2. Results of Observations

Figure 1

[9]  The experimental OS data obtained in two campaign are analyzed. The first campaign took place in the period 24 February to 11 March 1999 and the second campaign was carried out on 20-25 October and 28-31 October 2003. The OS path Magadan (59.7^oN, 150.5^oE) to Irkutsk (51.8^oN, 104^oE) was used in the first campaign. Coordinated observations at the Russian network of LFM ionosondes [Ivanov et al., 2003] were carried out in the second campaign from 20 to 25 of October 2003. The following paths were used: Khabarovsk (47.5^oN, 134.5^oE) to Rostov on Don (47.3^oN, 39.7^oE), Magadan-Rostov on Don, Irkutsk-Rostov on Don, Khabarovsk-Irkutsk, Magadan-Irkutsk, Noril'sk (69.4^oN, 88.1^oE) to Irkutsk, Noril'sk-Rostov on Don, Khabarovsk-Nizhny Novgorod (56.1^oN, 44.1^oE), Magadan-Nizhny Novgorod, Irkutsk-Nizhny Novgorod, and Noril'sk-Nizhny Novgorod. In the period from 28 to 31 October the Inskip (England, 53.8^oN, 2.8^oW) to Rostov on Don path was also used. The round the clock observations were conducted. The LFM ionosondes at Magadan, Khabarovsk, Irkutsk, and Noril'sk operated every 15 min in the frequency range 4-30 MHz and the ionosonde in Inskip radiated every 5 min in the frequency range 4.2-30 MHz. The rate of the frequency variations for all ionosondes was about 100 kHz s^-1. The geometry of the paths is shown in Figure 1. Some results obtained in this experiment are presented below.

2.1. First Campaign

[11]  A typical feature of the OS ionograms at the considered path during geomagnetic disturbances was registration of the signals propagating outside the great circle arc. Additional signals with anomalously long delay began to be observed since 1200-1400 UT when the path was in the dusk-night sector. The time interval of registration of these signals reached 11 hours from the moment of their appearance. In quiet conditions on 26 and 27 February the anomalous signals were registered during a short interval of time from 1245 to 1545 UT. Their delay relative to the standard modes was 3 ms and more. The anomalous signals was characterized by a strong spread and the frequencies exceeded MOF of the standard propagation modes. In the substorm conditions on 24-25 February the anomalous signals were registered in the more broad time interval from 1200 to 2000 UT. In the beginning of this interval "quasi-multiple" signals relative to the 2F mode (relatively weakly spread) and then the "quasi-multiple" signals relative to the 1F mode appeared. The distance-frequency characteristics (DFC) of the "quasi-multiple" signals were repeating the shape of DFC of the standard mode in the range of higher frequencies. The delay of the anomalous signals relative to the standard propagation modes exceeded 3 ms. In the process of the disturbance development this delay was decreasing down to 1-2 ms. The anomalous signals were becoming more spread conserving the DFC shape. This shows that the main mechanism of anomalous signal formation is the refraction in the three-dimensional inhomogeneous ionosphere. The scatter at irregularities of various scales intensifies only a signal spread.

[12]  A typical feature of the small magnetic storms on 3-4 March and 6-8 March is a significant increase of the auroral index of magnetic activity after 1800 UT (in the night hours of the local time). The auroral oval shifts in this time down to the Magadan latitudes and the great circle arc of the path is located completely within the main ionospheric trough. The MOF and amplitude of the standard signal are sharply decreasing. The anomalous signals at these hours are formed due to the scatter at irregularities of different scales located within the auroral zone.

Figure 2

[13]  In the conditions of a moderate magnetic storm, starting from 1100 to 1200 UT the anomalous signals appear with a delay of more 3 ms relative to the standard modes propagating along the great circle arc. The analysis of the DMSP satellites data showed that in these hours the equatorial wall of the auroral oval approaches the Magadan-Irkutsk propagation path. In this situation transverse gradients of ionospheric parameters begin to influence the propagation conditions. Figure 2 shows the oblique sounding ionogram and the geometrical position of the auroral oval for the moment 1202 UT on 1 March 1999. Together with the propagation modes 1F2 and 2F2 there presents in the ionogram an additional signal with the delay exceeding the delays of the main propagation modes. The appearance of such signals may be due to the refraction in the region of the polar wall of the main ionospheric trough, because the main direction of the antenna pattern of the transmitting station Magadan is oriented to Moscow.

Figure 3

[14]  As far as the location of the main through relative to the propagation path changes, the structure and delay of the additional signals change considerably. Figure 3 shows the oblique sounding ionogram for 1432 UT on 1 March 1999 and the data on the geographical location of the auroral oval. Quasi-multiple propagation modes with the frequency dependence typical for the main modes at night are observed. As far as the ionospheric trough approaches the propagation path, the MOF of anomalous signals begins to exceed MOF of the standard propagation modes. The latter fact may be considered as an indirect confirmation of the assumption that these signals are reflected in the regions of the polar wall of the main ionospheric trough. The analysis of the vertical sounding (VS) data during the magnetic storm showed that the middle points of the propagation path of the main modes were located within the ionospheric trough.

Figure 4

Figure 5

[15]  In the vicinity of the equatorial edge of the auroral oval at the boundary with the northern wall of the trough there exist structural regions of the ionosphere with dimensions of about 100 km with a sharp horizontal gradient (so called globules) which present a source of a wide spectrum of irregularities. The strong spread of the signal in Figure 4 may be due to the radio wave scatter at intense irregularities existing within such structures [Basler et al., 1988; Uryadov et al., 2004a]. In the morning hours at a shift of the auroral oval into the region of higher latitudes, the delay of spread anomalous signals again increases up to 2-3 ms relative to the standard modes. Figure 5 shows the oblique sounding ionogram for 2232 UT on 1 March 1999 and the data on the geographic position of the auroral oval.

[16]  Thus the joint analysis of the ionospheric VS and OS data and satellite data shows that the appearance of the additional signals may be caused by the refraction of the ray beam of the main lobe of the transmitting antenna at the transverse gradients of the electron concentration in the vicinity of the northern wall of the ionospheric trough as well as to the scatter at irregularities existing in the vicinity of the equatorial boundary of the auroral oval and the trough northern wall. The performed numerical evaluations of the localization of the scattering region confirm possibility of such mechanism of side signals formation.

2.2. Second Campaign

Figure 6

[17]  In the second campaign the monitoring observations at the Russian LFM ionosondes network were carried out. The signals of the LFM ionosonde located in Inskip (England) were also received. Thus a vast longitudinal sector from England to Magadan was controlled by the oblique ionospheric sounding. The observations began in the moderately disturbed period. On 24 October 2003 at 1500 UT a moderate substorm disturbance was detected (see Figure 6). The disturbance was caused by the sharp jump of the solar wind velocity from 400 km s^-1 to 600 km s^-1 and the increase of the interplanetary magnetic field (IMF) from 10 nT to 25 nT after a series of flares of M and X classes what occurred on 22-23 October. Especially strong disturbance of the magnetic field took place on 29-31 October and was caused by a series of prominent flares. The flare of the X17.2/4B class, which began on 28 October at 0951 UT, should be specially mentioned. A dense and quick eruption of the substance with a velocity exceeding 2100 km s^-1 was observed during this flare. In the evening of 29 October one more proton flare X10.0/2B S15W02 occurred [Panasyuk et al., 2004]. As a result a series of extremely strong magnetic storms began. According to the data by Panasyuk et al. [2004], at 0612 UT on 29 October 2003 there was a sharp increase in the IMF magnitude from 10 to 35 nT and a sharp change of the B_z component orientation to southward with the value of - 25 nT. This effect was caused by a strong solar flare of class X17.2. Figure 6 shows the geomagnetic activity level expressed in the Kp (a) and Dst (b) indices. The observations were conducted on the background of this geomagnetic situation.

Figure 7

Figure 8

[18]  Now we analyze the most prominent events characterizing propagation of HF signals at midlatitude paths during geomagnetic disturbances. They were mainly observed at Rostov on Don where the reception of the signals was conducted at an oblique V antenna with the diagram oriented to the azimuth of A 20^o. The ionograms presented in Figures 7a, 7b, and 8a illustrate appearance at the Irkutsk-Rostov on Don, Khabarovsk-Rostov on Don, and Inskip-Rostov on Don of additional signals during geomagnetic disturbances. The propagation time of the additional signal exceeds the propagation time of the main mode and on the ionogram these signals are marked by SS (side signals). The direct signals propagating along the great circle arc are marked by DS. One can see that all SS are spread signals, their intensity being comparable to the intensity of DS.

[19]  The OS ionograms registered at the Irkutsk-Nizhny Novgorod and Khabarovsk-Nizhny Novgorod paths for the same moments of time as the ionograms in Figures 7a and 7b are shown in Figures 7c and 7d for comparison. One can see in Figures 7c and 7d that only direct spread signals are observed in the ionograms, whereas additional (side) signals are absent. Different appearance of the side signals in the reception points with different latitudes (in this case those are Rostov on Don and Nizhny Novgorod) may be caused by several factors including the following. We have already mentioned that the auroral oval is shifted to middle latitudes during a magnetic storm. Small-scale irregularities localized at its southern boundary and responsible for the aspect scatter may appear outside the "visibility" zone for the more northern configuration of LFM HF radars (in our case it is true for the paths Irkutsk-Nizhny Novgorod and Khabarovsk- Nizhny Novgorod in comparison with more southern paths Irkutsk-Rostov on Don and Khabarovsk-Rostov on Don). This concept is confirmed by the satellite observations of the auroral oval. According to the data of the site (http://sd-www.jhuapl.edu/Aurora/ovation/ovation_display.html) for the above indicated moments of time the paths Khabarovsk-Nizhny Novgorod and Irkutsk-Nizhny Novgorod cross the oval. In this case the irregularities would influence the signal spread due to the radio wave forward scatter at intermediate-scaled irregularities of the auroral zone. This effect is seen in OS ionograms as a halo around of the main propagation modes (see Figures 7c and 7d). It is worth noting that similar situation is observed at the Magadan-Irkutsk path at 2047 UT on 1 March 1999 (see Figure 4).

Figure 9

[20]  The detailed picture of the spread side signal dynamics with time during the magnetic storm on 29 October 2003 is visually seen in Figure 9 which shows a sequence of ionograms at the Inskip-Rostov on Don path. The ionograms were recorded every 5 min, time in the ionograms goes from left to right and from top to bottom. The SS signals are presented at all ionograms though not all of them are marked to avoid overloading Figure 9. One can see that SS demonstrate a strong dynamics: the distance-frequency characteristics of SS change every 5-10 min. The latter manifests a rapid variability of the position of the region responsible for the spread side signal formation. From 1337 to 1437 UT a decrease of the SS delay relative to DS from about 3 ms to 2 ms was observed. Estimates show that the rate of the motion of this region in the latitudinal direction was on the average of 2.5^o per hour.

Figure 10

Figure 11

[21]  One should note that during the main phase of the magnetic storm from 0812 to 1002 UT DFC of the direct signal was modulated by traveling ionospheric disturbances (TID) descending in the coarse of time downward along the Pedersen mode. Figure 10 illustrates the dynamics of the DFC in the presence of TID. Moreover, from 1100 to 1300 UT deep quasiperiodic oscillations of MOF with a period of about 1 hour and amplitude of about 2 MHz were observed. They are visually seen in Figure 11 where the time behavior of MOF at the Inskip-Rostov on Don path from 0600 to 1800 UT on 29 October 2003 is shown. In Figure 11 the intervals are hatched of observations of ionospheric disturbances with the DFC modulation in the vicinity of MOF (light hatching) and spread side signals (dark hatching). It is worth noting that these time intervals correspond to the main phase of the magnetic storm development characterized by a depletion of the Dst index of the geomagnetic field (see Figure 6b).