3. Interpretation

Figure 2

[7]  No systematic seasonal differences in the Ap 50 index distribution exist (Figure 2). The Ap distributions are peaking at Ap 5 regardless the season; therefore the revealed annual threshold variations cannot be related to a peculiarity in the Ap index distribution. A well-known experimental fact that running average Ap indices exhibit a pronounced equinoctial maximum [e.g., Roosen, 1966] does not explain the threshold decrease during equinoxes (Figure 1); moreover, we consider the five smallest thresholds corresponding to low level of geomagnetic activity.

Figure 3

[8]  Negative long-duration F2-layer storms are known to result from thermospheric neutral composition variations, namely, O/N₂ decrease [e.g., Prölss, 1980, 1995]. Therefore there should exist some seasonal difference in the thermosphere reaction to geomagnetic disturbances. The MSIS 86 model [Hedin, 1987] gives very small seasonal differences in D [O/N₂] at the F2 region heights at moderate high latitudes ( j=55^o ) both for daytime and nighttime disturbances, with summer O/N₂ perturbations being larger than winter ones. This contradicts the European Space Research Organization (ESRO 4) observations by Prölss and von Zahn [1977], who found a pronounced seasonal difference in D (N₂/O) at 280 km for middle and high latitudes. The most important for present analysis result of Prölss and von Zahn [1977], observations is that "During summer the perturbations are of moderate magnitude compared with the larger disturbance effects commonly seen in the winter hemisphere". This seasonal disturbance effect is also present in the disturbed thermosphere composition model by Zuzic et al. [1997] based on the ESRO 4 observations. Figure 3, which can be obtained from the Zuzic et al. [1997] model, gives the R( N₂/O) = (N₂/O) _dist/ ( N₂/O) _quiet dependence on geomagnetic activity. The postmidnight (0000-0600) LT sector where negative F2 -layer disturbances are known to be the most frequent and pronounced was chosen for this illustration. Seasonal difference in the thermosphere reaction to geomagnetic activity is well presented in this model for enhanced geomagnetic activity. The ESRO 4 observed seasonal N₂/O disturbance variations can explain the revealed threshold seasonal changes (Table 2 and Figure 1). Indeed, large winter and equinoctial N₂/O disturbance effect (steep dependence on Kp ) needs lower level of geomagnetic activity to overcome the same dN_mF2 disturbance threshold (40% in our case). In summer when N₂/O perturbations are small (gently sloping dependence on Kp ), higher geomagnetic activity level is needed to obtain the same F2 -layer negative storm effect.

[9]  The seasonal (winter/summer) threshold variations seem to comprise two parts. The first one reflects seasonal changes of neutral temperature. For the sake of simplicity we may suppose that the thermosphere is isothermal and neutral species [O] and [N₂] are distributed in accordance with the barometric law: [O] = [O]₀ exp(-h/H) and [N₂] = [N₂]₀ exp(-1.75 h/H), where H = kT_n/mg is the atomic oxygen scale height. This gives [O]/[N₂] exp(0.75h/H ). Neutral temperature, T_n increases during disturbed periods, so we can write down d (O/N₂)/dH - exp(0.75h/H)/H². This expression tells the higher the background neutral temperature T_n (and corresponding H ), the smaller the [O]/[N₂] storm-induced changes. Neutral temperature is maximal in summer and minimal in winter (e.g., MSIS 86); therefore summer storm-induced [O]/[N₂] variations should be less compared to winter ones, and this is in line with the ESRO 4 observations.

[10]  The second part of the seasonal effect may be attributed to the seasonal difference in the spatial distribution of the perturbed neutral composition [Prölss and von Zahn, 1977]. In summer the [O]/[N₂] disturbance zone may extend all the way from the polar to the low latitudes, while in winter it is restricted to high latitudes only. This means that the same energy deposited in the auroral zone during a geomagnetic storm and resulted in the thermosphere perturbation is smeared over the whole hemisphere in summer, but it is localized only at higher latitudes in winter. The effect is known to be due the interaction of seasonal (background) and storm-induced thermospheric circulation [Duncan, 1969; Field et al., 1998; Forbes et al., 1996; Mayr and Volland, 1972]. This needs stronger geomagnetic disturbances (higher threshold) in summer compared to winter to have the same perturbation effect in neutral composition.