|Back to INAG Homepage||Back to UAG-104 Contents page|
SOME EVIDENCE SUPPORTING SHORT-TERM FORECASTING OF MID-LATITUDE IONOSPHERIC DISTURBANCE CONDITIONS
G. G. Bowman
Department of Physics, The University of Queensland, Brisbane, Qld 4072, Australia
The relationships (determined statistically) that have been reported in recent papers between auroral-zone geomagnetic activity and spread-F occurrence in mid-latitude regions, are illustrated by a few specific events. These relationships include a 3 day delay as well as delays measured in hours, thought to result from upper-atmosphere neutral-particle density (UA-NPD) changes which occur in mid-latitudes following geomagnetic activity. Increased geomagnetic activity in the early-evening hours results in a suppression of spread-F occurrence in the hours which follow, while reduced activity produces the opposite effect of enhanced spread-F occurrence. It is suggested that these short-term delays may be useful in forecasting ionospheric disturbance conditions.
A considerable amount of experimental evidence has accumulated over the years to support the idea that the primary cause of mid-latitude spread-F traces on ionograms is the existence of medium-scale travelling ionospheric disturbances (MS-TIDs) (see review by Bowman, 1990). It seems likely that these MS-TIDs are produced by the passage of ionospheric-level atmospheric gravity waves (IL-AGWs) (Dyson et al., 1970). Commenting on IL-AGWs Hines (1963) states "Because of their increase in amplitude with height (in the absence of strong dissipation) these waves would be more effective in the production of spread-F when the F layer is high." Of course the reason for this is the reduction with height of the upper-atmosphere neutral-particle density (UA-NPD). However, a reduction (or increase) in the UA-NPD at any specific height can also be expected to influence the IL-AGW wave amplitudes (and the corresponding MS-TID wave amplitudes). Recent results (Bowman 1992, 1993a) have shown that for mid-latitudes the diurnal, annual and sunspot-cycle variations of spread-F occurrence, as well as the annual and sunspot-cycle variations of equatorial spread-F, are closely related to UA-NPD changes. These associations are completely consistent with the concept of wave amplitudes (which need to exceed certain values for spread-F to be recorded) being modulated by UA-NPD changes.
Thus, it was not surprising to find from more recent analyses (Bowman, 1993b) (confirmation of earlier results - Bowman, 1984) that UA-NPD increases in mid-latitude regions which are expected approximately 6 hours after enhanced geomagnetic activity, seem to be effective in suppressing the level of spread-F occurrence. The reverse situation is also noted (at least statistically) where high levels of spread-F occurrence are found to be associated with low levels of geomagnetic activity (Bowman, 1993b). Since mid-latitude daytime disturbances are found to be associated with nighttime spread-F (Bowman, et al., 1987; Bowman, 1993c) the fact that these disturbances also appear to respond to geomagnetic activity a few hours earlier, is consistent with the spread-F results. An additional result indicates that UA-NPD changes directly associated with changes in the level of the 10.7 cm solar flux, also appears to influence the level of mid-latitude spread-F occurrence when considered on a world-wide scale (Bowman, 1993b). Another result involving delays of days rather than hours indicates that mid-latitude spread-F is enhanced after a delay of about 3 days following increased geomagnetic activity around midnight and also in the early morning hours around 0400 (local time) (Bowman 1993c). It has been proposed that delayed D-region absorption events in sub-auroral regions seem likely to be responsible, resulting in the generation of IL-AGWs with appropriately large wave amplitudes (Bowman, 1991a,b). These IL-AGWs are responsible in turn for the MS-TIDs which are associated with spread-F occurrence.
The Possibility of Short Term Forecasting
These results which have just been explained briefly, may allow estimates to be made of mid-latitude disturbance conditions a few days and also a few hours after noting the level of geomagnetic activity at particular local times. Thus short-term forecasting becomes a possibility. These recent results (Bowman, 1993b) have used spread-F controls to look for associated geomagnetic-activity levels. However, to obtain reasonable estimates of the reliability of these relationships for forecast purposes, the controls need to be obtained from geomagnetic-activity levels. With these controls the anticipated changes in spread-F occurrence levels can then be investigated. The statistical significance of these relationships using geomagnetic-field controls is indicated in Figure 4(b) of Bowman (1984). Spread-F occurrence levels at Moggill (35.7° S, 226.9° E geomagnetic (Gm) - near Brisbane, Australia), Norfolk Island (34.7° S, 243.2° E Gm) and Wakkanai (35.3° N, 206.0° E Gm) are considered. It is shown that for K indices at Macquarie Island (61.1° S, 243.2° E Gm) having values of 5 or 6, significantly suppressed spread-F activity after midnight is specifically related to this level of geomagnetic activity for interval 3 (1600 - 1900 Moggill local time, LT = UT + 10h). Furthermore, enhanced spread-F activity with a delay of 3 days is associated with interval 6 (0100 - 0400 Moggill local time). For very-low geomagnetic activity (ie. for K indices equal to 0 or 1) Figure 4(a) of Bowman (1984) records significantly enhanced post-midnight spread-F occurrence when using the K-index interval 4 (1900 - 2200 Moggill local time). It should be noted that minor time adjustments are needed for the local times of Norfolk Island and Wakkanai. Macquarie Island is in the same longitude zone as the spread-F stations.
The statistical significance of relationships between geomagnetic activity and mid-latitude spread-F occurrence has been discussed in the previous paragraphs. However the primary aim of this present paper is to consider a couple of specific examples which on some occasions illustrate clearly these relationships.
The most favourable conditions for the increased recording of spread-F traces in mid-latitudes would seem to be enhanced geomagnetic activity in the early-morning hours on a specific date coupled with conditions of low-level activity in the early-evening hours on the nights the delayed spread-F occurrence is expected (ie. after a delay of 3 days for Brisbane (Moggill)). When geomagnetic activity is low the associated UA-NPD levels will be lower than average levels for the particular time of the year being considered. Therefore the probability of spread-F being recorded should be increased due to the anticipated MS-TID wave-amplitude increases. Since for any location the date changes at midnight, each nighttime recording of spread-F involves two dates in local time (LT). Consequently since for Brisbane LT = UT + 10h it has been found more convenient to use Universal Time, thus assigning the same date to pre-midnight and post-midnight spread-F occurrence. On any night 1800 to 0600 in local time becomes 0800 to 2000 in Universal Time. Similarly for geomagnetic activity early-evening levels are recorded in Universal Time on the same date as early-morning levels the following day in local time.
Figure 1 indicates (see dashed lines) average levels of spread-F occurrence at Brisbane for sunspot-minimum periods for (a) June-solstice months (b) equinoctial months and (c) December-solstice months. Average occurrence levels (quoted in hours per night) are found to be consistent with an inverse relationship with the semi-annual variation of the UA-NPD (Bowman, 1992). In Figure 1(a) the average level is 8.9 hours per night, in Figure 1(b) it is 2.2 hours while in Figure 1(c) it is 4.5 hours. However it can be seen that, from night to night, significant changes can occur in this level of activity due, it seems likely, to the short-term effects which have been discussed.
Figure 2 illustrates two examples of enhanced spread-F occurrence at both Brisbane and a higher latitude station, Canberra (44.0S, 224.3E Gm) relative to geomagnetic activity levels at Macquarie Island as recorded by K indices. The spread-F is recorded a few days after early-morning (0100 to 0400 LT) enhanced geomagnetic activity (see labels C and H) combined with suppressed activity (see labels D and I) in the early-evening (1900 to 2200 LT) a few days after the early-morning increases. Also, the 10.7 cm solar flux (see Figures 2(e) and 2(j)) is shown to be below the average monthly levels for both examples. The observation from the diagrams in Figure 2 that Canberra spread-F occurrence precedes the occurrence at Brisbane (see labels A,B and F,G) by about a day is consistent with earlier results which indicated that following enhanced geomagnetic activity, the delays recorded increase for stations further from auroral-zone regions. (Bowman, 1979, 1982). Figure 3 illustrates two occasions when the level of Brisbane spread-F occurrence is reduced significantly. Associated AE indices and Macquarie Island K indices are both plotted. For the Macquarie Island records, the highest index for the local time interval 1600 to 0100 has been plotted for each day. Since the UA-NPD can increase to several times its initial value in the hours following severe geomagnetic conditions (Jacchia, 1965) it is not surprising that at these times the IL-AGWs wave amplitudes (and the corresponding MS-TID wave amplitudes) are not able to increase significantly at F2-layer levels to allow spread-F to be recorded. Consequently although, as Figure 1 shows, during some months of the year in a sunspot-minimum period spread-F at Brisbane is frequently recorded all night long, nevertheless even at these times increases in the UA-NPD can reduce the occurrence rate to a very-low level (Figure 3) on some nights.
If it is important to know in advance what the level to mid-latitude spread-F occurrence might be (relative to the expected level for the season - see Figure 1) on any particular night, it is proposed that real-time monitoring of geomagnetic activity at an auroral-zone station in the same longitude zone, should be helpful. Reasonable estimates can then probably be made of expected changes in nighttime disturbance conditions. Of less importance, but still exerting some measure of influence, is the 27-day variation of the 10.7 solar flux, because of its effect on UA-NPD levels (Jacchia, 1965).
The literature on the forecasting of ionospheric conditions seems to be mainly concerned with ionospheric parameters particularly foF2 (Ivanov-Kholodny and Mikhailov, 1986; Houminer et al., 1993). However this present paper (and the papers to which it refers) has been concerned specifically with short-term forecasting of mid-latitude ionospheric disturbance conditions. Concerning ionospheric parameters it should be noted that during the passage of MS-TIDs which produce disturbance conditions, there are changes to foF2 and h'F (Bowman, 1990) over short time intervals.
Apart from the 3 day delays in mid-latitude spread-F occurrence, which have been explained in other ways, it is proposed here that UA-NPD changes due to geomagnetic activity at specific local times, are effective in modifying spread-F occurrence levels after delays of a few hours. This modification occurs relative to spread-F occurrence levels due (it seems likely - Bowman, 1992) to the semi-annual and sunspot-cycle variations in the UA-NPD levels, as is illustrated by the dashed lines in Figure 1. Short-term forecasting appears a possibility.
Bowman, G. G., Latitude dependence of the time delay in spread-F occurrence following geomagnetic activity, J. Atmos. Terr. Phys., 41, 999-1004, 1979.
Bowman, G. G., Spread-F occurrence in mid- and low-latitude regions related to various levels of geomagnetic activityJ. Atmos. Terr. Phys., 44, 585-589, 1982.
Bowman, G. G., Some effects of geomagnetic activity and ionospheric height rises on mid-latitude spread-F occurrence, J. Atmos. Terr. Phys., 46, 55-64, 1984.
Bowman, G. G., A review of some recent work on mid-latitude spread-F occurrence as detected by ionosondes, J. Geomag. Geoelectr., 42, 109-138, 1990.
Bowman, G. G., Delayed ionospheric absorption events following enhanced geomagnetic activity, Indian J. Radio Space Phys., 21, 73-79, 1991a.
Bowman, G. G., Delayed mid-latitude spread-F occurrence following enhanced geomagnetic activity, Indian J. Radio Space Phys., 21, 80-88 1991b.
Bowman, G. G., Upper atmosphere neutral-particle density variations compared with spread-F occurrence rates at locations around the world, Ann. Geophysicae, 10, 676-682, 1992.
Bowman, G. G., The influence of the upper-atmosphere neutral particle density on the occurrence of equatorial spread-F, Ann. Geophysicae, 11, 624-633, 1993a.
Bowman, G. G., Short-term delays in the occurrence of mid- latitude ionospheric disturbances following other geophysical and solar events, in press J. Geomag. Geoelectr. 1993b.
Bowman, G. G., Daytime and nighttime ionospheric disturbances and their delayed occurrence after geomagnetic activity, Indian J. Radio Space Phys., 22, 1-10, 1993c.
Bowman, G. G., G. S. Dunne and D. W. Hainsworth, Mid-latitude spread-F occurrence during daylight hours, J. Atmos.Terr. Phys., 49 165-176, 1987.
Dyson, P. L., G. P. Newton, and L. H. Brace, In situ measurements of neutral and electron density wave structure from the Explorer 32 satellite, J. Geophys. Res., 75, 3200-3210, 1970.
Hines, C.O., The upper atmosphere in motion, Quart. J.R.Met. Soc., 89, 1-42, 1963.
Houminer, Z., J. A. Bennett and P. L. Dyson, real-time ionospheric model updating Electr. Electron. Eng. Aust., 13, 99-104, 1993.
Ivanov-Kholodny, G. S. and A. V. Mikhailov, The prediction of ionospheric conditions, Geophysics and Astrophysics Monographs, D. Reidel Publishing Coy, Dordrecht, 1986.
Jacchia, L. G., Atmospheric structure and its variations at heights above 200 km, Cospar International Reference Atmosphere (CIRA) 1965. North-Holland Publishing Coy. - Amsterdam, Appendix II, pp.293-313, 1965.