PEDOGENIC IRON OXIDES, SOIL WATER BALANCE, AND CLIMATE

Since the fist pioneering observations by Le Borgne (1955) it is known that magnetic iron oxides accumulate in many types of soils, producing what is known as magnetic enhancement with respect to a weakly magnetic parent rock. The best known case is that of the Chinese Loess Plateau (CLP), where the mean annual rainfall (MAR) and the enhancement of magnetic susceptibility (MS) of modern soils are correlated. This correlation has been used to reconstruct MAR values over the last ~1.1 Myr from CLP loess/paleosol sequences. The reconstruction contrasts with δ¹⁸O records from caves located S of the CLP. In order to solve the apparent contradiction between magnetic and isotope records, it is important to evaluate critically the magnetism/rainfall and δ¹⁸O/rainfall transfer functions. The focus of this presentation is on the magnetism/rainfall transfer function, which consists of four main elements: 1) the conversion of magnetic parameters, such as MS, into mass fluxes of known iron oxides, 2) the dependence of the magnetic iron oxides accumulation rate on soil moisture, and in particular on its stochastic nature, 3) the link between moisture and soil water balance, and 4) the expression of soil water balance in terms of selected climatic parameters. I will show how an appropriate model for the magnetism/rainfall transfer function based on these four elements resolves apparent contradictions between magnetic enhancement data for modern loessic soils from China and the Midwestern US. Data obtained so far indicate that loessic soils are characterized by a universal magnetism/rainfall transfer function that provides a well-defined link between the accumulation of ferrimagnetic iron oxides and climate, as long as the ratio between water input by rainfall and water loss by evapotranspiration is sufficiently low to avoid saturation.