The Geomagnetic Polarity Time Scale

4.2.1 Polarity Dating of Lava Flows 0-6 Ma

Mercanton (1926) first realized that if rocks containing reverse magnetizations were due to reversals of the Earth's magnetic field, then this should be registered in rocks worldwide and so he obtained samples from Spitsbergen, Greenland, Iceland, the Faroe Islands, Mull, Jan Mayen Land and Australia as a test. He found that some were magnetized in the same sense as the present Earth's field and others were roughly reversed from it. Matuyama (1929) observed similar effects in lavas covering the past 2 million years from Japan and Manchuria. He noticed that the reverse lavas were always older than the lavas magnetized in the same sense as the present field, which was the first suggestion of a time sequence associated with reversely magnetized rocks. Based on observations of volcanic rocks from the ChaTne des Puys in France, Roche (1951, 1956) concluded that the most recent reversal of the Earth's magnetic field took place in the middle of the Early Pleistocene. Similar observations were made by Hospers (1953, 1954) in lava sequences in Iceland, by Opdyke and Runcorn (1956) in rocks from the United States, and by Khramov (1955, 1957, 1958) in sedimentary sequences in western Turkmenia. These observations suggested that an ordered sequence of polarity inversions might exist in the geological record. This conclusion was further emphasized by the evidence that almost all rocks of Permian age are reversely magnetized (Irving and Parry, 1963).

In the above studies the assessment of the ages of magnetizations relied on the relatively imprecise methods used for dating rocks on the basis of fossil occurrences. It was only in the early 1960s that developments (by Everaden, McDougall and Dalrymple) in the K-Ar isotopic dating method enabled the dating of quite young volcanic rocks with some precision. Rutten (1959) was the first to use K-Ar dating to assess the age of magnetic polarities. He concluded that the present normal polarity had existed since at least 0.47 million years ago and that an earlier period of normal polarity existed about 2.4 million years ago.

In an attempt to define a polarity time scale, systematic studies using joint magnetic polarity and K-Ar age determinations on young lava flows were undertaken in both the United States and Australia. The first time scale put forward by Cox et al. (1963a) appeared to be consistent with a periodicity of magnetic reversals at about 1 million year intervals. However, as new data appeared in the literature (Cox et al., 1963b, 1964a; McDougall and Tarling, 1963, 1964) it rapidly became apparent that there was no simple periodicity; the lengths of successive polarity intervals varied haphazardly, some being long (~1 Myr) and others short (-0.1 Myr). Cox et al. (1964b) proposed that within intervals of predominantly one polarity lasting of the order of 1 Myr, there were short intervals of opposite polarity of the order of 0.1 Myr. The longer intervals were termed magnetic polarity epochs and the shorter intervals were called events. The epochs were named after pioneering scientists in geomagnetism (Brunhes, Matuyama, Gauss, and Gilbert) whereas the events were labeled from the location of their discovery (e.g., Jaramillo, Olduvai, Kaena, and Sidufjall). The terms chron and subchron have subsequently been officially adopted to replace the terms epochs and events (see Table 4.3). However, the terms "epoch" and "event" are still used but with somewhat different meaning. For example, the term reversal event is often used to describe the phenomenon of two reversals that occur close to each other in time, whereas the term subchron refers to the time interval of the stratigraphic record of an event (see §4.3.1).

A few of the earliest compilations, covering the years 1959-1966, of the GPTS for the past 4 million years are shown in Fig. 4.3. These early studies combined conventional K-Ar dating with measurements of magnetic polarity from widely spaced localities worldwide; they were not carried out on continuous sequences so the ordering relies on the accuracy of the K-Ar dates. In general, the evolution of this time scale has been marked by the inclusion of an increasing number of reversals. Glen (1982), in an excellent history of this evolution, notes that the classic work by Khramov (1955, 1957, 1958) on sedimentary sequences in what is now western Turkmenistan clearly influenced the early igneous rock work on polarity time scales of Cox, Doell, and Dalrymple in the United States and of McDougall and Tarling in Australia in the early 1960s.

The development of high-precision 40Ar/39Ar dating techniques has shown that many of the ages determined by the conventional K-Ar method were too young, so those ages have had to be corrected. Furthermore, as increasingly more measurements have been made, questions have arisen regarding the identification of some of the short subchrons. The problem that arises with some of these is to decide whether they represent true reversals or some other geomagnetic field behavior such as changes in intensity.

K-Ar age (millions of years)

Oct. 1955

June 1963

June 1964

Aug. 1965

May 1966

K-Ar age (millions of years)

Oct. 1955

June 1963

June 1964

Aug. 1965

May 1966

Fig. 4.3. Early evolution of the geomagnetic polarity time scale 1959-1966. Black represents normal polarity, white represents reverse polarity, and gray indicates uncertain polarity. Abstracted with permission from Cox (1969). © American Association for the Advancement of Science.

Figure 4.4 shows the present state of the GPTS for the past 6 Myr following Cande and Kent (1995). This represents the time interval during which all the polarity chrons and subchrons have been named. Cande and Kent (1995) also identified various cryptochrons (see Table 4.3), which may be due to geomagnetic behavior such as geomagnetic intensity changes that occur at 0.49, 1.20, (Cobb Mountain) and 2.42 Ma.

Although natural with the evolution of the GPTS, the development of the terminology of polarity chrons (such as the reverse Matuyama chron) containing polarity subchrons (such as the normal Jaramillo, Olduvai, and Réunion subchrons) was perhaps unfortunate. The problem is that the nomenclature suggests a reasonable interval of time (the chron) during which the polarity was biased toward one value while containing short, aberrant intervals (subchrons) of the opposite polarity. This concept of polarity bias persists in the literature today (e.g., Johnson et al., 1995; Algeo, 1996). As discussed in §4.5.4, the individual normal and reverse intervals (whether they be intervals of the chron polarity or subchrons of the opposite polarity) are equivalent, independent intervals drawn from a random process.

Age Epoch Polarity Polarity (Ma) chrons

Age Epoch Polarity Polarity (Ma) chrons

Gilbert Magnetic Epoch

Fig. 4.4. Geomagnetic polarity time scale for the past 6 Myr based mainly on 3'Ar/40Ar and paleomagnetic data on igneous rocks. Black represents normal polarity, and white represents reverse polarity. From Merrill etal. (1996).

Polarity subchrons

Fig. 4.4. Geomagnetic polarity time scale for the past 6 Myr based mainly on 3'Ar/40Ar and paleomagnetic data on igneous rocks. Black represents normal polarity, and white represents reverse polarity. From Merrill etal. (1996).

4.2.2 Geochronometry of Ocean Sediment Cores

Volcanic activity is intermittent so the study of lava successions does not produce a continuous sequence of polarity information. In contrast, continuous sequences can be obtained using ocean-bottom cores from deep-sea sediments, providing an independent method of determining the polarity time scale. The cores are not usually oriented but they are taken nearly vertically into the ocean bottom so that changes in sign of the magnetic inclination measured in the cores or changes of 180° in declination can easily be identified as records of polarity change. Deposition rates are relatively low in oceanic sediments, typically being in the range 1-10 mm per 1000 years. This means that the Matuyama-Brunhes boundary at 0.78 Ma will typically be found at a depth of between 0.78 and 7.8 m and the Gilbert-Gauss boundary at 3.58 Ma will typically be found at a depth of between 3.58 and 35.8 m. This slow sedimentation rate smooths the magnetic signal but does allow one to go further back in time for a given length of core. Oceanic sedimentation rates elsewhere (e.g., continental margins) can be much higher and provide a more detailed record of magnetic changes but over a shorter interval of time.

The earliest investigations of marine sediment cores rapidly confirmed the reality of the land-based polarity time scale (Ninkovich et al., 1966; Opdyke and Glass, 1969). Two examples from the North Pacific are given in Fig. 4.5, with the sedimentation rates differing by about 50%. In both cases there is positive (downward pointing) inclination (normal polarity) in the upper part of the core corresponding to the Brunhes chron. An abrupt change to negative (upward

Inclination (")

Was this article helpful?

+3 -1

Post a comment