RUSSIAN JOURNAL OF EARTH SCIENCES VOL. 8, ES1004, doi:10.2205/2006ES000191, 2006

Variations in the Geomagnetic Field Direction

[7]  The total magnitude of the geomagnetic field direction variations was determined using the standard angular deviation S=81/ K1/2, where K is the precision parameter in the sphere statistics [Khramov et al., 1982]. Using the same method, we determined S values for the whole of the Neogaea [Pechersky, 1998] and for the vicinity of the Mz-Kz boundary [Pechersky and Garbuzenko, 2005].

[8]  In contrast to the other geomagnetic field characteristics, briefly reported above, the abundant data available for the magnitude of the field direction variations, both in time and space, are sufficient to analyze not only the global behavior of this field characteristic, but also some local features, relative, in particular, to the center of the Siberian traps.

[9]  In order to describe the S behavior in the vicinity of the Paleozoic-Mesozoic (P/T) boundary, I used the Paleomagnetic Database (GPMDB-2005), and chose the paleomagnetic data, ranging roughly from 200 Ma to 340 Ma in age. This choice was based on the following reasons: (a) the basic impulse of the Siberian trap activity can be dated 251 Ma, some less intensive magmatic activity was recorded later, during the Triassic [Gurevich et al., 2004; Ivanov et al., 2005]. In order to reconstruct a more complete pattern of the geomagnetic field variation, I used the time interval after the end of the Siberian trap magmatism, namely, 250 Ma to 200 Ma; (b) earlier, using the examples of the modern igneous activity of the plumes [Pechersky, 2001], and of the plumes in the vicinity of the Mz-Kz boundary [Pechersky and Garbuzenko, 2005], we found that closer to the plume epicenter the S value and its scatter increase notably, this growth being especially significant during the period of ~20-50 million years before the beginning of the high magmatic activity of the plume at the Earth surface. In this connection I used the time interval overlapping the potential time of the plume rise, namely, 270-300 Ma. For comparison, I added the time interval of 310-340 Ma, preceding the beginning of the Siberian Plume formation.

[10]  The next step was to sort out the data chosen from the paleomagnetic data base into four categories:

[11]  (1) the unreliable paleomagnetic data obtained for not more than ten samples, the thermal demagnetization was not higher than 200o C, the AF demagnetization was not higher than 15 mT, the precision parameter K<7, the confidence angle a95>25, the coordinates of the paleopole were different greatly from the average pole for the given time for the continent in question, the paleomagnetic determinations being considered as the results of the remagnetization of the older rocks. The determinations of this kind were discarded;

[12]  (2) the low reliability of the paleomagnetic determinations with the number of the samples not more than 20, the cleaning temperature lower than 400o C, the AF cleaning being lower than 30 mT. The index of the paleomagnetic reliability was 0.1 for this group of samples. This index was used as a weight value in calculating the average S values;

[13]  (3) the intermediate reliability of the paleomagnetic determinations with the number of the samples higher than 20, the thermal cleaning was performed at temperatures not lower than 500o C, the AF cleaning was not less than 50 mT. There were examples of positive geophysical tests (fold, pebble, reversal or baking tests). The index of paleomagnetic reliability was 0.5. This index was used as the weight in calculating the average S values.

[14]  (4) the high reliability of paleomagnetic determinations with the number of samples larger than 20, the obligatory complete thermal demagnetization and AF demagnetization with the component analysis, the identification of the characteristic NRM component, and the positive fold, pebble, baking, and reversal tests. The index of paleomagnetic reliability, as high as 1.0, was used as the weight in calculating the average S values.

[15]  It should be emphasized that the geophysical tests of paleomagnetic reliability, such as the fold, pebble, reversal, and baking tests, are important to prove the identification of the primary NRM component, yet, they cannot be used to prove that the S value corresponds to the variation magnitude of the geomagnetic field direction. As will be demonstrated below, the scatter of the S values is very large, primarily, because of some technical and methodological reasons, such as not complete cleaning, measurement errors, magnetic biasing in the course cleaning, the uncertainty of the age interval during the calculation of the paleomagnetic direction, and the like. For this reason we can be sure only of the average and modal S values.

2006ES000191-fig02
Figure 2
[16]  Based on the types of the rocks we classified our determinations into four groups: sedimentary, redbeds, volcanic, and intrusive rocks. Most of our data belong to the sedimentary rocks. For this reason, the results of this study are combined into groups in terms of their ages and paleomagnetic reliability. A separate group combines the data for Australia, which are characterized, in the age interval discussed, by the elevated S values (Figure 2), which are unrelated to the Siberian plume.

2006ES000191-fig03
Figure 3
[17]  The paleomagnetic determinations, which remained after our data sorting were classified into the following age groups: 340-315, 310-290, 285-270, 265-245, and 240-200 Ma. Each of these intervals was characterized by the respective maps of paleotectonic reconstructions and APW paths for all continents, mainly after [Torsvik and Van der Voo, 2002] with the additions for China and other regions after [Bretstein and Klimova, 2005; Didenko et al., 1994; McElhinny and McFadden, 2000; Pechersky and Didenko, 1995; Scotese and McKerrow, 1990; Smethurst et al., 1998]. The paleomagnetic data points available were plotted on these maps, and after that were determined their ancient coordinates. There are two models of paleotectonic reconstructions: the GAD model, where the geomagnetic field is assumed to be a dipole one and the G3 model which takes into account a nondipole, octupolar component [Torsvik and Van der Voo, 2002]. Without dwelling on the substantiation of these models, I determined the ancient coordinates of the paleomagnetic observation sites for both models. The coordinates of the Siberian traps center, reconstructed for the time of 250 Ma, were found to be 57o N, 30o E, using the GAD model, and 60o N, 45o E, using the G3 model. In both cases we determined the distances, along the great-circle arc, from each of the paleomagnetic determination site to the center of the Siberian traps. I assumed that during the igneous activity of the Siberian plume the position of its epicenter had not changed notably. In terms of clarifying the dependence of the total field direction variation on the distance of the observation site from the trap center, the difference between these two models was insignificant (Figure 3): it was usually not higher than 10o. It was only in the case of the distances to the center of the Siberian traps amounting to more than 80o, it was sometimes as large as 20o and more, in the region where the association of the Siberian plume and the field direction variation was hardly probable. Therefore it is enough to consider only one model, in this paper we will dwell on the G3 model.

[18]  In determining the coordinates of the paleomagnetic observation site, we used two factors: the paleolatitude, calculated from the paleomagnetic inclination of the given site, and the position of the site in the map of the "suitable" age reconstruction. The fact is that in many cases the Database offers a fairly broad range of rock and paleomagnetic determinations ages. For this reason, I selected a map where the position of the point in question corresponded best of all to its paleolatitude determined using the paleomagnetic inclination. Also, we took into account the agreement of the polarity of the given paleomagnetic determination with the time-scale of geomagnetic polarity [Gradstein et al., 2004], within a five-million year averaging.

2006ES000191-fig04
Figure 4
[19]  The S value is known to vary with the latitude: it diminishes roughly by two times slowly from the equator to the pole. This is valid for the intervals of the steady-state geomagnetic field throughout the Neogaea [Pechersky, 1996]. According to the dependence of the S values with the latitude, all of the S determinations were reduced to one paleolatitude, namely, to the latitude of the pole ( S p ). Figure 4a clearly shows the S dependence on the paleolatitude of the observation site, which vanishes after its reduction to the latitude of the pole (Figure 4b).

2006ES000191-fig05
Figure 5
[20]  Depending on the model used (GAD or G3), the paleolatitudes of the observation points were found to be somewhat different and, hence, the S p values were found to be different, too. This difference is illustrated in Figure 5, which shows that in most cases the difference between the S p values of both models is not higher than 1-2o.

2006ES000191-fig06
Figure 6
2006ES000191-fig07
Figure 7
2006ES000191-fig08
Figure 8
[21]  All of our paleomagnetic determinations were distributed, with some or other accuracy, over five-million year intervals (Figure 6a). For each of the five-million year interval I calculated its average value and performed a smoothing operation for the time intervals of 10 million years, the smoothing step being 5 million years (Figure 6b). The resulting pattern (Figure 6b) shows the behavior of the variation direction magnitude of the "normal" geomagnetic field, where the scatters of different origin are smoothed off. This Figure shows that the direction of the "normal" field varies insignificantly, even without smoothing from 6o to 10o, the changes being minimal in the Kiama hyperchron during the stable state of the geomagnetic field, and more notable in the time interval of 320-300 Ma, which preceded the Kiama hyperchron (Figures 1, 6b). The P/T boundary is not expressed in the S p behavior, being restricted to the region of some S p value growth which began at the time of 260 Ma (Figure 6b). Against a background of the "normal" field there is some tendency toward the S p growth in the direction closer to the center of the Siberian traps by a distance less than 40o (Figure 7). This trend was also recorded in the summarized S p distribution histogram (Figure 8). This Figure clearly shows the main distribution with the 8o mode, against the background of which clearly shows up the second group of the higher S p values (10-13o) with a mode of sim11o, related mainly to the sites located closer than 40o to the center of the Siberian traps.

[22]  Discussed below in more detail is the behavior of the variation of the geomagnetic field direction during the main time intervals relative to the P/T boundary and to the center of the Siberian trap, which is believed to have been the epicenter of the plume. It should be emphasized that the distance from the plume epicenter to the region of its origin at the core-mantle boundary was about 3000 km, and its rising path being often inclined [Ernst and Buchan, 2003]. Combined with the insufficient accuracy of the paleomagnetic direction determinations and reconstructions, this fact might have caused the significant scatter of the S p values. It is pertinent to remind that each of these time intervals embraces 15 to 40 million years. Consequently, with the above mentioned uncertainty of the distance to the source of the plume and the long period of time compared to the time of the unstable state at the core-mantle boundary and, hence, the unstable state of the geomagnetic field, the conditions closer to the plume epicenter must be reflected, first, in the growth of the variation magnitude, that is, in the S p value and, secondly, in the growth of the S p value scatter. Moreover, the S p distribution relative to the Siberian plume epicenter could be disturbed notably by the other plumes of different ages, which might have been active during the time period discussed. Apart from the reasons mentioned above, one should not forget technical errors, the inaccuracies of the measurements, the "purity" of identifying the primary NRM components (the presence of stable secondary components), and the like. All of them augmenting the scatter of the S p unit values, the mean and modal values being more objective quantitative characteristics.

2006ES000191-fig09
Figure 9
2006ES000191-fig10
Figure 10
[23]  (A) 340-315 Ma. The distribution of the S p is characterized by the predominance of low values ranging from 4o to 8o with a distinct mode of 8o (Figure 9e), corresponding to the "normal" field in this time interval (Figure 6b). The distribution of the S p values is not controlled by the distance to the center of the Siberian traps (Figure 10).

2006ES000191-fig11
Figure 11
[24]  (B) 310-290 Ma. The S p values show a bimodal distribution (Figure 9d). The first 7o mode corresponds to the "normal" field in this time interval (Figure 6b). The second 10o mode depend on the distance to the center of the Siberian traps, which is indicated clearly by the average S p values (Figure 11). The largest S p values were found for the time of 290-300 Ma.

2006ES000191-fig12
Figure 12
[25]  (C) 285-270 Ma. The S p values show a bimodal distribution, including one distinct mode of 8o and a less distinct mode of 11o (Figure 9c). The first mode corresponds to the "normal" field, the second mode being associated with the S p growth closer to the center of the Siberian traps, as seen from the behavior of the average S p values (Figure 12). The highest S p values were recorded for the time interval of 270-280 Ma.

2006ES000191-fig13
Figure 13
2006ES000191-fig14
Figure 14
[26]  (D) 265-245 Ma. The S p values showed a unimodal distribution with a mode of 8o, which corresponds to the "normal" field, the mode associated with the Siberian Plume is absent (Figure 9b). This was confirmed by the histograms of the unit S p data selection obtained for the data points residing not far than 30o from the center of the Siberian traps (Figure 13a) and by the selection of the Siberian traps alone (Figure 13b). Some S p raise at the distance of less than 20o (Figure 14) is proved by a mere couple of reliable data points and averagely does not exceed the "normal" field (Figure 6b).

2006ES000191-fig15
Figure 15
[27]  (E) 240-200 Ma. This time interval is characterized by the S p unimodal distribution (Figure 9a) with a mode of 8o, which corresponds to the "normal" field, its somewhat unstable distribution being obviously associated with the variation of the "normal" field at that time (Figure 6b). As could be expected, the S p does not show any connection with the distance to the Siberian traps (Figure 15).


RJES

Citation: Pechersky, D. M. (2006), Geomagnetic field in the vicinity of the Paleozoic-Mesozoic boundary and the Siberian superplume, Russ. J. Earth Sci., 8, ES1004, doi:10.2205/2006ES000191.

Copyright 2006 by the Russian Journal of Earth Sciences

Powered by TeXWeb (Win32, v.2.0).