Neogaean paleomagnetism constraints on the processes at the core and surface of the Earth
D. M. Pechersky

Part 2. Comparison of Processes at the Core and Earth's Surface

According to the present concept of geodynamics, processes at the Earth's surface and in its lithosphere, mantle, and core are closely interrelated (see Introduction), which gave rise to the concept of geonomic periodicity [Khramov and Kravchinskii, 1984; Kravchinskii, 1977] stating that all phenomena of the same rank are essentially equivalent. This provides means for a quantitative description of one phenomenon or process in terms of another (conjugation principle). Unfortunately, most surface processes either cannot be quantified or available information does not embrace even the Phanerozoic. Paleomagnetic data provide the opportunity for the most complete quantitative description of the movement of large continental blocks in the Neogaea. In view of this, I addressed the best documented, measurable geological information, namely biota evolution, chronostratigraphy, and motion of continental plates. Since the frequency of geomagnetic field sign changes, predominance of one of polarities, and relative changes in paleointensity correlate with each other and are synchronous (see Figure 1, part 1; also [Pechersky and Didenko, 1995], we will use data on the reversal frequency which are most representative.

Changes in the Organic World and Geomagnetic Reversal Frequency

fig08 To compare the geomagnetic field behavior with changes in the organic world, I used (1) changes in the number of units in the chronostratigraphic scale (Figure 8a), (2) changes in the diversity of families of marine organisms, and (3) extinctions of marine organism families that occurred every 10 Myr (Figure 8c). The chronostratigraphic scale of Harland et al. [1990] with certain modifications (see Part 1) was used for the discussion of topic (1), and topics (2) and (3) will be discussed on the basis of the Phanerozoic summary presented by Benton [1995] (unfortunately, data on the Riphean are not so representative); only the data on marine organisms were used because they are more representative of the whole Phanerozoic. All data are correlated with the general geological time scale.

The following regularities on three time scales have been revealed.

(1) Regularity of the first order (the whole Neogaea): a sharp difference in the differentiation degree of the chronostratigraphic and magnetostratigraphic scales in the Riphean and Vendian-Phanerozoic (Figures 1 and 8), which is consistent with a rapid progress in the development of diverse life forms beginning from the Vendian-Cambrian.

(2) Regularity of the second order (geological eras): geological eras begin later than reversal frequency minimums all over the Neogaea [Khramov et al., 1982; Molostovskii et al., 1976; Pechersky and Didenko, 1995]. This delay ranging from 35 to 60 Myr (Figure 1) averages 35pm 10 Myr, which yields 4-10 cm/yr for the energy transfer rate from the core-mantle boundary to surface, i.e., for the mantle convection velocity. This value agrees with average velocities estimated for the drift of main continental plates (see below) [Jurdy et al., 1995; Zonenshain et al., 1987]. Therefore boundaries of geological eras are primarily related to an inner geodynamic mechanism.

(3) Regularity of third order (comparable with geological periods): minimums and maximums nearly coincide, which implies changes in the organic world to be synchronous with variations in the polarity reversal frequency and geomagnetic paleointensity; this synchronism is most distinct in the Phanerozoic (Figures 1 and 8). Thus, the difference between maximums (minimums) of reversal frequency and the nearest maximums (minimums) of numbers of stratigraphic units, diversity changes, and extinctions of organisms amounts to 3.1pm 9.6, 3.5pm 6.3 , and 1.4pm 9.5 Myr, respectively. Moreover, many boundaries of geological periods coincide with narrow minimums less than 10 Myr wide or with slight kinks in reversal frequency and/or paleointensity variations (Figure 1) which are smoothed away upon the 30-Myr averaging (Figure 1). Some differences in variation rates of the organic world and reversal frequency are likely to be related to deficient magnetostratigraphic information and dating uncertainties.

Thus, acceleration and deceleration of core-mantle boundary processes and evolution of the organic world occur synchronously; boundaries between geological periods are mostly associated with drops in the reversal frequency and geomagnetic paleointensity (but not with individual reversals!) and themselves often mark a slowdown in the evolution of organic world (Figures 1 and 8).

fig09 If large long-period field variations directly affect changes in the organic world, strong (direct or inverse) correlation should exist between synchronous variations in characteristics of organic world and geomagnetic field, in particular, reversal frequency. However, numerical estimates do not confirm such a correlation, as is evident from Figure 9. Consequently, an immediate effect of long-period field variations on changes in the organic world is either absent or insignificant, and only their cyclicities coincide.

Thus, the comparison of variation rates of organic world and geomagnetic field demonstrates that processes on a geological time scale, as well as shorter processes, which occur near the core and surface, are virtually synchronous, but they are not connected with each other or, speaking more cautiously, their interrelation is obviously secondary and not causative: the processes "respond" to a common external mechanism producing, for example, changes in the rotation axis angle and/or angular velocity of the Earth, which in turn synchronously affect movements in the core and lower mantle, producing changes in the geomagnetic field, and tectonic, climatic and other processes at the surface, producing changes in the state of biosphere. It is only geodynamic processes on a scale of geological eras that are related to the internal mechanism of upward energy transfer from the lowermost mantle (possibly through convection, plume ascent, etc.)

Horizontal Velocities of Continental Plates

The analysis of movements of tectonic blocks usually includes the following procedures: construction of the apparent pole wander path (APWP) of a given block over a specific geological time interval from paleomagnetic directions; spatial paleoreconstructions of a block (and adjacent blocks) for various time intervals, including the positioning of Euler poles of rotation; and determination of the block velocity for various time intervals from the motion of a certain point of the block. The reconstructions are based on tectonic, paleogeographic, paleoclimatic, and other evidence, position of hotspots, and balance of forces that affect the plate motion (slab pull in subduction zones, pull-apart forces at mid-ocean ridges, etc.) [Zonenshain et al., 1987]. Phanerozoic position of six continental plates (Africa, Antarctic, Europe, Siberia, India, and North America) has been recently reconstructed, and their velocities have been estimated [Jurdy et al., 1995]. However, paleoreconstruction of relative and especially absolute positions of the plates are rather ambiguous (particularly in the Precambrian). This is one of reasons why paleotectonic reconstructions of continental plates are not used here. However, the main reason is that this work addresses global regularities in tectonic motions rather than movements of separate blocks. The motion of continents across the Earth's surface resembles that of ice blocks during an ice drift: as a whole, those move downstream, although individual blocks may have different velocities, rotate in various senses and at various angular velocities, stop, and even move upstream, but the movements of each separate ice block have little in common with general regularities of the flow which are main concern of this paper. Therefore, to simplify the problem I will consider separately the latitude (constrained by paleomagnetic inclination) and rotational (constrained by paleomagnetic declination) components of the block velocity with respect to a point in the vicinity of its center as is readily done from its APWP. I used primarily the longest APWPs of various continental blocks, reconstructed by various authors and published reviews of paleomagnetic pole positions. The data collection included the following regions (the time series data and coordinates of the plate "center" are presented in parentheses): Australia (1700-0 Ma; 25o S, 135o E), Africa (720, 550-0 Ma; 0o S, 30o E), Europe (1700-0 Ma; 55o N, 35o E), India (1000, 820-750, 600-0 Ma; 20o N, 80o E), North America+Greenland (1300-1140, 1120-0 Ma; 40o N, 255o E), northern China (580-0 Ma; 45o N, 120o E), southern China (600-550, 450-0 Ma; 30o N, 110o E), and Siberia (1100-840, 730, 640-0 Ma; 60o N, 105o E) [Dawson and Hargraves, 1994; Elming et al., 1993; Embleton, 1984; Enkin et al., 1992; Gordon et al., 1984; Hyodo and Dunlop, 1993; Idnurm and Giddings, 1988; Khramov, 1991; Klootvijk, 1984; Lin et al., 1985; Meert and Van der Voo, 1996; Park et al., 1995; Park and Gower, 1996; Pechersky and Didenko, 1995; Piper, 1995; Pisarevsky et al., 1997; Radhakrishna and Joseph, 1996; Radhakrishna and Mathew, 1996; Saradeth et al., 1989; Seguin and Zhai, 1992; Shapiro et al., 1997; Smethurst et al., 1998; Tarling and Abdeldayem, 1996; Torsvik et al., 1992; Van der Voo, 1988, 1990; Van der Voo and Meert, 1991; Wu et al., 1993; Zhao et al., 1990, 1993].

fig10 The data collected were revised and generalized. Paleomagnetic determinations close in age are sometimes strongly divergent and give abnormally high velocities. These are mostly data from the Early and Middle Riphean of Africa, India, and North America. Such intervals were rejected. Paleolatitude, paleoinclination, paleolatitude component of the plate motion velocity, and angular velocity of paleoinclination variation (rotation about a fixed point of a plate) were calculated at centers of each plate for every 10-Myr APWP interval. To remove possible errors, the results were averaged with a smoothing window of 30 Myr. These data were used for the calculation of mean paleolatitude velocities of plate motion Vpl (Figure 10a) and rotation VD (Figure 10b), as well as standard deviations from these means dVpl and VD. The means were calculated for eight plates in the Phanerozoic, five plates in the Vendian-Late Riphean, three in the Middle Riphean, and two in the Early Riphean.

fig11 The above comparison with an ice drift is clearly illustrated in Figure 11: the majority continents move, with slight variations, consistently, which is most evident beginning from the mid-Paleozoic, but opposite movements (for example, Europe, Siberia, North America, and Australia in a 1100-900-Ma interval) are also observed. The tendency mentioned in the Introduction is also evident from the figure: concentration of most continents in the equatorial zone at the time of existence of Riphean Pangea and their divergence in the Early Paleozoic (breakup of Paleozoic Pangea) followed by movement of Laurasian continents to northern latitudes and movement of Gondwana continents to the equator and southern latitudes in the Mesozoic-Cenozoic.

The velocity pattern of the paleolatitude plate motion component is most distinct (Figure 10a), virtually coinciding with that of the full mean velocities calculated from the spatial reconstruction of Phanerozoic continents (Figure 10c) [Jurdy et al., 1995], although the sets of continents (and their number) are different. Hence, Vpl is likely to rather adequately reflect the variation in the velocity moduli of continents in the Phanerozoic and Precambrian. Figure 10a clearly shows a cyclicity in Vpl disturbed by less regular plate rotations (Figure 10b).

A sharp difference in the behavior of VD and dVD in intervals of 1700-900 and 900-0 Ma. In the first interval, VD has a low scatter and varies rather periodically (with maximums of dVD at 1100 and 980 Ma); in the second interval, both velocity and its scatter gradually increase toward the center of the interval and then gradually decrease. The scatters appear to be minimum during the existence of supercontinents. The Vpl pattern is different: its behavior, with minor deviations, is similar throughout the Neogaea and exhibits a regular change in mean velocities from their minimum (10-20 km/Myr) to maximum (40-60 km/Myr) values rarely (in the Riphean) reaching 80 km/Myr. On the other hand, specific behavior of dVpl is observed in the Riphean, where the scatter varies in the same manner as Vpl, and in the upper half of the Late Riphean-Phanerozoic, where their variations are different, but the most important is the relation dVplll Vpl. Against this background, an interval of "disturbed cyclicity" of Vpl and high dVpl values is clearly observed in the Vendian-uppermost Riphean (650-530 Ma). Based on direct measurements of intervals between adjacent maximums (minimums) and wavelet analysis results (D. K. Galyagin and P. G. Frik), the following characteristic times are recognized in the cyclicity of Vpl, dVpl, VD, and dVD : 20-30, 40-50, 70-80, approx 100, and approx 130 Myr.

Comparison Between Continental Plate Velocities and Geomagnetic Reversal Frequency in the Neogaea

First, I compare kinks in the plate APWPs with field reversal frequency (F) extremums (Figures 1a and 1b). As seen from Table 2, 14 kinks coincide with minimums of F, six coincide with its maximums, and four are behind Fmin by 20-40 Myr; i.e., similar to biota variations, sharp changes in the paleomagnetic pole position are mostly synchronous with reversal frequency extremums, although there are more complex cases of delayed variations.

fig12 Now I compare the maximums (minimums) of Vpl, dVpl, VD, and dVD with the closest reversal frequency maximums (minimums) of the field Fm. Unlike similar comparison with variation rates of the organic world (see above), a scatter of the extremums is broader (Figure 12), which may be related to various uncertainties. The differences Vm-Fm and dVm-Fm mostly form two groups: they are either close to zero ( 0pm 20 Myr) or negative (mainly in the interval of -30 to -60 Myr); therefore, the velocity extremums and their scatter are either synchronous with the reversal frequency extremums or behind those by the value close to delay times of geological era onsets with respect to reversal frequency minimums (Figure 1a). In spite of a large scatter, alternation of intervals dominated by minimum differences Vm-Fmapprox 0 and those with the prevailing values Vm-Fmapprox -40pm 20 Myr is evident (Figure 13). The characteristic time of this alternation is 350-400 Myr.


The above data may be interpreted as evidence of two active mechanisms: the outer, operating synchronously at the core and Earth's surface, and inner, responsible for the values Vm- Fm = -40pm 20 Myr. Negative values of Vm- Fm are a result of the energy transfer from the core-mantle boundary (D'' layer) to the surface (plumes, mantle convection, and so on). A delay of -40pm 20 Myr indicates the velocity of the energy transfer to be 5-10 cm/yr. Such a velocity is consistent with Figure 8 and known estimates of mean velocities of major plates [Jurdy et al., 1995; Zonenshain et al., 1987, 1990].

fig14 The mechanisms may be further specified from the correlation between Vpl and Vm- Fm. If the inner mechanism is dominated by convective motions in the mantle, these values should be inversely correlated, whereas the plume mode of energy transfer probably implies such a correlation to be absent, because continental plate motions are nearly independent of the plume ascent, but they are significantly affected by convective motions in the upper mantle; no correlation should exist in the case of operation of the external mechanism. The values of Vpl and Vm- Fm are likely to be uncorrelated in both intervals 0pm 20 Myr and -40pm 20 Myr (an inverse correlation seems to exist in the case of minimum values of Vpl, see Figure 14). Whereas the absence of correlation in a 0pm 20-Myr interval supports the action of the external mechanism, the absence of correlation in the second interval implies that the mantle convection is uninvolved in the interrelation between the processes in the core and lowermost mantle (D'' layer) and movements in the lithosphere. This may be explained by the fact that processes in the D'' layer and mantle convection are independent, as is indicated, for example, immobility of hotspots relative to moving plates (which is used for estimation of absolute plate movements) [Jurdy et al., 1995; Zonenshain et al., 1987]. The correlation may be disturbed by two-layer convection (in upper and lower mantle) which accounts for not more than 10% heat and mass transfer [Allegre, 1997].

As seen from Figures 13 and 14, the variation ranges of mean plate velocities in both mechanisms are very close (10-60 and 20-55 km/Myr) and overlap the range of possible dependence of Vpl on Vm- Fm. Therefore, the variations in the Vpl means are caused by general factors that affect both external and internal mechanisms. The rate of core-to-surface energy transfer varies within comparatively narrow limits as is evident from a -40pm 20-Myr range of delay times between velocity and reversal frequency extremums and from overlapping ranges of variation in maximum and minimum values of mean velocities of paleolatitude plate movements (these ranges are, respectively, 30-60 and 10-45 km/Myr for the external mechanism and 40-55 and 20-40 km/Myr for the internal one (Figure 14). Thus, irrespective of geomagnetic field generation mechanisms and plate movements, the rates of core-to-surface energy transfer and plate motions are close, which is evidently due to properties of the medium. As seen from Figure 13, there exist time intervals dominated by the action of the external synchronous mechanism ( Vm- Fm=0pm 20 Myr) or internal mechanism with a characteristic delay time of -40pm 20  Myr. In particular, the aforementioned Vendian interval with "anomalous" Vpl and dVpl lies within the longest interval dominated by the action of the external mechanism (Figures 1 and 13).

Generalization of results

The above data imply that the relations between processes in the core and lowermost mantle on the one hand and at the Earth's surface on the other hand are realized at least at three time scale levels.

First time scale level

is the whole Neogaea. At the Earth's surface, a relevant characteristic feature is a different degree of differentiation of the chronostratigraphic scale in the Riphean and Vendian-Phanerozoic (Figure 8), which is evidence of an intense rise in the development of various life forms that began in the Vendian-Cambrian. The final interval of the Riphean-Vendian (650-530 Ma) is characterized by disturbed cyclicity of the continental plate velocity (Vpl) variation (Figure 10), and the Vendian interval of anomalous Vpl lies within the longest interval dominated by action of the external mechanism (Figures 1 and 13). The conclusive stage of existence of the Riphean supercontinent Pangea and its breakup also took place at that time. The peak of tectonic and thermal activity, lying approximately between 570 and 500 Ma, was associated with a higher heat flow, the widest occurrence of the granulite facies of metamorphism, reworking of the crust, and epeirogenic uplifts and basalt outflows possibly related to mantle plumes [Williams, 1994]. At the same time, essential changes at the core-mantle boundary markedly affected all characteristics of the geomagnetic field, particularly, its reversal frequency whose pattern is very close to the "frequency" of startigraphic stages (Figure 8). The geomagnetic field behavior differs in its fractality. A difference between the Riphean and Phanerozoic intervals and uniqueness of their boundary which manifested themselves at both the Earth's surface and core may be explained by resonance-type enhancement of core motions, caused by critical duration of the Earth's diurnal rotation ( approx22.2 hours) due to the tidal decrease in its angular velocity by a factor of about 1.5 from the Archean to present time. This deceleration resulted in a considerable temperature increase near the core-mantle boundary, destabilization of the D'' layer, and ascent of mantle plumes [Williams, 1994].

Second level

is the scale of geological eras, characterized by their onset delay with respect to reversal frequency minimums throughout the Neogaea. On this scale, long intervals of stable geomagnetic polarity, typically preceding the beginning of geological eras occur cyclically with a 160-200-Myr period close to the era duration (except for two anomalies between 1150 and 1100 Ma and between 700 and 630 Ma). This regularity is supported by a fractal dimension of reversal frequency of about 0.9 in the Neogaea. These stable polarity intervals are likely to coincide with plate velocity minimums, which is natural, given stable conditions at the core and Earth's surface. The aforementioned delay amounts to 35-60 Myr (Figure 1) and yields a value of 4-10 cm/yr for the energy transfer rate from the core-mantle boundary to the Earth's surface, i.e., for the mantle convection velocity. This value is consistent with mean velocity estimates for the drift of major continental plates (Figure 10) [Jurdy et al., 1995; Zonenshain et al., 1987]. Therefore, boundaries of geological eras are mainly related to the action of inner mechanism.

Third level

is the scale of geological periods. First, it is characterized by nearly exact coincidence of minimums or maximums; i.e., variations in the organic world and continental plate velocities occur synchronously with the polarity reversal frequency and variations in the geomagnetic paleointensity (Figures 1, 8, and 10). Moreover, geomagnetic events are synchronous with such "lithospheric" events as trap outflows, jumps in the spreading velocity, stratigraphic unconformities in geological sections (reflecting fluctuations in the sea level), folding phases, appearance of evaporites and tillites (climatic changes), occurrence of black shales (redox conditions). Most prominent is the coincidence of the above events at times of 250-245, 196-190, 148-144, 113-110, and 97-91 Ma [Rampino, 1988; Rampino and Caldeira, 1993; also see Introduction]. The cyclicity of all aforementioned processes exhibits periods ranging from 20 to 100 Myr (periods of ~20-30, ~50, and ~100 are best defined). In addition, many boundaries of geological periods coincide with narrow minimums less than 10 Myr long and small kinks in the reversal frequency and/or paleointensity variations (Figure 1). Consequently, accelerations or decelerations core-mantle boundary processes, variations in the organic world, plate motions, and other processes at the Earth's surface occur synchronously, and boundaries of geological periods are often associated with a decrease in the reversal frequency and geomagnetic paleointensity and often manifest a decline in the organic world development (Figures 1 and 8).

Second, in addition to synchronism (coincidence of Vm and Fm extremums), Vm extremums are behind the reversal frequency extremums mainly by 30-60 Myr. A similar delay is observed in the onsets of geological eras with respect to reversal frequency minimums (Figures 1 and 13). Moreover, intervals dominated by the differences Vm- Fm=0 alternate with those dominated by Vm- Fm=40pm 20 Myr (Figure 13). The characteristic time of their alternation is 350-400 Myr, which is a twofold cyclicity period of constant polarity intervals.

These data may be explained in terms of active mechanisms of two types: an external mechanism that acts synchronously at the core and Earth's surface and an internal mechanism responsible for a -40pm 20-Myr delay of processes at the Earth's surface with respect to those at the core. In all of the cases, this is only a qualitative relationship, whereas no numerical relation between processes at the surface and core virtually exists (Figures 9 and 14). Consequently, there is no causal connection between these processes; rather, they are subjected to the action of a general mechanism.

The aim of this paper is to constrain current hypotheses using available information rather than to determine a specific mechanism (or mechanisms).

On the one hand, there exist processes that are useful for reasonable explanation of synchronism of events at the Earth's surface and core such as the tidal evolution of the Earth-Moon system, evolution of the Earth as a part of the Solar system, and general evolution of the Galaxy. Thus, tidal movements produce periodic variations in the angular velocity of the Earth, Earth-Moon distance, and inclination of the rotation axis; as a result, the rotation poles of the Earth markedly change their position (rotations of the planet as a whole relative to the ecliptic and/or rotations of the mantle relative to the core). The same group of processes includes changes in the rotation axis position caused by the continental drift. The above phenomena in turn lead to changes in climate, sea level, and position of the core, which should affect the geomagnetic field. Thus deceleration-acceleration regimes of the Earth's rotation should result in polarity reversals or preferred polarity intervals. Furthermore, the intervals of maximum gradients in mean velocities of the continental drift, summary amplitude extremums of direction paleovariations, and maximums of prevailing reversed polarity are very close to the epochs of the Solar system passage through the Galaxy plane; the main period of such oscillations is close to 30 Myr and virtually coincides with the main period of fauna extinctions, sea level lowering, large basalt outflows, folding, abrupt drops in spreading rate, and enhancement episodes of reduction conditions over the past 250 Myr [Rampino, 1988; Rampino and Caldeira, 1993], but it is also the half-period of variations in main parameters of the geomagnetic field in the Neogaea; finally, one of the periods of field characteristics and plate motions, considered above, is close to the Galactic year.

On the other hand, there are surface processes (geological era onsets, variations in mean velocities of continental plate motion, etc.) that proceed later than those at the core-mantle boundary (Figure 13); this delay is most naturally treated in terms of the energy transfer from the core-mantle boundary (D'' layer) to the Earth's surface (plumes, mantle convection, etc.). The delay time constraint provides a value of 5-10 cm/yr for the energy transfer rate, which is consistent with velocities of major plates. Lacking (or very weak) numerical correlation implies that the mantle convection cannot be responsible for the interrelation between processes in the D'' layer and lithospheric motions. This may be explained in terms of two-layer convection (in the lower and upper mantle) with very limited heat and mass transfer between these layers (not more than 10%) [Allegre, 1997] and/or by the fact that processes in the D'' layer and mantle are independent.

General action of the two mechanisms may be described as follows. The external mechanism causes processes in the D'' layer (activity, instability, etc.) which stimulate the heat and mass transfer in the mantle, i.e., bring about the action of internal mechanism. Furthermore, mantle mass movements (convection, plumes, subduction), caused by the activity in the D'' layer and related to plate motions, change the planetary moment of inertia, i.e., bring about synchronous action of the external mechanism, and so on. Such a relation is supported by very close limits of variations in mean velocities of plate motion as constrained by each of the mechanisms (10-60 and 20-55 km/Myr, respectively).

In general, irrespective of mechanisms responsible for the geomagnetic field generation and plate movements, the energy transfer from the core to Earth's surface and plate movements have close velocities, so that these processes are evidently controlled by properties of the medium.

This document was generated by TeXWeb (Win32, v.1.0) on April 7, 1999.