Magnetism of loess/palaeosol sequences: recent developments

Investigation of the global loess record-primarily as a palaeoclimatic arch ive-continues unabated. One important aspect of this research effort is the determination of the magnetic properties of the loess itself and of the pa laeosols developed within it. In earlier work, these have provided a record of major long-term past climatic changes; however, recent studies are now revealing sub-Milankovitch signals with teleconnections from the Chinese Lo ess Plateau (CLP) to the Greenland and Antarctic ice caps as well as to cli mate-related events in sediments recovered from the North Atlantic. These h igh-resolution profiles cover to the last similar to 130,000 years; however , progress is also being made in extending the record down below the loess (sensu stricto) of the CLP into the so-called Red Clay, so that the aeolian record can now be carried back beyond 7 Ma BP. Several authors have attempted to obtain quantitative estimates of palaeopr ecipitation by comparing magnetic susceptibility of palaeosols with values for modem soils. Although the broad picture is qualitatively clear (wetter interglacials, dryer glacials), the actual quantitative values provided by the various techniques differ significantly. In the classic Chinese sites, palaeosols are always magnetically enhanced, due to in situ pedogenic processes. However, evidence is now emerging to in dicate that this is not the only magnetoclimatological model. In cases (suc h as Siberia) where wind vigour dominates over pedogenesis, the opposite pa ttern is found, with magnetic susceptibility minima in palaeosols. This als o occurs in situations where the sediments are water-logged for significant amounts of time leading to "gleying" and concomitant loss of magnetic mine rals. The use of magnetic remanence as a geomagnetic polarity clock requires caut ion. Early interpretations were acceptable for establishing broad chronolog ical control; however, recent scrutiny suggests that remanence "lock-in" ma y be delayed by as much as 20,000 years, Furthermore, synthetic magnetostra tigraphies modelled on the basis of realistic sediment characteristics indi cate that there are limits to the amount of geomagnetic detail likely to be resolved. This seems to be the case for the Blake Event, which manifests i tself in a very variable way at different sites in the CLP. (C) 2001 Elsevi er Science B.V. All rights reserved.