Cosmic Ray Spectra and the Solar Magnetic Polarity:
Preliminary Results from 1994-2002
J. W. Bieber1, J. Clem1, M. L. Duldig2, P. Evenson1, J. E. Humble3, and R. Pyle1
1 Bartol Research Institute, University of Delaware, Newark, DE 19716, U.S.A.
2 Australian Antarctic Division, Kingston, Tasmania 7050, Australia.
3 School of Mathematics and Physics, University of Tasmania, GPO Box 252-21, Hobart, Tasmania 7001, Australia.
Abstract. Each year, beginning in 1994, a U.S./Australia collaboration has conducted a neutron monitor latitude survey
from the United States to McMurdo, Antarctica and back over a ~6-month period. This experiment permits us to observe
the high-energy cosmic ray spectrum as it evolves from the last solar activity minimum, through the recent maximum and
magnetic polarity reversal, and into the early stages of the new polarity epoch. In this work, we focus on the controversial
issue of whether a high-energy (~8 GeV) ''cross-over'' exists in spectra observed during opposite magnetic polarities. We
report on a preliminary analysis of the data and discuss our findings. In particular, we investigate the sensitivity of the
modulated cosmic ray spectrum to the sign of the Sun's magnetic field.
1. INTRODUCTION.
2. DATA AND METHODS.
Data were taken on eight separate trips from Seattle to McMurdo and return. These voyages are plotted
in Fig. 1.; also shown are selected 1980 vertical geomagnetic cutoff contours. Counts from the three
counter tubes are recorded once a second, together
with data from pitch and roll inclinometers. Once a
minute, pressure data and the GPS-derived latitude,
longitude and time are recorded. In this preliminary
study, we have not yet corrected for any possible effects resulting from non-level operation; however, the
data we are utilizing in this paper are from regions
where the geomagnetic cutoff is greater than 2 GV,
which eliminates most periods of rough seas, especially near Antarctica.
We have calculated the effective geomagnetic
cutoff for each hour of the surveys, at the exact time
and location of the measurement, taking into account
the applicable DGRF magnetic field model, geomagnetic activity level and tilt angle of the geomagnetic
dipole (using the Tsyganenko model magnetosphere);
[5, 6, 7, 2, 8].
During each survey, the monitor spent several
weeks in the harbor at McMurdo, near the McMurdo
neutron monitor. We used this period to normalize the
total counting rate to the McMurdo monitor. This
compensates for any instrumental changes which may
have occurred from year to year. In December 2001
we installed the monitor in a new shipping container
Over the past eight years we have conducted an
annual latitude survey which traversed the Pacific
Ocean from Seattle, USA to McMurdo, Antarctica and
return during a ~6-month interval each year. The monitor, a standard 3-NM64 design, is carried aboard one of
two U.S. Coast Guard icebreakers, the Polar Sea or the
Polar Star. The data from the surveys cover a wide
range of cutoff rigidities, from ~0 GV at McMurdo to
over 14 GV in the mid-Pacific. The survey technique
([1, 2] and references therein) has been used for many
years to improve our knowledge of the neutron monitor
response function and to test geomagnetic cutoff models. Differentiating the curve relating counting rate and
cutoff rigidity produces the neutron monitor
“differential response,” which is a measure of the cosmic ray spectrum. There have been recent reports of
one or more ‘crossovers’ in the spectral forms from two
opposite magnetic polarity epochs (e.g. [1, 3, 4]. One of
the major goals of this series of surveys is to try to observe such a crossover by deriving a series of cosmic
ray spectra through a solar magnetic polarity change (as
occurred in 1999/2000). We believe that this is the first
set of observations which span the entire period from
solar minimum to solar maximum at semi-regular intervals. Data reported in this paper extend through the first
half of the 2001/2002 survey.
CP679, Solar Wind Ten: Proceedings of the Tenth International Solar Wind Conference,
edited by M. Velli, R. Bruno, and F. Malara
© 2003 American Institute of Physics 0-7354-0148-9/03/$20.00
628
Joint UD-UTas-AAD Latitude Surveys 1994-2002
Ship Courses and 1980 Vertical Cutoff Contours
88
8 6
9
13
4
97
14
13
Surveys
11
7
5
44
Code Years Segments
4 1994/95
1
5 1995/96
2-4
6 1996/97
5-8
7 1997/98
9-12
8 1998/99
13-16
9 1999/00
17-20
0 2000/01
21-24
1 2001/02
25-26
3
1
4
290
9
280
130
120
110
100
90
80
-80
5
270
4
-70
3
260
-60
210
-50
200
-40
190
6
8
6
-30
180
-20
170
-10
17
160
66
6 88
6
0
150
10
1
9
8
7
0
76
00
9
61
5
16
78 5
15
0
916
5 0
9
7
5 00 9 9
1
9
7
6
9 05
8 0
5
6
0
9
0
9
1
5
8
7
008 5
8
6
6
9
04
59
8
9
5
8
0
6
7
1
7 7
1
8
6
9
7
588
90
80
68
485
7 17
4 8
5
409
56
98
96
0
9 8
8
5
5
6
6
7
4
0
9
87 0 6
8
49
5
7
6
1
4
1
5
7
6
9
4
1
7
0
140
88 6
20
1
0
1
8
6
7
96 9
60
7
9
7
5
0
8
7
1
0 95
7
6
9 9
7 7
11
97
250
8
30
Latitude
6
240
40
230
50
8
8
220
60
Longitude E of Greenwich
FIGURE 1. Course plots for the 8 surveys used in this paper. Each is labeled at 5-day intervals by the start year of the
survey (e.g. 7 for 1997/98). “Segment” codes are given in the inset. Selected vertical cutoff contours are labeled.
629
square fit to a three-parameter Dorman function, (e.g.
[1]) was performed for all data above 2 GV. The resulting fit was then differentiated to give the differen10000
600
Dorman Function:
9500
9000
c1 ( 1 − e
8500
( −c2 pc
( −c3 )
)
)
8000
7500
7000
500
Sample Fit for Pc > 2GV
Segment 9
1997/10/14 - 1997/11/29
D(Dorman)/D(pc)
6500
6000
5500
5000
Cts/Hr/100
during a port call in Hobart, Australia, and used the
nearby Kingston neutron monitor to renormalize the
data. Otherwise, during each survey year
(approximately November-May), care was taken not to
make any instrumental changes which might have affected the normalization.
In order to remove various noise problems encountered during the trips, the counting rate data were corrected on a minute-by-minute basis, time-corrected using onboard GPS clock data, and then pressurecorrected to 760 mmHg using a pressure coefficient varying with cutoff rigidity as follows: β = 0.983515 0.00698286*Pc, where β is in percent per mmHg and Pc
is in GV [8]. Since this series of observations was conducted during a period of frequent and often extreme
changes in modulation level, we have organized the data
to yield the highest time resolution possible, consistent
with a significant sweep over a large range of cutoff
rigidities. Therefore, we have divided the 8 surveys into
26 segments, with each traverse to and from the Magnetic Equator (or highest vertical cutoff value) treated
separately. Some segments were adjusted to avoid the
inclusion of major Forbush decreases. Each voyage,
therefore, would be expected to yield 4 data segments,
but equipment failures, etc, result in four of the possible
segments not being available. For each segment, the
hourly data points were plotted against the effective
vertical cutoff rigidity at the center of the hour. A least-
400
300
4500
D(Dorman)/D(pc):
4000
−
c1 c2 pc
3500
( −c3 )
c3 e
pc
( −c2 pc
( −c3 )
)
200
Effective Vertical Cutoff Rigidity (GV)
3000
2
4
6
8
10
12 14 16 18 20 2224
100
FIGURE 2. Sample fit of a segment’s data to a Dorman
function, along with the corresponding derivative.
21
21
1
242
21
1
3
21
2
242
1
434343
3
23
343
1
11
3
21 4 24
1424
3
13
1 24
3
2
3
4
1
2
3
1
1 24
3
8000
1
24
3
31
132 4
1 2
2 4
3 24
3
1
1
3 24
7000
3
1 24
2
12
3
13 4
1
3 24
32 4
1
3 24
1
3 12
6000
12
3
32 4
14
3
1
2
31
3
5000
Effective Vertical Cutoff Rigidity
1 2 4
2
1
4
3
2
1
3
1
3
Cts/hr/100
2
Flux (arbitrary units)
9000
3
2
(GV), Using 1980 Field Model
2 4
4
6
Ade>Equ: 1997 02/21-03/31 2:
Ade>Equ: 1999 02/27-03/22 4:
1:
3:
2
1
3
Cts/hr/100
6000
24
1
3
1
2
4
3
3 12
1
2
4
3
1
2
4
3
2
31
1
4
3
Effective Vertical Cutoff Rigidity
(GV), Using Tsyganenko Field Model
1
4
3
2
4
6
8
10
12
350
1
2
4
3
2
4 1
3
14 16 18 20
3. RESULTS.
250
200
150
100
350
300
250
200
150
100
FIGURE 3. Western Pacific Segments: using 1980
Cutoffs (top) vs Tsyganenko Cutoffs (bottom).
80
60
1997 3/22-4/17
40
1999 3/22-4/11
20
In our previous analysis of these data [9], we used
constant 1980 vertical cutoff values. We found that all 4
segments which went west of Australia and through the
western Pacific (segments 7, 8, 15, 16 from the 1996/7
and 1998/9 surveys - see Fig. 1) formed a separate and
distinct set from those segments east of Australia and in
the mid-Pacific. These four spectra did not agree among
themselves, even at high cutoff rigidities (>10 GV). We
attributed this difference to the secular drift of the
earth’s geomagnetic field and/or changes in the overall
level of geomagnetic activity, in combination with our
use of a simple vertical cutoff.
With the use of the new cutoff calculations, this discrepancy has been resolved. In Figure 3 we depict the
results from our earlier paper (based on the 1980 effective cutoffs with a static field model) as well as the same
spectra analyzed using the Tsyganenko model magnetosphere. Figure 4 displays the two surveys (segments 7,
8, 15, and 16) together with a plot of the McMurdo
counting rate with the times of the segments indicated.
The key point is that the differential response curves
should converge at high rigidity, where solar modulation
becomes weak. This convergence was not obtained
with the 1980 cutoffs (top panel of Figure 3), but it does
occur with the new Tsyganenko cutoffs (bottom panel).
In Figure 5, the rigidity of the response function
7.0
0
-20
1999 2/27-3/22
6.6
1997 2/21-3/22
-40
6.2
80
100
120
140
160
180
200
220
6.0
240
Longitude E. of Greenwich
5.8
10400
5.6
10200
07 08
5.4
10000
9800
9400
9200
15 16
Cts/Hr/100
9600
19
24
20
22
6.4
-60
5.2
5.0
9000
4.8
1999/10/01
1999/07/01
1999/04/01
1999/01/01
1998/10/01
1998/07/01
1998/04/01
1998/01/01
1997/10/01
1997/07/01
1997/04/01
8800
1997/01/01
13
26
6.8
Rigidity at Spectrum Peak (GV)
Latitude
tial response. For one segment, a sample set of results is
shown in Fig. 2.
300
400
1
24
3
50002
400
450
1
Flux (arbitrary units)
8000
2
1
3
8
10
12 14 16 18 20
Equ>Sea: 1997 03/31-04/17
Equ>Sea: 1999 03/22-04/11
22
2
22
1
1
1
4
14
4
1
2
34
34
32
4
1
1
3
34
22 2
2
3
4
34
1
1
2
34
23 4
31 2
1
1
34
2
3
1
3
24
1
34
1
2
1
2
3
34
24
2
1
1
3
3
3
1
2
1
4
2
24
32
1
1
3
4
3
3
1
2
24
1
4
1
2
3
3
2
1
4
9000 3
7000
2
4
450
2
21
17
25
18
12
23
15 316
7
149 18 11
5 4
Solar Dipole Polarity:
Positive (Segments 1-20)
Negative (Segments 21-26)
6
4.6
Relative Flux at Spectrum Peak
300
320
340
360
380
400
420
440
10
460
FIGURE 5. Rigidity at the spectrum maximum is plotted
versus relative flux at the maximum for each of the 26
segments.
FIGURE 4. Course plots for the 4 western Pacific segments (top). Times of the segments on a plot of the
McMurdo counting rate (bottom).
630
maximum is plotted versus the relative flux at the maximum for each of the 26 segments. As solar activity increased from 1994 to 2000, the points move towards
upper left. With increased modulation, the spectrum
peak is lower, but it occurs at a higher rigidity. The
more recent surveys occurred during negative solar
magnetic polarity (shown red in the plot), and they appear to have moved off the trend line established during
positive magnetic polarity (shown green). The results
are suggestive of a nascent “hysteresis loop.” If confirmed by further analysis, this hysteresis effect would
be evidence for effects of drift and/or magnetic helicity
in solar modulation.
Moraal et al. [1] previously reported evidence of an
8 GeV spectrum cross-over between positive and negative solar polarity, with harder spectra in negative polarity. If the negative (red) polarity points in Figure 5
continue to stay to the left of the positive (green) polarity points as solar activity decreases, this would however be consistent with harder spectra in positive polarity (i.e., for a fixed peak flux, the rigidity of the peak is
higher in positive polarity), which is in the opposite
direction of the effect reported by Moraal et al.
4. DISCUSSION.
The existence of a 22-year hysteresis loop or of a
spectrum cross-over at 8 GeV would indicate the influence of transport effects sensitive to magnetic polarity.
Only two such transport mechanisms are known: drift
and scattering by turbulence containing nonzero magnetic helicity. Drift models of solar modulation can
rather easily produce a spectrum cross-over near 400
MeV, but have difficulty producing one as high as 8
GeV [4]. On the other hand, the low-frequency turbulence responsible for scattering > 10 GeV cosmic rays
has been shown to possess finite magnetic helicity with
a definite dominant sign [10]. However quantitative
modeling of this effect has yet to be performed.
In the near future we will be incorporating additional (small) corrections in the analysis by taking into
account 1) the tilt of the monitor, especially important
during periods of rough seas; 2) corrections due to anisotropy (by fitting the Spaceship Earth data to measure
the first-order anisotropy [11]) and the effects of spectral variations (by using a network of stations at various
cutoff rigidities); 3) examination of the contribution of
off-vertical directions to the calculated cosmic ray cutoffs .
We plan to continue these yearly surveys at least
until the next solar minimum, so that a complete 11year modulation cycle can be studied in detail.
631
5. ACKNOWLEDGMENTS.
This work was supported in part by NSF grant
ATM- 0000315 (JB, JC, PE and RP). L. Shulman, J.
Roth assisted with the onboard equipment. We thank
the officers and crew of the United States Coast
Guard Cutters Polar Star and Polar Sea for their assistance in these surveys.
6. REFERENCES.
1. Moraal, H., Potgieter, M.S., Stoker, P.H., and van der
Walt, A.J., “NM Latitude Survey of the Cosmic Ray
Intensity During the 1986/87 Solar Minimum”, J. Geophys. Res. 94, 1,459–1,464, 1989.
2. Bieber, J.W., Evenson, P.E., Humble J.E., and Duldig,
M.., 1997, “Cosmic Ray Spectra Deduced from Neutron Monitor Surveys”, Proc. 25th Intl. Cosmic Ray
Conf. (Durban) 2, 45– 48, 1997.
3. Lockwood, J.A. and Webber, W.R., “Comparison of the
rigidity dependence of the 11-year cosmic ray variation
at the Earth in two solar cycles of opposite magnetic
polarity”, J. Geophys. Res., 101, 21573-21580, 1996.
4. Reinecke, J.P.L., Moraal, H., Potgieter, M.S., McDonald, F.B., and Webber, W.R., “Different Crossovers”, Proc. 25th Int. Cosmic Ray Conf. (Durban) 2,
49–52, 1997.
5. Lin, Z., J.W. Bieber and P. Evenson, “Electron trajectories in a model magnetosphere: Simulation and observation under active conditions”, J. Geophys. Res., 100,
23,543-23,549, 1995.
6. Flückiger, E.O. and Kobel, E., “Aspects of Combining
Models of the Earth's Internal and External Magnetic
Fields”, J. Geomag. GeoElec., 42, 1123-1128, 1990.
7. Cooke, D.J., Humble, J.E., Shea, M. A., Smart, D.F.,
Lund, N., Rasmussen, I.L., Byrnak, P., Goret, P. and
Petrou, N., “On Cosmic Ray Cut-off Terminology”,
Nuovo Cimento Soc. Ital. Fis. C, 14, 213, 1991.
8. Clem, J.M., Bieber, J.W., Evenson, P., Hall, D., Humble, J.E., and Duldig, M., “Contribution of Obliquely
Incident Particles to Neutron Monitor Counting Rate”,
J. Geophys. Res., 102, 26919, 1997.
9. Bieber, J.W., Clem, J., Duldig, M.L., Evenson, P.,
Humble, J.E. and Pyle, R., “A continuing yearly neutron monitor yearly survey: Preliminary results from
1994-2001”, Proc. 27th Intl. Cosmic Ray Conf.
(Hamburg) 10, 4087-4090, 2001.
10. Smith, C. W., and Bieber, J. W., “Detection of Steady
Magnetic Helicity in Low-Frequency IMF Turbulence”, Proc. 23rd Internat. Cosmic Ray Conf.
(Calgary), 3, 493-496, 1993.
11. Bieber, J.W. and Evenson, P., “Spaceship Earth — an
Optimized Network of Neutron Monitors”, Proc. 24th
Intl. Cosmic Ray Conf. (Rome) 4, 1078-1081, 1995.