2006 Philadelphia Annual Meeting (22–25 October 2006)

Paper No. 2
Presentation Time: 8:15 AM

LOESS MAGNETOCLIMATOLOGY


EVANS, Michael Edwin, Institute for Geophysical Research, University of Alberta, Edmonton, AB T6G2J1, evans@phys.ualberta.ca

Aeolian sediments in north-central China (the Chinese Loess Plateau, CLP) have been widely exploited as a rich source of magnetoclimatological information. A major result has been the emergence of a pedological model in which the neoformation of magnetic minerals during interglacial periods leads to an increase in magnetic susceptibility, typically by a factor of two or more. Some authors have proposed transfer functions that attempt to convert the observed susceptibility variations directly into quantitative estimates of paleoprecipitation, but there is no general agreement on the resulting values. The pedological model is by no means restricted to the CLP; it applies equally to most loess/paleosol sections falling in a broad, elongated band stretching through central Asia to eastern Europe. But it is not universal. At several important sites, the complete opposite is found, with magnetic peaks in glacial intervals. This alternative pattern has been attributed to stronger average wind vigour during glacial periods leading to increased entrainment of dense iron-oxide particles. In the Ob River drainage, southern Siberia, this latter model tracks the last glacial-interglacial cycle and also captures the signatures of the abrupt cold pulses responsible for the Heinrich layers in North Atlantic marine sediments. Although the broad picture of loess magnetoclimatology is now clear, many problems remain, in particular the derivation of an accurate chronology. The recognition in China (some 20 years ago) of the main geomagnetic polarity reversals re-invigorated loess research, but closer inspection revealed important mismatches with the chronology of the marine oxygen isotope stages. Moreover, when detailed sampling is carried out across reversal boundaries, instead of a single transition it is often found that multiple swings back and forth are observed. Such complexities are currently being modeled in terms of variable lock-in processes that inevitably lead to stratigraphic offsets and ultimately limit the temporal resolution of loess records.