Monitoring aurorally-generated large scale travelling ionospheric disturbances (TIDs) over a sunspot cycle (1980-85)

 L.A. Hajkowicz and B. Zimkowski
Department of Physics, University of Queensland, Qld.4072, Australia.
Abstract.
The onset of auroral substorms is invariably associated with the generation of large scale ionospheric disturbances (LSTIDs) which modify the virtual height (h'F) of the F-region, primarily at sub-auroral locations. Five sub-auroral ionosonde stations, located in three largely different diurnal sectors: Australian-Asian (Hobart-Yakutsk), Indian Ocean (Kerguelen) and European (Uppsala-Dourbes) were used to monitor the aurorally-generated LSTIDs over a sunspot cycle (1980-85). The virtual height enhancements (Dh'F) were analysed in conjunction with auroral substorms onsets deduced from the auroral electrojet index AE. Significant statistical results were obtained from the analysis of auroral-ionospheric data obtained during 473 disturbed days (for AE-index 400 nT). It is evident that substorm onsets are most frequent in a time interval: 11-19 UT, peaking at 15 UT. Since LSTIDs are strongly attenuated during daytime, due to the ion drag, the most advantageous geographical location for monitoring the disturbances is in the Australian-Asian sector which is then well into the nighttime sector of the earth, with the northern station Yakutsk being closest to the auroral source region. The other westward-displaced sectors are progressively less suitable for monitoring of LSTIDs, as the propagation of LSTIDs occurs in the sunlit sector of the earth. Symmetrical propagation of LSTIDs in both hemispheres is observed mainly in the equinoctial months.

1. INTRODUCTION

 Large scale ionospheric disturbances (LSTIDs), which are ionospheric manifestation of aurorally-generated atmospheric gravity waves, have attracted a considerable attention as a principal energy sink of auroral substorms (Walker et al., 1988; Walker and Wong, 1993; Huang and Cheng, 1993; Prolss, 1993; Yeh et al., 1992). Unlike other manifestations of auroral disturbances this wave phenomenon has a truly global character, spanning both hemispheres from high to equatorial latitudes and over a large range of longitudes. An extensive study has been undertaken on the global occurrence of LSTIDs simultaneously launched from the northern and southern auroral ovals at the onset of auroral substorms. A particularly useful and readily available ionospheric tracer of LSTIDs was found to be the minimum height or the virtual height (h'F) of the F region (Walker et al., 1988). A sequential change in the ionospheric height is indicative of equatorward propagating trains of LSTIDs in both hemispheres, with a typical propagation velocity between 500 - 1300 m/s, originating at the auroral locations characterised by a shell with L-value of about 5 (Hajkowicz, 1983a and b; Hajkowicz, 1990; Hajkowicz, 1991a and b; Hajkowicz and Hunsucker, 1987). The disturbances reach the highest amplitude at sub-auroral locations (Hajkowicz, 1991a and b), rapidly decrease at mid-latitudes, and remain fairly constant at low latitudes with a slight tendency to increase at the equator (Hajkowicz, 1991b).

A consistent factor in the propagation of LSTIDs is their rapid attenuation in daytime which sometimes results in a gross asymmetry in the occurrences of disturbances in northern and southern hemispheres, following conjugate onsets of auroral substorms (Hajkowicz, 1990 and 1991a). This attenuation is discussed in detail by Hajkowicz (1990) who attributed it to the ion drag retarding the gravity wave propagation, associated with the high ionization density of the F-region in daytime. The ion drag appears also to be responsible for pronounced difference in the occurrence of LSTIDs at widely separated longitudinal (or diurnal) sectors, associated with the so-called "universal time effect" of the onset of auroral substorms (Hajkowicz, 1992). The effect was only studied at a relatively high sunspot number (1980-82) but because of its global significance was also extended to the low sunspot activity interval (1983-85) in the present considerations. The present study, which completes the previous results, clearly demonstrates the preference of certain geographical locations in monitoring the global morphology of large scale gravity waves.

2. RESULTS

The ionospheric height enhancements Dh'F were obtained by subtracting monthly median values of h'F from the current height variations, thus removing the diurnal thermospheric wind effect. Numerous examples of sequential ionospheric height rises Dh'F, occurring progressively later as LSTIDs propagated towards the equator, has been presented in the previous reports (e.g. Hajkowicz, 1992). In all these cases the virtual height enhancement invariably followed an onset of auroral substorms as inferred from a rapid rise in the auroral electrojet index AE. It is evident that the statistical study of h'F enhancement, particularly at the locations close to the source region (Hajkowicz 1991b), is of considerable importance in the long- time evaluation of the ionospheric response to substorm activity. In the initial phase of the project six sub-auroral ionosonde stations, as specified in Table 1, were used in the analysis. It became evident that average auroral activity had a distinct diurnal distribution, peaking between 13-19 UT and the most distinct height rises tended to occur for the stations which were then at nighttime, i.e. in the Australian-Asian sector (Hobart-Yakutsk). Other stations, west of this sector and towards the sunlit region of the earth (i.e. Kerguelen, Uppsala, Dourbes and Ottawa) showed progressively smaller LSTID's activity. Thus, the average height enhancements were smaller in the European sector (Dourbes-Uppsala) and were very much reduced in the North American sector (Ottawa) which was in the full daylight during the period of maximum auroral activity.

Table 1. Ionosonde stations used in the height analysis.

Average night-time (at an altitude of 300 km), geographic latitude and East longitude, invariant latitude and L-value are indicated.

  
Geographic
 
Invariant
 
Night-time interval (UT)
 
Station
 
Lat.
 
Long.
 
Lat.
 
L
 
Summer
 
Autumn
 
Winter
 
Spring
  
(deg.)
 
(deg.)
           
Hobart
 
-42.9
 
147.3
 
53.6
 
2.8
 
12-17
 
10-19
 
08-20
 
10-19
 
Yakutsk 
62.0
 
129.6
 
55.1
 
3.1
 
08-22
 
12-18
 
Day
 
12-18
 
Kerguelen
 
-48.8
 
70.0
 
58.2
 
3.6
 
18-20
 
14-23
 
13-01
 
14-23
 
Uppsala
 
59.8
 
17.6
 
56.6
 
3.3
 
16-05
 
20-02
 
Day
 
20-02
 
Dourbes
 
50.8
 
4.4
 
48.5
 
2.3
 
18-06
 
20-04
 
23-01
 
20-04
 
Ottawa
 
45.4
 
284.3
 
57.7
 
3.5
 
01-13
 
01-09
 
03-07
 
01-09

The southern seasons, (Summer: Nov.-Jan., Autumn: Feb.-Apr., Winter: May-Jul., Spring: Aug.-Oct.), are used throughout unless otherwise specified.

seasonally averaged ionospheric response

Figure 1 Examples of seasonally averaged ionospheric responses for individual stations: Hobart (H-solid line) and Yakutsk (Y-broken line) and the corresponding average responses for the Australian-Asian sector (DH'F, solid line) and the associated AE-index (scaled down by a factor of 5). The southern seasons and years (eg. SU81,AU81 refers to summer 1981 and autumn 1981) are indicated (the number of the disturbed days used for averaging are indicated in the brackets). The solid and broken lines indicate the corresponding nighttime (at an altitude of 300 km) for each season.

Australian-Asian sector (a-d) Australian-Asian sector (e-h) Australian-Asian sector (k-n)

Figure 2 The complete ionospheric responses in the Australian-Asian sector, averaged for each season at sunspot maximum (1980-82) and minimum (1983-85).( The symbols as in Fig.1. Note: no 2i, 2j figures).

The same stations (except for Ottawa for which much of the data was missing) were used in the analysis of ionospheric response in the sunspot minimum period (1983-85). Examples of ionospheric responses to individual substorms over a wide range of longitudes have been presented in Figs. 1 and 2 by Hajkowicz (1992). The examples clearly indicate a delay effect between the occurrence of auroral disturbances and the subsequent ionospheric height enhancements, characteristic of LSTID's propagation. A large number of such individual cases was analysed, resulting in an average variation of AE-index and the associated height enhancement Dh'F for specific seasons. The high degree of correlation of the ionospheric responses (in the Asian-Australian sector) with the auroral electrojet surges can be seen in the diurnal variations of Dh'F (derived from averaging responses at Hobart and Yakutsk, symbolised as DH'F) for selected seasons in Fig.1. A number (indicated in brackets) of disturbed days were used to derive the average Dh'F and corresponding AE. It can be also seen that in all the cases Dh'F for Hobart (H) and Yakutsk (Y) closely resembles the average trend in the auroral electrojet activity peaking between 12-17 UT, within the local nighttime for these stations. The responses at Yakutsk and Hobart were of similar duration in the equinoctial months, with similar nighttime (at the height of the F-region) periods at both stations. In summer (northern winter) the response was more extended at Yakutsk than at Hobart as the former station was in the winter hemisphere (Fig. 1 b and d), characterised by long nighttime, facilitating propagation of LSTIDs the northern hemisphere. It is also evident that the responses at Yakutsk tend to be of a larger amplitude than those at Hobart as the former station is closer to the source region (larger L-value in Table 1).

Note that the average ionospheric responses (DH'F) in the Australian-Asian sector follow almost exactly the trend in the auroral electrojet index in autumns 1981 and 1982 (Fig.1 e and g). This has been noted for individual disturbances by Hajkowicz (1991b) who reported that the correlation coefficient between AE-index and the corresponding height enhancements was on average 0.7, on occasions reaching 0.9.

The complete average ionospheric response for all seasons (autumn and spring are combined in the equinoctial period) can be seen in Fig.2 for a large number of disturbed day in each season at sunspot maximum (1980-82) and minimum (1983-85). Altogether 473 disturbed days, with a minimum onset value of AE-index 400 nT, were used in the analysis. It is now evident that the universal time effect in the diurnal preference in the occurrence of substorm onsets is maintained throughout the solar cycle for all seasons except in winter ( Fig. 2 e and g) where the substorm onsets are more evenly distributed throughout the day. Thus the majority of auroral surges, as signified by the rapid variations in AE-index, were in an interval 11-19 UT coinciding with nighttime at Hobart (local midnight 14 UT) and at Yakutsk (local midnight at 15 UT). It should be noted that during southern summer (northern winter), particularly during sunspot maximum, response at Yakutsk is wider and larger than at Hobart as Yakutsk has then a long nighttime, resulting in more favourable propagation conditions for LSTIDs (also demonstrated for the selected summers in Fig.1). This was particularly evident at sunspot maximum (Fig.2 a and b). In southern winter (northern summer) Yakutsk is in continuous daytime and shows less extended ionospheric response than Hobart (cf. Fig.2 e and f, and g and h). It is of considerable interest that LSTIDs recorded in this season at Yakutsk propagate through the sunlit atmosphere, with the most of the disturbances reaching the station close to local midnight (Fig.2f and h), i.e. at the time when the sun reaches the lowest elevation at this station. Generally the response at Yakutsk tends to be larger in magnitude than that at Hobart particularly at equinox at sunspot minimum, and in the post-midnight sector ( Fig.2 m and n).It is clearly evident (from the large statistical samples) that the symmetrical propagation of LSTIDs in both hemispheres, following of the onset of auroral substorms, is characteristic of the equinoctial periods for the solar cycle (Fig.2 k and l, and m and n). It should be also noted that a sharp peak in the ionospheric response at Yakutsk is observed at almost the same time, 15-16 UT (i.e near local midnight), in northern winter (Fig.2b and d) and in the equinox (Fig.2l and n) for the sunspot maximum and minimum. This sharp peak is mostly missing in the responses recorded at Hobart.

Indian & European sectors
Figure 3 The complete ionospheric responses in the Indian (for Kerguelen-K) and European (for Uppsala-U and Dourbes-D) sectors for different seasons at sunspot maximum and minimum.(The line symbols for three stations are given in e and f).

The ionospheric response is less distinct in two other longitudinal sectors: Indian Ocean (Kerguelen- K) and Europe (Uppsala-U and Dourbes-D ) positioned towards the evening side of the auroral electrojet maximum (Fig.3). The maximum frequency of auroral substorm onsets occurs close to the sunset at Kerguelen but the peak of auroral activity is at local afternoon at Uppsala and Dourbes. Consequently, a large proportion of aurorally-generated LSTIDs are attenuated in these sectors resulting in a narrower and weaker ionospheric response than that for the Australian-Asian sector. It should be noted the response at Kerguelen is quite high at equinox in the post-sunset sector (particularly during sunspot maximum, Fig.3 e) as this station, positioned at L-value = 3.6, is nearest (from all the stations) to the auroral source region. The European sector shows a much weaker response both in magnitude and duration over the sunspot cycle. In the equinoctial months there is a consistent decline in the ionospheric response, from Kerguelen to Dourbes at sunspot maximum and minimum (Fig3. e and f). This again is in agreement with the attenuation of LSTIDs as they propagate from the source region to lower latitudes. Finally, in summer most LSTID's propagation at higher latitudes (particularly at Kerguelen, Fig.3a) takes place in the sunlit atmosphere as was the case for Yakutsk.

COMBINED SOLAR CYCLE RESPONSE COMBINED SOLAR CYCLE RESPONSE

Figure 4 The sunspot maximum and minimum (a-f) and combined sunspot cycle (g-n) responses. The symbols AE-L,AE-H,Y-H etc. refer to the average values of AE-index at low and high sunspot numbers, average Dh'F for Yakutsk for high sunspot numbers etc. The arrows indicate the local midnight (Note: fig 4i & 4j are not used).

The comparative average ionospheric responses over the sunspot maximum and minimum, and over the entire cycle are given in Fig.4. It is evident that the trend in the onset of auroral substorms shows the characteristic universal time effect, with substorm activity mainly concentrated between 11-19 UT (with a peak at about 15 UT), over the solar cycle (Fig.4 a and g). The peak in AE is closely followed by a peak in Dh'f at Yakutsk (Fig.4b), at sunspot minimum (Y-L) and maximum (Y-H).A sharp maximum is reached at Yakutsk at local midnight for the entire sunspot cycle (Fig.4h). On the whole the response at Yakutsk is more pronounced than that at Hobart (Fig.4h), particularly in the post-midnight sector. The response is is still pronounced at Kerguelen being of the same magnitude as that of Hobart but shifted away from the center of auroral activity, towards local nighttime (Fig.4d and k). The European sector is characterised by a much weaker response at Uppsala and Dourbes (Fig.4e and f) as compared to that of other sectors (Fig.4m).

It is evident that the responses at sunspot minimum are consistently lower than that at sunspot maximum for all the sub-auroral stations (L-value 2.8; Fig.4 b-e) despite that the average associated AE-index is larger at sunspot minimum than maximum (Fig.4a).

Fig.4m is of particular interest as it gives a concise classification of the three sectors according to their ability to monitor aurorally-generated LSTIDs. Thus, the most effective sector in detecting the disturbance is the Australian-Asian region (A), followed by the Indian Ocean (B) region and then by the European region (C). The combined response (averaged A, B and C) provides monitoring facilities of LSTIDs from about 12 to 23 UT (Fig.4n).

3. RESULTS AND CONCLUSION

The statistical analysis of ionospheric response (using 473 disturbed days), associated with the propagation of large scale ionospheric disturbances (LSTIDs), was conducted at three largely different longitudinal sectors at northern and southern sub-auroral locations. For the first time it was possible to establish the longitudinal preference in monitoring LSTIDs over a sunspot cycle. The following is evident:

  1. The substorm onsets, as inferred from the auroral electrojet surges (AE-index 400 nT), are most frequently encountered in a time interval: 11-19 UT, peaking at about 15 UT. This trend in auroral activity appears to be present in other solar cycles as discussed by Hajkowicz (1992).

  2. The propagation of o LSTIDs is affected by the ion drag and critically depends on the solar illumination conditions. Thus, the Australian-Asian sector which is close to local midnight during the peak of auroral activity records LSTIDs considerably more frequently than other westward- displaced sectors. The results obtained for the American sector at sunspot maximum (Hajkowicz, 1992) gives a global classification of the geographic locations suitable for monitoring of LSTIDs (from the most to the least suitable): Australian-Asian sector (Hobart, Yakutsk), Indian Ocean sector (Kerguelen), European sector (Uppsala, Dourbes) and American sector (Ottawa).

  3. Since monitoring of LSTIDs is also most effective at sub-auroral locations the Asian station Yakutsk (L-value=3.1) appears to be the most suitable location for monitoring LSTIDs.(The locations shifted even closer to the source region would be impractical due to frequent absorption of ionosonde signals). The southern equivalent of this station in Australia, Hobart also has an important role in monitoring simultaneous LSTIDs propagating equatorwards from the southern auroral oval.

  4. There is a consistent decline in the amplitude of LSTIDs from sunspot minimum to maximum for all sub-auroral stations (L-value 2.8), especially in the post-midnight sector. This decline was particularly noticeable for the southern sub-auroral station Hobart.

The results herewith presented support the previous findings on the attenuation of large scale gravity waves due to the ion drag, which retards the propagation of the disturbance in daytime. It should be noted however that this attenuation does not prevent the gravity wave from reaching a sub-auroral station, such as Yakutsk, even at daytime as can be seen from the relatively large response at Yakutsk during northern summer when both the station and source region are in continuous daylight (Fig.2f). This is supported by a case study ( during the great storm of 13 March 1989, by Hajkowicz, 1991a) when a large auroral surge was followed by the global height enhancement, including a large height rise at Uppsala which was then in full daylight. The attenuation along the disturbance's propagation path is however very rapid during daytime and little or no disturbance is observed at lower latitude stations (Hajkowicz, 1990a and 1991a). On the other hand, LSTIDs propagate almost with a constant amplitude to low latitudes at nighttime (Hajkowicz and Hunsucker, 1987).

A separate and unknown aspect of the present study concerns the consistent decline in the amplitude of LSTIDs during sunspot minimum. This can be associated with a decline in the median height h'F from sunspot maximum to minimum. For example, a study of the median height of the F-region in the Asian - Australian over a sunspot cycle indicates a gradual decrease in the background h'F at all seasons (Hajkowicz, 1991c). Such background height variations may have a substantial effect on the amplitude of LSTIDs as the gravity wave amplitude grows exponentially with altitude, i.e. with the height of the F-region. This could lead to lower amplitudes of LSTIDs at sunspot minimum as observed.

ACKNOWLEDGEMENTS

The ionosonde data were supplied by the Ionospheric Prediction Service (IPS), Sydney, Australia, the World Data Center B2 , Moscow, Russia, Institut Royal Meteorologique de Belgique, Brussel, Belgium, Inst. of National Defence, Stockholm, Sweden and Centre National d'Etudes des Telecommunications, Lannion, France. We are grateful to Mrs. D.J. Dearden for her assistance with the data analysis and to Dr. G.G. Bowman for his discussion on some aspects of this work.

REFERENCES

 Hajkowicz,L.A., Auroral riometer absorption and the F-region disturbances observed over a wide range of latitudes, J.atmos.terr.Phys.,45,175-179, 1983a.

Hajkowicz,L.A., Conjugate effects in the generation of travelling ionospheric disturbances (TIDs) in the F-region. Planet.Space Sci.,31,1409-1413, 1983b.

Hajkowicz, L.A., A global study of large scale travelling ionospheric disturbances (TIDs) following a step-like onset of auroral substorms in both hemispheres, Planet.Space Sci., 38,913-923, 1990.

Hajkowicz, L.A., Global onset and propagation of large scale travelling ionospheric disturbances as a result of the great storm of 13 March, 1989, Planet.Space Sci., 39,583-593, 1991a.

Hajkowicz, L.A., Auroral electrojet effect on the global occurrence pattern of large scale travelling ionospheric disturbances, Planet.Space Sci.,39,1189-1196, 1991b.

Hajkowicz, L.A., Median ionospheric height variations over a sunspot cycle in the Australian-Japanese longitudinal sector, Planet. Space Sci.,39,1607-1615, 1991c.

Hajkowicz,L.A., Universal time effect in the occurrences of large-scale ionospheric disturbances, Planet. Space Sci.,40,1093-1099, 1992.

Hajkowicz,L.A. Ionospheric response to auroral substorms during sunspot maximum (1980-82), Anns.Geophys., In print.1994.

Hajkowicz,L.A. and R.D. Hunsucker, A simultaneous observation of large-scale periodic TIDs in both hemispheres following an onset of auroral disturbances, Planet.Space Sci., 35, 785-791, 1987.

Huang, Y.-N. and K. Cheng, Ionospheric disturbances around East Asian region during the 20 October 1989 magnetic storm, J.atmos.terr.Phys., 55 ,1009-1020, 1993.

Prolss, G.W., On explaining the local time variation of ionospheric storm effects, Ann.Geophys., 11, 1-9, 1993.

Walker,G.O. and Y.W. Wong, Ionospheric effects observed throughout East Asia of the large magnetic storm of 13-15 March 1988, J.atmos.terr.Phys.,55,995-1008, 1993.

Walker,G.O.,Wong,Y.W.,Ma,J.H.K.,Kikuchi,T.,Nozaki,K.,Huang,Y.N., and V.Badillo, Propagating ionospheric waves observed throughout east Asia during the WAGS October 1985 campaign, Radio Sci., 23,867-878, 1988.

Yeh,K.C.,Liu,K.H., and R.O.Conkright, The global behavior of the March 1989 ionospheric storm, Can.J.Phys.,70,532-543, 1992.

CONTENTS
Formating comments (E-mail: phil@ips.gov.au)
Date last altered: 2 Nov 99