On the construction of geomagnetic timescales from non‐prejudicial treatment of magnetic anomaly data from multiple ridges

SUMMARY When marine magnetic‐anomaly data are used to construct geomagnetic polarity timescales, the usual assumption of a smooth spreading‐rate function at one seafloor spreading ridge forces much more erratic rate functions at other ridges. To eliminate this problem, we propose a formalism for the timescale problem that penalizes non‐smooth spreading behaviour equally for all ridges. Specifically, we establish a non‐linear Lagrange multiplier optimization problem for finding the timescale that (1) agrees with known chron ages and with anomaly‐interval distance data from multiple ridges and (2) allows the rate functions for each ridge to be as nearly constant as possible, according to a cumulative penalty function. The method is applied to a synthetic data set reconstructed from the timescale and rate functions for seven ridges, derived by Cande & Kent (1992) under the assumption of smooth spreading in the South Atlantic. We find that only modest changes in the timescale (less than 5 per cent for each reversal) are needed if no one ridge is singled out for the preferential assumption of smoothness. Future implementation of this non‐prejudicial treatment of spreading‐rate data from multiple ridges to large anomaly‐distance data sets should lead to the next incremental improvement to the pre‐Quaternary geomagnetic polarity timescale, as well as allow a more accurate assessment of global and local changes in seafloor spreading rates over time.