Spectral analysis of Pennocarboniferous geomagnetic variation data from glacial rhythmites

Geophys. J. R. asb. SOC.(1981) 67, 641-647

Spectral analysis of Pennocarboniferous geomagnetic variation data from glacial rhythmites

M.Ernest0 and I. G . Pacca Institute AstronBmico e Geofisico da Universidade d e Sdo Paulo, Olixa Postal 30. 627, 01000-SEo Paulo, Brazil

Received 1981 February 27; in original form 1979 October 23

Summary. Palaeomagnetic results for a sequence of Permocarboniferous rhythmites presented in the previous paper have been submitted to maximum entropy spectral analysis to test whether these palaeomagnetic data could supply information on geomagnetic variations. There is a good correlation between the thickness of the rhythmites and sunspot spectra, suggesting that these sediments are really seasonal. The palaeomagnetic spectra are compared with those of observatory records. Periods of approximately 24.4, 12.4,8.6, 6.7 and 5.5 found for palaeomagnetic data have corresponding values in the geomagnetic spectrum. Most of these periods, however, are the same as those found in the thickness data, implying that magnetization can be influenced by the sedimentation process as suggested by other investigators. On the other hand, both geomagnetic and climatic (thickness) variations seem to be related to solar activity. Therefore, at least indirectly, palaeomagnetic data may reflect geomagnetic variations. 1 Introduction

Results of a palaeomagnetic analysis of a series of Permocarboniferous glacial rhythmites have been presented by Sinito et al. (198 1, referred to as Paper I). If the rhythmites are annual, the palaeomagnetic record could be used for studying the fine structure of geomagnetic field variations in the geological past. However, Paper I shows that, even if the rhythmites are varves, the palaeomagnetic record contains fluctuations which do not correspond to the present-day geomagnetic field behaviour. The question is: to what extent can palaeomagnetic data from varved sediments be used to study geomagnetic field variations in the past? The main purpose of this work is to find out whether the present data support the hypothesis that the Itu rhythmites are varves and, if so, to establish how palaeomagnetic variations reflect variations of the Permocarboniferous geomagnetic field. All the series of the Itu palaeomagnetic and varve thickness data were submitted to the Maximum Entropy Spectral Analysis (MESA). A similar treatment was also applied to another rhythmite sequence belonging to the same geological formation (Itarark Subgroup) and located at Potreiro Grande, about 1000km from Itu. Thicknesses were measured by Delaney (1964) and are plotted on Fig. 1.

642

M. Ernest0 and I. G.Pacca THICKNESS

(rnm)

Figure 1. Potreiro Grande varve thickness plotted against pair number, in stratigraphic order.

2 Spectral analysis and results The advantages of the MESA method proposed by Burg (1967), especially its high resolution, have been pointed out by several authors, e.g. Radoski, Fougere & Zawalick (1975), Radoski, Zawalick & Fougere (1976), Chen & Stegen (1974) and Ulrych (1972). MESA is particularly useful when the periods to be investigated are of the same order of magnitude as the total series length, as often happens in geophysical studies. However, peak shifting at the low-frequency end of the spectrum may still occur. The time series submitted to MESA were annual palaeomagnetic data ( N = 167) and varve thicknesses (N=181 for the Itu varves, and N = 187 for the Potreiro Grande varves) given in

643

Permocarboniferous geomagnetic variations Periodlyl

1000

250

145

100

77

63

,--vi-T-

("1 53

P*llOd

I000

53

250

I45

100

63

77

7 T

-

-

I

\

--As.-05

0

10

05

I5 Frequency lcwi

Figure 2. MESA spectra for (a) inclination (M= 20); (b) declination ( M =27); (c) intensity (M=33); (d) susceptibility ( M = 2 0 ) ; (e) Itu thicknesses ( M = 35) and (f) Potreiro Grande thicknesses ( M = 30). The ordinate is a power density function.

Paper I. As discussed in Paper I, all the series are affected by noise which may alter the quality of spectra (Chen & Stegen 1974). They were smoothed following the same procedure as in Paper I but now the elementary binomial smoothing was applied only once. Trends in the series, which affect the low-frequency ends of spectra, were removed by subtracting values from the straight line passing through the first and last points (Alldredge 1976). A clear physical justification cannot be presented for this procedure since the processes which cause such trends are not known. Effects of removing a trend from the data are discussed by Courtillot, Le Moue1 & Mayaud (1977). Spectra were computed using the estimation procedure presented by Andersen (1974). Periods exceeding the series length have not been investigated and a high-frequency cut-off at f=0.200 cycle yr-' was used since periods of less than five years were not considered meaningful. The use of MESA requires a suitable choice of the number of coefficients ( M ) for the prediction error filter (PEF). The final prediction error method (FPE) proposed by Akaike (1970) was applied, as described by Ulrych & Bishop (1975). However for the analysed data it was found that in some cases the FPE first minimum occurred for M between 2 and 4.

-

Frequency 7 -

50

67

I00

200

Icpyi

m

Per od

-,

~

200

-

I00

I -

61

(years)

Figure 3. MESA spectrum for the complex plane direction data ( M = 29).

50

M. Ernest0 and I. G. Pacca 644 These values were considered unacceptable since the corresponding spectra were highly smoothed. M was then selected at the next outstanding minimum, or at a slightly larger value, but not exceedingN/2 (Courtillot et al. 1977). The spectra obtained for the various time series are shown on Fig. 2 and the periods found are summarized in Table 1 . Declination and inclination were analysed separately (Fig. 2a and b) and also jointly as a complex equivalent series, according to the method proposed by Denham (1975). In this case the spectrum is calculated both for positive and for negative frequencies (Fig. 3). The complex series periods are found in the declination spectrum but not in the inclination spectrum, even i f M is increased up to 80 when only very weak peaks appear at 1 1 . 1 and 6.7 yr. A possible explanation for this could be fluctuations in inclination errors due to grain size and compaction effect variations which would tend to mask the highest frequency variations. 3 Discussion

Periodicities in the Itu and Potreiro Grande thickness series (Table 1) show a correspondence which would be improbable if each layer which was taken to represent one year corresponded to a variable number of years. It appears that both thickness series follow a general law since the two sequences are about 10" of latitude apart and are separated by a time interval of the order of a few million years: microflora determinations indicate a late Carboniferous age for the base and an early Permian age for the top of the Itarare Subgroup (Daemon & Quadros 1970; Rocha-Campos & Rosler 1978). The Itu and Potreiro Grande sequences belong respectively t o the base and the top of the Subgroup. Solar-climate relationships have been suggested for example by Willet (1967) who found that periodicities for some climate-related parameters were consistent with the solar activity spectrum. Therefore if the rhythmites are seasonal, it is reasonable to compare their spectra with those of sunspot and other climate-related parameters. Table 2 shows a consistency between these results and the periodicities in the Itu and Potreiro Grande rhythmites series which supports the idea that the thicknesses are climatically controlled, in annual deposition cycles. Spectral analysis results for the magnetic elements are compared with thickness results in Table 1. Magnetization intensity depends on deposition through changes in shape, size and quantity of magnetic grains and also on changes in the physico-chemical properties of the environment as well as the palaeomagnetic field strength. Magnetic susceptibility is also grain size- and shape-dependent since it is determined by the magnetic mineralogy of the sediment. Their correspondence with thickness should be even better than observed. However Barton, McElhinny & Edwards (1980) have observed a demagnetization associated with compaction which could obliterate some periodical variations. Susceptibility periods of 83.3 y r and its harmonic 14.1yr (83.3/6) are not found at the thickness spectrum and it seems improbable that there would be a mechanism which would Table 1. Period (yr) Inclination Declination Complex data Intensity Susceptibility Itu thickness Potreiro Grande thickness

I 1.4

23.8 24.4

11.6 12.4 17.2 14.1

83.3

40.0 435

11.5 12.0

8.4 8.6 8.0 7.6 8.3 8.6

6.1 6.7 6.4

6.8

5.5 5.4 5.6 5.5

Pennocarboniferousgeomagnetic variations

645

Table 2. Period (yr)

Itu thickness Potreiro Grande thickness Sunspot number data (Fraser-Smith 1972) Sunspot number data (Currie 1973b) Sunspot number data (Cohen & Lintz 1974) Glacial varves (Anderson & Koopmans 1963) Evaporite varves (Anderson & Koopmans 1963) Clastic marine varves (Anderson & Koopmans 1963) Glacier recession (Antevs 1929) Strandlines (Fairbridge & HiliaireMarcel 1977) Modern temperatures (Landsberg, Mitchell & Crutcher 1959)

40 .O 43.5 44.1 23.6

11.5 12.0

8.3 8.6

6.5

5.6 5.5

16.2

10.2

7.2

6.8

5.3

14.7

11.1

9.9

8.3

6.8

5.7

10.9

9.7

5 6

45 25-22

12

8

5

6

55

45

11.0

5 .O

alter susceptibility periodically, without affecting thickness. The 83.3 yr susceptibility period could correspond on the other hand to the 80-90 yr Gleissberg sunspot cycle found in sun-climate correlation (Willet 1967) as well as in evaporitic varves data (Anderson & Koopmans 1963). The latter result would indicate that this period could exist in the thickness spectra although it has not been found in this analysis. Table 3. Complex data periods (yr)

Geomagnetic spectrum Periods (yr) References

24.4

22 21.4 22.9 22-24.6 11 10.2 10.5 11.5 10 11-12 7.04 7.1 6.07 6.9 6.4-6.5 5.5 5.1 5.15 5.6 5.5 5.4

12.4

8.6 6.7

5.5

Bhargava & Yacob (1969) Currie (1973a) Currie (1976) Courtillot et 4Z. (1977) Bhargava & Yacob (1969) Fraser-Smit h ( 1972) Currie (1973a) Currie (1976) Courtillot & Le Mouel (1976) Courtillot ef 4Z. (1977) Fraser-Smith (1972) Currie (1973a) Currie (1973a) Currie (1976) Courtillot el 4Z. (1977) Bhargava & Yacob (1969) Fraser-Smith (1972) Currie (1973a) Currie (1976) Courtillot & Le Mouel (1976) Courtillot et al. (1977)

646

M. Ernest0 and I. G. Pacca

Table 1 shows that except for the 24.4 and 6.7 yr periods all the direction periods have corresponding values in the Itu thickness spectrum. All direction periods have, on the other hand, been found in the geomagnetic spectrum, as shown in Table 3. The Gleissberg period has been found in the geomagnetic spectrum (Bhargava & Yacob 1969; Currie 1976) and although it does not appear at the declination and complex spectra, there is a clear peak at the inclination spectrum (Fig. la) at 71.4 yr. Similarly for the magnetization intensity 17.2 yr period, already mentioned by Bhargava & Yacob (1969) and by Fraser-Smith (1972) who also found a 35.6 yr period in the activity index Ap spectrum which could correspond to the 40-43 thickness period. It should be added that this period was rather unstable with changes o f M for the spectral analysis. In conclusion, it can be said that both thickness and magnetic periodicities in the varve data correspond to the geomagnetic periodicities obtained by various authors. However, it does not seem possible to separate direct geomagnetic effects from those caused by other factors during the acquisition of magnetization. This may be a consequence of both the geomagnetic field and glacier process being affected by solar activity. It is improbable that short-period geomagnetic variations can be detected in rocks which have been subjected to weathering and secondary magnetization. However, variations in the deposition process of varves, which can be inferred from their thickness, could act as an amplifying agent for geomagnetic variations, as already suggested by Verosub (1977) and Morner (1978).

Acknowledgments This work has been supported by FAPESP grants 741024 and 7510413. The authors wish to thank Dr Paul F. Fougere for some suggestions on the MESA method, and Dr C. Barton for his kind suggestions and revision of this paper.

References Akaike, H., 1970. Statistical predictor identification, Ann. Inst. Statist. Math., 22,203-217. Alldredge, L. R., 1976. Effects of solar activity on annual means of geomagnetic components, J. geophys. Res., 81,2990-2996. Andersen, N., 1974. On the calculation of filter coefficients for maximum entropy spectral analysis, Geophysics, 39,69-72. Anderson, R. Y. & Koopmans, L. H., 1963. Harmonic analysis of varve time series,J. geophys. Res., 68, 877 -893. Antevs, E., 1929. Cycles in variations of glaciers and ice sheets and in ice melting, Reports of the Conference on Cycles, pp. 1-83, Carnegie Institute, Washington. Barton, C. E., McElhinny, M. W. & Edwards, D. J., 1980. Laboratory studies of depositional DRM, Geophys. J. R . asp. Soc., 61,355-377. Bhargava, B. N. & Yacob, A., 1969. Solar cycle response in the horizontal force of the earth’smagnetic field, J. Geomagn. Geoelect., 21, 385-391. Burg, J. R., 1967. Maximum entropy spectral analysis, paper presented at the 37th Annual International Meeting, Society of Exploration Geophysicists, Oklahoma City, Oklahoma. Chen, W. Y. & Stegen, G. R., 1974. Experiments with maximum entropy power spectra of sinusoids,J. geophys. Res., 79,3019-3022. Cohen, T. J. & Lintz, P. R., 1974. Long term periodicities in the sunspot cycle, Nature, 250, 398-399. Courtillot, V. & Le Mouel, J. L., 1976. On the long-period variations of the earth’s magnetic field from 2 months t o 20 years, J. geophys. Res., 81,2941-2950. Courtillot, V., Le Mouel, J. L. & Mayaud, P. N., 1977. Maximum entropy spectral analysis of the geomagnetic activity index aa over a 107-year interval, J. geophys. Res., 82,2641 -2649. Currie, R. G., 1973a. Geomagnetic line spectra-2 to 70 years, Astrophys. Space Sci., 21,425-438.

Pernocarboniferous geomagnetic variations

647

Currie, R. G., 1973b. Fine structure in the sunspot spectrum - 2 to 70 years, Astrophys. Space Sci., 20, 509 -5 18. Currie, R. G., 1976. Long period magnetic activity - 2 to 100 years, Astrophys. Space Scb, 39, 251-254. Daemon, R. & Quadros, L. P., 1970. Bioestratigrafia d o Neopaleozbico da Bacia do Paranh, An. X X I V Congr. Bras. Geol., pp. 359-412. Delaney, P. I. V., 1964. Itarar6 outliers in Rio Grande do Sul, Brazil, BoZ. Paran. Geogr., 1011.5, 161-173. Denham, C. R., 1975. Spectral analysis of palaeomagnetic time series, .r. geophys. Res., 80, 1897-1901. Fairbridge, R. W. & Hillaire-Marcel, C., 1977. An 8,000-yr palaeoclimatic record of the ‘Double-Hale’ 45-yr solar cycle,Nature, 268,413-416. Fraser-Smith, A. C., 1972. Spectrum of the geomagnetic activity index Ap, J. geophys. Res., 77,42094221. Landsberg, H. E., Mitchell Jr, J. M. & Crutcher, H. I., 1959. Power spectrum analysis of climatological data for Woodstock College, Maryland, Mon. Weath. Rev., 87,233-298. Morner, N. A., 1978. Paleomagnetism of varved clays, IAS 10th Congress, Jerusalem. Radoski, H. R., Fougere, P. F. & Zawalick, E. J., 1975. A comparison of power spectral estimates and applications of the maximum entropy method,J. geophys. Res., 80,619-625. Radoski, H. R., Zawalick, E. J. & Fougere, P. F., 1976. The superiority of maximum entropy power spectrum techniques applied t o geomagnetic micropulsations, Phys. Earth planet. Int., 12, 208216. Rocha-Campos, A. C. & Rosler, O., 1978. Late paleozoic faunal and floral successions in the Parana Basin, Southeastern Brazil, Bol. IG., Inst. Ceoc., University of Sio Paulo, 9, 1-15. Sinito, A. M., Valencio, D. A., Ernesto, M. & Pacca, I. G., 1981. Palaeomagnetic study of Permocarboniferous glacial varves from the Itarari: Subgroup, Southern Brazil, Geophys. J. R. astr. SOC.,67, 635-640. Ulrych, T. J., 1972. Maximum entropy power spectrum of truncated sinusoids, J. geophys. Res., 77, 1396-1400. Ulrych, T. J. & Bishop, T. N., 1975. Maximum entropy spectral analysis and autoregressive decompostion, Rev. Geophys. Space Phys., 13,183-200. Verosub, K. L., 1977. Depositional and Postdepositional Processes in the magnetization of sediments, Rev. Geophys. Space Phys., 15,129-143. Willet, H. C., 1967. Solar-climate relationships, The Encyclopedia of Atmospheric Science and Astrogeology (Reinhold), p. 869.