Abstract. Paleomagnetic and environmental magnetic studies are commonly conducted on samples containing mixtures of magnetic minerals and/or grain sizes. Major hysteresis loops are routinely used to provide information about variations in magnetic mineralogy and grain size. Standard hysteresis parameters, however, provide a measure of the bulk magnetic properties, rather than enabling discrimination between the magnetic components that contribute to the magnetization of a sample. By contrast, first-order reversal curve (FORC) diagrams, which we describe here, can be used to identify and discriminate between the different components in a mixed magnetic mineral assemblage. We use magnetization data from a class of partial hysteresis curves known as first-order reversal curves (FORCs) and transform the data into contour plots (FORC diagrams) of a two-dimensional distribution function. The FORC distribution provides information about particle switching fields and local interaction fields for the assemblage of magnetic particles within a sample. Superparamagnetic, single-domain, and multidomain grains, as well as magnetostatic interactions, all produce characteristic and distinct manifestations on a FORC diagram. Our results indicate that FORC diagrams can be used to characterize a wide range of natural samples and that they provide more detailed information about the magnetic particles in a sample than standard interpretational schemes which employ hysteresis data. It will be necessary to further develop the technique to enable a more quantitative interpretation of magnetic assemblages; however, even qualitative interpretation of FORC diagrams removes many of'the ambiguities that are inherent to hysteresis data.

We demonstrate a powerful and practical method of characterizing interactions in fine magnetic particle systems utilizing a class of hysteresis curves known as first order reversal curves. This method is tested on samples of highly dispersed magnetic particles, where it leads to a more detailed understanding of interactions than has previously been possible. In a quantitative comparison between this method and the ␦M method, which is based on the Wohlfarth relation, our method provides a more precise measure of the strength of the interactions. Our method also has the advantage that it can be used to decouple the effects of the mean interaction field from the effects of local interaction field variance.

Research on global ice-volume changes during Pleistocene glacial cycles is hindered by a lack of detailed sea-level records for time intervals older than the last interglacial. Here we present the first robustly dated, continuous and highly resolved records of Red Sea sea level and rates of sea-level change over the last 500,000 years, based on tight synchronization to an Asian monsoon record. We observe maximum 'natural' (pre-anthropogenic forcing) sea-level rise rates below 2 m per century following periods with up to twice present-day ice volumes, and substantially higher rise rates for greater ice volumes. We also find that maximum sea-level rise rates were attained within 2 kyr of the onset of deglaciations, for 85% of such events. Finally, multivariate regressions of orbital parameters, sea-level and monsoon records suggest that major meltwater pulses account for millennial-scale variability and insolation-lagged responses in Asian monsoon records.

No abstract