Geoscience Research Institute


Norm Smith
Ph.D. in Math and Statistics


Jane Smith
Research Associate
7129-C Rock Ridge Lane
Alexandria, Virginia  22310

Origins 20(1):6-21 (1993).


The oceanic patterns of magnetic anomalies are thought to be one significant evidence supporting the general notions of plate tectonics. Arguments are presented suggesting that important characteristics of the anomaly patterns could be due to the measurement process itself rather than being a direct reflection of geomagnetic reversals and plate movement, as is usually claimed. While it is certainly true that this reinterpretation does not disprove the popular explanation of anomaly pattern formation, it does open the door to exploration of various alternative explanations.


    Over the last several decades, a general theory of large-scale geophysical behavior of the earth, known as plate tectonics, has gained wide acceptance. This theory has been able to provide a unified explanation of a wide variety of geophysical phenomena, doing so within the world view held by the majority of the academic community. The general theory is an integration of several concepts which fit consistently together (see for example, Uyeda 1978, and Takeuchi, Uyeda and Kanamori 1970). This article concerns itself with only one of these, the origins of the oceanic magnetic anomalies. Before describing the motivation for and the main thought of this article, we will give a brief introduction to the general theory of plate tectonics. This introduction will focus on four of the prominent concepts in the general theory.
    The concept which first gained prominence was that of continental drift. Although it is often associated with its most active proponent, Alfred Wegener, anyone looking at a world map could observe that if the American continents could somehow slide up against Europe and Africa, the coastlines would fit together remarkably well. According to the continental-drift concept, the continents originated from the breakup of one large land mass (presumably a couple hundred million years ago) and subsequently, slowly "drifted" to their present positions. In the early part of this century, Wegener and his associates marshaled a great deal of evidence supporting this idea. Although the concept was widely discussed, interest later waned, mainly because of a lack of independent evidence which would corroborate continental-plate motion. Toward the middle of the century, several areas of investigation produced data that revived interest.
    The thought that continents in the past were arranged differently is of interest in itself, but the concepts of plate tectonics in general, and continental drift in particular, play a central role in modern geophysics far beyond mere historical interest. The motion and collision of the continental segments provide a rather unified explanation, either directly or indirectly, for most geological features. For example, the Himalayas are thought to have arisen from the collision of India with Asia.
    One area of investigation that produced data which revived interest in continental drift is continental paleomagnetism. This data collection began in the 1950s. Samples from rock formations, most notably those taken from bedrock, generally have a low-level "remanent" magnetism. The orientation of this remanent magnetism is thought to indicate the orientation of Earth's magnetic field at the time of emplacement of the rock, specifically, the orientation at the time of cooling. Sample orientations from local deposits are usually averaged to obtain a magnetic-pole position for that formation. Averaged pole positions obtained from rocks of roughly similar geologic age in a given region tend to be similar, except for one modification. That modification is a 180 flip in the polarity indicated for about half of the averaged pole positions. From these observations has come the notion of a historical sequence of polarity reversals of Earth's magnetic field. The reversals are thought to occur randomly at an average rate of a few per million years. [However, there are indications that these reversals can occur in a few months (Brown 1989).]
    A plot of a continent's indicated pole position as a function of geologic age is useful in reconstructing past continental motion. However, the concept of an historical sequence of geomagnetic reversals, not geomagnetic-pole positions, is of most significance to the topic of this article.
    The notion of seafloor spreading played an important part in the wide acceptance of the general plate-tectonics concept. Seafloor spreading provides a reasonable mechanism for continental drift.
    The crust of the earth can be thought of as divided into several relatively rigid plates slowly moving relative to one another. The boundaries of these plates correspond to a network of faults, oceanic trenches and mid-ocean ridges. The relative plate motions are believed to be driven by patterns of circulation deep in the underlying semifluid mantle. This motion is analogous to the motion of various segments of scum on the surface of a pan of syrup that is heated enough to cause circulation, but short of the boiling point. The scum stays on the surface but is moved about by the circulation of the syrup. The mantle circulation is thought to rise under the mid-ocean ridges and then to flow away from the ridges in both directions. As the crust is carried away from a ridge in both directions by riding on this circulation, new crust is formed at the ridge by mantle material cooling in the "fissure" — hence the phrase sea-floor spreading. As new crust is formed at the mid-ocean ridges, old crust must be destroyed elsewhere. In the plate-tectonics theory, this destruction occurs at a downturn of the crust into the mantle at subduction zones generally corresponding to oceanic trenches. Lateral motion of plate edges forms transverse faults that complete the boundary network. Continental drift occurs as the continents are carried along as parts of the moving plates.
    Of special interest to our discussion is the formation of the "oceanic magnetic anomalies." The currently accepted interpretation of these magnetic anomalies provides a wealth of detail for refining the reconstruction of past plate movements, and for many it provides the "clinching" argument in support of the general plate tectonics theory.
    Fairly detailed surveys of the intensity of the magnetic field at the ocean surface have been made over most of the world's oceans. This magnetic intensity shows an interesting pattern of variation about its local average value. These variations are often called magnetic anomalies. When maps are made showing regions of above average and below average intensity, the regions form a striped pattern as illustrated in Figure 1. The stripes are usually a few tens of kilometers wide. Most interestingly, the stripes usually are roughly parallel to the nearest ocean ridge.

FIGURE 1. Magnetic anomaly patterns in the East Pacific off Vancouver Island (a) and southwest of Iceland (b). Positive anomalies are shown in black. After Raff and Mason (1961) and Vine (1966) (a), reproduced with permission of the publishers; and after Heirtzler, Le Pichon and Baron (1966) (b).

    Vine and Matthews (1963) provided the "ah ha" insight that fit these oceanic magnetic anomalies into the general scheme of plate tectonics. They observed that as new crust cools at the ridge, it would take on the magnetic orientation and polarity of the then-current geomagnetic field. By this new crust moving gradually away from the ridge, a band of crust of this polarity is formed adjacent to the ridge. When the geomagnetic polarity reverses, a new band of opposite polarity is started. As time goes on stripes of alternating polarity move away from the ridge center. In this way the ocean floor becomes a sort of "tape recording" of the historical sequence of geomagnetic polarity reversals. One could expect a mild disturbance of the striped patterns due to faulting, and such is observed.
    The Vine and Matthews explanation of oceanic magnetic anomalies is one of the major supports for the notion of seafloor spreading. However, even if the concept of large-magnitude seafloor spreading is accepted on the basis of other evidence, the interpretation of the magnetic anomalies is of considerable significance. The manner and global coordination of seafloor spreading would be only vaguely conceptualized without the magnetic anomaly data. The point of the Vine and Matthews explanation is not just that geomagnetic reversals have somehow been involved in the formation of the magnetic anomalies, but that there is a rather precise one-to-one correspondence between the magnetic anomaly sequence and a historical sequence of geomagnetic reversals. Because of the global nature of the geomagnetic reversals, acceptance of this correspondence provides a wealth of relatively precise detail on the global coordination of seafloor spreading and the manner in which it occurred. For example, one would know that the thirtieth anomaly bands away from ridge centers were at one time being simultaneously formed at the ridges. Thus, if accurate age data were available, a plot of continental position versus time could be constructed. If this picture of spreading history is accepted as established fact, the possible scenarios for geophysical history are considerably restricted.
    Although there are other notions and a great amount of further elaboration involved, the above concepts are the "core" of the plate-tectonics theory.


    This article will suggest, and to a degree explore, an alternative to the generally accepted explanation for the origin of the oceanic magnetic anomalies. We do not claim to show that the generally accepted explanation is false. Therefore, additional discussion of the motivation for this article (alluded to in the Introduction) is provided here.
    To a person who is exploring general theories of origins which are alternatives to those popularly accepted in academic circles, it is important not to unduly restrict the range of possible explanations open for consideration in each of the many areas which a general theory must address. In a given area of study, one usually finds a group of alternatives which the physical evidence itself would eliminate from further consideration; however, there often remains a range of possibilities for each of which the physical evidence is not conclusive pro or con. To those interested only in theories fitting within the conventionally accepted world view, many of these remaining possibilities may be dropped from consideration. But such possibilities may actually be essential elements of some alternative theory of origins.
    Such is the case discussed in this article. The claim is made here that an alternative against which current physical evidence is not conclusive has not been adequately explored. The authors suggest that acceptance of the conventional explanation as proven fact, as seems to be the case in most recent treatments of the subject, is at present an unwarranted restriction of the possibilities open for serious consideration.


    In the general literature, little attempt is made to distinguish between the case for the conventional explanation of the anomalies and that for the general concept of seafloor spreading. Because the claim of this article is not directly related to seafloor spreading, care will be taken to distinguish between the evidence related to seafloor spreading and the evidence related to origins of magnetic anomalies.
    It would seem that the primary reason for the wide acceptance of the conventional explanation of the origin of the oceanic magnetic anomalies, is that it nicely ties together the notions of seafloor spreading and a historical sequence of geomagnetic reversals. There is, however, also other evidence for seafloor spreading. Among these are the remarkable (yet not precise) fit of the continental shelf boundaries; the layering pattern of oceanic sediments; a degree of symmetry of geologic features across ocean ridges; a pronounced parallelism of oceanic features to the ocean ridges; the plate movements detected by intercontinental interferometry; the geophysical characteristics of transverse faults and oceanic trenches and the coincidence of the pattern of earthquake epicenters with the global network of ridges, transverse faults and trenches.
    It is conceivable that the most compelling evidence for the conventional explanation of the anomalies could come from a direct examination of the magnetic orientation of the oceanic basement rock beneath the sedimentary layers. (It is generally agreed that the origin of the anomalous effects must be in the upper few kilometers of the basement.) What if the basement rock of each anomaly band were widely sampled and directly found to have a uniform remanent magnetic polarity and that adjacent bands were of opposite polarity? What if in addition, the only geologic feature distinguishing adjacent anomaly bands were shown to be this magnetic polarity? If such were the case, the conventional explanation of sequential formation of the anomaly bands in one-to-one correspondence with geomagnetic reversals, would be a strong contender indeed. Unfortunately, as one might expect, such information is not currently available. The Deep Sea Drilling Project has drilled the ocean floor in many locations but generally these holes are confined to the sedimentary layers (see Hall and Robinson 1979, Hall and Ryall 1977, Smith 1985, and Smith and Banerjee 1986). Only three or four holes have penetrated the basement to any great degree. These few holes have shown a wide scatter of magnetic inclinations within each hole and have not demonstrated lateral continuity. In the real world one would not expect the simplistic uniformity pictured at the beginning of this paragraph, yet one might hope for considerably greater uniformity than has been observed. While explanations can be found for the lack of uniformity, one might contend that direct examination of basement rock magnetic orientation has not currently provided strong support for the conventional explanation of magnetic anomalies.
    Close examination of seafloor magnetic orientation has also been done by using deep towed sensors and to a very limited extent by submersibles such as ALVIN. The results have been equivocal and are summarized in another manuscript (Smith and Smith 1993; see also MacDonald et al. 1983; Luyendyk 1969; and Klitgord et al. 1975).
    Another area from which one might expect "direct" evidence relating to the interpretation of magnetic anomalies is the correspondence of the anomalies with independently determined geologic ages. Correspondence might be expected in two forms if the conventional explanation is correct. First, one would expect independently determined ages for the anomaly bands to be consistent with each other globally. Second, one would expect such ages determined for anomaly bands to be consistent with the ages determined for the corresponding geomagnetic reversals from continental paleomagnetic studies.
    There is a scarcity of studies addressing the degree of such correspondence. A lack of interest in such studies would be understandable since the conventional explanation of anomalies is usually regarded as established fact. Relevant data are not abundant (relative to the task) since usable age determinations need to be taken from cores taken down to the basement rock. Regarding global age consistency of the anomaly bands, one might presume there is at least an approximate consistency since the literature is not replete with cases to the contrary. However, is the correspondence close enough to justify full confidence? If one grants that seafloor spreading has occurred, one could expect an approximate correspondence of age with distance from a ridge. Since the anomalies have a fairly even spatial distribution, one could expect an approximate consistency of anomalies with age, whether or not geomagnetic reversals were involved. It would require a rather tight correspondence to provide evidence of some other global factor as the cause, such as the geomagnetic reversal sequence.
    Consistency of the oceanic anomaly reversal dates with reversal dates determined by continental paleomagnetism was addressed in an early paper by Heirtzler et al. (1968). A practical difficulty arises since continental samples are not sequenced in time by some independent process as could be the case with seafloor spreading. Continental reversal ages cannot be determined past about five million radioisotope years, the age at which the reversal durations are about equal to age resolution. Due to this limitation, Heirtzler et al. (1968) had to extrapolate by an order of magnitude using an assumed uniform spreading rate. That the correspondence came within a factor of two would seem strongly in favor of the conventional explanation of magnetic anomalies. That it was not even better might be mildly against it.
    An additional area of direct evidence relating to the interpretation of oceanic magnetic anomalies involves widespread comparisons of magnetic anomaly patterns. This is the area perhaps most often cited as yielding support for the conventional explanation of Vine and Matthews. If the formation of oceanic magnetic anomalies is indeed due to the global effect of geomagnetic reversals, one would expect a global similarity of the anomaly patterns produced. One would also expect a symmetric similarity from opposite sides of a ridge. The matching of anomaly patterns is called correlation.
    As is usually the case, the comparison is corrupted by other considerations. Binary anomaly patterns such as shown in Figure 1 are developed from intensity traces such as those in Figures 2 and 3. One of the most objective methods for generating the binary pattern is to compare the intensity to some threshold (usually average) value. Still, it is possible for small changes in the threshold used to cause dramatic changes in the binary sequence in areas of low intensity variation. Also, it is generally agreed that some local expansions and compressions should be allowed to accommodate local variation in the spreading rate. Allowances for such effects could in itself allow binary sequences to appear quite similar. For correlation comparisons, it is better to use the intensity traces.

FIGURE 2. Magnetic profiles from the South Atlantic east of the mid-Atlantic ridge and a profile from the North Pacific. Vertical scale in gammas. From Dickson, Pitman III, and Heirtzler (1968).

FIGURE 3. Magnetic profiles from the Pacific. From Pitman, Herron and Heirtzler (1968).

    A glance at Figures 2 and 3 reveals that any correlation of intensity traces is not exact (see also Brozena 1986). This should not be surprising even if the conventional explanation is true, since this is real world geologic data that is often corrupted by variation introduced from a multitude of sources. In this case the main source could be the intermittent nature of crust formation.
    Maps of oceanic geologic features other than magnetism also reveal a striking orientation parallel to the ridges. Faults, dikes and linear volcanoes roughly parallel to the ridge are thought to be common. That such features could contribute to the similarity of traces crossing a ridge separated by considerable distances would not seem an unreasonable possibility. Also one can observe a degree of symmetry of geologic features across the ridges. Obviously, the extent of such symmetry, especially below the floor, is not clearly known. The degree to which such symmetry and parallelism increase the similarity between traces globally and across ridges is surely open to conjecture, further clouding the conclusions to be drawn from trace correlation.
    In Figures 2 and 3 the correlations are indicated by the vertical lines between traces. These figures are rather representative of traces near to the ridges. Additional traces are shown in the companion article (Smith and Smith 1993). Generally it can be noted that the shapes of individual peaks (and valleys) are of little help in correlations. Sudden changes in overall amplitude that extend for several peaks and individual isolated large peaks are easier to match. At times groups of peaks have a similar spacing. Generally speaking, it appears that correlation is better in the Pacific, Figure 3, than in the Atlantic, Figure 2, and it appears that the gross features of the traces are better correlated than the individual peaks. The remarkable symmetry of trace EL19N, Figure 3 (compared with EL19N reversed), across the ridge appears more the exception than the rule.
    There is a question as to how close such correlations must be to provide strong evidence that the conventional explanation of the anomalies is correct. The evaluation of such correlations and the assessment of their implications is a subjective matter. The simulated correlations shown later are intended to assist in the evaluation of degrees of correlation in a comparison of the conventional explanation with the alternative explanation to be suggested shortly.
    In summary, there appears to be nothing in the directly related evidence that would preclude acceptance of the conventional Vine and Matthews explanation of the origin of the oceanic magnetic anomalies. That the evidence permits the conventional explanation is, of course, significant. It is the subjective view of these writers, however, that the direct evidence is less than compelling.


    As explained earlier, the motivation of this article is to help delineate the widest range of possible explanations that are not directly precluded by the physical evidence. This attitude prompts a search for alternative explanations. Two observations have led to the alternative suggested below.
    The first observation is that the appearance of the oceanic magnetic intensity traces is quite similar to what in harmonic analysis is called band limited noise. In other words, the traces look like the result of some system that is varying randomly but is constrained to vary neither too slowly or too quickly with distance.
    The second observation is related. One can think of several possible causes of spatial variation in magnetic intensity as one moves away from the ridge. What could give such variations their observed appearance of moderate regularity? It is known that a fairly irregular fine-grained variation can be made to appear coarser and more regularly varying by averaging over local regions. [In harmonic analysis talk, this is saying that broad band noise can be made band limited by applying a moving average filter (see Lee 1960 and Bracewell 1978).] If this could be the case, what could be causing the averaging? The clear candidate here is the measurement process itself. The intensity measurements are (generally) made from surface vessels which travel several kilometers above the ocean floor, and more above the basement rock. The intensity measured at any surface point is thus actually the (weighted) average effect of the intensities in the basement rock for a horizontal distance of several times the vertical distance to the measuring device. This is indeed about the width of the anomaly peaks in the intensity traces. Is this a coincidence? If so, perhaps it is as interesting a "coincidence" as the "within a factor of two" match between observed anomaly geologic ages and the extrapolation from continental data mentioned earlier as evidence for the conventional explanation.
    The alternative explanation suggested here is, then, that the observed oceanic magnetic anomaly surface traces are the result of the measurement "filter" applied to an in-part finer variation not in one-to-one correspondence with a historical sequence of geomagnetic reversals.
    The question that immediately comes to mind regarding this alternative explanation is what could cause the pronounced orientation parallel to a ridge. The recognizable anomaly stripes are a few tens of kilometers wide but are many tens even hundreds of kilometers long parallel to a ridge. The stresses and strains associated with a ridge appear to be oriented to produce geophysical features parallel to a ridge. Could these forces produce faults and dikes many tens even hundreds of kilometers long parallel to a ridge yet separated by only a few kilometers perpendicular to a ridge? The answer is not clearly known. While the matter certainly invites further investigations, at present it would appear to be an open possibility.
    Another question might come up of whether this suggestion has enough substance to deserve being dubbed an alternative hypothesis. It does explain the observed behavior in terms of the physical measurement process but it also resorts to random variation of speculative origin. Speculation is, however, not uncommon in this area of study. Perhaps what it is called is not so important as that it gives insight for another possibility that invites further investigation.


    Two questions prompt the simulation described in this section. The first is how similar in general appearance to the observed intensity traces, are traces produced by the measurement process acting upon randomly generated intensity variation of primarily finer grain. The second is whether the correlations observed between widely separated observed traces are impressively better than are the correlations between randomly generated and averaged traces. Both questions address whether the alternative described above could be a reasonable explanation for the oceanic magnetic anomalies. The answer to the second question will give a rough idea of how good correlations would need to be in order to be impressive evidence favoring the conventional explanation.
    Considerable ugly detail about this simulation is presented in the companion article. It will not be repeated here. The simulation first generated a sequence of magnetic polarity variations of finer grain (i.e., spacing) than the grain of the magnetic anomalies. Variations of grain size and the nature of the random generating process did not make much difference in the appearance of the resulting traces. The fine variation was then filtered by a model of the measurement process. Again reasonable variations in the measurement model made no drastic difference in the appearance of the results. (The variation in the modeled ocean depth was an exception, of course.) Four traces were simulated using different seeds for the random generator to obtain Figure 4. In generating this figure, large-scale amplitude changes (between "regions") were modeled using nine arbitrary amplitude steps common to all four traces; these are fairly clear in the figure. Other examples are available in the companion article.

FIGURE 4. Illustration of degree of apparent correlation between randomly generated 800 km traces with a few overall amplitude step changes.

    It should be emphasized that this simulation is not an attempt to fit any particular observed trace by applying the measurement model to a sequence of blocked variations adjusted to obtain a fit. This exercise has been done repeatedly in the literature with coarse blocks as input; it could certainly be done more easily with fine. Here the input was randomly generated and not adjusted to obtain a fit. The fitting type of exercise is appropriately done by one who already accepts the conventional hypothesis, to obtain detailed information on the underlying reversal sequence. The simulation here was done to illustrate general behavior.
    The few correlations drawn in Figure 4 were selected visually. One would suppose that there is a heavy visual component also in the selection of correlations shown in the literature. Individual peak correlations were not drawn in Figure 4 in order to mimic the appearance of Figures 2 and 3.
    Evaluation of the questions at the beginning of this section is, of course, very subjective; here is the judgment of these writers. First, the general appearance of the simulated traces is quite similar to the observed traces. The peaks have about the same degree of regularity in both. Degree of variation in peak shape is similar in both. Second, with the exception of the EL19N example and its reverse in Figure 3, the correlations between the random simulated traces are of about the same quality as those between the observed traces. In both, individual peak shape is of little use in correlation. In both, to achieve a one-to-one peak correspondence, the peak counts must be fudged a little by counting minor peaks where necessary or by lumping some somewhat distinct peaks into one. In both, there are regions of little obvious similarity. In both, pronounced correlations between peak groups "fizzle out" upon passing to yet other traces.
    From these evaluations, these writers would draw the following subjective conclusions. As far as is revealed by this simulation, the alternative explanation described earlier is a viable possible explanation of the origin of oceanic magnetic anomalies. With isolated exceptions, the correlations drawn between observed magnetic traces are not of sufficient quality to constitute a strong support for the conventional Vine and Matthews explanation of the origin of oceanic magnetic anomalies.


    No claim has been made in this article that either the "plate tectonics theory" in general or the Vine and Matthews explanation in particular has been shown to be false. The admittedly subjective evaluation has been made here that the presently documented direct evidence relating to the Vine and Matthews explanation of oceanic magnetic anomalies is equivocal. If the jury is not still out, it would seem that it should be.




All contents copyright Geoscience Research Institute. All rights reserved.
Send comments and questions to

| Home | News |
| About Us | Contact Us |