Removal of spurious normal polarity directions in the Kanapoi Formation with progressive thermal demagnetization: implications for dating Pliocene hominin fossils from the Turkana Basin of Kenya

The Kanapoi Formation is a 75-m-thick sedimentary unit in northern Kenya whose fossil record documents the evolution of the genus Australopithecus and habitual bipedalism in Pliocene hominins. Published 40Ar/39Ar dates of 4.20 to 4.10 Ma for three aqueously reworked tuffs from the lower and middle stratigraphic intervals of the formation and a whole-rock K-Ar date of 3.4 ± 0.1 Ma on the Kalokwanya Basalt, which caps the succession, coupled with paleomagnetic data from an unpublished thesis have been used to constrain the age of the unit (McDougall+2012 J Geol Soc). Normal polarities at the base of the formation were correlated to normal subchron C3n.1n (Cochiti Subchron) and would extend the Kanapoi chronostratigraphy back to between 4.29 and 4.18 Ma. The credibility of the normal polarities at the base and elsewhere in the formation as records of the geomagnetic field at the time of deposition, however, is called into question by the ubiquity of the high-coercivity mineral hematite in the Kanapoi sediments and the alternating field demagnetization methods exclusively used to generate the polarity directions. We tested this proposed chronostratigraphy of the Kanapoi Fm by collecting nearly 70 independently orientated block samples from 10 outcrop localities, which sampled roughly two-thirds of the total stratigraphic thickness of the formation and covered the entire vertical extent of the supposed basal normal polarity magnetozone. For most every specimen, stepwise thermal demagnetization (TD) resolved well-defined characteristic magnetizations of uniformly reverse polarity and were carried by a high-temperature component consistent with hematite. TD of the Kalokwanya Basalt confirmed its reverse magnetic polarity, which implies emplacement most likely prior to the base of normal Chron C2An (Gauss Chron) at 3.58 Ma. Kanapoi sediments thus accumulated sometime during reverse subchron C2Ar (4.18-3.58 Ma) but over a period of much less than 600 kyr. Basal age control on the formation is imprecise and complicated by the inconsistency between the radiometric date of the lowest reworked tuff (4.20 ± 0.03 Ma) and its new magnetostratigraphic context. We use these new lines of evidence to assess the chronostratigraphic data that help to implicate A. anamensis as one part of an anagenetically evolving hominin lineage.