Comment on “Propagation of the preliminary reverse impulse of sudden commencements to low latitudes” by P. J. Chi et al.

[1] Chi et al. [2001] suggested that the arrival of the preliminary reverse impulse (PRI) of the geomagnetic sudden commencement was earlier at lower latitudes by defining the arrival time of the PRI as the time of the maximum amplitude of the PRI, in contrast to the simultaneous onset of the PRI at all latitudes including the dayside dip equator as concluded by Araki [1977]. Chi et al. [2001] without analyzing the equatorial PRI insisted that Araki [1977] came to an incorrect conclusion regarding the instantaneous propagation of the polar electric field to the equator. To interpret the latitude dependence of the arrival time, Chi et al. [2001] applied the model of the three-dimensional propagation of the compressional mode waves in the uniform magnetosphere [Tamao, 1964a, 1964b] instead of the Earth-ionosphere-waveguide model for the instantaneous propagation from the polar ionosphere to the equator [Kikuchi and Araki, 1979]. There are several questions surrounding the definition of the arrival time and the interpretation of the arrival of the PRI based on the Tamao [1964a, 1964b] model. In this comment, we point out: (1) Araki [1977] defined the arrival of the PRI as the onset of the first inflection of the PRI, which is completely different from that defined by Chi et al. [2001]. (2) The equatorial PRI observed during the same event as analyzed by Chi et al. [2001] does not match their scenario, but suggests that the maximum amplitude of the PRI must be determined by a superposition of the magnetic effects of the Chapman-Ferraro currents and the ionospheric currents associated with the PRI and the main impulse (MI) as modeled by Araki [1994]. (3) Any theory and model should explain the global features of the PRI, in particular the simultaneous occurrence of the ionospheric plasma motion at low latitudes in the midnight as detected by the HF Doppler observation [Kikuchi, 1986]. Finally, we confirm that the TM0 mode in the Earth-ionosphere waveguide well explains the global PRI signatures including the instantaneous onset of the equatorial PRI and the ionospheric plasma motion in the midnight. [2] The preliminary reverse impulse (PRI) of the geomagnetic sudden commencement (SC) often occurred at the high latitude in the afternoon sector and at the dayside dip equator, while rarely occurred at the low latitude [Matsushita, 1962]. Araki [1977] found that the occurrence of the PRI at the dayside dip equator: Koror (KO) and at the afternoon high latitude: Point Barrow (PB), College (CO), Sitka (SI) was simultaneous within the temporal resolution of 10 s as shown in Figure 1 (from Figure 7 of Araki [1977]), by defining the onset time as the time of the first inflection of the PRI. Basing on these observations, Araki [1977] suggested that the polar electric field propagated to the equator near-instantaneously, inducing appreciable ionospheric currents because of the enhanced conductivity in the equatorial ionosphere. The ionospheric current system of the PRI consists of the twin-vortex Hall currents at high latitudes and the zonal Pedersen currents amplified by the Cowling effect at the dayside dip equator [Araki, 1994]. Araki et al. [1984] confirmed the ionospheric currents responsible for the high latitude PRI by showing the anti-correlated magnetic vectors measured by MAGSAT above the ionosphere and by magnetometers on the ground. Figure 1 also indicates that the DL defined as the magnetic increase due to the Chapman-Ferraro currents started at mid and low latitudes: Fredericksburg (FR) and Tucson (TU) about one minute after the PRI started at the high and equatorial latitudes. [3] To explain the instantaneous propagation of the polar electric field to the equator, Kikuchi et al. [1978], Kikuchi and Araki [1979] proposed the Earth-ionosphere waveguide model where the zeroth-order transverse magnetic (TM0) mode propagates at the speed of light accompanying the electric field and current in the ionosphere. Kikuchi et al. [1978] suggested that the electric field decreased its intensity with decreasing latitude because of the geometrical attenuation, but would be strong enough to induce the ionospheric currents at the equator. Later, Kikuchi et al. [1985] detected the PRI electric field in the low latitude ionosphere on the dayside and nightside using the HF Doppler observations. The onset of the HF Doppler frequency deviations was found to be simultaneous with the onset of the high latitude PRI within the temporal resolution of 10s as shown in Figure 2 (from Figure 1 of Kikuchi [1986]). The HF Doppler observations provided us with evidence of the instantaneous propagation of the dusk-to-dawn electric field from the polar ionosphere to the low latitude on the nightside as well as on the dayside. All the magnetometer and HF Doppler observations agree with the TM0 mode of the Earth-ionosphere waveguide. After the onset of the PRI, the ground magnetic field is determined by a superposition of the magnetic fields caused by the Chapman-Ferraro currents (DL part) and ionospheric currents (DP part) associated with the PRI and the main impulse (MI) of the SC [Araki, 1994]. The maximum amplitude of the PRI does not necessarily occur simultaneously at different latitudes. For example, the maximum amplitude of the PRI studied by Araki [1977] had occurred more than 1 min earlier at the equator (Figure 1). [4] In contrast to the simultaneous onset of the PRI concluded by Araki [1977], Chi et al. [2001] suggested that the PRI occurred earlier at lower latitudes by defining the arrival time of the PRI as the time of the maximum amplitude of the PRI. Chi et al. [2001] insisted that Araki [1977] incorrectly concluded the instantaneous propagation of the PRI electric field from the polar ionosphere to the equatorial ionosphere. However, it should be noted that the arrival time defined by Chi et al. [2001] is absolutely different from the onset time defined by Araki [1977]. Indeed, the time of the maximum amplitude of the PRI studied by Araki [1977] is delayed at higher latitudes in agreement with the data presented by Chi et al. [2001], although the onset of the inflection was simultaneous at high and equatorial latitudes (Figure 1). [5] The maximum amplitude of the PRI focused by Chi et al. [2001] must be caused by the combined effects of the ionospheric currents in addition to the Chapman-Ferraro currents as proposed by Araki [1994]. Figure 3 shows the magnetometer data from the equatorial stations, Guam (GAM, 13.58°N, 144.87°E, geographic coordinate (GR), 4.89°N, 215.26°E geomagnetic coordinate (GM)) and Yap (YAP, 9.3°N, 138.5°E GR, 0.3°S, 209.0°E GM) located in the morning sector for the PRI event analyzed by Chi et al. [2001]. As clearly seen in Figure 3, the amplitude of the PRI is greater at Yap since the ionospheric conductivity is larger nearer the dip equator. It should be noted that the first inflection of the PRI occurred simultaneously at Guam and Yap within the temporal resolution of a few seconds. We also note that the maximum amplitude of the PRI tends to be delayed at Yap by several seconds, and the onset of the MI is delayed at Yap by 15 sec, if the onset of the MI is defined as the time when the magnetic field strength becomes greater than the pre-event level. The distinct features of the PRI at the two equatorial stations may not be attributed to the MHD waves propagating through the magnetosphere, but be a result of the superposition of the magnetic fields due to two consecutive ionospheric currents (DP currents) induced by the dusk-to-dawn and dawn-to-dusk electric fields transported from the outer magnetosphere along the magnetic field lines [Araki, 1994]. From the comparison of the PRI at Guam and Yap, we may deduce that the DP currents of the PRI continued to flow at 2346 UT when the MI reached the first positive peek. It should be noted that the amplitude of the first peek was smaller at Yap than at Guam, whereas the amplitude of the major part of the MI at 2348UT was much greater at Yap as expected for the normal equatorial MI. Indeed, the PRI still continued at 2346 UT at the auroral stations, GILL and RABB in 17 MLT as shown in Figure 2 of Chi et al. [2001]. Furthermore, it should be noted that the preliminary impulse (PI) of the SC at mid latitudes is not always the PRI but appears as the PPI (positive impulse) when the PRI occurs at the dayside dip equator [Kikuchi and Araki, 1985; Kikuchi et al., 2001]. The midlatitude PPI was explained as a superposition of the Biot-Savart's effect of the field-aligned currents on the ground magnetic perturbations due to the DP currents [Kikuchi et al., 2001]. Consequently, the temporal and spatial features of the observed PRI are not necessarily determined by the MHD waves, but by several current components in the magnetosphere and ionosphere. We would suggest not using the time of the maximum amplitude of the PRI, but to use the time of the first inflection to identify the onset of the PRI as was done by Araki [1977]. [6] Similar analysis can be applied to the latitudinal tendency of the arrival time obtained by Chi et al. [2001]. Chi et al. [2001] concluded that the PRI arrived earlier at lower latitudes. If we assume that the observed PRI is a result of the superposition of the DP (PRI), DP (MI) and DL as proposed by Araki [1977], the DP (PRI) would be overcome by the DP (MI) + DL to be vanished earlier at lower latitudes. This in turn means that the maximum amplitude of the PRI occurs earlier at lower latitude since the DP (PRI) decreases its amplitude with decreasing latitude. Chi et al. [2001] demonstrated that the maximum of the PRI shifted longitudinally, and attributed this longitudinal shift to the three-dimensional propagation of the compressional mode waves in the magnetosphere. It should be noted that the four stations used are ordered not only from the noon to the afternoon, but also from the low latitude to the mid latitude. The longitudinal tendency of the arrival time could be explained in the same way as mentioned above. [7] We note again that the onset of the PRI analyzed by Araki [1977] was simultaneous at high latitudes and at the equator within the temporal resolution of 10 sec. In studying the onset of the PRI, the onset should be sharp enough to be identified within a resolution of 10 sec, since the PRIs often started several tens of seconds earlier than the onset of the DL [Araki, 1977; Kikuchi, 1986]. When we select PRI events for the analysis of the onset of the PRI, we need to consult the equatorial magnetometer data. Araki [1977] analyzed quite a few PRIs with the onset sharp enough to be identified within the temporal resolution of 10s. The onset of the equatorial PRI shown in Figure 3 is also identified within the temporal resolution of a few seconds. In contrast, it is rather difficult to read the onset of the first inflection recorded at the low latitude (see EWA in Figure 2 of Chi et al. [2001]) because of the too-small amplitude of the low latitude PRI. When we read the onset of the high latitude PRI, we should examine the background fluctuations and fluctuations excited by the sudden compression of the magnetosphere. The high latitude magnetometer data shown by Chi et al. [2001] indicate fluctuations with various durations depending on the latitude and local time. The complex features of the high latitude PRI would be due to the local ionospheric disturbances and the motion of the foot of the field-aligned currents as predicted from the analysis of the equivalent current of the PI [Nagano et al., 1985] and from the MHD simulations [Lysak and Lee, 1992; Slinker et al., 1999]. It should be stressed that the equatorial magnetometer often detects clear PRIs with well-defined onsets, because the small-scale high latitude disturbances suffer from severe attenuation during their propagation to the equator [Kikuchi et al., 1978]. The large-scale electric field originating in the magnetosphere such as the PRI, MI, DP2 and so on can reach the equator with appreciable amplitude to induce the electric currents at the dayside dip equator amplified by the Cowling effect. [8] Chi et al. [2001] attempted to explain the delayed arrival of the maximum amplitude by calculating the propagation time of the transverse mode waves converted from the three-dimensional compressional mode waves (CT mode: converted transverse mode) in the uniform magnetosphere [Tamao, 1964a, 1964b]. Since the onset of the PRI is simultaneous at all latitudes as defined by the first inflection [Araki, 1977], we cannot apply the Tamao [1964a, 1964b] model for the interpretation of the onset of the PRI. Instead, we explain the instantaneous propagation of the polar electric field to the low latitude using the zeroth-order TM mode in the Earth-ionosphere waveguide proposed by Kikuchi et al. [1978] and Kikuchi and Araki [1979]. In addition to the simultaneous occurrence of the PRI at all latitudes, we should remember that the PRI electric field is detected not only with the magnetometers on the dayside but also with the HF Doppler observation on the nightside [Kikuchi et al., 1985; Kikuchi, 1986]. The Earth-ionosphere waveguide model does explain the propagation of the electric field to the low latitude on the nightside as well as on the dayside. The MHD wave propagation would need several tens of seconds to reach the low latitude on the nightside [e.g., Wilken et al., 1982]. This never explains the instantaneous occurrence of the ionospheric plasma motion in the midnight coherent with the high latitude PRI as found by Kikuchi [1986]. [9] We now make a brief comment on the applicability of the CT mode of Tamao [1964a, 1964b] to the ground magnetic perturbations. The CT mode is characterized by the solenoidal (rotational) electric field and vortex-free plasma flow in the plane perpendicular to the magnetic field lines. Tamao [1964b] attempted to explain the twin-vortex equivalent currents responsible for the high latitude PRI [Nagata and Abe, 1955] by mapping the streamline of the CT mode on the anisotropic conducting ionosphere. If the ionospheric Hall currents were induced by the CT mode, we could apply the propagation time as calculated by Chi et al. [2001], and could expect that the Hall current developed earlier at lower latitudes. However, the Hall current element developed first at lower latitude would accumulate electric charges at the ends of the current element. The accumulated charge would produce Pedersen currents in opposite direction to the primary Hall current and a Hall current in the direction of the primary electric field of the CT mode. The resultant ionospheric current would not be the Hall current vortices responsible for the PRI. Another important point is that the Hall currents always accompany the Pedersen currents in the collision-dominated ionosphere. The energy dissipated by the Pedersen currents could not be supplied by the CT mode because the CT mode does not carry the divergence of the electric field, i.e., field-aligned currents into the ionosphere [Tamao, 1964b]. Only the PT (pure transverse) mode carries the electric currents and charges into the ionosphere, supplying the energy dissipated by the Pedersen currents. The PT mode does not transport electric charges to the mid and low latitude ionosphere across the magnetic field lines. However, the electric potential carried by the PT mode would excite the TM0 mode in the Earth-ionosphere waveguide, causing the instantaneous propagation of the ionospheric electric fields and currents responsible for the PRI (Figure 4). The TM0 mode propagates at the speed of light, and does not suffer from attenuation except for the geometrical attenuation [Kikuchi et al., 1978]. If we assume the nighttime condition for the ionospheric conductivity, we could expect some amount of attenuation in its propagation. However, even if we assume the height-integrated ionospheric conductivity as low as 0.1 mho, the ionosphere would act as a good conductor for the time period of the PRI compared with the intrinsic impedance of vacuum (377 ohm) below the ionosphere. We should make a remark on the curvature of the Earth-ionosphere waveguide, which will be significant for the propagation over long distance. The Poynting flux transported by the TM0 mode is determined by the transverse magnetic field and the vertical electric field. Since the electric field is kept vertical at the surface of good conductor, the Poynting flux propagates along the surface of the conductors, even when the waveguide is spherical as is the case for the propagation from the polar ionosphere to the equator. [10] The arrival time of the PRI defined by Chi et al. [2001] is the time of the maximum amplitude of the PRI, which is absolutely different from the onset time defined by Araki [1977] as the time of the first inflection of the PRI. By showing the magnetometer data recorded at two equatorial stations during the PRI event analyzed by Chi et al. [2001], we suggest that the maximum amplitude of the PRI does not represent the arrival of the PRI signal in the magnetosphere, but is determined by the superposition of the DP (PRI), DP (MI) and DL components as modeled by Araki [1994]. Furthermore, the PI of the SC at mid latitudes would be the PPI (reversed PRI) if the DP (PRI) is overcome by the magnetic effects of the field-aligned currents associated with the PRI [Kikuchi et al., 2001]. The onset of the PRI defined by Araki [1977] indicates the first appearance of the DP (PRI) caused by the dusk-to-dawn electric field carried by the field-aligned currents. We confirm that the PRI occurs simultaneously at all latitudes as concluded by Araki [1977]. We should also remember that the electric field associated with the PRI drives the ionospheric plasma motion in the midnight simultaneously with the high latitude PRI [Kikuchi, 1986]. The global features of the PRI are explained by the TM0 mode in the Earth-ionosphere waveguide proposed by Kikuchi et al. [1978] and Kikuchi and Araki [1979]. In order to identify the onset of the PRI, we need to select PRIs having a clearly defined onset within the temporal resolution of 10 sec, since the PRI tends to begin earlier than the onset of the DL by some tens of seconds [Araki, 1977; Kikuchi, 1986]. We would not be able to apply the Tamao [1964a, 1964b] model to the PRI at low latitudes because the rotational electric field of the CT mode does not generate Hall currents that close in the ionosphere. The Hall current must be generated by a divergent electric field caused by the electric charge carried by the PT mode. The electric charge will be transported to low latitudes by the TM0 mode in the Earth-ionosphere waveguide or the TEM mode in the transmission line (Figure 4), resulting in the simultaneous onset of the PRI at all latitudes both on the day- and nightside. [11] The author (T. K.) thanks J. Galen of the U.S. NOAA/Weather Service Office, Yap, P. Hattori at the USGS/Guam Magnetic Observatory, and K. Nozaki at the Communications Research Laboratory for their help in operating the magnetometers at Yap and Guam. [12] Arthur Richmond thanks Mark Engebretson for his assistance in evaluating this paper.