Source of Acquisition
4ASA Marshall Space Flight Center
GEOPHYSICAL RESEARCH LETTERS, VOL. 33, L18101, doi: 10.1029/2006GL027053, 2006
e Ot 4
David H. Hathaway' and Robert M. Wilson'
Received 31 May 2006; revised 1 August 2006; accepted 10 August 2006; published 16 September 2006.
[i] The level of geomagnetic activity near the time of solar [ 19931 noted that the number of geomagnetically disturbed
activity minimum has been shown to be a reliable indicator days during a cycle (as indicated by the number of days
for the amplitude of the following solar activity maximum. with Ap index 225 between solar activity minima) is
The geomagnetic activity index aa can be split into two proportional to the sum of the amplitudes of that cycle
components: one associated with solar flares, prominence and the following cycle. While this method for predicting
eruptions, and coronal mass ejections which follows the the amplitude of the next cycle requires waiting until cycle
solar activity cycle and a second component associated with minimum occurs, it is also found to be a very reliable
recurrent high speed solar wind streams which is out of method [Hathaway et al., 19991. All three of these geo-
phase with the solar activity cycle. This second component magnetic precursor methods indicated a larger than
often peaks before solar activity minimum and has been one *
achieved cycle for cycle 23 (Oh1 - 135 35; Feynman -
of the most reliable indicators for the amplitude of the *
154 25; Thompson - 153 f 35; achieved - 121) but with
following maximum. The size of the recent maximum in the predictions within or just outside the 2 a errors.
this second component indicates that solar activity cycle 24 [4] The use of geomagnetic activity as a predictor for
will be much higher than average - similar in size to cycles future solar activity seems counter-intuitive. The Sun is the
2 1 and 22 with a peak smoothed sunspot number,of 160 Ifr 25. source of the solar wind disturbances that drive geomag-
Citation: Hathaway, D. H., and R. M. Wilson (2006), netic activity and thus it would seem that solar activity
Geomagnetic activity indicates large amplitude for sunspot cycle should predict geomagnetic activity, not the other way
24, Geophys. Res. Lett., 33, L18101, doi:l0.1029/2006GL027053. around. Nonetheless, geomagnetic activity near the time
of solar activity minimum has proved to be a reliable
1. Introduction predictor for future solar activity. A likely explanation for
this connection comes fi-om the sources of geomagnetic
[2] Knowing the level of solar activity years in advance activity. Solar eruptions such as flares, filament eruptions
has important consequences. High levels of solar activity and coronal mass ejections are active producers of geomag-
can heat and inflate the Earth's outer atmosphere. This netic activity. The frequency of these eruptions rises and
increases the drag on satellites in low Earth orbit and can falls with the solar activity cycle as indicated by the number
lead to their early reentry. Placing a satellite in a low orbit of sunspots. These eruptions represent the solar cycle
can lead to a costly early end of its mission. Placing a component of geomagnetic activity as described by
satellite in a higher orbit either increases the launch costs or Feynman [1982]. Additional drivers of geomagnetic activity
decreases the payload weight. Accurate and reliable pre- include interplanetary shocks from high-speed solar wind
dictions for solar activity levels are needed by the people, streams associated with coronal holes that are out of phase
companies, and organizations that build, operate, and use with the sunspot cycle [cf. Luhmann et al., 20021. As the
satellites. polar coronal holes expand during the approach to sunspot
[3] Geomagnetic activity near the time of sunspot cycle minimum their low-latitude extensions produce recurrent
minimum has been shown to be a good indicator for the high-speed streams that give rise to geomagnetic activity.
level of maximum activity during the following cycle. Oh1 The magnetic field strengths and configurations that give
[1966] noted that the geomagnetic index aa reaches a rise to these coronal structures may provide a prelude to the
minimum near (but usually after) the time of solar activity strength of the ensuing sunspot cycle.
minimum and that this minimum in the aa index is well [ 5 ] Models for the Sun's magnetic dynamo may help to
correlated with the amplitude of the following activity explain this connection. Recent models [cf. Dikpati and
maximum. Feynman [19821 suggested that the geomagnetic Charbonneau, 19991 incorporate the Sun's meridional cir-
activity indicated by the index aa could be split into two culation to transport strong, sunspot forming, magnetic field
components - one in phase and proportional to the solar toward the equator at the base of the convection zone. This
activity cycle and a second (residual) component associated provides a simple explanation for the equatorward drift of
with interplanetary disturbances that is out of phase with the the sunspot latitudes. It also suggests that evidence of the
activity cycle. This second, interplanetary component often next cycle might be seen in the mid-latitudes prior to the
peaks just before solar activity minimum and has been appearance of sunspots. This "extended" solar cycle was
shown to be an even better indicator for the amplitude of suggested earlier by n%lson et al. 1119881 based on observa-
the following cycle [Hathaway et al., 19991. Thompson tions of ephemeral active regions, coronal emission-line
structures, and the torsional oscillation signal. These struc-
tures appear at mid-latitudes prior to the first appearance of
'National Space Science and Technology Center, NASA, Huntsville, the sunspots of the new cycle and may very well contribute
Alabama, USA. to geomagnetic activity at that time.
[6] Predictions for the size for cycle 24 have already been
This paper is not subject to U S . copyright.
Published in 2006 by the American Geophysical Union. published. Hathaway and Wilson [2004] predicted a large
L18101 1 of3
L18101 HATHAWAY AND WILSON: LARGE AMPLITUDE FOR SUNSPOT CYCLE 24 18101
351 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' j
x
4 30
-
S
Q
25
.-
c
a
C
9 20
El
d- 15
(D
3
a5 10
-aa, = 10.9+ 0.097R
0 50 100 150 200
Annual InternationalSunspot Number
Figure 1. Annual values for the geomagnetic index aa as a
function of the corresponding annual International Sunspot
Number for the years 1868 to 2005. Both monthly activity
indices are smoothed with a 24-month FWHM Gaussian
and then sampled at yearly intervals. At a given sunspot
number there is a baseline level of geomagnetic activity
which is proportional to the sunspot number. Geomagnetic
activity extends above this level - particularly late in each
solar cycle. This baseline level, aaR, is determined by fitting
a line through the lower boundary of activity.
amplitude (145 f 30 for the maximum of the smoothed
monthly sunspot number) for cycle 24 based on the equa-
torward drift rate of the active latitudes during cycle 22.
Svalgaard et al. [2005] predicted a small amplitude (75 8) *
based on the weak polar fields observed on the Sun during
the decline of sunspot cycle 23. A significant new devel-
opment in predicting the solar activity cycle is the use of a
dynamo model with assimilated sunspot data as described
by Dikpati et al. [2006]. Using historical records for sunspot
areas and positions over the last 130 years as input for the
35-' a ' " I a ' ' ' " " ' * ' I ' ' - I ' -
-aaR= 10.9+ 0.097R
I I 1 ,
source of the surface magnetic fields that seed the dynamo, I
they predict an amplitude of 150-180 for cycle 24. 30 1 .............. aa, = aa - aaR 19 -
[7] In this letter we examine recent geomagnetic activity
using the methods described by Feynman [1982] and find
that this activity indicates a much larger than average cycle
for solar cycle 24 - on par with the prediction of Hathaway
and Wilson [2004] and Dikpati et al. [2006].
2. Data and Methodology
[SI Geomagnetic activity is measured by noting the rapid
changes in the geomagnetic field strength and direction.
Numerous observatories have made these measurements
and a number of indices have been constructed to charac-
terize the level of activity. The most widely used long-term 1860 1880 1900 1920 1940 1960 1980 2000
index is the aa index. This index is produced using two Date
observatories at nearly antipodal positions on the Earth's
Figure 2. Solar cycle, aaR, and interplanetary, aaI,
surface. The index is computed fiom the weighted average components of geomagnetic activity as functions of time.
of the amplitude of the field variations at the two sites over
The solar cycle component is directly proportional to the
three-hour intervals. Monthly averages of this index began
sunspot number and illustrates the sequence of cycle
in January of 1868 with Greenwich, England and
amplitudes for the last 13 solar cycles (numbered). The
Melbourne, Australia as the two sites. Greenwich was interplanetary component has a similar sequence of peaks
replaced by Abinger, England in 1926 and by Hartland,
but they occur several years earlier - usually before the
England in 1957. Melbourne was replaced by Toolangui, time of cycle minimum.
18101 HATHAWAY AND WILSON: LARGE AMPLITUDE FOR SUNSPOT CYCLE 24 L18101
250 maximum for cycle 24 of about 160 i 14. This prediction is
very much in line with the predictions of a large cycle 24 by
Hathaway and Wilson [2004] and by Dikpati et al. [2006]
200 but in contrast to the prediction of a small cycle 24 by
Svalgaard et al. [2005].
150 [15] All four of these predictions are based on different
methods. The prediction of a small cycle 24 by Svalgaard et
al. [2005] is based on a correlation observed between
100 directly measured polar fields and sunspot number for the
last three cycles, following the method of Schatten et al.
[1978]. The Dikpati et al. [2006] prediction is on the firmest
50
physical ground - a dynamo model based on our current
knowledge of the dynamics of the Sun's convection zone
0 (note, however, comments by Tobais et al. [2006]). All
0 5 10 15 20 three of the high predictions are consistent with each other
Max(aa,) near Minimum and are based on data from the last 12-13 sunspot cycles.
The Dikpati et al. [2006] prediction is based on a dynamo
Figure 3. Maximum sunspot number for the following
model with observational data on sunspot areas and posi-
cycle as a function of the maximum in the interplanetary
tions. The Hathaway and Wilson [2004] prediction is based
component of the aa index. The following cycle amplitude
on an observed, significant but loose, relationship between
is well correlated with the earlier peak in aaz with little
the equatorward drift rate of the active latitudes and future
scatter about a linear relationship (solid line with 1-sigma
solar activity. The prediction presented in this letter is based
limits shown with dotted lines). The prediction for sunspot
on an observed, significant and tight, relationship between
cycle 24 is shown with the circled number 24 - a maximum
geomagnetic activity and future solar activity. The consis-
sunspot number of 160 f 25.
tent results with these three methods strongly suggests that
cycle 24 will indeed be a large cycle.
similar sequence of peaks but shifted back in time -
usually to before sunspot cycle minimum. It is this References
behavior of the interplanetary component that provides Dikpati, M., and P. Charbonneau (1999), A Babcock-Leighton flux trans-
port dynamo with solar-like differential rotation, Astrophys. J., 518,508-
predictive capability. 520.
[n] The relationship between the peaks in the interplan- Dikpati, M., G. de Toma, and P. A. Gilman (2006), Predicting the strength
etary component of the aa index and the amplitude of the of solar cycle 24 using a flux-transport dynamo-based tool, Geophys. Res.
Lett., 33, LO5 102, doi:10.1029/2005GL025221.
following sunspot cycle is shown in Figure 3. There is a Feynman, J. (1982), Geomagnetic and solar wind cycles, 1900-1975,
strong positive correlation (correlation coefficient 0.94) J. Geophys. Res., 87, 6153-6162.
between the two quantities and the chance of getting this Hathaway, D. H., and R. M. Wilson (2004), What the sunspot record tells
us about space climate, Sol. Phys., 224, 5-19.
relationship from uncorrelated quantities is less than about Hathaway, D. H., R. M. Wilson, and E. J. Reichmann (1999), A synthesis
0.1%. of solar cycle prediction techniques, J. Geophys. Res., 104, 22,375-
[i3] The relationship shown in Figure 3 can be used to 22.388.
predict the size of the next solar cycle. The smoothed L&ann, J. G., Y. Li, C. N. Arge, P. R. Gazis, and R. Ulrich (2002), Solar
cycle changes in coronal holes and space weather cycles, J. Geophys.
interplanetary component of the aa index peaked at a level Res., 107(A8), 1154, doi:10.1029/2001 JA007550.
of 13.7 in October of 2003. This indicates a maximum Ohl, A. I. (1966), Forecast of sunspot maximum number of cycle 20, Solice
sunspot number of about 160 for sunspot cycle 24. This is Danie, 9, 84.
Schatten, K. H., P. H. Scherrer, L. Svalgaard, and J. M. Wilcox (1978),
similar to the amplitudes of cycles 21 and 22 but less than Using dynamo theory to predict the sunspot number during solar cycle
that of cycle 19. The 90% prediction interval is 160 f 25; 21, Geophys. Res. Lett., 5, 411-414.
therefore, there is only a 5% chance that cycle 24 maximum Svalgaard, L., E. W. Cliver, and P. Le Sager (2004), IHV A new long-term
geomagnetic index, Adv. Space Res., 34, 10.1016/j.asr.2003.01.029.
amplitude will be smaller than 135. Svalgaard, L., E. W. Cliver, and Y. Kamide (2005), Sunspot cycle 24:
Smallest cycle in 100 years?, Geophys. Res. Lett., 32, LO1104,
doi: 10.1029/2004GL021664.
3. Conclusions Thompson, R. J. (1993), A technique for predicting the amplitude of the
solar cycle, Sol. Phys., 148, 383-388.
[i4] The geomagnetic index aa can be split into two
Tobais, S . , D. Hughes, and N. Weiss (2006), Unpredictable Sun leaves
components - one proportional to, and in phase with, the researchers in the dark, Nature, 442, 26.
sunspot number and another, interplanetary, component aaI Wilson, P. R., R. C. Altrock, K. L. Harvey, S. F. Martin, and H. B.
which is out of phase with the sunspot cycle. This second Snodgrass (1988), The extended solar activity cycle, Nature, 333,
748-750.
component has peaks in activity that mimic those seen in
the sunspot number but shifted in time several years earlier.
These peaks in aaI usually occur before the time of sunspot D. H. Hathaway and R. M. Wilson, National Space Science and
cycle minimum and provide an accurate prediction for the Technology Center, NASA, MC W62, Huntsville, AL 35812, USA.
(david.hathaway@nasa.gov)
amplitude of the following sunspot cycle. The recent
(October 2003) peak in aaI indicates a sunspot number
3of3
