Using the data of the DMSP F6 and F7 satellites, we show that there is a systematic difference between ion precipitation in the auroral oval and in the polar cap. The oval ion precipitation is smooth, and the ratio of the total electron number flux to the total ion flux is less than 20. The polar cap ion precipitation is patchy, and the ratio is more than 20. The boundary between these two types of ion precipitation usually can be detected for both northward and southward interplanetary magnetic field by a sharp fall of the ion total number flux below the level 5 x 105-106 ions (cm2s st) -1. Since this determination of the poleward boundary of the auroral oval is efficient for any conditions, we propose to define the polar cap as a region bounded by this boundary. The polar cap determined in this way has roughly the shape of an ellipse aligned along the 1100-2300 MLT meridian with daytime and nighttime boundaries at 80 ø and 72 ø , respectively. Four patterns of electron precipitation in the polar cap are identified for northward
IMF conditions: (1) symmetric precipitation in both polarcaps with no distinct regularity. (2) symmetric precipitation with a prevalence of electron fluxes in the morning sector of both polar caps, (3) asymmetric distribution with no correspondence between the patterns in the opposite polar regions, and (4) asymmetric distribution where the increased electron fluxes are observed in opposite sectors (morning or evening) of the opposite polar caps depending on the polarity of the By IMF component. The energy spectra of electrons even in neighboring spikes in the same polar cap may be greatly different, and therefore discrepancy in electron spectra in opposite polar caps cannot be regarded as unique proof for the open character of the magnetic field lines. tions. Following Meng's [1981] suggestion that the polar cap arcs are generated by soft precipitation of the widened auroral oval, Lassen et al. [1988] have proposed defining the polar cap as a near-pole region free of aurora. In this case the polar cap is thought to be teardrop or pear shaped. The horse collar type aurora introduced by Hones et al. [1989] on the basis of satellite data also contains a "slot" bordered by bright arcs with no aurora inside it. Rich and Gussenhoven [1987], using electron measurements, claimed that the polar cap is a small region confined almost entirely to the nightside when IMF is near zero or slightly northward. However, such a concept is evidently not in conformity with great many experimentM data showing the appearance of Sun-aligned arcs along the noon-midnight meridian just in the teardrop-shaped region [Meng and Akasofu, 1976; Ismail et al., 1977; Lassen and Danielsen, 1978; Gussenhoven, 1982; Ismail and Meng, 1982; Gusev and Troshichev, 1986, 1990]. According to the results of Troshichev et al. [1988] the occurrence of Sun-aligned arcs along the noon-midnight 8337 8338 TROSHICHEV AND NISHIDA: PRECIPITATION PATTERNS IN THE POLAR CAP meridian is normal for conditions when Bz >0 and By •0. Another auror...