2 5 1 9 8 7
r.
O B S E R V A T I O N S
OF
A R T I F I C I A L S A T E L L I T E SOF
T H E E A R T H N° 2 51 9 8 7
M E S T E R S É G E S H O L D A K M E G F I G Y E L É S E I N ° 2 5
1 9 8 7
On the Dependence of NPSE on Different Geophysical Parameters E. Illés-Almár and I. Almár
Konkoly Observatory, H-1525 Budapest, Box 67, Hungary
Earlier we have demonstrated that the density (<5 ) deduced from atmospheric drag of satellites indicates increased values after geomagnetic storms as compared to semiempirical models
(CIRA-72 in particular) [l] . The phenomenon was called by us
"Neutral Post-Storm Effect": NPSE. The density maxima compared to CIRA-72 occured four to six days after the onset of geo-
satellite separately. After eliminating all remaining long terms from the f-curves, all f-values have been devided by the mean of geomagnetic storms. An earlier investigation .indicated that
there are somé extra maxima on the f-curves nőt belonging to
lected D st minima as indicator of geomagnetic disturbances.
In [3] we examined the profile of NPSE in different posi- tions of the perigee, as well as high or low A^ maxima with or without an ionospheric post-storm effect (characterized by a maximum in MIA - Mean Ionospheric Absorption) e t c . The same investigation has been repated later transforming the position of the perigee intő geomagnetic coordinates, but no significant selected by chance and the corresponding MSE curves constructed fór 35 - 35 days before and after storm onset. These curves, together with a mean MSE curve (heavy line), are plotted on Figure 1. It is evident that if a mean curve of at least 50 events is available then accidental fluctuations are cancelled out and the NPSE clearly vis'ible.
-4 5 / 372
-In order to stress this asymmetry, the pre-storm part of the curves was reflected with respect to the storm-onset .time and traced by thin lines on Figure 2. On all other figures the curves are plotted only 20 - 20 days before and after the storm, since more distant fluctuations have obviously little influence on the results of our investigations. we considered essential to investigate whether an accidental clustering of the disturbances can lead to an apparent effect the storm was free from another significant disturbance (curves marked by "c"). On Figure 2 the three basic curves were plotted and the post-storm effect (with respect to the corresponding reflected curves) accentuatéd by shading. We observed with satisfaction that though in case "a" we have an NPSE of ~1%,
3 7 3
-all cases, or our observational matériái is nőt always good enough to detect such a delicate phenomenon.
DEPENDENCE OF NPSE ON SOMÉ GEOPHYSICAL PHENOMENA
It is evident from our previous results that in thé time interval under consideration there are considerably more maxima on the f-curves than geomagnetic storms as indicated by observed, the geomagnetic disturbance was attributed to coro- tating streams (42 cases, 32%). Sometimes a solar flare appeared unexpectedly enough, they are reál and important disturbances:
the small group of these events does nőt coincide with the group
The difference between disturbances of different origin (shock waves and corotating streams respectively) has been
analysed from the point of view of NPSE. Since there were enough
"shock wave cases" only, their MSE curves have been compared to totál MSE curves on Figure 5 , and to MSE curves of the com-plementary group on Figure 6 . Both figures illustrate that NPSE
W f f i 374 te
is larger by cca 1% in shock wave cases than in all other cases and the p o s t - s t o r m effect pro b a b l y starts somewhat earlier as w e l l .
C O N C L U S I O N
It seems to be c o n v i n c i n g that the neutral post- s t o r m effect is a generál consequence of geoma g n e t i c d i s t u rbances - as in d i c a t e d by m i n i m a - and its a m p l itude is larger if the d i s t u r b a n c e is c o r m ected to somé kind of shock w a v e . The a m p l i t u d e of the NPSE is 1 - 3%.
R E F E R E N C E S
1 E .I I l é s - A l m á r , P.Bencze, F:Márz: Is there any "after effect" in density v a r i a t i o n of the neutral a t m osphere?
Nabl. ISZ 23 pp 333-337 Praha 1984
2 E. Illés-Almár: I n v e s t i g a t i o n of the 27-day peri o d i c i t y in the t h e r m o s p h e r i c dens i t y f l u c t uations
Space R e s e a r c h 1_9 PP 207-210 P e r g a m o n Press 1979
3 E .I l l é s - A l m á r , I.Almár, P.Bencze, Á.Horváth: I n v e s t i g a t i o n of the t h e r m o s p h e r e — i onosphere i n t e r a c t i o n by means of the n e u t r a l post storm effect
COSPAR, Toulouse C.1.2.6 in print 1986
- 3 7 6
-132
- 77
20. day
F i g u r e 3
132 103
100 81
69 17
F i g u r e 4
3 7 7
-0.1
-
132
-
87
- 100
-
47
a
2 0. day
F i g u r e 5
0.1
05 67
47 53
20. day
F i g u r e 6
P
AÍV. Space Rés. Vol. 9, No. 12, pp. (12)205-(12)208, 1989
**G eodetical and G eophysical Research Institute, H -9401 Sopron, B ox 5, Hungary
ABSTRACT
The neutral post-storm effect is reconsidered by means of accelerometrio data. Sinoe Ap has proved to be different function of Kp during and outside recovery phases, but a unique func- tion of Dst, the latter is considered as a better index fór correcting the effect of geomag
netic activity in models, i.e. it seems that the ring current plays an important role in the geomagnetic effect of the equatorial thermosphere.
INTR0DUCTI0N
It has been demonstrated in former papers that a density increase nőt previously considered in upper-atmospheric models occurs in the upper atmosphere after geomagnetic storms /1,2/.
This phenomenon has been called "neutral post-storm effect" (NPSE) and has been found by means of f-values representing the ratio of density values determined from satellite orbital drag to model values corrected even fór geomagnetic activity. The excess density has been at
tributed to heating due to energetic particles precipitating intő the upper atmosphere as a consequence of charge exchange between ring current ions and neutral H atoms in the geocorona (ENA) /3/.
It is well known that atmospheric data based on the orbital decay method have a limited time resolution. In the present paper the NPSE is reconsidered by means of acoelerometric data of much better time resolution.
DATA, METHOD AND RESULTS
Our recent investigations weré based on the measurements of the French CASTOR/CACTUS accel- erometer around minimum solar activity. The acoelerometric density data have been compared with corresponding totál density p values of the DTM model /4/ substituting Kp=0; the differ- ence between observed and calculated densities was formed. In the first step, CACTUS data of the interval MJD 42590-43010 (July 1975 through August 1976)' referring to altitudes between 400 and 403 km were selécted and analysed (their time-resolution consequently^decreased to only 8-15 data per day). Since the orbital inclination of the satellite was ~ 30 , our inves
tigations refer only to the reaction of the equatorial region to geomagnetic heating. The observed values belong to two groups according to LST (Local Solar Time), because upleg and downleg crossings are separated by~ 6 hours. The variation of LST within each lég is negligi- ble, therefore daily average means could be calculated giving two separate Ap values each day.
Such series of mean values correspond better to the frequency óf Kp and Dst and consequently to deduced model density values.
At first our intention was to verify the existence of the NPSE - based on this kind of obser- vational data as well. The cross-correlation between the Ap values and the geomagnetic indices Kp and Dst indicate an increase in the correlation if a shift of 6-8 days was applied (Fig. 1a). In the case of Dst the correlation proved to be significant at the 99% signifi
cance level. (As it is known the Dst index indicates the intensity of the ring ourrent.) The autooorrelation functions of the indices Kp and Dst alsó indicate a small increase on the 6th day (Fig. 1b), i .e. the geomagnetic activity has a slight recurrence tendency of 6 days during the NPSE time interval - the cross-correlation analysis can nőt decide which part of the excess density is due to the recurrence and which part to the NPSE.
Therefore the Ap values have been separated intő two groups according to their epoch with respect to geomagnetic storms, viz.: those belonging to one óf the recovery phases and those
(12)205
(12)206 E- Illés-Almár et al.
in any other remaining time interval (excépt the main phase). In each group the dependence on Kp and Dst has been studied separately forming mean values of Ap in appropriate Kp as well as Dst intervals (Ap). The results are demonstrated in Fig. 2a.
Fig. 1a. Cross-oorrelation functions between measured deviations from DTM model totál density (putting Kp=0): Ap and Kp, Dst geomagnetic indices respectively.
1b. Autocorrelation functions fór Kp and Dst values.
Fig. 2a. Ap versus Kp and Dst fór time intervals inside and outside recovery phases separately 2b. Separation of Ap(Kp) function (see first part of 2a) according to local solar time:
morning (4-10 hours); day (10-18 hours); evening (18-22 hours); night (22-4 hours).
n
Upper Atmosphere Ring Current Effect l'O j (12)207
The two ip curves are different in the case of Kp; Ap has a steeper increase with Kp during the recovery phase than in the remaining time interval:
Ap = [0.054 (K - 1 ) 2 - 0.047] 10”12 kg.nf3 Ap = [0.030 (K - 1)2 - 0.079] 10^12 kg.m-3 while the dependence of Ap ón Dst in the recovery phase the remaining time interval.
Since the density is a double valued function of Kp, bút a unique function of Dat, it is ob- vious that at low latitudes Dst is a better index with regard to the geomagnetic effect in the neutfal thermosphere than Kp. This can alsó be seen from Table 1, that gives the correlation coefficients between Ap measured on the n-th day after geomagnetic disturbances and the cor- responding Ap or Dst respectively.
In the case of Dst the correlation is almost constant until the 7th day, creases fór Ap. It means that in the recovery phase the correlation with vanishes with time, bút remains almost constant fór Dst.
bút it strongly de- Ap of the same day
The material was further separated according to diurnal phase (LST). The dependence on local time is plotted -in Fig. 2b. The Ap(Kp) curves diverge more in the evening hours and less during daytime. It is alsó obvious that the response of the atmosphere to strong geomagnetic heating is more pronounced in the daytime than at night - a conclusion stated previously by Berger et al. /5/. A detailed analysis of the diurnal behaviour of the geomagnetic effect will be the topic of another paper.
0 -50 -100 -150 Dst
Fig. 3- Ap versus Dst. A least squares linear fit is alsó given.
(12)208 E. Illés-Almár et al.
USE OF Dst FÓR MODELLING
Using the single valued dependence of Ap on Dst (Fig. 3) the Kp=0 version of the DTM model can be complemented as a first approximation by a simple linear term fór the geomagnetic ac
tivity effect. (The model is, however, limited to the altitude and latitude interval in ques- tion.) The proposed term is
Ap = (-0.0125 Dst - 0.110).10“12 kg.nf3 /3/
determined as a best fit to points in Fig. 3. Using equation (3) a histogram of the residuals is plotted a./ fór the original 420 days (Fig. 4a)
b./ fór an additional 309 day control interval (Fig. 4b).
The oontrol interval indicates that equation (3) can be extrapolated in time, hence at low latitudes Dst is a comprehensive and appropriate index fór the geomagnetic activity effect.
Fig. 4. Histogram of deviations from
In our former studies it was found that in post-storm periods a density excess ocours compared with model values (using the Kp index to consider the geomagnetic effect). Therefőre it has been named a neutral post-storm effect and attributed to an additional heating process. The present investigation indicates, however, that the geomagnetic effect can, at low latitudes, be described as a function of the Dst index alsó in the post-storm period (in contrast to the Kp index). It seems that in this case there is somé sort of process which is linked with the ring current and thus it is nőt restricted solely to post-storm periods. The double valued character of the Ap (Kp) function in Fig. 2a indicates different behaviour during and outside recovery phases respectively. Therefőre it is clear that straightforward use of the Kp or Ap index is nőt sufficient to characterize the geomagnetic effect in atmospheric models. On the other hand we have good reason to believe that at least at low latitudes a more appropriate description of the geomagnetic effect is possible utilizing the Dst index, because of the better correlation of Ap with Dst, and, furthermore, because of the identical dependence of Ap on Dst inside and outside the recovery phase. Our results alsó imply that a more suitable cor- rection fór the geomagnetic effect in the neutral upper atmosphere is necessary - considering nőt only high bút alsó médium and low latitudes. This might be realized by taking intő account the complex natúré of the geomagnetic activity consisting of polar and equatorial sources.
We plán to extend the time and height interval of our analysis as well as thoroughly inves- tigate the connection between the diurnal and geomagnetic effect in the upper atmosphere.
We should like to express our thanks to CNES and to Prof. F. Barlier in particular fór making the CACTUS material available to us. Mrs. M. Nagy and Mr. P. Decsy are thanked fór their able help in the preparation of this paper.
[17]
c