A. K. Svirzhevskaya and N. S. Svirzhevsky
P. N. Lebedev Physical Institute, Moscow, Russia
The north-south asymmetry in cosmic ray fluxes in the lower atmosphere established on the basis of long-term sonde measurements at the high-latitude stations Murmansk and Mirny (Antarctic) has been discussed by many authors [Charakhchyan, 1986a, 1986b; Charakhchyan and Charakhchyan, 1979; Charakhchyan and Stozhkov, 1983; Stozhkov et al., 1996; Svirzhevskaya et al., 1997]. Nevertheless, the physical mechanism responsible for different particle fluxes in the Northern and Southern hemispheres incident at large angles to the ecliptic is still unclear. The attempts to find a parameter in the solar wind or interplanetary magnetic field (IMF) that would correlate with the asymmetry in the particle fluxes for a long time period have also failed.
Probably, there are several causes for the N-S asymmetry. Any north-south asymmetry in the solar wind and IMF parameters can lead to asymmetry in the particle fluxes. Various scenarios for development of asymmetry also are in favor of the fact that there are different sources of the asymmetry. Most often, the asymmetry arises when time variations in the particle fluxes in the north and south are in phase, but somewhat shifted in time and differ in magnitude. However, sometimes the particle fluxes at the two stations are observed to be in antiphase, as, for instance, in February-June 1973.
Charakhchyan [1986b] put forward the hypothesis that properties of the N-S asymmetry are determined by the sign of the total magnetic field of the Sun and, hence, can experience 22-year variations. At first glance, this hypothesis was not confirmed because a large positive N-S asymmetry (like in the 1970s) was expected to take place after the solar magnetic field polarity reversal in 1991, when the IMF direction became the same as in 1971-1979. In 1991-1998, the north-south asymmetry was indeed large, but it changed the sign [Svirzhevskaya et al., 1997]. Probably, the 22-year periodicity is manifested only in the magnitude of the asymmetry. So it was nearly zero under negative IMF in the Northern hemisphere of the heliosphere (N - S + ) in 1963-1970 and 1980-1988 and large under the IMF of opposite sign (N + S - ) in 1971-1979 and 1991-1998.
Svirzhevskaya et al.  showed, using a large positive asymmetry of 1973 as an example, that the magnitude of the N-S asymmetry ANS does not change the sign when the Earth passes from one IMF sector to another. The parameter, which is related to the interplanetary plasma and IMF and behaves in a similar manner at that time, is the difference between the solar wind velocities in neighboring sectors, dVsw. There was an evident anticorrelation between ANS and dVsw in the first half of 1973. Svirzhevskaya et al.  proposed a possible scheme for formation of the N-S asymmetry involving the diffusion current of particles which has the same direction in neighboring sectors. In case of different solar wind velocities at opposite sides of the HCS, an asymmetric convective particle outflow takes place. It leads to a lower density of particles in the sector with a higher wind velocity. In this case a density gradient of cosmic rays perpendicular to the HCS and the diffusion current associated with this gradient arise. The current flows from the sector with a lower Vsw to the higher-velocity sector and does not change its direction at the sector boundary. The diffusion current supports the asymmetry as long as the difference between the solar wind velocities dVsw in neighboring sectors remains sufficiently large. It is quite clear what sign the asymmetry has in this situation: the additional particle flux is directed from the sector with a low solar wind velocity toward the higher-velocity sector. The verification of the ascribed above scheme with the data on the N-S asymmetry obtained in 1990-1998 when its sign was changed is the main goal of the present work.
We have also wanted to describe some characteristics of the N-S asymmetry in the particle fluxes in the Earth's atmosphere. These characteristics are the next: (1) the N-S asymmetry in the particle fluxes does not change its sign when the Earth crosses the boundary of the IMF sector; (2) the averaged over an 11-year cycle N-S asymmetry is large in magnitude under the positive IMF in the Northern hemisphere of the heliosphere (N+S- or A>0 ) and it is nearly zero under the IMF of opposite sign (N-S+ or A<0 ), while the sign of the N-S asymmetry does not depend on the sign of the IMF; (3) the magnitude of the N-S asymmetry ANS is within the limits of errors at the cutoff rigidity R<2 GV, while at R=3 -6 GV ANS equals 5-8% sometimes; (4) the part of the N-S asymmetry in the particle fluxes in the atmosphere arises from the season temperature variations in antiphase at the polar stations in the south and north hemispheres.
The characteristic feature of the N-S asymmetry in the particle fluxes observed in the measurements in the atmosphere is its large magnitude in comparison with that obtained from the high-latitude neutron monitor data [Pomerantz and Bieber, 1984; Belov and Oleneva, 1997]. We would like to note that in the last time the data on the N-S asymmetry in the atmosphere were experimentally confirmed by Gorchakov et al. , where the N-S asymmetries of the proton and helium fluxes with R>2 GV measured on Kosmos 600 satellite deep in the polar caps were defined. In some periods in 1973-1978 the magnitude of the N-S asymmetry ANS=4 -6% that agrees with our data. The magnitude of the N-S asymmetry ANS=0.05 -0.2% obtained from the neutron monitor data can be explained by the drift mechanism proposed by Swinson . Pomerantz and Bieber  and Belov and Oleneva  derived the quite reasonable radial gradients of the galactic cosmic rays (GCR) from the data on the N-S asymmetry. It is obvious that the N-S asymmetry of the cosmic ray fluxes observed in the atmosphere which by the order of magnitude exceeds that for the neutron monitors cannot be explained by drift mechanism and can not be consistent with the GCR radial gradients. We think that the sources of the N-S asymmetry in the cosmic ray fluxes in the atmosphere are the asymmetries of the basic parameters of the solar plasma (the solar wind velocity, and the values and disturbances of the IMF) with respect to the HCS as well as the asymmetric position of the HCS itself in the heliosphere.
The N-S asymmetry manifests itself also in the particle fluxes averaged over an 11-year cycle (Figure 3). In the 1970s and 1990s, the value of the asymmetry ANS was 2-4%. On the average, in the 1970s the particle fluxes from the northern direction of the heliosphere were larger than from the southern direction. In the 1990s, the opposite situation took place. It can be expected that some physical parameters on the Sun or in the heliosphere varied in a similar fashion. The value of the asymmetry ANS in the 1980s was by an order of magnitude lower than in the 1970s and 1990s. Note that there is no pronounced correlation between the solar wind velocity and the asymmetry averaged over an 11-year cycle. The average solar wind velocity Vsw near the Earth in the 1970s, 1980s, and 1990s was 443.6, 443.3, and 438.9 km s-1, respectively.
As it was shown earlier the ANS does not change the sign when the Earth passes from one IMF sector to another [Svirzhevskaya et al., 1997]. The parameter, which is related to the interplanetary plasma and IMF and behaves in a similar manner, is the difference between the solar wind velocities in neighboring sectors, dVsw = V+ - V-. In this work the value ANS and sign of the N-S asymmetry are compared with the dVsw in 1990-1998. The OMNI files were used as initial data on the IMF and solar wind [NASA, 1998]. The magnetic field in the Northern hemisphere of the heliosphere directed away from the Sun forms a positive sector near the helioequator if the azimuthal IMF component Bj falls into the longitudinal interval 45o-225o with the average value of 135o in the ecliptic (right-handed) coordinate system with the Ox axis pointed from the Earth's center toward the Sun. The field in the Southern hemisphere directed toward the Sun forms a negative sector near the helioequator in the longitudinal interval 225o-45deg with the average value of 315o. For the positive and negative sectors, monthly average velocities V+ and V- were derived, and then dVsw was calculated.
In 1991-1997 (with the exception of 1995), the annual wave was observed in the N-S asymmetry. The oscillations in ANS were not symmetric with respect to zero, they were shifted toward negative values. During this period, the annual wave was also observed in the velocity difference dVsw. In 1994-1997, it was probably associated with the change in the Earth's heliolatitude. In March, the Earth occupies the extreme southern position in heliolatitude and the solar wind velocity Vsw in the southern sectors is considerably higher than in the northern ones. In half a year, in early September, the Earth's heliolatitude is ~7o N, and Vsw is higher in the northern sectors than in the southern ones. In 1992-1993, the annual wave in dVsw had a different phase. Note also the asymmetry of dVsw with respect to its zero value. The solar wind velocity to the south of the HCS in 1994, 1995, and 1996 was higher than to the north of the sheet, as if the Earth moved to greater distances from the heliospheric current sheet in the negative (southern) sector than in the positive (northern) sector. This could happen if the HCS extended northward from the helioequator. The northward extension of the HCS was confirmed by the Ulysses data in 1996 [Forsyth et al., 1997] but there are reasons to believe that in 1994 the HCS extended southward from the helioequator by approximately 10o [Simpson, 1996].
Below we discuss some properties of the N-S asymmetry in charged particle fluxes in the lower atmosphere. The source of secondary radiation in the lower atmosphere (under atmospheric pressure x > 300 g cm -2 ) are the galactic cosmic rays, mainly protons with a threshold rigidity greater than 4 GV and the median rigidity of spectra ~15-20 GV.
The N-S asymmetry is typically described by invoking the drift currents B (grad n)r, where B is the IMF strength and (grad n)r is the radial component of gradient of the distribution function of cosmic ray density n [Pomerantz and Bieber, 1984]. These currents reverse the sign when the Earth crosses the sector boundary. The fact that the N-S asymmetry does not change its sign when the Earth traverses the HCS means that the drift currents of the B (grad n)r type do not give a significant contribution to ANS in our case.
Svirzhevskaya et al.  discussed a possible scheme for formation of the N-S asymmetry involving the diffusion current of particles which has the same direction in neighboring sectors. In case of different solar wind velocities at opposite sides of the HCS, an asymmetric convective particle outflow takes place that leads to a lower density of particles in the sector with a higher wind velocity. As a consequence of that, the diffusion current flows from the sector with a lower Vsw to the higher-velocity sector. The diffusion current supports the asymmetry as long as the difference between the solar wind velocities dVsw in neighboring sectors remains sufficiently large. This scheme describes well the large asymmetry in February-June 1973. When high-velocity solar wind fluxes were observed in the Southern hemisphere of the heliosphere, dVsw was negative and ANS was positive. One can see from Figure 4 that this scheme correctly predicts the asymmetry sign for the greater part of the 1990s, but during some periods (1991, the second half of 1994 and of 1995) the anticorrelation is violated. The condition dVsw > 0 (or dVsw < 0 ) is not sufficient for the onset of negative or positive asymmetry in these cases.
A large stable asymmetry during two years (1996-1997) should be noted. In the first half of 1996 the sector structure at the Earth's orbit was nearly absent: for the most part of the period the Earth was in the southern sector, that is, in the magnetic field of the same sign. The asymmetry in this case could be caused by the equatorward drift of particles from the South pole. The drift current in this case could be detected only at Mirny where asymptotic reception cones pointed southward.
As follows from the drift theories of modulation, in the 1970s and 1990s the drift currents were directed from the heliosphere poles toward the equator and further, along the HCS, to the heliosphere boundary [Potgieter and Moraal, 1985]. In the 1980s, the drift currents flew along the HCS toward the Sun and then to the poles. The longitudes of the asymptotic directions for particles with a rigidity of 10 GV for Murmansk and Mirny stations during the whole year were within the angle of 90o directed from the Earth toward the Sun. Therefore, detectors in the lower atmosphere at these stations can efficiently detect particles only in case of the (N+S- ) IMF orientation when particles move from the poles toward the helioequator. For the opposite IMF, detectors were pointed in the drift direction, and, hence, were not sensitive to particles.
Thus the mutual orientation of asymptotic directions of Murmansk and Mirny stations and drift currents depends on the IMF direction. In the 1970s and 1990s, drift particle fluxes were efficiently detected in the lower atmosphere. The conditions for development of asymmetry were, therefore, favorable, and ANS (by modulus) was indeed large during those years. In the 1980s, detectors in the lower atmosphere were not sensitive to the drift particle fluxes, and ANS 0.
The basic objective of present work was to test the assumption about the association of the north-south asymmetry of cosmic rays with the difference dVsw in the solar wind velocities above and under the heliospheric current sheet near the Earth. In the suggested simple scheme of the possible origin of north-south asymmetry its magnitude ANS and sign are predicted by the magnitude and sign of the velocities difference dVsw. The crucial test of this scheme was made with the data on the N-S asymmetry obtained in 1990-1998 when its sign was changed on the opposite one. The definitive evidence of the dependence of the magnitude ANS and the sign of the north-south asymmetry on the dVsw was not found.
It is believed that the cause of the N-S asymmetry observed in the Earth's atmosphere is the GCR anisotropic distribution beyond the magnetosphere. As noted above, the attempts to find one characteristic in the solar wind or IMF that would correlate with the N-S asymmetry in the particle fluxes in the atmosphere for a sufficiently long period of time have failed. Any asymmetry in basic characteristics of the solar plasma (solar wind velocity responsible for convection and also the IMF strengths and disturbances causing particle drift and diffusion) must lead to asymmetry in particle fluxes. Therefore, to quantitatively describe the asymmetry, the transport equation of the GCR for distributions of the solar wind velocities and IMF characteristics in the Northern and Southern hemispheres of the heliosphere should be solved for specific time periods.
Belov, A. V., and V. A. Oleneva, Long-term variations of the cosmic ray north-south anisotropy in 1973-1977 and 1984-1997, Geomagn. Aeron. (in Russian), 37 (1), 174, 1997.
Charakhchyan, T. N., Zonal modulation of cosmic rays and meteorological effects in sounding measurements of cosmic rays in the stratosphere, in Variations in Cosmic Rays and Space Research (in Russian), pp. 110-125, IZMIR AN, Moscow, 1986a.
Charakhchyan, T. N., Mysteries of zonal modulation of cosmic rays, Geomagn. Aeron. (in Russian), 26 (1), 10, 1986b.
Charakhchyan, A. N., and T. N. Charakhchyan, Specific features of zonal modulation of cosmic rays, in Interaction of Cosmic Rays with Medium (in Russian), pp. 379-394, Leningrad Inst. of Nucl. Phys., Leningrad, 1979.
Charakhchyan, T. N., and Yu. I. Stozhkov, Zonal modulation and anomalous component of cosmic rays, Izv. Akad. Nauk SSSR Ser. Fiz. (in Russian), 47 (9), 1674, 1983.
Forsyth, R. J., A. Balogh, E. J. Smith, and J. T. Gosling, Ulysses observations of the northward extension of the heliospheric current sheet, Geophys. Res. Lett., 24 (23), 3101, 1997.
Gorchakov, E. V., V. A. Iozenas, Yu. V. Mineev, and M. V. Ternovskaya, The cosmic ray north-south asymmetry in dependence of the interplanetary magnetic field on the sector structure, Izv. Ross. Akad. Nauk, Ser. Fiz. (in Russian), 59 (4), 113, 1995.
NASA, A compilation of 1963-1997 solar wind field and plasma data, Near Earth and Deep Space, Helio 001, Greenbelt, Md., 1998.
Pomerantz, M. A., and J. W. Bieber, Cosmic ray north-south anisotropy: Steady state characteristics, in Proceedings of the International Symposium on Cosmic Ray Modulation in the Heliosphere, pp. 39-47, Morioka, Japan, 1984.
Potgieter, M. S., and H. Moraal, A drift model for the modulation of galactic cosmic rays, Astrophys. J., 294 (2), 425, 1985.
Simpson, J. A., Ulysses cosmic-ray investigations extending from the south to the north polar regions of the Sun and heliosphere, Nuovo Cimento C, 19 (6), 935, 1996.
Stozhkov, Y. I., G. A. Bazilevskaya, P. E. Pokrevsky, N. S. Svirzhevsky, I. M. Martin, and A. Turtelli, Jr., Cosmic rays in the atmosphere: North-south asymmetry, J. Geophys. Res., 101 (A2), 2523, 1996.
Svirzhevskaya, A. K., G. A. Bazilevskaya, N. S. Svirzhevsky, and M. B. Krainev, North-south cosmic ray asymmetry and high speed solar wind, Adv. Space Res., 19 (6), 929, 1997.
Swinson, D. B., "Sidereal" cosmic-ray diurnal variations, J. Geophys. Res., 74 (24), 5591, 1969.