Abstract:
Constraints on the timing of geomagnetic polarity reversals have come mainly from K-Ar, or more recently40Ar39Ar , age determinations of lavas or their K-rich phenocrysts that erupted prior or subsequent to particular geomagnetic events. We have obtained 40Ar39Ar isochron ages using incremental heating techniques on groundmass separates, phenocryst-poor whole rock samples, or plagioclase, from eight basaltic to andesitic lavas that erupted during the Matuyama-Brunhes (M-B) polarity transition at four geographically dispersed sites. These eight lavas range from 784.6 +/- 7.1 ka to 770.8 +/- 5.2 ka (1σ errors); the weighted mean, 778.7 +/- 1.9 ka, gives a high-precision age that is remarkably consistent with revised astronomical age estimates for the M-B polarity transition [6,12,13]. Despite uncertainties in absolute calibration of40Ar39Ar ages relative to mineral standards used as neutron fluence monitors, our age determinations are consistent with five other 40Ar39Ar studies focused on the M-B transition. These results confirm that the earlier K-Ar based geomagnetic polarity time scale underestimated the age of the M-B reversal by about 6%. None of the eight isochron ages are distinguishable from one another at the 95% confidence level. However, we are tantalizingly close to testing for the duration of this reversal. One lava at the base of a sequence of transitionally magnetized flows in Chile and the uppermost lava in a similar sequence on Maui are only just indistinguishable in age at the 95% confidence level and preserve different magnetic orientations. We suggest that the # 12 kyr difference in age represents an upper limit for the duration of the reversal and is similar to the period of low magnetic field intensity associated with records of the M-B reversal from deep sea sediment cores. Together with the short duration (# 2 kyr) of the directional reversal observed in several different marine sediment sections, our data suggest that reversal of the field's direction could have occurred at slightly different times depending on the position of the recording site.