Using ground-based observations of the ultralow-frequency (ULF) (f = 10-3-1 Hz) waves in Earth's magnetic field as indirect diagnostics of the state of the magnetosphere and the solar wind has been popular because it has both scientific and practical return. First of all, these observations enhance our understanding of the physics of magnetospheric processes. Second, obtaining surface diagnostics of the state of the solar wind and magnetosphere might complement costly satellites when their direct, continuous, and detailed measurements are unavailable.
Remarkable correlations are found between ULF wave characteristics in the 10-s to 150-s range, or equivalently Pc3-4 band, and solar wind parameters [Bol'shakova and Troitskaya, 1968; Troitskaya et al., 1971; Gul'elmi, 1974; Greenstadt and Olson, 1976, 1979]. Pc3-4 activity occurs predominantly when the cone angle of the interplanetary magnetic field (IMF) is low. The wave frequency of this activity is also correlated with the IMF magnitude. These correlations indicate that a class of ULF waves in the magnetosphere originate from the upstream waves in the foreshock region when the IMF is approximately aligned with the Earth-Sun line. Those upstream waves can be carried by the solar wind convection into the magnetosheath, impinge upon the magnetopause, and then cause oscillations in the magnetosphere.
It has also been found that Pc3-4 wave power is correlated with the solar wind velocity [e.g., Singer et al., 1977], which suggests that the enhancement of wave power at high solar wind speed conditions might be due to the Kelvin-Helmholtz instability on the magnetospheric boundary. A joint correlation which incorporates both the solar wind velocity and the IMF cone angle reduces the overall scatter and suggests that both bow shock origin and the Kelvin-Helmholtz amplification for the pulsations are present [Greenstadt et al., 1979].
The propagation of the Pc3-4 wave energy from the upstream solar wind to the magnetosphere is still an important topic not fully understood. Perhaps owing to the clear correlation between the upstream waves and low-latitude observations, nearly all early efforts to model the transport of wave signals from the solar wind toward Earth have concentrated on the physical processes near the equator. It is possible that the wave energy may enter the magnetosphere through some other paths. Several studies have shown that Pc3 pulsations are most intense in the dayside cusp regions [e.g., Bol'shakova and Troitskaya, 1984; Morris and Cole, 1987]. Engebretson et al.  proposed that the pulsation-modulated precipitation of magnetosheath/boundary layer electrons may be responsible for the propagation of upstream wave power into the high-latitude ionosphere, from where the wave energy can be transported throughout the dayside outer magnetosphere. One of the purposes of this study is to examine the efficacy of this cusp entry hypothesis on the Pc3 pulsations observed at a low-latitude station.
The energy source of low-latitude Pc5 is also an interesting topic that requires more investigations. At latitudes for which , the fundamental frequency of field line oscillations is, in general, higher than Pc5 frequencies [e.g., Nishida, 1978, Figure 98; Menk et al., 1994, Table 1]. However, Pc5 pulsations are still observed at these latitudes [Lilley and Bennett, 1973; Ziesolleck and Chamalaun, 1993; Bloom and Singer, 1995]. The spatial structure, polarization, and diurnal variation of Pc5 waves in the low-latitude plasmasphere has been studied by using multiple ground magnetometers. Without the information of the upstream solar wind conditions, the early work considered that low-latitude Pc3-4 and Pc5 waves may be generated by a common or a similar solar wind-driven source mechanism [Ziesolleck and Chamalaun, 1993]. In this study, we will also study the correlation between the low-frequency (4-8 mHz) wave power and the solar wind parameters to find possible energy sources.
It is worth mentioning that there have been numerous attempts to determine the exact relationship between the wave power on the ground and the solar wind parameters [e.g., Wolfe et al., 1980; Wolfe, 1980]. The solar wind serves as the input energy, and the wave power is the output. The intermediate processes, albeit extremely complicated, can be described by a multivariate function to be obtained by a statistical regression analysis with a large amount of data. This type of study can further be inverted in an attempt to predict the state of the interplanetary medium from ground-based measurements. We will briefly discuss the possibility of such an approach.
In this study, we use 2 years (1978 and 1980) of the wave power data calculated by Bloom and Singer  (henceforth their paper is referred to as BS1) from the magnetic field observations at Mount Clemens, Michigan. The details of their wave power data are described in section 2. The main purpose of BS1 was to characterize the regular daily variation of the geomagnetic background spectrum. This study extends their analysis further to examine the important relationships between the solar wind properties and the wave power at different frequencies.