GEOPWSICAL RESEARCH LETTERS, VOL. , XXXX,DOI:10.1029/,

The ‘‘- 150 Day Quasi-Periodicity” in Interplanetary and Solar Phenomena During Cycle 23 I I. G. Richardson” and H. V. Cane”

An intermittent “quasi-periodicity” of N 150 days in period is similar to that of the “- 150- (typically 155) day various solar and interplanetary phenomena has been re- periodicity” first recognized during solar cycle 21 in the OC- ported in earlier solar cycles. We suggest that variations cmrence of gamma ray flares [Riegeret aL, 19841 and sub quently identified other solar and interplanetary phenom- in the occurrence of solar energetic particle events, inter- in ena. These include other flare data sets [Bai and Stumk, planetary coronal mass ejections, and geomagnetic storm 1993, and d&isEsunspot of this quasi-periodicity, which is also present in the sunspot areas [Lean, 19901, the soft X-ray background [Aschwandeq number, in particular in the northern solar hemisphere. It 19941, solar proton events [Bai and Cliver, 1990; Cane et d., is not, however, prominent in the interplanetary magnetic 19981, the interplanetary magnetic field [Cane et aL, 19981, field strength. and the geomagnetic Ap index [Gonzalez et al., 19931. We refer here to the 150-day periodicity” but recognize that, like the ll-year” sunspot cycle, this phenomenon is only quasi-periodic. For example, Lean I19901 noted that it only 1. Introduction occurred intermittently in sunspot areas for intervals of - 1 - 3 years duration around the maxima of cycles 12 - 21, During the current solar cycle, variations in the level of while the exact “period” varied from 130 to 185 days, and energetic solar activity have occmed that are superposcd often changed significantly during an individual cycle. The on the genera1 solar cycle variation. For example, Figure origin of the ”- 150-day periodicity” is unclear, in partic- l(a) shows the intensity of 19 - 28 MeV solar protons ob- ular whether it originates near the surface of the Sun [Bo< served by the WIND spacecraft in 1996 - late-2004. Note 1987, 1988; Lou, 20001, reflects changes in the rate of so- that in late-1997 - late-1999, during the ascending phase lar magnetic flux emergence [e.g., Cane et d., 19981 or is of the solar cycle (cf. the monthly sunspot numbers in the a global phenomenon [e.g., Bai and Sturrock, 1987; Wol8 northern and southern solar hemispheres in Figures l(d - 19921. Bai (19871 and LRan [1990] noted an association with e)), the proton enhancements are not evenly distributed in complex active regions containing large sunspots (“superac- time, but tend to occur in clusters (in some cases extending tive regions”). over several solar rotations, such as that in late1998) sep Evidence for a “- 150-day periodicity“ during cycle 23 arated by intervals during which such events are relatively has been ambiguous. A similar period (- 140 days) was re- rare. A similar pattern may be discerned during the de ported by Ddla et al. [2001] in solar energetic particle events clining phase of the cycle, from late 2002 until at least the during 1998 - 1999, and by Hdl et al. [2001] (151 days) in time of writing. In particular, the intense events in Octo- anomalous cosmic ray intensities in the outer heliosphere, ber - November, 2003 were preceded and followed by N 4 also during 1998 - 1999. Ballester et al. [ZOO41 identitied - 5 month intervals without major particle events. Around a 163 day period (nearly identical to that reported in the solar maximum, the frequency of events increases, and a IChlE rate) in the Mount Wilson solar index, which empha- similar pattern is less apparent. sizes strong photospheric magpetic flux, around the max- Variations in the level of solar activity are also sug- imum of cycle 23. However, Ozgiic et al. 120021 found no gested by the rate of interplanetary coronal mas ejections evidence of a “- 150-day” period in the flare index up to the (ICMEs), the manifestations in the solar wind of coronal end of 2000. In this paper, we consider the relationship be mas ejections (CMEs) at the Sun, passing the Earth. Fig- tween the - 166day “period” inferd by Cane and Richard- ure l(b) shows the number of ICMEs/solar rotation, up son [2003] from the occurrence rate of ICMEs, and variations dated from Cane and Richardson [2003]. This rate shows in the intensity of energetic solar particle events, the sunspot prominent variations on timdes of several months. Even number, and interplanetary magnetic field (IMF) strength around solar maximum, there are depressions in the ICME at 1 AU, during the ascending, maximum, and descending rate with durations of one or more rotations that suggest phases of solar cycle 23. temporary reductions in the CME rate at the Sun. ’ Cane and Richardson 120031 noted that the power spec- trum of the ICME rate in 1996 - 2002 had one prominent 2. Observations peak in the 50 - 300 day range, at a period of 166 days. This 2.1. Interplanetary Observations We first examine further the occurrence rate of ICMEs, in Figure l(b). The dashed vertical in Figure 1 are drawn Laboratory for High Energy Astrophysics, NASA lines Goddard Space Flight Center, Greenbelt, Maryland, USA (arbitrarily relative to January 1, 2002) at intervals of the 2Also Department of Astronomy, University of Maryland, 166-day period inferred by Cane and Richardson [2003] from College Park, Maryland, USA. the ICME rate in 1996 - 2002. We suggest that the dashed 3Also School of Mathematics and Physics, University of lines tend to align with the temporary minima in the ICME Tasmania, Hobart, Australia rate during the - 3-year interval from around mid-1999 to late 2002, encompassing the cycle maximum. Outside this interval, the relationship is less clear. Thus, the “166day Copyright 2004 by the American Geophysical Union. period” inferred in the ICME rate appears to be related to OO948276/O4fS5.00 the variations around solar maximum. 1 x-2 RICHARDSON AND CANE: “N 150-DAY PERIODIC’ ACTIVITY

Wavelet Analysis [e.g., Torrence and Compo, 19981 prc- particular, enhancements in the IMF were associated with vides a method of examining “quasi-periodic” features in the increased energetic particle intensities, most conspicuously ICME rate and other data sets in Figure 1, in particular their during the declining phase of the cycle. Figure l(f) shows variation in time. Figure 2 shows wavelet power spectra ob- 27-day averages of the interplanetary magnetic field strength tained using a Morelet wavelet with dimensionless frequency at 1 AU (from the NSSDC OMNI data base). Overall, vari- wo = 12 [cf. Torrence and Campo, 19981 which gives rea- ations in the IMF strength during the current cycle are not sonable spectral and temporal resolution. The power level closely related to those in the ICME rate or SEP intensity, scales are arbitrary and linear, and power at periods of 50 - 300 days is shown. Edge effects may be present in the “cone or ordered by the vertical lines separated at 166 day inter- , vals, although we note that the distinct local minima in the of influence” indicated by hatched areas. Regions with sig- nificance levels of 70%, 90% and 95% are indicated by in- IMF in mid-1999, early 2001 and mid-2002 are coincident creasingly thicker overlaid contours and assume a red-noise with temporarily low ICME rates. Wavelet analysis of the background spectrum (for details, see Torrence and Compo IMF strength (Figure 2(f)) indicates that longer periods, [1998]). Figure 2(b) shows that greatest power in the ICME centered around N 270 days which have no counterpart in rate occurred in an “island” of enhanced power centered at the proton intensity or ICME rate, are dominant. Power is

N 160 days and extending from N 140 to 180 days, dur- less strongly enhanced at periods associated with the “150- ing N 1999 to 2002. This is consistent with the 166-day day periodicity”, although there is evidence of intermittent “period” indicated by the power spectral analysis, and the weak signals in this range. above conclusions based on inspection of Figure l(b). We now consider the 19 - 28 MeV proton intensity. In- 2.2. Solar Observations spection of Figure l(a) suggests that the clusters of energetic proton events during the ascending and descending phases of We now briefly compare the variations observed in these the solar cycle discussed in the Introduction (including the interplanetary phenomena with those at the Sun, specifi- October - November, 2003 activity) tend to be organized cally in the monthly sunspot number (SSN) in the northern around the vertical dashed lines, suggesting that these are (SSN(N)) and southern (SSN(S)) solar hemispheres (Fig- not randomly distributed in time, but have a quasi-periodic ures l(d - e)). Variations in SSN(N) tend to follow to those component similar to N 166 days. We should emphasize that in the ICME rate and energetic proton intensity (though these proton data are dominated by solar events. At lower with occasional exceptions) and show some organization rel- energies, this pattern is less conspicuous because interplane ative to the vertical dashed lines (e.g., the lines tend to align tary processes (including particle acceleration at corotating with local minima in SSN(N) around solar maximum, again and ICME shocks) contribute more prominently. Wavelet with exceptions). Wavelet analysis for SSN(N) shows an is- analysis of log(19 - 28 MeV proton intensity) (Figure 2(a)) land of enhanced power in the - 120 - 200 day range in shows maximum power centered on 150 days in 2000 - - -late-1998 to 2002, with maximum power at N 150 days 2001. Evidence for this can be seen in Figure l(a) where N Weaker signals the larger particle events tend to occur between the verti- in 1999, increasing to 170 days in 2000. centered on N 170 days are present until 2002. Interest- cal lines in 2000 - 2001. In late 2001 - 2002, the period of - ingly, SSN(S) (Figure 2(e); contours are at the same levels peak power shifts to N 220 days, associated with the multi- ple particle events during this time. Intermittent signals at as for SSN(N)) shows much weaker signals than SSN(N) in N - periods of N 2, 3 and 4 solar rotations are also present in the 100 200 day range, with few features consistent Figure 2(a), associated with activity that persists for several with a “150-day” periodicity other than possibly a peak at solar rotations. Thus, despite the complexity of the 19 - 28 N 130 days in late-1999 - early 2000. Consistent with this MeV proton data around solar maximum, a “N 150 day” conclusion, SSN(S) shows little organization by the vertical component was present, at least temporarily. Figure 2(a) lines in Figure 1. A northern-hemisphere bias in the “N 150- also shows a weak component centered on N 150 days in day periodicity”, as suggested by these observations, was late1997 - early 1998, consistent with the clustering noted also reported by Lean [1990] in sunspot areas during cycles in the introduction, and the DaZZa et al. [2001] result. The 15 - 21. results for late 2003 - 2004 are compromised by edge effects, We have examined the heliolatitudes of the sources of though there is an indication of a weak power enhancement all > 20 MeV solar proton events during this cycle (ex- at N 135 days that may be related to the event clustering tending Table 1 of Cane et al. [2002]) to see if this bias during this period, including the October-November, 2003 activity, discussed above. is reflected in the particle events. We find that northern Figure l(c) shows that variations in the rate(/month) SEP events outnumber southern, by 3.1:1, in 1999 - 2000. of geomagnetic storm sudden commencements (SSCs; data However, southern events then dominate in 2001 - 2002 (ra- from the National Geophysical Data Center) usually associ- tio 0.6:1), giving a ratio of 1.1:l for the entire 1999 - 2002 ated with the Earth passage of interplanetary shocks driven period. The stronger “150-day” periodicity in the proton by ICMEs (and less frequently with corotating shocks) are intensity in 2000, when a similar “periodicity” was still ev- somewhat similar to those in the ICME rate. In particular, ident in SSN(N) (cf. Figures 2(a) and 2(d)) and northern minima in the SSC rate also tend to line up with the dashed hemipshere particle events dominated, and weaker period- lines in -mid-1999 - late 2002. Wavelet analysis (Figure icity in 2002, when southern particle events were predomi- 2(c)) shows an “island” of enhanced power, with maximum nant, might be consistent with the northern bias in the SSN power at N 150 days in 2000 - mid-2001. As for the proton periodicity. However, we note that there were few particle data, the signal weakens and shifts to longer periods in late events in 1999 (and hence no clear “periodicity” is evident) 2001. Although affected by edge effects, data for 2004 indi- when peak power at 150 days occurred in SSN(N). Fur- cate a strong signal at N 90 days which is also evident in the - time series in Figure l(c). In summary, the dominant spec- thermore, it is evident from Figure 2 that peak power at tral features in the energetic proton, ICME and SSC data “N 150 days” in the sunspot numbers occured over a year occur around solar maximum and have periods consistent earlier than in the energetic particle data or SSC rate, and nearly 2 years earlier than in the ICME rate. Hence, the with those of the “N 150 day periodicity”. In cycle 21, - 150 day variations were observed in both “150-day” periodicity was not most conspicuous during the the interplanetary magnetic field strength and the occur- same intervals (and at exactly the same periods) in each of rence of energetic particle events [Cane et al., 19981. In the data sets considered here. RICHARDSON AND CANE &- 150-DAY PERIODIC“ ACTIVITY x-3

3. Summary and Discussion Ballester, J. L., R. Oliver, and M. Carbonell, Return of the near 160 day periodicity in the photospheric magnetic flux during Inspection of variations in time-series data, and wavelet solar cycle 23, Astmphys. J. (Lett.), 173, 2004. analysis, suggests that the u- 150-day (quasi)-periodicity” Cane, H. V, and I. G. Richardson, Interplanetary coronal ma was present in solar and interplanetary activity levels dur- ejections in the near-Earth solar wind during 1996 - 2002, ing cycle 23. However, as is typical of such quasi-periodic J. Geophys. Res., 108(A4), 1156, doi:lO.l029/2002JA009817, phenomena, it was only present intermittently, and varied 2003. in period, both with time and between the data sets consid- Cane, H. V., I. G. Richardson, and T. T. von Rosenvinge, Inter- ered. It is manifested most clearly in the wavelet analysis planetary Magnetic Field Periodicity of - 153 days, Gwphys. during - 3 years around solar maximum. At lower activity Res. Lett., 25,4437, 1998. levels, both during the ascending and descending phase of Cane, H. V., W. C. Erickson, and N. P. Prestage, Solar flares, type the cycle, it appears to organize clusters of energetic parti- 111 radio bursts, coronal mass ejections, and energetic parti- cle events, including the intense October - November, 2003 cles, J. Gwphys. Res., 107(10), doi:10.1029/2001JA000320, 2002. events. In the sunspot it was more prominent in number, Dalla, S., A. Balogh, B. Heber, and C. Lopate, Further indica- the northern hemisphere. In contrast to the situation in cy- ,. n, tions of a -140 day recurrence in energetic particle fluxes at 1 LE LA, the “i56-d~y-peiiodiciiy”was not piesent in the I?vW and 5 AU from the Sun, Gwphys. Res., 106, 5721,2001. strength, which was dominated by longer-period variations. J. Gonzalez, A. L. C., W. D. Gonzalez, S. L. G. Dutra, and B. T. It has been suggested [e,g., Bai 19891 that identifyingpatr Tsurutani, Periodic variation in the geomagnetic activity: a terns of “periodic”, or “quasi-periodic” activity in solar and study based on the Ap index, J. Gwphys. Res., 98,9215,1993. interplanetary phenomena, may provide a tool for solar ac- Hill, M. E., D. C. Harhilton, and S. M. Krimigis, Periodicity of tivity prediction on timescales of several months. However, 151 days in outer heliosphere anomalous cosmic ray fluxes, J. given the intermittency and variations in period. which also Geophys. Res., 106, 8315, 2001. differ between data sets, and the fact that intervals of en- Kile, J. N., and E. W. Cliver, A search for the 154 day periodic- hanced activity can persist for several rotations, an accurate ity in the occurrence rate of solar flares using Ottawa 2.8 GHz (to within a few days) prediction of the onset of future ac- burst data, 19551990: Astrophys. J., 370,442, 1991. tivity is certainly unrealistic. At best, at times when the Lean, J. L., Evolution of the 155 day periodicity in sunspot areas “periodicity” appears to be is present, a general prediction during solar cycles 12 to 21, Astrophys. J., 363, 718, 1990. of pmibie increased activity - 150 days (iN 1 - few so- Lean, J. L., and G. E. Brueckner, Intermediate-term solar peri- lar rotations) following a period of energetic solar activity odicities: 100 - 500 days, Astruphys. J., 337, 568, 1989. might be made. Lou, Y.-Q.,Rossby-type wave-induced periodicities in flare activ- Acknowledgments. We gratefully admowledge use of SSC ities and sunspot areas or groups during solar maxima, Astm- data from the National Geophvsical Data Center. sunsDot num- phys. J., 540, 1102, 2000.

. ,. 1146, DO1 10.1029/2001JAbo9080, 2002. The wkelet analysis software by Tomand Compo [1998] Rieger, E., G. H. Share, D. J. Forrest, G. Kanbach, C. Reppin, available from http://paos.colorado.edu/research/wavele. IGR and E. L. Chupp, A 154-day periodicity in the occurrence of i5 supported by NASA grant NCC 5-180 and HVC by a NASA hard solar flares?, Nature, 312, 623, 1984. contract with USRA. Torrence, C., and G. P. Compo, A practical guide to wavelet analysis, Bull. Amr. Meteor. Soc., 79, 61, 1998. Wolf€, C. L., ‘Intermittent’ solar periodicities, Solar Phys., 142, 187, 1992. References Antalova, A., Periodicities of the LDEtype flare occurrence H. V. Cane ahd I. G. Richardson, Code 661, NASA (1969-1992), Adv. Space Re. 14(10), 721-724, 1994. Goddard Space Flight Center, Greenbelt, MD, 20771 (hi- Aschwanden, M. J., Irradiance observations of the 1-8 A solar [email protected]; [email protected]). soft X-ray flux from GOES, Sol. Phys., 152, 53, 1994. Bai, T., Distribution of flares on the sun - Superactive regions and active zones of 1980 - 1985, Astrophys. J., 314, 795,1987. Bai, T.,Distribution of flares on the sun during 19531985 - ‘Hot Figure 1. Solar and interplanetary data for 1996 - spots’ (active zones) lasting for 30 years, Astruphys. J., 328, late-2004: (a) WIND/EPACT 19 - 28 MeV proton in- 860,1988. tensities (MeV cm2 sr)-l, bhour averages); (b) rate of Bai, T., When and where to look to observe major solar flares, in NASA, Goddnrd Space Flight Center, Maz ’91 Work.shop ICMEs/solar rotation, updated from Cane and Richard- 2: Developments in Observations and Theory for Solar Cycle son [2003]; (c) number of SSCs/month; monthly north- 22, eds. R. M. Winglee and B. R. Dennis, p. 46, 1989. em (d) and southern hemisphere (e) sunspot numbers; (f) Bai, T. and E. W. Cliver, A 154 day periodicity in the occurrence IMF strength (27-day averages). Vertical dashed lines are rate of proton flares, Astwphys. J., 363, 299, 1990. drawn at 166day intervals relative to January 1, 2002. Bai, T. and P. A. Sturrock, The 152-day periodicity of the solar flare occurrence rate, Nature, 327, 601, 1987. Bai, T., and P. A. Sturrodc, Evidence for a fundamental period of the Sun and its relation to the 154 day complex of periodicities, Figure 2. Wavelet analyzes for the parameters in Fig- Astrophys. J., 409, 476, 1993. ure 1, in the range 50 - 300 days.

moo 0 5: 00: 00 0 0 OOQ, -0 m 0 om om OCUN-r ON s51 gm-