Signature of the 29 March 2006 eclipse on the ionosphere over an equatorial station

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112,, doi: /2006ja012197, 2007 Signature of the 29 March 2006 eclipse on the ionosphere over an equatorial station J. O. Adeniyi, 1,2 S. M. Radicella, 1 I. A. Adimula, 2 A. A. Willoughby, 2 O. A. Oladipo, 2 and O. Olawepo 2 Received 29 November 2006; accepted 30 January 2007; published 16 June [1] This study documents some results of the effect of the 29 March 2006 eclipse on the ionosphere over Ilorin, Nigeria (longitude 4.57 E, latitude 8.53 N, dip 4.1 S), an equatorial station in the West African region. The maximum obscuration of the eclipse at this station was 99 percent and it occurred before midday. True height electron density profile analysis below the F2 peak was employed in the study. The effect on the E and F1 layers was a drastic decrease in electron density, with maximum decrease percentages of 60 and 68 for the E and F1 layers, respectively. A decrease in fof2 began at about 1 hour 20 min after those in the lower layers had started. Variation of electron density with height showed that the decrease in the electron density occurred through out the E and F1 heights at about the same time while that of the F2 region began at lower heights and extended progressively toward the peak of electron density height of the layer. The recovery in the E and F1 layers has already reached an advanced stage before the effect of the eclipse got to the maximum in the F2 region. A major departure of hmf2 from the normal variation was observed and discussed. Citation: Adeniyi, J. O., S. M. Radicella, I. A. Adimula, A. A. Willoughby, O. A. Oladipo, and O. Olawepo (2007), Signature of the 29 March 2006 eclipse on the ionosphere over an equatorial station, J. Geophys. Res., 112,, doi: /2006ja Introduction [2] The region of West Africa had the rare experience of having the occurrence of a total eclipse of the Sun before midday on 29 March The occurrence of an eclipse of this nature generally gives atmospheric scientist an opportunity of making special observations with regards to the solar control of the Earth s atmosphere. For example, the effects of 11 August 1999 solar eclipse on the ionosphere over Western Europe were documented by a number of authors [Baran et al., 2003; Farges et al., 2003; Farges et al., 2001; Altadill et al., 2001]. The annular eclipse of 23 September 1987 gave an opportunity to study the ionospheric effect around the equatorial anomaly crest in the west Pacific region [Cheng et al., 1992]. [3] The path of the total solar eclipse of 29 March 2006 across parts of Benin and Nigeria located in West Africa is shown in Figure 1 [Espenak and Anderson, 2004]. The central path of the occurrence of complete (i.e., 100%) obscuration is indicated by the thick line, almost diagonal across the map. The northern and southern limits are shown by the thinner lines, parallel to the central one. The time (in UT) of occurrence of the eclipse totality, its duration (minutes and seconds), and the Sun s 1 Aeronomy and Radio Propagation Laboratory, Abdus Salam International Centre for Theoretical Physics, Trieste, Italy. 2 Physics Department, University of Ilorin, Ilorin, Nigeria. Copyright 2007 by the American Geophysical Union /07/2006JA altitude in degrees are shown in the squares beside the line of the northern limit. The duration of totality decreases with distance on both sides of the central line and this is illustrated with parallel lines labeled in minutes on both sides of the central lines. The ionosonde used in taking measurements for this study is located at Ilorin in Nigeria (geographic latitude N, longitude E, dip 4.1 ) and is indicated on the map by an arrow. This station as can be seen on the map is about 38 km from the southern limit of the path of the predicted total eclipse. For the local location of Ilorin, the details of the occurrence time in UT on the ground for this eclipse are as follows: the beginning of the eclipse at 0809:28.7, maximum magnitude of the eclipse at 0923:02.8, and the end of eclipse at 1043:34. The maximum obscuration was 99.36% at this station (S. Bell, Total eclipse of the Sun: 2006 March 29 global circumstances and animation, 2006, available at eclipse.org.uk/eclipse/ /). [4] A study of the effect of 2 October 1959 eclipse on the ionosphere in Nigeria was reported in two papers by Skinner [1967, 1969]. The observations were made in two locations within Nigeria, namely Ibadan (geographic latitude N, longitude E, dip 6 ) and Maiduguri (geographic latitude N, longitude E, dip +4 ). Maximum obscuration was 73.6 percent for Ibadan and 100 percent for Maiduguri. The study which was confined mainly to the F region was focused on the calculation of photoionization rate and loss coefficient. Our study includes the E, F1, and F2 regions and is focused on morphological changes. Another analysis is planned for 1of10

2 Figure 1. Map showing the path of the 29 March 2006 solar eclipse over Benin and Nigeria. The location of the station where the data for this study was taken is indicated by an arrow (after Espenak and Anderson [2004]). 2of10

3 Figure 2. The comparison of the critical frequencies of E, F1, and F2 layers on the eclipse day 29 March with those of the control days 28 and 30 March, The dashed line shows the variation of the percentage obscuration of the eclipse. the investigation on photoionization and loss rate among other things. 2. Data Analysis [5] An Ionospheric Prediction Service (IPS) 42 ionosonde was used for recording the ionograms at the ionospheric station at Ilorin, Nigeria. The antenna of the ionosonde is an equilateral delta antenna whose sampling is vertical. It has a coverage of about 200 km radius, at about the heights of the peak electron densities of the ionosphere. The ionograms recorded during the eclipse show no indication of oblique echoes, indicating the absence of tilts due to horizontal gradients. This excludes the possibility of ambiguity in the interpretation of the ionograms in terms of virtual heights and critical frequencies. [6] Data for the eclipse day 29 March 2006 were compared with those of 28 and 30 March 2006, a day before and a day after the eclipse day, respectively, chosen as the control days. The geomagnetic AP indices for 28, 29, and 30 March 2006 are 6, 5, and 4, respectively. This indicates that the 3 days were very quiet with regards to geomagnetic activity. The sunspot numbers R for those 3 days are 11, 31, and 35, respectively. The period of the eclipse falls within a Figure 3. Ionograms recorded at 1020 LT (a) close to the time of maximum obscuration on 29 March, the eclipse day (only the F2 trace was present), and (b) on the control day 30 March (all three layers were present). 3of10

4 Figure 4. (a f) Comparison of true height electron density profiles on the eclipse day 29 March with those of 28 and 30 March, the control days, from 0915 LT to 1030 LT. (g l) Comparison of true height electron density profiles on the eclipse day 29 March with those of 28 and 30 March, the control days, from 1045 LT to 1200 LT. low solar activity period. The critical frequencies fof2, fof1, and foe of the F2, F1, and E layers, respectively, were obtain for the 3 days under consideration. The available records for 28 March were those taken at 15 min interval and the critical frequencies at that interval were used. We used 5 min interval records of critical frequencies for 29 and 30 March from 0730 to 1250 LT and 15 min interval record thereafter. The closer intervals data covered the period before, of, and after the occurrence of the eclipse. [7] Foratrueheightprofileanalysis,ionogramsat15min interval from the 3 days under consideration were scaled manually on the computer. The POLAN technique [Titheridge, 1985] was used for the ionogram inversion in order to obtain the true height profiles. 3. Results 3.1. Critical Frequency [8] The results of the comparisons of the critical frequencies on the eclipse day with those of the control days for each of F2, F1, and E layers are shown in Figure 2. The percentage obscuration at the E and F1 region heights ( km) at the station is also shown in the figure. 4of10

5 Figure 4. (continued) The effects on foe and fof1 are quite similar and they become noticeable at about 0920 LT, which is about 12 min after the onset of the eclipse, and is seen as decreases in the critical frequencies. Observation of consecutive ionograms showed that these two layers virtually fade out around 1020 LT and begin to appear again about 5 min later. A full recovery occurred at about 1125 LT which is about 15 min before the end of the eclipse period. The ionogram recorded at 1020 LT on the eclipse day is shown in Figure 3a and reveals the absence of the E and F1 traces. The record at the same time on the following day is shown in Figure 3b for the purpose of comparison, to indicate what the ionogram at that time of the day could have been. [9] The observed effect on fof2 is quite different from those of the lower layers. A consistent decrease in the critical frequency of the layer began at 1040 LT which is about 80 min after the decrease in the lower layers started. The minimum value, which fell into the normal magnitude of the F1 layer critical frequency, occurred around 1100 LT, about 40 min after the maximum obscuration. The magnitude of fof2 got back to the normal value at about 1200 LT, 18 min after the end of the eclipse. The duration of the observed decrease for this layer was 1 hour 20 min, while that of the lower layers lasted 2 hours and the eclipse duration was about 2 hours 15 min. 5of10

6 Figure 5. (a) An example of an ionogram recorded on the eclipse day during the rise of the F2 layer to higher heights, compared to (b) the record taken at the same time of the day on the control day True Height Profile [10] Figures 4a to 4l show the results from true height profile analysis from 0915 to 1200 LT at 15 min interval. The electron density profiles for all 3 days under consideration are presented except for some few occasions on 28 March where there were no data due to E layer blanketing. A consistent change begins to be noticed from 0930 LT (Figure 4b) around the F1 region below 200 km in the form of a decrease in electron density. (Figures 4c 0945 LT and 4d 1000 LT) show that the decrease in electron density intensified but remained below the F2 peak. At 1015 LT (Figure 4e), only the F2 region is seen and the decrease Figure 6. Variation of the heights of peak electron density at the E, F1, and F2 regions from 0900 to 1200 LT on the eclipse day and the control days. 6of10

7 Figure 7. Simultaneous variations of hmf2 and fof2 (a) on the eclipse day and (b) on a control day. has progressed to the F2 peak. (Figures 4f 1030 LT to 4h 1100 LT) show the decrease in electron density clearly from the E layer up to the F2 peak. (Figures 4i 1115 LT to 4l 1200 LT) indicate that the electron density recovered in the E and F1 layers before it does so in the F2 layer. [11] Another feature that could be seen in these profiles is the increase in the height of F2 layer. This can be observed from Figure 4e (1015LT) to 4h (1100). Figure 5a shows an example of an ionogram that falls into the period of the rise in hmf2. When this is compared with Figure 5b which is the record at the same time on a control day (30 March), one can see the large difference in the virtual heights of the F2 layers. In order to make this clearer, the variations of the true heights of the peaks of the three layers are presented in Figure 6. The vertical lines labeled B, M, and E are used to indicate the time of the beginning, occurrence of maximum obscuration, and end of the eclipse. Generally, there seems to be no noticeable changes in the peak electron density heights of the E and F1 layers. A look at hmf2 on the control days indicate the normal morning steady rise in hmf2 from 0730 and gets to a peak around A decrease of about km occurs after this peak and hmf2 enters the period of low range of variation. This is the normal pattern of variation for hmf2 [Radicella and Adeniyi, 1999]. On the eclipse day, hmf2 followed the normal pattern from 0730 to the period when a peak is normally attained but departed from the normal variation after this time. Instead of a decrease, the steady rise continued up to about 140 km above the normal height. The maximum height of 465 km occurred at about 1100 LT and is followed by a sudden fall to a minimum of 240 km which is about 65 km below the normal value around 1115 before it goes back to the control days magnitude. 4. Discussions of Results 4.1. E and F1 Layers [12] The inverse correlation of the variation of the percentage obscuration of the Sun during the eclipse with the critical frequencies of the E and F1 layers confirms that solar radiation is the major controlling factor of the variation of electron density in these two layers. The disappearance of the E and F1 traces on the ionograms around 1020 LT during the eclipse may not necessarily mean that these layers disappeared completely. It could just be that the electron density in the layers became so thin that the sounder could not detect them during that period as is the case for the E layer at night on normal days. The threshold frequency of the sounder is about 1.8 MHz in the morning hours which is similar to the condition around the eclipse period. The minimum percentage decrease in peak electron density as deduced from the critical frequencies that were measured for the E and F1 are 60 and 68 percent, respectively. 7of10

8 Figure 8. Samples ionograms around 1100 LT when there was a sudden change in HmF2 at (a) 1059 LT when F2 layer was at a very high altitude, (b) 1104 LT when the F2 layer at the high altitude has almost disappeared and another F2 layer is beginning to appear just above the F1 layer, (c) 1108 LT when the F2 trace at the higher altitude is no longer there but the one above the F1 has moved further up and (d) 1118 LT when the new F2 layer has become prominent F2 Region [13] The maximum decrease in NmF2 observed in this study is 54 percent. This is less than the 68 percent mentioned earlier on for the F1 region. The results of Skinner [1967] were for a period of high solar activity when the F1 region is usually not very distinct from the F2. Combining our results for the maximum percentage reduction in F1 and F2, the maximum percentage reduction at about 170 km is 68 and that at about 250 km is 54. These results compare well in magnitude and trend with 71 and 58 percent for 200 and 300 km, respectively, obtained by Skinner. A more detailed analysis at fixed heights with determination of loss coefficients will be undertaken in another study. [14] It is generally known that the vertical drift of ionization is a major controlling factor in the F2 region. The electric field E in the equatorial E region transmitted to the F region heights by the Earth s geomagnetic field lines B causes a vertical drift of ions and electrons in the F region with an E B force. This causes electrons to diffuse along magnetic field lines away from equatorial latitudes to higher latitudes. This is what leads to the well-known equatorial fountain effect [Rush and Richmond, 1973; Balan and Bailey, 1996; Anderson and Anghel, 2002]. At sunrise, the drift of ionization is upward and increases with time along with NmF2 and hmf2 until the time of occurrence of the morning maximum in hmf2 which is about 0945 LT as observed in this study. After this time, the drift velocity decreases at a rate that is very much slower than the rate of the morning rapid rise. The variation of hmf2 follows a similar pattern as that of the vertical drift [Scherliess and Fejer, 1999; Radicella and Adeniyi, 1999]. The occurrence of the eclipse began around the time of the normal occurrence of the morning peak in hmf2 and continued up to the time period when the range of variation in hmf2 is usually very low, a period of quasi-equilibrium state on a normal 8of10

9 day. The reduction of solar radiation intensity due to the eclipse upset this quasi-equilibrium between production by photo ionization and loss of ionization due to recombination and diffusion of electrons away from the equator. On the eclipse day, the increase in hmf2 between 0945 and 1100 LT was accompanied by a rapid decrease in fof2, which indicates a decrease in electron density. The peak in hmf2 coincides with the minimum electron density. This observation can be seen in Figures 4g and 4h but for the purpose of clarity, we show this in Figures 7a and 7b. Figure 7a shows the eclipse day situation while Figure 7b shows the case on a control day for the purpose of comparison. The rate of change of hmf2 within this time interval ( LT) is about 24 m/s, which is of the same order of magnitude of equatorial vertical drift velocity normally observed [Fejer et al., 1991; Scherliess and Fejer, 1999]. The disruption of the quasi-equilibrium state must have occurred as a result of decrease in production of electrons, the continuation of the depletion of electron by the drift and loss of electrons due to dissociative recombination. The dissociative recombination [Skinner, 1967] increases with increase of neutral molecules which are molecular oxygen and nitrogen. Loss rate due to recombination therefore is greater at lower than at higher altitude since the densities of the neutral molecules decrease with height. The observation made on the profile analysis showed that the decrease in electron density in the F region started from the lower heights of the region and progress upward. Such an upward progressive decrease coupled with the upward drift of ionization would shift the position of the peak density progressively upward along the magnetic field lines. [15] The apparent sharp drop in the F2 layer height after the peak at 1100 LT is not easy to explain in terms of changes in vertical drift of ionization because of the sudden change and the high magnitude of the drift velocity that would be involved. A change in hmf2 of about 225 km within 5 min gives a rate of change of 750 km/s which is completely out of the range of magnitude of observed equatorial vertical drift. The observed post sunset peak in vertical drift velocity, which is the maximum observed, does not even come close to this value. In order to account for this rapid change in hmf2, we examine the trend of changes in the ionograms around the 1100 LT. Figures 8a 8d show samples of the ionograms around this period. Figure 8a shows the record at 1059 LT which is similar to Figure 5a, the record at Figure 8a shows an F2 layer with a higher virtual height (wider valley between F1 and F2 layer) and lower fof2 in comparison with Figure 5a. In Figure 8b (1104 LT), the F2 layer that had a progressively rising hmf2 has almost disappeared and a new F2 layer at a lower altitude above the F1 layer is beginning to emerge. In (Figures 8c 1108 LT and 8d) the F2 layer that was at the higher altitude has completely disappeared (or the electron density was too low to be detected by the sounder as indicated earlier on) while the lower one continues to develop. In short a sudden fall in height did not take place. The initial F2 layer at higher altitude thinned out progressively while another one gradually formed at lower altitude thus replacing the former one. The effect of the gradual decrease in percentage eclipse obscuration leads to the recovery of the F2 layer by the production of ionization at lower altitude. The density of the ionizable specie, which is atomic oxygen decreases with altitude and the rate of production decreases exponentially upward. [16] The altitude of the Sun as the eclipse passed through Ilorin is close to 52 degrees and the speed of the velocity of the shadow is km/s [Espenak and Anderson, 2004]. A rough estimate of the time difference between the shadow contact at the F1 peak density height and that of the F2 layer gives about 5 min. This cannot account for the time lag between the occurrence of the eclipse maximum obscuration and the occurrence of the maximum decrease in fof2. Rishbeth [1963] used the computer model of Briggs and Rishbeth [1961] to solve the continuity equation for electron density in the F region of the ionosphere. Account was taken of production, loss, and vertical diffusion of electrons. The study among other things investigated the speed with which the F layer responds to changes in ionization radiation. The results include the time interval between eclipse totality and the maximum reduction in electron density (eclipse lag). An eclipse with a duration of about 3 hours and with the totality occurring near noon was assumed and the experiment were performed for large, small, and zero diffusion coefficients. From the results of the study, presented as height versus eclipse lag, the eclipse lag increased with height. For large and small diffusion coefficient, the eclipse lag varied from about 30 to 90 min and 1 hour to 4 hours, respectively, from about 250 to 500 km heights. Our observation of an eclipse lag of about 1 hour falls within the range of Rishbeth s results for large diffusion coefficient. The increase of eclipse lag with height is also in agreement with our observation. [17] Acknowledgments. One of the authors (J.O.A) would like to acknowledge the support of the ICTP program for Training and Research in Italian Laboratories for this work. The Institute of Atmospheric Research, Science Academy of Czech Republic is acknowledged for the donation of the IPS 42 ionosonde used in this study to the University of Ilorin, Nigeria, under the collaboration between the Ionospheric Research Group of Physics Department, University of Ilorin, Nigeria, and the ICTP, Aeronomy and Radio Propagation Laboratory, Trieste, Italy. [18] Amitava Bhattacharjee thanks Carl L. Siefring and another reviewer for their assistance in evaluating this paper. References Altadill, D., F. Gauthier, P. Vila, J. G. Sole, G. Miro, and R. Berranger (2001), The solar eclipse and the ionosphere: a search for the distant bow-wave, J. Atmos. Sol. Terr. Phys, 63, Anderson, A., and A. Anghel (2002), Estimating daytime vertical EXB drift velocity in the equatorial F-region using ground-based magnetometer observation, Geophys. Res. Lett., 29(12), 1596, doi: / 2001GL Balan, N., and G. J. Bailey (1996), Modelling studies of equatorial plasma fountain and equatorial anomaly, Adv. Space Res., 18(3), Baran, L. W., I. I. Ephishov, I. I. Shagimuratov, V. P. Ivanov, and A. F. Lagovsky (2003), The response of the ionospheric total electron content to the solar eclipse on August 11, 1999, Adv. Space Res., 31, Briggs, B. H., and H. Rishbeth (1961), An analog solution of the continuity equation of the ionopheric F region, Proc. Phys. Soc., 78(3), Cheng, K., Y.-N. Huang, and S.-W. Chen (1992), Ionospheric effects of solar eclipse of September 23, 1987, around the equatorial anomaly crest region, J. Geophys. Res., 97(A1), Espenak, F., and J. Anderson (2004), Total solar eclipse of 2006 March 29, Tech. Publ. NASA/TP , 55, NASA, Washington, D. C. Farges, T., J. C. Jodogne, R. Bamford, Y. Le Roux, F. Gauthier, P. M. Vila, D. Altadill, J. G. Sole, and G. Miro (2001), Disturbances of the western European ionosphere during the total solar eclipse of 11 August 1999 measured by a wide ionosonde and radar network, J. Atmos. Sol. Terr. Phys., 63, of10

10 Farges, T., A. Le Pichon, E. Blanc, S. Perez, and B. Alcoverro (2003), Response of the lower atmosphere and the ionosphere to the eclipse of August 11, 1999, J. Atmos. Sol. Terr. Phys, 65, Fejer, B. G., E. R. de Paula, S. A. Gonzalez, and R. F. Woodman (1991), Average vertical and zonal F region plasma drifts over Jicamarca, J. Geophy. Res., 96(A8), 13,901 13,906. Radicella, S. M., and J. O. Adeniyi (1999), Equatorial ionospheric electron density below the F2 peak, Radio Sci., 34, Rishbeth, H. (1963), Further analogue studies of the ionospheric F layer, Proc. Phys. Soc., 81, Rush, C. M., and A. D. Richmond (1973), The relationship between the structure of the equatorial anomaly and the strength of the equatorial electrojet, J. Atmos. Terr. Phys., 35, Scherliess, L., and B. G. Fejer (1999), Radar and satellite global equatorial F region vertical drift model, J. Geophys. Res., 104(A4), Skinner, N. J. (1967), Eclipse effects in the equatorial F-region, J. Atmos. Terr. Phys., 29, Skinner, N. J. (1969), A new analysis of eclipse effects in the equatorial F-region, J. Atmos. Terr., 31, Titheridge, J. (1985), Ionogram analysis with the generalized Program POLAN, Rep. UAG-93, World Data Cent. for Sol. Terr. Phys., Boulder, Colo. J. O. Adeniyi, I. A. Adimula, O. A. Oladipo, O. Olawepo, and A. A. Willoughby, Physics Department, University of Ilorin, P.M.B. 1515, Ilorin, Nigeria. (segun47@yahoo.com) S. M. Radicella, Aeronomy and Radio Propagation Laboratory, Abdus Salam International Center for Theoretical Physics, Strada Costiera 11, I Trieste, Italy. 10 of 10