Digital power spectral analysis and coherency analysis are powerful techniques for studying ultra‐low‐frequency (ULF) waves in the earth's magnetosphere. Wave polarization parameters provided by these techniques are important in the development of theoretical models for wave generation. Because of this, it is important to understand the capabilities of the digital analysis techniques. Three different techniques of using the spectral matrix to do wave analysis have been presented in the literature. Because data for wave studies involve measurement in arbitrary coordinate systems, it is necessary to transform the spectral matrix to the principal plane of the wave before coherency analysis can be performed. The fundamental differences in the three techniques lie in how they determine the transformation to the principal plane. A comparative study of these three techniques was done using simulated data involving known wave and noise properties and real ULF wave event data from the geosynchronous satellites ATS 1 and ATS 6. In general, the quality of performance of the three different techniques on both simulated and real wave events was approximately the same.

Power spectra of geomagnetic field variations measured at synchronous equatorial altitude (geomagnetic shell parameter L ∼ 6.6) in the magnetosphere are used to calculate the time dependence of the radial diffusion coefficient for particles in the radiation belts. The diffusion coefficients calculated are mainly applicable for relativistic electrons. The magnitudes of the derived diffusion coefficients using data only from local day observations are consistent with those reported from analyses of most particle observations and thus are slightly larger than those derived from magnetic sudden commencements. They are consistent with the diffusion coefficients calculated from power spectra of ground‐based geomagnetic data measured near L = 4.

A statistical study of 215 magnetic pulsation events identified on the basis of waveform and period as Pc 4 (45‐ to 150‐s period) observed at synchronous orbit by ATS 6 is reported. The study was designed to be similar to that done with Pc 3 (10‐ to 45‐s period) by Arthur et al. (1977). These Pc 4 are found to occur most often near 1800 local time but are observed at all local times. The dominant period is ∼100 s, but a secondary peak in the distribution occurred at ∼53 s. They tend to be nearly linearly polarized, principally transverse to the ambient magnetic field. The Pc 4 are found to divide naturally into two classes (radial and azimuthal) on the basis of the azimuth of the major axis of polarization. These classes divide further into low‐frequency (≤0.015 Hz) and high‐frequency (≥0.015 Hz) classes. These four classes organize the remaining wave characteristics quite well. Division of the Pc 4 events into these classes does not lead to identification of a single source mechanism as it did with Pc 3. However, some correspondence between the classes and various proposed mechanisms does exist.

For 253 of the 420 ATS 6 Pc 3 events discussed in a previous paper, simultaneous IMF data from Imp J are available. By combining these two data sets an examination of the relationships between pulsation and IMF parameters is done for the first time using pulsation observations by a spacecraft in the magnetosphere. The results of this study indicate that within statistical errors, Pc 3 occur with approximately equal probability for normal IM F conditions. There is an indication of a preference for the toward spiral angle. No relationship between the pulsation frequency and the IMF magnitude was seen, but there is a clear, although weak, relationship between the angle of the IMF from the sun-earth line and the amplitude of the pulsations. Comparison of these results to those obtained by other researchers indicates that the different methods of analysis used are not yet sophisticated enough to deal with the obviously complex phenomenon involved.

A search for simultaneous 10‐ to 45‐s waves at the synchronous orbit at the location of ATS 1 and at Tungsten, Northwest Territories, Canada (near the foot of the ATS 1 field line), during 1969 resulted in the identification of only one event. This event occurred near 0900–1000 LT and was identified as a Pc 3 event. By combining 1967 data from College, Alaska (near the same L shell as ATS 1), with ATS 1 data, 90 simultaneous events were identified. Analysis of 60 of these events led to the conclusion that Pi 1 are rarely, if ever, observed on the magnetic equator at synchronous orbit, even when Pi 1 activity is present on the ground near the conjugate point. Pc 3, on the other hand, are common at synchronous orbit. A statistical study of the Pc 3 occurring at ATS 1 during simultaneous College wave activity indicated characteristics similar to those observed at ATS 6.

Large-amplitude Pc 5 waves with 7-8 min quasi-periodic variations were observed almost continuously for the 48 hours of November 14-15, 1979, by several spacecraft in or near geostationary orbit on the dayside portions of their orbits. The waves were observed as large modulations in both the electron and ion fluxes over a wide range of energies (-1 to 500 keV) by the spacecraft 1976-059, 1977-007, 1979-053, P78-2, and GEOS 2 and as magnetic field peak-to-peak modulations of 15-25 •, by the P78-2 and GOES 2 and 3 magnetometers. The remarkably long persistence of these waves contrasts substantially with observations of typical flux modulation events which usually last less than 1 hour and which typically show little modulation of the -• 150 keV proton fluxes. Data taken concurrently by the ISEE spacecraft in the solar wind and outer magnetosphere indicate that the solar wind also had unusual properties. ISEE 3 measurements indicate that the solar wind velocity (-350 km/s) and density (-2 cm -3) were simultaneously very low for this period. The alpha-to-proton ratio for the solar wind plasma attained an extremely low value (<1%) early in the event. These solar wind properties imply such a much reduced dynamic pressure on the magnetosphere during this period. Consequently, the ISEE 1 and 2 spacecraft passed through the magnetopause at the uncommonly large radial distance of 18 Re at -•0830 local time where the typical magnetopause geocentric distance is 12 Re. The exceptional solar wind and outer magnetospheric conditions may have determined the unusual properties of the ULF event observed near geostationary orbit. Some candidate mechanisms for producing these oscillations are presented, but no definitive explanation for this event can be given at present. between 1530 UT on November 14 and 0330 UT on November 15.

Power spectra of geomagnetic field variations at synchronous orbit were calculated for the month of August 1974. Two segments of approximately 9 hours each, centered at local noon and local midnight, were used for the analysis. Each spectrum was parameterized as a straight line in the log‐log regime. The power spectral density at 1 mHz was higher for local night (particularly for the D component) than for local day and increased with geomagnetic activity as indicated by ∑Kp for the half day. The slopes of the straight line fits to the spectra ranged mainly from −1.5 to −3.0 and showed no clear difference night to day (except for the H component) and no clear relationship to geomagnetic activity. The day‐night difference in power can be explained mainly by broadband noise caused by substorms. The average power spectrum at synchronous orbit, as represented by the ATS 6 observations, was compared with reported observations from the interplanetary medium, the magnetosheath, the region inside the magnetosphere near the magnetopause, and the ground.

A study of several 10-to 45-sec wave events in the morning sector at Tungsten, Northwest Territories, Canada (approximately at the foot of the AT.S 1 fie•d line), reveals that at least two distinctly different types of wave phenomena occur. Digital power spectral and coherency analysis has been Used to determine the average characteristics of the two types of waves. The first, identified as Pi 1, is characterized by a broad-band spectral enhancement with an average center frequency of •0.07 Hz (•14 sec), nearly linear E-W polarization, and a close association with substorms. The second type, Pc 3, is characterized by narrow spectral lines at an average frequency of •0.04 Hz (~27 sec), nearly linear N-S polarization, and no close association with substorms. Although these two types of wave activity occur in the same local time sector, sometimes simultaneously. there is no indication of any relationship between them. Micropulsations of many types recorded at the earth's surface have been studied for many years, and a few types recorded in space have been studied more recently [Saito, i969; Gendrin, 1970; McPherron et al., 1972]. However, Pi 1 micropulsations have been one of the !e•s studied types, especially those in space. Only one possible observation in space has been discussed in the literature [McPherron and Coleman, 1971]. To remedy this situation, a study was initiated in which data from the UCLA magnetometers on the synchronous Satellite ATS I and at Tungsten, Northwest Territories, Canada (near the foot of the ATS 1 field line), were to be analyzed to determine (1) the characteristics of Pi 1 observed at the surface, (2) the characteristics of Pi I observed in space, and (3) the relationship between simultaneous events at the surface and the satellite. This paper will report the results of the first part of this study by using Tungsten (surface) data.As the first step in the study, events having irregular wave forms and 10-to 45-sec periodsx Publication 1199-25 of the Institute