On May 11, the day the solar wind almost disappeared, the WIND spacecraft was located 50 RE upstream from the Earth and 40 dawnward from the sun-Earth line. If the bow shock were in its nominal location, the Wind would have been in the undisturbed solar wind because of its large distance from the usual bow shock location given the IMF orientation on that day, but the bow shock moved to unusually large distances and was even seen by Wind briefly (1730-1936 UT) due partially to the expanded radius of the magnetopause and partially to the low Mach number. These placed the Wind spacecraft much closer to the bow shock and into the foreshock region before 1030 UT and after 1936 UT on May 11. Meanwhile, the IMP-8 spacecraft was also located close to the bow shock on May 11. The IMP-8 data show that the spacecraft was in the region upstream from the bow shock before 1050 UT and downstream from the bow shock after 1308 UT with a data gap between 1050 to 1308 UT on May 11. Since the IMP-8 spacecraft was located duskward from the sun-Earth line, it thus remained in the undisturbed solar wind before 1050 UT on May 11. Figure 1 shows the schematic of the foreshock geometry and the spacecraft locations in the plane (the plane that contains the solar wind velocity, the IMF, and the spacecraft) on May 11. The planes for the two spacecraft differ but have been superimposed here under the assumption that the upstream waves are dependent mainly on the relative location to the bow shock. Due to the outward motion of the bow shock during this period, we used average location of the bow shock in Figure 1. The Wind and IMP-8 locations in Figure 1 are for the interval 01 to 11 UT on May 11 when both spacecraft were in the upstream region, showing that they were on either side of the foreshock boundary.
We compare the magnetic field data observed simultaneously by Wind [Lepping et al., 1995] and IMP-8 as shown in Figure 2. The top panels of Figure 2 are time series of magnetic field for three sample intervals of simultaneous Wind and IMP-8 data at 3 s resolution. The bottom panels show the corresponding power spectra (the sum of power in all three components). The power spectra clearly show the existence of upstream waves seen by Wind, but they have extremely weak enhancements of power. In the Pc 3-4 band, the peak power seen by Wind is at 10 nT2/Hz. This is true for all upstream waves observed by Wind on May 11. In comparison, the peak power of upstream waves under nominal solar wind conditions is 100 nT2/Hz [Le and Russell, 1990]. Furthermore, the upstream waves had very little compressional power as evident in the magnetic field strength. Thus, the foreshock was unusually quiet and the upstream waves were about one order of magnitude weaker on May 11, when the Mach number became small. Based on the preliminary calculations, the magnetosonic Mach number was only slightly above the unity for most of the day and possibly fell below the unity for a few hours. The weak bow shock resulted in few backstreaming ions present in the foreshock, thus only weak wave generation.
We could not, however, maintain the same locations of our upstream monitors, IMP-8 and Wind on the control day, May 14, although the foreshock geometry was similar. On May 14, Wind was not in the foreshock region because it was too far away from the bow shock. Fortunately we know very well what waves are seen in the upstream region under normal solar wind conditions [Le, 1991]. Since both the solar wind and IMF conditions are nominal on May 14, we know that the subsolar upstream region was filled with large-amplitude waves for IMF cone angle based on numerous previous observations [Greenstadt et al., 1980, Hoppe et al., 1981, Le and Russell, 1992]. If the current paradigm for the association of Pc 3-4 waves in the magnetosphere with the foreshock generated waves is correct we would expect that the magnetosphere was quite quiet on May 11 when compared with May 14.