[2] Usual F2 -layer storms are related to geomagnetic disturbances which are characterized by various indices of geomagnetic activity. It is well known that the correlation between relative dfoF2 deviations and various indices of geomagnetic activity is not very high [e.g., Araujo-Pradere et al., 2002; Wrenn, 1987; Wu and Wilkinson, 1957; Zevakina et al., 1990]; however, needs of practice force researchers to develop methods for ionospheric foF2 short-term prediction based on indices of geomagnetic activity [e.g., Araujo-Pradere et al., 2002; Cander and Mihajlovic, 1998; Cander et al., 1998; Chan and Cannon, 2002; Francis et al., 2000, 2001; Kutiev and Muhtarov, 2001; Kutiev et al., 1999; Marin et al., 2000; Mikhailov, 1990; Muhtarov and Kutiev, 1999; Muhtarov et al., 1998; Wintoft and Cander, 2000; Wu and Wilkinson, 1995; Zevakina et al., 1990]. Planetary geomagnetic activity indices may serve as an indicator of an average (in a statistical sense) disturbance level over the globe, while F2 -layer perturbation picture is individual for each particular storm and depends on the geomagnetic storm intensity, season, latitude and longitude, UT and LT of storm onset etc. Ionospheric F2 -layer storms at middle latitudes mainly reflect the state of the disturbed thermosphere (winds, neutral temperature and composition) resulted from high-latitude energy deposition via magnetospheric electric fields and particle precipitation. On one hand, this energy input spatially is not uniform during a geomagnetic storm and different longitudinal sectors turn out to be in different situations, on the other hand the thermosphere reaction to one and the same external impact is different and depends on its current state and prehistory. Unfortunately, the latter is not known in principle and hardly any thermosphere monitoring will be implemented in observable future. All these complexities do not allow to produce a deliberate foF2 forecast with an acceptable accuracy for periods of geomagnetic storms [e.g., Wintoft and Cander, 2000] which are the most interesting and important from practical point of view. Because of a complex and indirect relationship between F2 -layer storm effects and the level of geomagnetic activity there is a wide range of estimates for the time delay between geomagnetic and ionospheric storm onsets: 0-6 hours for positive disturbances [Zevakina and Kiseleva, 1978], 12 hours [Wrenn et al., 1987], 15 hours [Wu and Wilkinson, 1995], 6-12 hours [Forbes et al., 2000]; 16-18 hours [Kutiev and Muhtarov, 2001], and 8-20 hours depending on season [Pant and Sridharan, 2001].
[3] However, researchers are working in this direction and any empirical regularities revealed in the F2 -layer storm morphology are important for creating the methods of foF2 short-term prediction. Apart from this practical aspect of the problem, physical interpretation of the revealed regularities in storm features help understand F2 -layer storm formation mechanisms and this is important for ionospheric physics. This paper is devoted to analysis of the seasonal variations in the geomagnetic activity threshold for the ionospheric F2 -layer negative storms onset. The results of statistical analysis based on foF2 ionosonde observations are interpreted in the framework of the F2 -layer formation mechanism.