The magnetic properties of paleosols developed in Quaternary sequences of loess have been used for: stratigraphic definition; correlation with other terrestrial and deep-sea sequences; and paleoclimatic (paleorainfall) reconstruction. In some loess/paleosol sequences, including those of the Chinese Loess Plateau, Tajikistan and the Czech Republic, maxima in magnetic susceptibility values correspond with the paleosol horizons, and minima with the least-weathered loess layers. In other loess/soil sequences, including those of Siberia, Alaska and Argentina, the relationship is completely opposite, with susceptibility minima associated with the most developed paleosols. To account for these opposite relationships, the respective roles of: (1) magnetic enhancement and (2) magnetic depletion and/or dilution in determining soil magnetic properties are investigated for a range of modern soil types. Most magnetic enhancement is seen in the upper horizons of well drained cambisols. Absence or loss of magnetic iron oxides is apparent in acid, podsol profiles and waterlogged soils. For the cambisol profiles, significant correlation is found between susceptibility and organic carbon, cation exchange capacity and clay content. The mineralogy, morphology and grain size of soil magnetic carriers, extracted from three modern enhanced soils and a paleosol from the Chinese Loess Plateau, are also identified by independent petrographic means (microscopy and X-ray diffraction of magnetic extracts). Magnetite and maghemite of ultrafine grain size (from 0.4 to < 0.001 μm) are the major contributors to the magnetically enhanced soils. Weathering can concentrate detrital magnetic grains, especially in the fine silt size fractions of soils. The magnetic data from the modern soils indicate that interpretation of paleosol magnetic properties must be done on a site-specific basis, taking into account the possibilities of pedogenic enhancement, pedogenic dilution or depletion, and allochthonous inputs of magnetic minerals. Excessively arid, wet or acid soils are unable to form significant amounts of pedogenic ferrimagnets. Well drained, intermittently wet/dry soils, with reasonable buffering capacity and a substrate source of Fe, show most magnetic enhancement. For those soils which favour magnetic enhancement processes, correlation has been found between the maximum value of the pedogenic susceptibility and the annual rainfall. The almost unique pedogenic system in the Chinese loess plateau, where variation in soil-forming factors other than climate is reduced to a global minimum, allows use of the paleo-susceptibility values as proxy paleorainfall values. To identify the mechanism of any link between magnetic properties, for example, susceptibility, and climate change, the mineralogy and grain size of soil magnetic carriers, requires detailed investigation. Given the prospect of: (1) quantitative paleorainfall records from continental loess records; (2) predictions of present and future climate change; and (3) the value of paleoclimate data for testing of numerical climate models; such magnetic investigations can contribute significantly to our understanding of past and future environmental change.
