the dip angle at Gadanki and WD the plasma drift velocity due to plasma diffusion and is given by g/υin, where g is the acceleration due to gravity and υin is the ion neutral collision frequency. The error in the meridional wind estimation using this method is estimated to be about ∼25 m s−1 [Krishna Murthy et al., 1990]. This method is used generally for estimating meridional winds at nighttime, in the absence of production. In the present study, we have used this method for estimating daytime winds (during the time of solar eclipse) as the conditions are similar to nighttime because of the solar eclipse induced cutoff of solar insolation. The accuracy of the wind estimates increases progressively as we approach the time of maximum obscuration (∼13:30 IST).


[10] The GPS data from low-latitude station Bangalore (13°N, 78°E), quite close to Gadanki is used as a proxy for the TEC over Gadanki. In addition to this GPSTEC data from the stations of Trivandrum (8.5°N, 77°E), Hyderabad (17.5°N, 78°E) and Lucknow (26°N, 81°E) are used in this study. Carrier phase delays and pseudo ranges of the GPS signals at L1 and L2 frequencies are used to obtain the Absolute Slant GPSTEC (STEC). The STEC are then converted to Absolute Vertical TEC (VTEC) using the mapping function as given below:

display math

where, χ is the zenith angle at ionospheric pierce point (IPP) which is estimated from the satellite elevation angle. The shell height is taken as 350 km. Rama Rao et al. [2006] have shown that an elevation angle cutoff of >50° is ideally suited to represent the TEC over the Indian sector and hence the present analysis is based on this criterion. Averaged VTEC values are obtained by averaging every 15 min values of VTEC from satellites with elevation angle above 50°.

[11] The geographic and geomagnetic coordinates [Finlay et al., 2010] of all the stations along with the instrument used for the present study are given in Table 1.

Table 1. Geographic Coordinates, Geomagnetic Latitudes, and Instruments Used From Various Stations
StationGeographic Latitude (N)Geographic Longitude (E)Geomagnetic Latitude (N)Instrument

3. Results and Discussion

3.1. Temporal Evolution of ΔH on Eclipse Day Vis-à-Vis Mean Control Day Pattern

[12] Figure 1 depicts the temporal variation of ΔH on the eclipse day with respect to the mean control day pattern. For the control day mean, the ΔH values increase from around 11 nT at 08:00 IST to reach a peak value of around 54 nT at 10:45 IST. Thereafter it continuously decreases up to 19:00 IST. As against this normal pattern, on the eclipse day it increases from a value of 10 nT at 08:00 IST and steeply rises to around 28 nT ∼09:00 IST. After this ΔH starts decreasing gradually up to around 12:00 IST, when there is a short enhancement followed by a steep decrease up to around 14:30 IST which is the time when the counter electrojet [St.-Maurice et al., 2010] attains maximum strength. After 14:30 IST the magnetic field gradually starts recovering toward normal electrojet conditions. The ΔH value which represents the overhead currents in the E region is a function of the conductivity and the ionospheric zonal electric field. It is clear from the eclipse day variations that there is a gradual weakening followed by the reversal of electric field resulting in the counter electrojet. This ionospheric electric field is the driving force of the EIA. This implies a weakening/inhibition of the fountain effect over most part of the day. Therefore by the time of peak eclipse, the fountain effect is significantly weakened compared to the control day pattern.

Figure 1.

Temporal variation of ΔH of eclipse day at Thirunelveli vis-à-vis the control day mean pattern.

3.2. Response of the Thermospheric Neutral Wind to the Eclipse

[13] A schematic illustrating the movement of the maximum obscuration region and consequent expected pattern of the meridional wind is shown in Figure 2a. The estimated meridional wind for the period from 12:00 to 14:00 IST is shown in Figure 2b. In the time period close to 12:00 IST, the maximum obscuration region is south of the geographic equator and hence the heating up of the geographic equatorial region is continuing unabated. During this time a poleward wind is expected in the northern hemisphere and this is validated by the estimated winds also. Around 12:45 IST, the maximum obscuration region comes over the geographic equator and the resultant cooling occurs there. This is expected to produce an equator ward wind in both hemispheres and the estimated winds again show the anticipated pattern. In the time period ∼13:45 IST, when the maximum obscuration region is already north of Gadanki, the convergence in the northern hemisphere is expected to be poleward and this is borne out by the estimated winds which turn strongly poleward during this time. In fact, the poleward wind with magnitude of ∼250 m s−1 is observed during this time. Such large magnitude winds can significantly contribute to the movement of ionisation away from Gadanki.

Figure 2.

(a) A schematic illustrating the movement of the zone of maximum obscuration and the consequent response of the thermospheric meridional neutral wind on the eclipse day (not drawn to scale). (b) Temporal variation of thermospheric neutral wind close to the time of maximum obscuration on eclipse day at Gadanki.

3.3. Temporal Evolution of foF2 and EIA at Gadanki on Eclipse Day in Relation to the Mean Control Day Pattern

[14] Figure 3 (left) shows the time variations of foF2 at Gadanki on the eclipse day along with the mean control day pattern. The control mean pattern of foF2 shows an increase in the morning up to around 09:45 IST after which it decreases gradually to attain a minimum around noon hours. Later on the foF2 increases up to ∼15:30 IST before falling off toward late evening. This temporal evolution pattern is characteristic of the days when the ionospheric electric field is strong and attains a significantly high peak value close to noon (as seen in section 3.1). The high electric field on such days results in a significant noontime bite out in foF2 associated with the movement of the crest toward higher latitudes [Chandra et al., 2009].

Figure 3.

(left) The foF2 at Gadanki and (right) equatorial ionization anomaly (EIA) proxy variation on eclipse day along with the mean control day pattern.

[15] On the eclipse day, the foF2 is significantly higher than the control day pattern right after ∼09:00 IST up to around noon. This high value in the morning hours is attributed to the fact that the ionospheric electric field is significantly weaker on the eclipse day right from morning as shown in section 3.1 and this confines the EIA crest to lower latitudes around Gadanki unlike on the control day when the strengthening of the electric field after 09:00 h is expected to push the crest to higher latitudes [Chandra et al., 2009]. As for the cause of the weakened electric field from 09:00 h, this problem is already addressed by St.-Maurice et al. [2010]. They have discussed in detail on how the EEJ growth was interrupted after 09:00 IST, one half hour after the eclipse had started in Africa. They point out that by then the obscuration region was approaching Trivandrum from the west, and the influence of the Counter Electrojet (CEJ) was already being felt.

[16] Further, the foF2 remains more or less steady up to ∼13:45 IST indicating that pumping of ionization away from Gadanki is not taking place. The conspicuous absence of the noontime bite out in foF2 is indicative of the fact that the crest has not moved beyond Gadanki. Also, in order to confirm that the EIA is only weakened (such that the crest reaches only up to latitudes near Gadanki) and not fully inhibited with crest returning to Trivandrum, we have examined the time evolution of foF2 (Gadanki)/foF2 (TRV) which is shown in Figure 3 (right). This EIA proxy parameter (referred to as EIA proxy in Figure 3) exceeds 1 when the EIA crest is at or beyond Gadanki. It is evident that up to 14:00 IST EIA proxy is above 1, and 14:30 onward it is below 1. From the above discussion it is obvious that the crest is near Gadanki latitudes up to the peak phase of the eclipse. After 13:45 IST (post maximum obscurity) we see a steep fall in foF2. This fall in foF2 coincides with large increase in the poleward meridional neutral wind, which pushes the ionization to lower altitudes away from Gadanki. Between 13:45 and 14:30 IST, foF2 decreases by ∼21%. Part of this reduction is also due to the inhibition of the anomaly which is shown to occur after 14:00 IST in Figure 3 (right). Therefore, the decrease in foF2 (after maximum obscuration) is the combined effect of neutral dynamics and inhibited electrodynamics, while the higher values during the course of the eclipse is due to the effect of weakened electrodynamics. The maximum percentage increase in foF2 observed with respect to control day pattern in the period before the peak phase is 21% at ∼12:30 IST. In the post maximum obscuration phase the maximum percentage deviation in foF2 is 19% at 15:30 IST.

3.4. Altitude Structure of Electron Density on Eclipse Day in Relation to a Sample Quiet Day

[17] Altitude profiles of electron density up to the F2 peak on eclipse day and a quiet day are shown in Figure 4. It is readily seen that from 08:00 IST up to 11:00 IST, the electron density profiles encompass similar altitude regions up to the F2 peak. It is also clear from Figure 4 that the foF2 value is higher on the eclipse day even before the eclipse (profiles of 09:00–11:00 IST) as mentioned earlier. The profiles close to the eclipse peak phase (from 12:00 IST) show that although the foF2 is higher on the eclipse day, there is large reduction in electron density at lower altitudes below 270 km. This reduction in the lower-altitude regions is attributed to the dominant chemical recombination due to eclipse induced cutoff of solar insolation. This in turn gives rise to a significant reduction in the columnar content on the eclipse day in spite of the higher values of foF2. The dominance of recombination at altitude regions below 300 km, during nighttime when significant production is absent, has been shown by a number of researchers [Bittencourt and Abdu, 1981; Krishnamurthy et al.,1990]. During the eclipse similar conditions are prevailing with the production being drastically curtailed. In the post eclipse profiles (14:00 and 15:00 IST) also, the bottom side electron densities have not recovered substantially because of the absence of significant production in the late afternoon hours. Thus, the dominant chemical recombination contributes significantly to the reduced columnar content during the course of the eclipse and afterward.

Figure 4.

Electron density profile of the ionosphere during peak time of eclipse and at the same time during the control day at Gadanki.

3.5. Temporal Evolution of GPSTEC on Eclipse Day Vis-à-Vis Mean Control Day Pattern

[18] Figure 5a shows the time variation of GPSTEC at Bangalore on the eclipse day along with the mean control day pattern. The GPSTEC at Bangalore which is located very close to Gadanki is expected to follow similar pattern of temporal evolution as that at Gadanki. At Bangalore, GPSTEC is higher on eclipse day with respect to control day pattern around 09:30 IST because of the presence of EIA crest near Bangalore (as has been discussed in section 3.3). It is seen that GPSTEC decreases significantly from ∼12:00 IST. This is around the time when electron density shows significant depletion at lower altitudes (discussed in section 3.4). The reduction in electron density at lower altitudes continues in the post maximum obscuration phase because, by that time, it is well past noon and the solar zenith angle is increasing. Hence the ionization production is much less. Therefore, after 13:45 IST the TEC decreases more steeply because of the combined effects of neutral dynamics (section 3.2), inhibited electrodynamics (section 3.3) and dominant chemical recombination (section 3.4) at lower altitudes. The maximum percentage decrease in GPSTEC around the eclipse peak phase is ∼17% while that in the post eclipse phase is ∼30% (around 15:30 IST).

Figure 5.

(a) GPSTEC variation at Bangalore on the eclipse day in relation to control day mean. (b) GPSTEC variation at Hyderabad on the eclipse day in relation to control day mean. (c) GPSTEC variation at Lucknow on the eclipse day in relation to control day mean.

[19] The time variation of GPSTEC at Hyderabad and Lucknow are shown in Figures 5b and 5c. Hyderabad is a location situated 4.5° north of Gadanki and experienced 67.5% maximum obscuration during the eclipse, whereas Lucknow which is 13° north of Gadanki, is a station with less than 40% obscuration. It is evident that the reduced TEC observed with respect to the control day mean pattern during and after the peak phase is seen for Hyderabad also as was seen for Bangalore. For Lucknow, which is outside the eclipsed region, the TEC pattern is found to follow the mean control day pattern, thereby clearly indicating that the TEC decrease observed at the lower latitudes is indeed an eclipse induced effect.

4. Conclusions

[20] 1. Large increase in foF2 and simultaneous decrease in TEC is observed over low latitudes near Gadanki around the eclipse peak phase. The increased foF2 is attributed to the presence of EIA crest over Gadanki as a result of weakened electrodynamics. The dominant chemical recombination at lower altitudes due to the cutting off of solar insolation and consequent reduction in electron densities leads to a significant reduction in the TEC in spite of an increase in foF2.

[21] 2. Substantial reduction is observed in foF2 and TEC in the post peak phase compared to mean control day pattern over low latitudes near Gadanki. This effect is attributed to the combined effects of neutral dynamics, inhibited electrodynamics and persistent depletion of lower-altitude electron densities.


[22] This work was supported by Department of Space, Government of India. One of the authors, M. K. Madhav Haridas, gratefully acknowledges the financial assistance provided by the Indian Space Research Organization through Research Fellowship.

[23] Robert Lysak thanks the reviewers for their assistance in evaluating this paper.