2004 Articles

Geomagnetic Polarity Timescales and Reversal Frequency Regimes

Lowrie, William; Kent, Dennis V.

An analysis of geomagnetic reversal history is made for the most reliable polarity timescales covering the last 160 Myr. The timescale of Cande and Kent [1995] (CK95) is the optimum representation of Cenozoic and Late Cretaceous polarity history, and the timescale of Channell et al. [1995] (CENT94) best represents the Early Cretaceous and Late Jurassic. The CK95 timescale can be divided into two nearly linear segments at Chron C12r. The lengths of chrons in the younger segment have no systematic trend, and so this part of the polarity sequence is considered to be stationary for statistical analysis. The mean chron length is 0.248 Myr and the gamma index, k, for the distribution of chron lengths is 1.6 ± 0.4; inserting just 8 additional short subchrons that have been verified from magnetostratigraphic studies as polarity reversals reduces the mean chron length to 0.219 Myr and k to 1.3 ± 0.3. The older segment is stationary if the two long polarity chrons C33n and C33r adjacent to the Cretaceous Normal Polarity Superchron are omitted; in this case the mean chron length is 0.749 Myr and k is 1.2 ± 0.4. The chrons in the CENT94 timescale are stationary with mean length 0.415 Myr and k is 1.3 ± 0.3. The gamma indices of the chron distributions are not significantly different from a Poisson distribution (k = 1), which implies that the reversal process is essentially free of long-term memory. However, if the mean chron duration is an indicator of stability of the reversal process, it appears that long lasting episodes of stable behavior may be followed by abrupt change to another stable regime with a markedly different reversal frequency. There is no significant change of the gamma index from one regime to another although the mean polarity chron length changes by more than a factor of three. This would imply that the probability of a reversal may be constant within each regime but varies inversely with mean interval length from regime to regime.