Introduction

General considerations.

According to current notions, the geomagnetic field is generated by processes in the liquid core, and therefore information about its time behavior reflects the processes that occur both in the Earth's core and at its boundary with the mantle (D^'' layer). On the other hand, geological information characterizes processes at the Earth's surface and in the lithosphere, and correlation between geomagnetic and geological phenomena reflect processes in the Earth as a coherent system including core and lithosphere.

A unique source of information about the time evolution of geomagnetic field is the paleomagnetic record imprinted in ferrimagnetic minerals which are present in the majority of rock and preserve this information over the geological history of the Earth. Available paleomagnetic data enable the analysis of geomagnetic field behavior during long time intervals. In this paper, I focus on the Neogaea, the time period covering the past 1700 Myr I chose this time interval because rocks older than 2000 Ma rarely compose continuous stratigraphic sequences that were not subjected to significant transformations during their geological history and preserved the original paleomagnetic information. For this reason, we cannot now obtain an uninterrupted time series of main characteristics of the geomagnetic field and construct the geomagnetic polarity scale for the time older than 1.5-2 Ga.

Main characteristics of the geomagnetic field determined from paleomagnetic data.

The geomagnetic field intensity (the vector characterized by its magnitude, direction, and time variations) is derived from both direct observations at the Earth's surface and paleomagnetic records. It is desirable to have all characteristics of geomagnetic field, but the question is which of them are actually available.

Geomagnetic reversals, i.e.  180^o changes in the field direction, are most easily recognized. This global phenomenon underlies the construction of the global geomagnetic polarity time scale with the use of such methods as (1) magnetostratigraphic approach including paleomagnetic study of successive beds of stratigraphic sections, (2) magnetochronological approach combining paleomagnetic determinations and radiological datings of rocks, and (3) analysis of oceanic linear magnetic anomalies.

Scale intervals are most reliable if they are constrained by all of the three methods. Since method (3) is limited by a seafloor age of 170 Ma, this interval of the geomagnetic polarity scale is most reliable.

The scale is commonly used for the analysis of the reversal frequency variation and geomagnetic polarity asymmetry over the past hundreds of millions of years (see e.g., [Algeo, 1996; Courtilliot and Besse, 1987; Gaffin, 1989; Irving and Pulaiah, 1976; Johnson et al., 1995; Khramov et al., 1982; McElhinny, 1971; McFadden and Merrill, 1984, 1986; Pechersky and Didenko, 1995; Pechersky and Nechaeva, 1988]). Presently, several regional magnetostratigraphic scales have been published, providing a basis for the construction of the Neogaea polarity geomagnetic time scale [Pechersky, 1997, 1999].

The next characteristic is the magnitude of the field (intensity modulus) and its variation. The incorrect term paleointensity, standing for the intensity modulus, is widely accepted in the present-day paleomagnetic literature, although this term means the complete vector. Moreover, the term "paleointensity" is used even when data are given in tesla which is the measurement unit of induction. Following the "tradition", we are compelled to use the term "paleointensity" in the meaning of the paleointensity magnitude regardless of measurement units.

The paleointensity behavior analysis, usually including the past 400 Myr and occasionally the Phanerozoic and Proterozoic, yields evidence of preservation of the dipole field component and cyclicity of paleointensity variation close to the reversal frequency [Bolshakov and Solodovnikov, 1981; Khramov et al., 1982; Merrill and McElhinny, 1983; Pechersky, 1998; Pechersky and Nechaeva, 1988; Perrin and Shcherbakov, 1997; Petrova, 1989; Petrova et al., 1992]. The global paleointensity database was created in the 1990s [Tanaka and Kono, 1994], which considerably stimulated the work on data generalization including the Neogaea [Pechersky, 1998].

The last characteristic is variations in the geomagnetic field direction at an observation point or variations in the position of the virtual geomagnetic pole (VGP), ranging within several tens of degrees. The VGP analysis can be performed only for the past few millions of years, when relative displacements of lithospheric blocks were small and VGP position may be referred to the present coordinates of an observation point.

Magnetologists often address the analysis of paleomagnetic directions and VGPs over the past 5 Myr along with data of laboratory observations (see e.g., [McElhinny and McFadden, 1997; McElhinny et al., 1996; Merrill and McElhinny, 1983; Tsunakawa, 1988]).

Information about paleointensity and field direction variations is vital to the study of the mechanism of geomagnetic field generation and development of theoretical approaches, but deep studies of the variations (and, generally speaking, the fine structure of geomagnetic field) require detailed analysis of the field behavior based on the comprehensive investigation into sections composed of rapidly accumulating volcanic-sedimentary deposits and archaeological objects. This can be done for only short time intervals because of immense volume of work and scarcity of sections that were continuously accumulated over geologically large time intervals. An alternative approach, not requiring the direct study of paleovariations of various types and origin, focuses on the analysis of the summary amplitude of all paleovariations in a given time interval, rather than the behavior of the geomagnetic field direction proper. Instead of long sections, this approach requires representative data on the concentration of paleomagnetic directions (precision parameter), because the summary amplitude of geomagnetic variations is defined by the standard angular deviation S = 81/K^1/2 where K is the precision parameter of individual vectors in terms of the spherical projection statistics [Fisher et al., 1987].

The analysis of paleomagnetic data covering geologically long time intervals was very rarely applied to the summary amplitude of Phanerozoic direction paleovariations [Irving and Pulaiah, 1976; Pechersky and Nechaeva, 1988]. The creation of the computer global database of paleomagnetic directions and poles [McElhinny and Lock, 1993] largely facilitated the problem and enabled the amplitude analysis of the Neogaea geomagnetic variation directions [Pechersky, 1996, 1997].

Cyclicity analysis.

The long-period cyclicity of geomagnetic and other processes has been statistically estimated in a number of works (see e.g., [Keondzhyan and Monin, 1977; Loper et al., 1988; Marzocchi and Mulargia, 1992; McElhinny, 1971; Merrill and McElhinny, 1983; Pechersky and Didenko, 1995; Pechersky and Nechaeva, 1988]). Periods of about 15 to 430 Myr have been revealed. Different geomagnetic polarity regimes with characteristic times of 1, 10, 50, and 200 Myr have been recognized; they serve as a basis for the classification of magnetic zones of various ranks, their correlation and construction of magnetostratigraphic scales [Danukalov et al., 1983]. On the other hand, some authors note that no periodicity is observed in the intervals between reversals; i.e., this process is stochastic, with reversal frequency increasing in the Mesozoic-Cenozoic, and this general trend is modified by a periodicity mainly related to the thermal regime at the core-mantle boundary (in the boundary layer D^'' ) and possibly with its topography [Cox, 1981; Gubbins, 1989; Marzocchi and Mulargia, 1992; McFadden and Merrill, 1984; Ricou and Gibert, 1997].

Correlations between geomagnetic field variations (core) and surface processes.

During more than one decade, attempts have been made to generalize available paleomagnetic data and to find regularities in the behavior of main geomagnetic field characteristics and their correlations with other phenomena of the Earth and surrounding space. We dwell on examples of investigations into long-period processes.

Correlations of tectonic, magmatic, climatic, paleogeographic, and biostratigraphic events, as well as their cyclicity, with the geomagnetic field behavior and specifically with geomagnetic reversals have been noted by many authors [Aparin, 1982; Courtillot and Besse, 1987; Didenko, 1998; Eide and Torsvik, 1996; Gaffin, 1987; Khramov, 1978; Khramov et al., 1982; Kiselev and Aparin, 1987; Kravchinskii, 1977, 1987; Larson, 1991; Larson and Olson, 1991; Loper and McCartney, 1986; Loper et al., 1988; Marzocchi et al., 1992; Pechersky and Didenko, 1995; Pechersky and Nechaeva, 1988; Rampino, 1988; Rampino and Caldeira, 1993; Ricou and Gibert, 1997; Varygin and Aparin, 1989; Vogt, 1972, 1975]. This evidence supports the well-known concept of deep origin of tectonic movements and universality of the Earth's endogenic process. A pioneer of this concept was Yu. M. Sheinman. Developing the ideas of Sheinman, Kravchinskii [1977, 1987], Khramov [1978], Khramov et al. [1982], and Khramov and Kravchinskii [1984] compared the geomagnetic reversal behavior, paleomagnetic pole velocities, accelerations of continents, stages of plate subduction and obduction, stages of folding and continental basaltic volcanism (traps), seafloor accretion rate, and other events and found that events at the core and in the lithosphere correlate on a qualitative level. They concluded that global processes of the Earth are basically interrelated. Kravchinskii offered the concept of geonomic periodicity according to which all events of the same rank are essentially equivalent. This leads to the principle of conjugation which states the possibility of sufficiently accurate quantification of one event or process in terms of another. Large-scale cyclicity has been revealed in both the geomagnetic field behavior and plate motion, and three Vendian-Phanerozoic stages are recognized in the plate configuration [Khramov, 1991; Khramov et al., 1982; Zonenshain et al., 1987, 1990]: (1) Late Riphean-Vendian: continents form a supercontinent (Pangea); (2) Early-Middle Paleozoic: continents are concentrated near the equator in the southern hemisphere, and plate velocities, averaging high values, considerably fluctuate; and (3) Late Paleozoic-Cenozoic: at the beginning of this stage, continents were aligned along a meridian and formed a new Pangea; plate velocities were more uniform, reaching a minimum by the middle of the stage. Later, continents resume their E-W configuration, although by that time they are mostly located in the northern hemisphere. Similar to the geomagnetic field, their main reorganizations occurred at times of 600 and 260-300 Ma. Paleomagnetic estimates of plate velocities indicate 8-10 velocity pulses in the Phanerozoic so that each two pulses form an interval nearly coinciding with a geotectonic cycle. Chronological correlation between the velocity pulses of horizontal movements and tectonic activity in the lithosphere is evidence that these movements are the main factor of tectogenesis [Khramov and Kravchinskii, 1984]. Moreover, the time pattern of these events, explainable within the framework of plate tectonics, is consistent with actual models of plate motion based on paleomagnetic data (see e.g., [Khramov, 1991; Khramov et al., 1982; Zonenshain et al., 1987, 1990]). Common periods and correlation between the events supports the concept of a global evolutionary process and general geonomic sequence of stages.

Synchronism of processes.

Numerous data indicate the movements at the core-mantle boundary and in the lithosphere to be nearly synchronous. Thus, plate motion events at times of 42-45, 70-80, and 110-120 Ma were synchronous with reversal frequency changes; the stages of major changes in the biosphere and long- period variations in the seawater level, spreading rate, subduction and reversal frequency are very close in time (differing by no more than 10 Myr) [Aparin, 1982; Danukalov et al., 1983; Gaffin, 1987; Loper et al., 1988; McFadden and Merrill, 1986; Pechersky and Didenko, 1995; Pechersky and Nechaeva, 1988; Ricou and Gibert, 1997; Van der Voo, 1988; Vogt, 1972, 1975].

Varygin and Aparin [1989] noted a strong negative correlation of long-period variations in the seawater level with the geomagnetic reversal frequency in the Cambrian-Early Carboniferous and Late Cretaceous-Anthropogene ( Marzocchi and Mulargia [1992] confirmed this for the past 150 Myr), whereas such a correlation is absent in the Middle Carboniferous-Early Cretaceous. In the first two periods, continents were completely isolated, and the system of mid-ocean ridges was well developed, whereas during the period with no correlation a hypsometrically high supercontinent (Pangea) existed. Supposedly, variations in the sea level give rise to small variations in geoid heights, which affect the Earth's angular velocity and thereby the generation of geomagnetic field. Thus, the relation between variations in the reversal frequency and geoid heights (as a result of eustatic fluctuations and continental accretion processes) underlies the synchronism of global processes.

The Mesozoic-Cenozoic is characterized by synchronism between such events as trapp outflows, spreading rate jumps, stratigraphic unconformities in geological sections (reflecting eustatic sea level fluctuations), folding phases, appearance of evaporites and tillites (climatic changes), occurrence of black shales (redox conditions); their coincidence in time is most clear at 91-97, 110-113, 144-148, 190-196, and 245-250 Ma [Rampino, 1988; Rampino and Caldeira, 1993]. The cyclicity of all processes mentioned above exhibits periods approximately ranging from 20 to 100 Myr (most prominent are periods of ~20-30, ~50, and ~100 Myr).

The maximum entropy method was applied to the analysis of the Mesozoic periodicity in variations of the reversal frequency F, polarity asymmetry R, summary amplitude of direction variations S, paleointensity H_a, and continental drift velocity V [Pechersky and Nechaeva, 1988]. Mean ratios of neighboring periods of 1.4 ( S ), 1.67 ( H_a ), 1.5 ( F ), 1.5 ( R ), and 1.55 ( V ) are very close and similar to the ratios of short-period characteristics, which average 1.52 for two neighboring periods of secular variations and age differences of neighboring excursions [Petrova, 1989; Petrova et al., 1992]. Apart from the similarity in periods, the intervals of maximum gradients of mean drift velocities virtually coincide with summary amplitude extremums of secular variations S and maximums of the R curve [Pechersky and Nechaeva, 1988].

Synchronism of geological events that occur at the surfaces of core and Earth may be due, for example, to changes in the angular velocity and/or rotation axis angle of the Earth; such changes should affect the actual motion of the geographic pole which was as large as 10-30^o and more over the past 150-200 Myr [Andrews, 1985; Courtillot and Besse, 1987; Donn, 1989; Kerr, 1987; Sabadini and Yuen, 1989; Van Fossen and Kent, 1992], between the Late Ordovician and Late Devonian [Van der Voo, 1994], and in the Early Cambrian [Kirschvink et al., 1997] due to both movement of mantle relative to core and rotation of the whole Earth that might be caused in part by the continental drift [Keondzhyan and Monin, 1977]. Based on the general distribution of climates in the Phanerozoic, came to the conclusion on systematic changes in the orientation of the Earth (relative to its axis of rotation) and in its angular velocity, caused by changes in the planetary moments of inertia due to inner redistribution of masses.

Synchronic cyclicity of many processes in the Earth, including the geomagnetic field behavior, is close to the cyclicity inherent in the tidal evolutionary scheme of the Earth-Moon system [Avsyuk, 1986; Kiselev and Aparin, 1987] and cyclicity of impact meteorite craters, related to the half-period of oscillations of the solar system relative to the Galaxy plane. Analyzing bio- and lithostratigraphic data associated with polarity reversal intervals, Yu. I. Kats and A. I. Bereznyakov concluded in 1974 that deceleration and acceleration stages of Earth's rotation should have given rise to geomagnetic polarity reversals, with one polarity being preferable. On a larger scale, irregular tidal slowing-down of Earth's rotation (diurnal rotation time became two times longer over the period from Archean to the present time [Williams, 1994]) correlates with large intervals of a constant geomagnetic polarity; thus, a departure from the monotonic slowing-down in the Late Carboniferous-Permian almost exactly coincides with the Kiaman reversed polarity superchron. Afterward the angular velocity remained nearly constant until 75 Ma when it resumed its previous value, and the second jump in the slowing-down rate occurred at the end of the Cretaceous normal polarity Djalal superchron [Panella, 1972]. Many authors noted the cyclicity in geomagnetic field behavior close to the galactic year [Bolshakov and Solodovnikov, 1981; Irving and Pulaiah, 1976; Khramov et al., 1982; Loper et al., 1988; Negi and Tiwari, 1983; Pechersky and Didenko, 1995]. Also, one should not discard the hypothesis of precession origin of geomagnetism [Dolginov, 1977] which provides the possibility of using the huge reservoir of Earth's rotational energy.

D^'' layer and plumes.

Whereas the geomagnetic field generation is associated with processes in the liquid core with characteristic times not longer than tens of thousands of years [Merrill and McElhinny, 1983; Yanovskii, 1978], long-period variations of geomagnetic field, which are not directly related to the core, originate at the base of the mantle (boundary layer D^'' ) as a result of its interaction with the liquid core controlled by the heat and mass transfer between core and mantle. Temperature, density, and topography of the D^'' layer are laterally inhomogeneous, and the formation of plumes may be a result of episodic instability of this layer [Courtillot and Besse, 1987; Gubbins, 1987, 1989; Larson, 1991; Larson and Olson, 1991; Loper, 1991; Loper and McCartney, 1986; McFadden and Merrill, 1984, 1986; Stacey, 1992; Vogt, 1975; Zharkov et al., 1984]. The above considerations can explain the interrelation between movements in the lower mantle and surface volcanism, spreading processes (including their cyclicity), and so on. However, geochemical and seismological data indicate plumes to form mostly in the upper mantle at its boundary with the lower mantle, and the mass transfer between the upper and lower mantle is no more (in Cambrian even less) than 10% [Allegre, 1997]. Consequently, processes in the D^'' layer may account for only rare large-scale tectonomagmatic events, and it is not surprising that numerous attempts were made to associate long intervals of stable geomagnetic field of constant polarity with formation of large plumes [Courtillot and Besse, 1987; Jacobs, 1994; Larson and Olson, 1991; Loper and McCartney, 1986; McFadden and Merrill, 1984, 1996; Rampino, 1988; Vogt, 1972, 1975]. Also, long-period variations of the geomagnetic field may be related to prolonged subduction and penetration of lithospheric cold material into the lower mantle with the formation of cold anomalies at the core boundary [Eide and Torsvik, 1996]. Both types of the material exchange and energy transfer (plumes and subduction) from core to surface and vice versa imply the processes at the core and Earth's surface to be recurrent. Thus, according to various estimates, the ascent of plumes from the mantle base to surface takes 10 to 40 Myr [Courtillot and Besse, 1987; Loper, 1991; Richards et al., 1989], which is consistent with both spreading rates and continental drift velocities [Eide and Torsvik, 1996; Jurdy et al., 1995; Zonenshain et al., 1987]. Lithosphere material reaches the mantle base at even smaller rates. To explain the synchronism, Vogt [1972, 1975] supposed a very large rate of plume ascent amounting to 1-4 m/yr. Ricou and Gibert [1997] note that abrupt increases in the reversal frequency and moments of major plate reorganization stages (time lags of reversal frequency peaks average 3 Myr) occurred synchronously over the past 160 Myr, which precludes the transfer of the "thermal signal" by means of mantle convection. Both events are independent and are associated with the topography of the core-mantle boundary.

Surface traces of plumes are hotspots. They are virtually immobile with respect to moving plates and spreading of oceanic lithosphere (this fact is used for reconstruction of absolute positions and motions of lithospheric plates [Jurdy et al., 1995; Zonenshain et al., 1987]. Therefore, processes in the D^'' layer related to the formation and ascent of plumes are independent of mantle convective flows. Then, if long-period variations in the geomagnetic field are due to the processes in the D^'' layer, no correlation should be expected between plate movements and motions in the layer! For example, tectonic regimes preceding the Permian-Carboniferous Kiaman and Cretaceous Djalal superchrons of constant polarity are basically different: the former took place when continents converged to form Pangea, and the latter is associated with the time of its breakup [Eide and Torsvik, 1996].

Asynchronism of processes.

Along with the synchronism, a marked time lag of 15-60 Myr is noted between the starting moments of geological epochs and increase in the reversal frequency in the Phanerozoic [Khramov et al., 1982; Molostovsky et al., 1976], in the Paleozoic [Didenko, 1998], and throughout the Neogaea [Pechersky, 1999], as is mentioned above. This time lag is naturally associated with processes at the core-mantle boundary (D^'' layer), and their "signals" reach the surface at the velocity close to that of the continental drift, which is probably controlled by the velocities of mantle convection and plume ascent.

Thus, at least two different, external and internal, mechanisms appear to underlie processes in the Earth. The external mechanism is responsible for synchronous processes at its core and surface, and the second one is responsible for the time lag between surface and D^'' layer processes.

As is evident from the above review, rather numerous attempts have been undertaken to generalize paleomagnetic data for studying the relationships between processes in the lithosphere and at the core. My study is distinguished by the analysis of as many geomagnetic field characteristics as possible and encompasses a longer time interval, namely the whole Neogaea.

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