COMPARATIVE STUDY OF THE GEOMAGNETIC ACTIVITY EFFECT ON foF2 VARIATION AS DEFINED BY THE TWO CLASSIFICATION METHODS AT DAKAR STATION OVER SEASONS

Sibri Alphonse Sandwidi, Doua Allain Gnabahou and Frédéric Ouattara Laboratoire de Recherche en Energétique Et Météorologie de l’Espace (LAREME) de l’Université Norbert ZONGO, BP 376 Koudougou Burkina Faso. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History Received: 15 May 2020 Final Accepted: 20 June 2020 Published: July 2020

Here, we focus our attention to a comparison of both classifications profiles over solar cycle phases on the F2 layer critical frequency (foF2) time variation at Dakar station. The novelty of this study appears as it is a first paper where such comparison is made for both geomagnetic activity classifications methods. This paper aims, in short term, to highlight how the signature of the storm effect is shown by the geomagnetic events of both classifications methods and in mean and long term to contribute to improve the understanding of the sources of geomagnetic activities.
The outline of the paper is as follows: section 2 concerns materials and methods, section 3 is devoted to results and discussions, and the conclusion and research perspectives end the paper as its fourth section.
It's important to note that, this Legrand and Simon's classification or ancient classification (AC) clearly classify only 60% of the geomagnetic activity Amory-Mazaudier, 2008, 2009;Ouattara and Zerbo, 2011;Zerbo et al., 2012) whereas the scheme proposed by Zerbo et al. (2012) allows to clearly classify 80% of the geomagnetic activity. The new classification is characterized by: (1) the same limit with the first classification for the quiet activity (Aa < 20 ) and (2) The lowering of the limit (Aa ≥ 20 nT) for the shock and recurrent events for whom the causes from solar activity are more highlighted (Zerbo et al., 2012). Therefore: (1) recurrent activities (RA), characterised by Aa ≥ 20 nT, are now the recurrent activity of Legrand and Simon plus the Corotative Moderate Activity (CMA) and they are due to solar high wind stream and Corotating interaction Region (CIRs); (2) shock activities (SA) are now the shock activity of Legrand and Simon plus the Cloud Shock Activity (CSA) and they are caused by Coronal Mass Ejections (CMEs) and magnetic clouds, respectively, with Aa ≥ 20 nT and (3) fluctuating activities (FA) or unclear events represent the fluctuating activity of Legrand and Simon plus the Corotating Moderate Activity (CMA); they are consequences of the fluctuation of solar heliosheet. This paper thus concerns the comparison between the profiles of foF2 daily variation during disturbed activities of both schemes of geomagnetic activity classification as described.
The pixel diagram presents the variation of the geomagnetic activity by solar rotation, also called Bartels rotation ). This colour board is used to determine the geomagnetic data as a function of solar activity as described by solar rotation duration (27 days) which consists of: (1) plotting the aa daily values in Bartels diagrams, (2) exhibiting the different types of geomagnetic activities as described above according to both criteria. Figure 2 showed the pixel diagram highlighting the process to determine the different geomagnetic activities following the criteria of the two methods of classification.

Data analysis method:
This paper aims to study the difference of foF2 time variation, throughout the seasons for a given disturbed activity for both geomagnetic activity classifications. The data analysis will be done through two ways: (1) The analysis of the recurrence of the different geomagnetic activities according to both methods of classification over seasons; (2) The profiles obtained are analysed in comparison with the five standard profiles established by Faynot and Vila (1979) for African Equatorial Ionization Anomaly (EIA) region and (3) the quantitative analysis based on comparison between foF2 profiles from both classifications methods.
Firstly, the analysis of the occurrence of the geomagnetic activities is done by: (1) comparing the occurrence of the different geomagnetic activities in relation with both classification by using the equation 1; ∆ occ = GA occ / AC − GA occ / NC (1) Where ∆ occ is the gap of the occurrence of the concerning geomagnetic activity classes. GA occ / AC is the occurrence of the concerning geomagnetic activity class in relation with the ancient classification. GA occ / NC is the occurrence of the concerning geomagnetic activity class in relation with the new classification. and (2) putting error bars in data graphs. We can note that error bar is obtained by using the equation 2.
with x ̅ mean value and N the total number of observations for a particular dataset.
Secondly, the foF2 profiles (see Figure 1), linked to the nature, the strength or the absence of electric currents in E layer of ionosphere (Acharya et al., 2010(Acharya et al., , 2011Dunford, 1967;Vassal, 1982aVassal, , 1982b, are: Noon bite out or "B" profile characterized by a double peak (morning and evening) parted by a trough around midday; Reversed or "R" profile characterized by a single peak at afternoon; Morning pic or "M" profile characterized by a single maximum at morning; Plateau or "P" profile characterized by an ionization plateau during daytime and Dome profile or "D" profile characterized by a single maximum around noon.
This morphological analysis of the profiles will reveal the difference between the equatorial ionosphere characteristics highlighted by the profiles obtained by both classification methods. In fact those profiles express respectively a signature of a high electrojet, an important afternoon conter-electrojet, a moderate electrojet, a weak electrojet and an absence of electrojet. In addition Fejer, (1981); Fejer et al., (1979,1981), Abdu et al., (1981Abdu et al., ( , 1992) and Farley et al., (1986) reviewed on mechanism responsible of equatorial trough at noon induced by electrodynamics E × B process. panel "c" is "D" and that of panel "d" is "P" profile. These two profiles characterized the absence of electrojet.
Panel "e" is "R" profile and expresses the signature of strong counter electrojet.
In a third time, the quantitative analysis based on the appreciation of the difference between foF2 values of the ancient classification and those from the new one will be made through the relative deviation of foF2 defined by: Where foF2 AC and foF2 NC are the foF2 from the ancient classification and the new one respectively and σ foF2 is the relative deviation with the following appreciation: 1. σ foF2 > 10% the ancient classification overestimates the new standard classification value; 2. σ foF2 < −10% the ancient classification underestimates the new standard classification value; 3. 10 % < σ foF2 < 10% both classifications graphs are in agreement.

Results and Discussion:-
Occurrences of the geomagnetic activities over seasons Figure 3 highlights the occurrence of the different geomagnetic activities throughout seasons. Panels "a" and "b" concern the classes according to the ancient classification (AC) and the new classification (NC) respectively.
For each season of both classifications, the diagram columns represent respectively the fluctuant, the recurrent, the shock and the very quiet activities. For the occurrence of the fluctuant activity (FA occ ) over seasons, we have for the AC, 25.12%; 23.76%; 25.95% and 25.18% and for the NC, 25%; 22.6%; 28.7% and 26.8% respectively for spring, summer, autumn and winter seasons. Therefore, for both classifications, the fluctuant occurs more during the autumn season and less during summer. The maximum negative gap between both classifications occurrences is observed at autumn time with a value ∆ occ = −2.75% and the maximum positive occurrence is observed in summer season with a value ∆ occ = +1.16%. Besides, for spring and summer seasons, the FA occurrence of the AC is greater than that of the NC; and we observe the opposite for the autumn and the winter seasons. Consequently, during spring and summer times, the AC's FA occurrences are greater than those of the NC and the opposite is observed during autumn and winter. Winter and equinoctial anomalies are obviously highlighted by the FA of the AC while for those of the NC we observe a less marked winter anomaly and an absence of equinoctial anomaly For the RA, we have for the AC, 47.43%; 4%; 28.57% and 20% and for the NC, 32.7%; 20.1%, 28.7% and 18.5% respectively for spring, summer, autumn and winter seasons. The recurrent events occur more during the spring season for both classifications; and less during summer for AC and during winter for the NC. The maximum negative gap between both classifications occurrences is observed at summer time with a value of ∆ occ = −16.1% and the maximum positive occurrence is observed in spring season with a value of ∆ occ = +14.73%. Besides, for spring and winter seasons, the RA's occurrence of the AC is greater than that of the NC; and we observe the opposite for the summer and autumn seasons. Therefore, the RA of the AC occurs more during equinoxes than solstices than the NC's RA; that testify the semi-annual anomaly during RA events which is more marked for the AC. The maximum negative gap between both classification occurrences is observed at summer time with ∆ occ = −6.08% and the maximum positive one is observed in spring season with ∆ occ = +2.34%. Besides, for summer season, the SA's occurrence of the AC are lower than that of the NC; and we observe the opposite for the other seasons. Therefore, we observe that the AC's SA more occur during equinoxes than solstices; but for those of the NC, there is not an obvious difference between both seasons. In consequence, there is semi-annual anomaly, winter anomaly (due to the asymmetry between winter and summer) and the absence of equinoctial anomaly (due to the fact that the occurrences of spring and autumn season of the AC's SA don't show remarkable difference); that is not the case of the NC because the occurrences of the NC's SA don't differ considerably over seasons..

921
The NC recurrent events are those of the AC plus the corotating moderate activity (CMA) and its shock events correspond to those of the AC plus the cloud shock activity.  Figure 4 to Figure 6 show the foF2 profiles during fluctuant activity (FA), recurrent activity (RA) and shock activity (SA) respectively. The panels "a", "b", "c" and "d" present the graph of the spring, summer, autumn and winter seasons respectively. For all geomagnetic activities, the absence of equinoctial asymmetry is clearly highlighted and  Figure 4 shows that, during FA, both graphs show an "R" profile during all seasons but it shows a "M" profile for winter season. Therefore, the FA is characterised by a moderate electrojet during winter season, while it shows an important afternoon conter-electrojet in E layer of ionosphere during the other seasons. The two classification's curves show the same shape and the same amplitude for all seasons with respect to error bars shown in NC graph; thus, for FA the foF2 profiles of the AC and NC don't show difference during all seasons; it is so not important to estimate the relative deviation of foF2 during FA. This absence of difference between the two classifications can be explain by the fact that the FA is characterised by the unclear solar activity; perhaps the highlighting of the twenty percent (20%) of the geomagnetic activity which are mixed in the unclear activities (remember that the NC explain about 80% of the solar activity in relation of the geomagnetic activity) will enable to make difference with the AC.  Figure 5 shows the foF2 values' diurnal variation for the RA according to the two classification and the relative deviation percentage of both classifications foF2 values. For equinox seasons (see panels "a" and "c"), the two classification's curves show an "R" profile, with the same shape and the same amplitude with respect to error bars shown in NC graph. Therefore, the RA is characterised by a moderate electrojet during equinox seasons. Otherwise we observe a difference between both classifications curves in the meantime surrounding 0200 LT during spring months and in the meantime surrounding 0700 LT, 2100 LT and 0000 LT concerning autumn season. The right column of the panels "a" and "c" highlight this observation. In fact, for the spring season (panel "a"), the σfoF2 shows value within ±10% between all the time excepted in the meantime surrounding 0200 LT with a maximum negative value of −6.9% observed at 0800 LT and a maximum positive value of 21.55% observed at 0200 LT. And the graph of right column of the panel "c" present a σfoF2 value within ±10% between all the times excepted in the meantime surrounding 0700 LT, 2100 LT and 0000 LT; with a maximum positive value equal to +13,96, observed at 0000 LT and a maximum negative value equal to −10,28, observed at 0700 LT.

Profiles comparison for both classifications throughout seasons:
For the summer month (panel "b") we observe an obviousness difference between both curves; in fact, while we have an "R" profile for the NC, the AC curve show a "B" profile characterised by a late morning peak at about 1400 LT (~10.82 MHz), an afternoon peak at 1800 LT (~11.80 MHz), and a trough at 1500 LT (~9.87 MHz). The obviousness difference between the AC and the NC curves during summer season can also be explain by the large gap (∆ occ = −16.1%) between the occurrences of the RA for both classifications as shown at the paragraph 3.1. Besides, the right column of this panel highlights a σfoF2 value within ±10% just between after 0600 LT to after 0800 LT and between 1400 LT to 1600 LT; but for the remaining time, σfoF2 values are greater than +10% or small than −10%. Therefore both classification's RA present an almost different profile during summer season, except between after 0600 LT to after 0800 LT and between 1400 LT to 1600 LT with respect to the σfoF2 value. Panel d shows an "M" profile for both classifications curves during the winter month; therefore the RA profile highlights a moderate electrojet during winter time. Otherwise, we observe a difference between the two graphs after 1700 LT to before 0600 LT with respect to error bars shown in NC graph. Besides, the right column of this panel highlights a σfoF2 value greater than +10% between 2000 LT to 0600 LT and a value within ±10% the time remaining. The maximum positive value, equal to +15,62, is observed at 0500 LT and the maximum negative value, equal to −1,59, is observed at 0800 LT.   Figure 6 shows the foF2 values diurnal variation for the SA of the two classifications and the relative deviation percentage of foF2 diurnal values of both classifications. Panel "a" presents a "P" profile for the two curves during spring time. Therefore, the SA is characterised by a weak electrojet during spring season. Besides, we observe that the two curves present the same shape and the same amplitude with respect to error bars shown in NC graph. In fact, the right column of this panel highlight that the σfoF2 value is within ±10% between all the daily times; thus, the shock events foF2 daily profiles variation of both classifications are in good agreement during spring time. The maximum negative value of σfoF2, equals to -6.22%, is observed at 0400 LT and the maximum positive value is observed at 0200 LT and it is equal to +4.3%.
In panel "b" we observe that both graphs present a "D" profile which highlights the signature of an absence of electrojet during summer time for 2 classification's SA. Otherwise, the AC graph is always above that of the NC; the two classification's graphs are different between 1300 LT to 1700 LT and 2100 LT to 0400 LT with respect to error bars shown in NC graph. Besides, the right column of this panel highlights a σfoF2 value greater than +10% between 2100 LT to 0400 LT and a value within ±10% the time remaining. The maximum positive value is observed at 0000 LT and it is equal to +68.53%.
Panel "c" shows a "B" profile with a lower trough at 1200 LT for both curves. Therefore, the 2 classification's SA are characterised by a short living high electrojet located at 1200 LT during autumn time. Otherwise, the AC graph is always down that of the NC; the two classifications are different during a meantime surrounding 0100 LT and 0300 LT with respect to error bars shown in NC graph. Besides, the right column of this panel highlights a σfoF2 value negative all the daily time with values small than -10% at 0100 LT and at 0300 LT and values within ±10% the time remaining. The maximum negative value is observed at 0300 LT and it is equal to −12.67% and the maximum positive value, equal to +0,35% is observed at 2100 LT.
Panel "d" shows an "M" profile for both classifications curves. Therefore, the 2 classification's SA are characterised by a moderate electrojet during winter time. Otherwise, we observe that the AC graph is almost always above that of the NC, except between 0900 LT to 1100 LT. The two classifications are different between 0400 LT to after 0600 LT with respect to error bars shown in NC graph. Besides, the right column of this panel highlights negative value of σfoF2 between 0900 LT to 1100 LT and positive the time remaining. The maximum negative value of σfoF2, observed at 1000 LT, is equal to −1.08% and its maximum positive value, observed at 0600 LT, is equal to +13.59%.

Conclusion:-
This paper shows that both classification's geomagnetic classes, at Dakar station, are different in terms of occurrences throughout seasons. It emerges that the more the occurrences between two classes for a given season are different, the more the profiles show difference. Therefore, it's observed an obvious difference of the electrodynamics in the ionosphere layer during the disturbed geomagnetic activities according to both classifications method. In general the AC and NC curves show different profiles during the summer season for the RA and the SA. The FA profiles don't depend on the classification method at all.