Abstract. Several properties of the A/C index of polar cap variations introduced by L. Svalgaard in 1972 have been found to vary with time. In the presatellite era, C days, as measured by the Ap index, are almost twice as geomagnetically active as A days, while in the modern epoch they have essentially identical activity. Prior to 1962 there were over 40% more A days than C days per year, while during the modern epoch there are essentially equal numbers of A days and C days. In view of this strong bias to assigning a "toward" classification on geomagnetically active days we recommend against using the Svalgaard A/C classification in studies of geomagnetic activity. We further recommend that a full and thorough documentation of the index be prepared and/or that others undertake to compile such a classification separately.
While many studies have shown that the north-south component of the interplanetary magnetic field, or equivalently the dawn-dusk component of the interplanetary electric field, has a strong influence on ensuing geomagnetic activity [Rostoker and Falthammar 1967; Hirshberg and Colburn, 1969; Arnoldy, 1971: Aubry and McPherron, 1971; Foster et al., 1971; Burton et al., 1973; Caan et al., 1975], the influence of the polarity of the interplanetary magnetic field is much more subtle [Burch, 1973; Russell and McPherron, 1973]. On the long-term average the polarity is found to have no effect on the level of geomagnetic activity. However, when the semiannual variation of geomagnetic activity is examined for each polarity separately, the variation is split into two annual waves of opposite phase. Studies of such effects are presently restricted to the period of roughly 1964-1970 by the availability of in situ interplanetary magnetic field observations, a time short in comparison with the length of a solar magnetic cycle. The availability of interplanetary magnetic polarity for a much longer period is highly desirable for it would allow, for example, a test of the prediction that the double-sunspot cycle of geomagnetic activity, observed by Chernosky , is related to a reversal in the phase of the heliographic latitude dependence of the interplanetary magnetic polarity from one sunspot cycle to the next. We note that these polarity effects should not be confused with the behavior of geomagnetic activity at sector boundaries, which is independent of polarity or the change in polarity [Wilcox and Ness, 1965] and which is associated with the velocity structure at sector boundaries [Hirshberg and Colburn, 1973; Burton et al., 1974].
The possibility of extending such studies was presented by the publishing of the A/C classification of polar cap magnetograms for the period 1926-1971 by Svalgaard [1972a, b]. The basis for inferring the interplanetary magnetic polarity from these data is that the strength of the vertical component of the magnetic field at Thule, Greenland, is found to be strongly correlated with the solar magnetospheric Y component of the interplanetary magnetic field [Friis- Christensen et al., 1972]. It was also noted that the horizontal component measured at Godhavn at a lower geomagnetic latitude (77.5o N) correlated with the variations in the Thule vertical component. The classification based on data from these two stations has been tested for two different years, 1965 [Campbell and Matsushita, 1973] and 1969 [Friis- Christensen et al., 1971], and it was found that Svalgaard's identification agreed with the measured sector pattern about 70% of the time. While 30% would be a moderately large error rate if we were examining individual cases, we had hoped that predicted variations would be present in the long-term averages. Unfortunately, we were not able to extend our studies by using the Svalgaard A/C classification due to certain inherent characteristics. It is the purpose of this letter to describe those characteristics and to point out the limitations of the index for geomagnetic studies. We note that some of these characteristics have also been found independently in a recent study by Fougere .
The top two panels of Figure I show the annual variation of the Ap index for the years 1932-1956 and 1963-1969 separated according to the daily A/C classification. The reason for the choice of averaging periods will become clear in the next section. The top right-hand panel weakly exhibits the separation of the semiannual variations into two annual variations predicted by Russell and McPherron. The lack of a more clear separation may be in part due to the weakness of the semiannual variation in these data and in part due to the presence of errors in the polarity identification. There should be approximately one wrong identification for every two right identifications if the statistics for 1965 are valid throughout this interval.
The variation in the upper left-hand panel is truly amazing. Not only do both the A curve and the C curve exhibit the semiannual variation, but the average Ap index for C days is about twice that for A days. The difference between the 1932-1956 data and the 1963-1969 satellite era data suggests that there has been a significant change in the properties of the index from the early period to the later period. Further, it seems that the earlier data are in error because the satellite era data reproduce the variation observed when in situ measurements were used. It seems obvious that physical laws are less likely to change with time than the quality of an index hand-scaled from visual records.
Investigating further, we checked the annual variation of the number of C and A days. If the A and C identification really reflected the interplanetary polarity, we would expect equal numbers of A and C days on the long-term average. The annual variation over short periods such as that in the right-hand panel should reflect the heliographic latitude dependence of the dominant polarity of the interplanetary field [Rosenberg and Coleman, 1969], but a yearly average of this effect should be close to zero, as it is in the right-hand panel. From the left hand panel it is obvious that there is a strong, perhaps 50%, bias toward choosing an A day in the presatellite era data. Secondly, we note that the phase of the annual variation seen in the satellite era data has a simple explanation, since it is in step with the heliographic latitude of the earth's orbital motion, while the phase of the annual variation in the presatellite era data is not so simply explained.
|Fig. 1. (Top) Annual variation of Ap index for A days and C days [Svalgaard, 1972b] for two different epochs. (Bottom) Annual variation of ratio of number of A days to number of C days per month for same two epochs.|
The curves in Figure 1 show conclusively that the properties of the A/C classification have changed with time, but they do not pinpoint when. In particular, there is no a priori justification for assuming that the change is coincident with the advent of in situ observation of the interplanetary magnetic field. We have identified above three properties that changed: the high correlation between C days and strong geomagnetic activity, the average number of C or A days per month, and the phase of the annual variation of the number of C or A days. Each of these parameters can be used independently to assess the changing quality of the index with time. To make this assessment for each year from 1932 to 1969, we have found the average Ap index for A days and C days separately and have Fourier-analyzed the number of C days per 28-day month returning the average number of C days per month, the amplitude of the annual wave, and the phase of the annual wave. These are plotted in Figure 2. If it is accurate, the A/C classification should have the following properties: the average Ap index for A days should equal that for C days, the average number of C days per month should be close to 14 (we have used 28-day months), and the phase of the variation should be close to -115o; or 65o; (corresponding to September and March 5) if the Rosenberg-Coleman results for the last solar cycle are indeed generally true. There is only one period when all three criteria are true, 1963-1969. In all other periods at least two of the criteria are not true. For example, even though the period from 1948 through 1958 shows mainly the proper phase, C days are far more active than A days, and there are far fewer than 14 C days per month.
Figure 2 leaves little doubt that the present A/C classification has serious shortcomings, but nevertheless the technique for inferring interplanetary magnetic polarity from ground data should not be discarded. First of all, the data produced by Friis-Christensen et al.  are convincing, and Campbell and Matsushita  were able under proper circumstances (large polar cap variations during summer) to obtain 85% agreement with satellite determinations. Secondly, the classification even during the most distant presatellite era shows vestiges of the expected variation. The dashed vertical lines marked with stars in Figure 2 indicate the time when the Russell-McPherron interpretation of Chernosky's 22-yr geomagnetic cycle would predict a reversal in phase of the heliographic latitude dependence of the interplanetary magnetic field. (We note that due to the shortness of modern solar cycles, this has become a 20-yr cycle.) At each of these times the phase of the variation changes in the expected direction, a suggestion of some interplanetary control.
Fougere  has also found the A/C classification wanting and has pursued the correlation with geomagnetic activity intensively, calculating a correlation coefficient for each year. He obtains high correlation coefficients with geomagnetic activity every year before 1958, a highly variable correlation coefficient from 1958-1962, and then a rather low correlation coefficient from 1963 to the present. A hint of this behavior can be seen in Figure 2. Such a gradual deterioration of the index cannot be attributed to the availability of satellite data, but rather it must either be in the quality and/or availability of the ground-based data, or it must reflect a change in the techniques used to derive the index. Finally, we note that there is no apparent correlation between the stations used to create the A/C index (L. Svalgaard, personal communication, 1973) and the quality of the index.
The present A/C index cannot be used to infer the sector structure during the presatellite era because of a strong bias to assigning a toward classification on geomagnetically active days. It is not obvious why a technique that apparently is successful with modern data is unsuccessful prior to 1963. However, the index does exhibit some vestiges of interplanetary control during the earlier period, and the reported effects upon which the index is based have been confirmed by independent workers. Therefore it may be possible to derive a correct index for this period. We recommend first that Svalgaard prepare a full and thorough documentation on his preparation of the index and second that others undertake to compile such a classification independently. Finally, we recommend that the present index not be used in studies of geomagnetic activity during the presatellite era.
|Fig. 2. Yearly average Ap index for C and A days, number of C days per month, amplitude and phase of annual Variation in number of C days per month, and number of sunspot groups for years 1932-1969.|
Acknowledgments. This work was supported in part by the National Science Foundation under grant GA 34148-X and by the National Aeronautics and Space Administration under NASA grant NGR 05-007-004.
The Editor thanks J. P. Heppner for his assistance in evaluating this report.
Arnoldy, R. L., Signature in the interplanetary medium for substorms, J. Geophys. Res., 76, 5189, 1971.
Aubry, M. P., and R. L. McPherron, Magnetotail changes in relation to the solar wind magnetic field and magnetospheric substorms, J. Geophys. Res., 76(19), 4381-4401, 1971.
Burch, J. L., Effects of interplanetary magnetic sector structure on aurora] zone and polar cap magnetic activity, J. Geophys. Res., 78(7), 1047, 1973.
Burton, R. K., R. L. McPherron, and C. T. Russell, A study of the relationship between the interplanetary electric field and the main phase minimum of geomagnetic storms (abstract), Eos Trans. AGU, 54(4), 415, 1973.
Burton, R. K., C. T. Russell, and R. L. McPherron, Further comments on magnetic activity at sector boundaries, submitted to J. Geophys. Res., 1974.
Caan, M. N., R. L. McPherron, and C. T. Russell, Magnetospheric substorms and the geomagnetic tail lobes, J. Geophys. Res., 80, 191, 1975.
Campbell, W. H., and S. Matsushita, Correspondence of solar field sector direction and polar cap geomagnetic field changes for 1965, J. Geophys. Res., 78, 2079-2087, 1973.
Chernosky, E. J., Double sunspot-cycle variation in terrestrial magnetic activity, 1884-1963, J. Geophys. Res., 71(3), 965-974, 1966.
Foster, J. C., D. H. Fairfield, K. W. Ogilvie, and T. J. Rosenberg, Relationship of interplanetary parameters and occurrence of magnetospheric substorms, J. Geophys. Res., 76, 6971, 1971.
Fougere, P. F., Dependence of inferred magnetic sector structure upon geomagnetic and solar activity, Planet. Space Sci., 22, 1 173, 1974.
Friis-Christensen, F., E. K. Lassen, J. M. Wilcox, W. Gonzalez, and D. S. Colburn, Interplanetary magnetic sector polarity from polar geomagnetic field observations, Nature Phys. Sci., 233, 48, 1971.
Friis-Christensen, F., K. Lassen, J. Wilhjelm, J. M. Wilcox, W. Gonzalez, and D. S. Colburn, Critical component of the interplanetary magnetic field responsible for large geomagnetic effects in the polar cap, J. Geophys. Res., 77, 3371, 1972.
Hirshberg, J., and D. S. Colburn, Interplanetary field and geomagnetic variations: A unified view, Planet. Space Sci., 17, 1183, 1969.
Hirshberg, J., and D. S. Colburn, Geomagnetic activity at sector boundaries, J. Geophys. Res., 78, 3952, 1973.
Rosenberg, R. L., and P. J. Coleman, Jr., Heliographic latitude dependence of the dominant polarity of the interplanetary magnetic field, J. Geophys, Res., 74(24), 5611-5622, 1969.
Rostoker, G., and C. -G. Falthammar Relationship between changes in the interplanetary magnetic field and variations in the magnetic field at the earth's surface, J. Geophys. Res., 72(23), 5853-5863. 1967.
Russell, C. T., and R. L. McPherron, The semiannual variation of geomagnetic activity, J. Geophys. Res., 78(l), 92, 1973,
Svalgaard, L., Interplanetary magnetic sector structure 1926-1971, J. Geophys. Res., 77(22), 4027-4033, 1972a.
Svalgaard, L., Interplanetary magnetic sector structure 1926-1971, Geophys. Pap. R-29, Dan. Meteorol. Inst., Copenhagen, 1972b.
Wilcox, J. M., and N. F. Ness, Quasi-stationary corotating structure in the interplanetary medium, J. Geophys. Res., 70, 5793, 1965.