Pedogenesis and paleoclimate: Interpretation of the magnetic susceptibility record of Chinese loess-paleosol sequences: Comments and Reply COMMENT B. A. Maher School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom R. Thompson Department of Geology and Geophysics, University of Edinburgh, Edinburgh EH8 9YL, United Kingdom Verosub et al. (1993) studied 31 Chinese loess-paleosol samples provided by Kukla and Maher. Their conclusions can be summa- rized as follows: (1) The magnetic susceptibility is primarily pedo- genic in origin; (2) this information is a prerequisite for paleoclimatic interpretation of the susceptibility record; (3) special attention should be paid to the existence of maghemite; (4) a combined ap- proach using mineral magnetic methods and soil chemistry can be useful. We address these four conclusions in turn. 1. The primarily pedogenic origin of the susceptibility record of the Chinese loess has been widely recognized by European and Chi- nese workers and is not a new observation. Most recently, Hus and Han (1992) described detailed studies of susceptibility, anhysteretic and isothermal remanence, thermal demagnetization and hysteresis properties, supporting the pedogenic origin of the loess susceptibil- ity. Maher and Thompson (1992) went further, extracting the pedo- genic components for direct observation by transmission electron microscopy, then quantifying their contribution to both loess and paleosols. Using quantitative models of remanence acquisition, they showed that >90% of the susceptibility signal is contributed by ul- trafine pedogenic magnetic grains. 2. Heller et al. (1993) and Beer et al. (1993) have already paid close attention to the pedogenic nature of the loess record. They used Be 10 and susceptibility measurements not just to estimate the pedogenic contribution but to reconstruct paleorainfall. Their fully quantitative approach has already focused "on the critical role of climate in the development of the paleosols," directly related "dif- ferences in pedogenesis between paleosols . . . to climate varia- tions," and precisely characterized "the climate at different times," tasks that Verosub et al (1993) suggested have yet to be addressed. 3. The citrate bicarbonate-dithionite (CBD) method is depend- ent on grain size as well as composition and so is unable to distin- guish between submicrometre magnetite and maghemite. Vanden- berghe et al. (1992) applied Mossbauer spectroscopy to Chinese loess and paleosols and answered the question of the importance of maghemite. Loess and paleosol samples from Jixian contain pre- dominantly haematite and magnetite, with some maghemite, partic- ularly in the soil. Relative contributions of magnetite and maghemite to loess magnetic properties have also been identified by thermo- magnetic work. Figure 1 shows two thermomagnetic curves for pa- leosol samples from the more humid, central area of the Loess Pla- teau and the drier, western area. For the eastern sample, we attribute the marked drop in magnetization at —580 °C to the mag- netite Curie temperature, and the inflection around 400 °C to the oxidation of maghemite. Magnetite is clearly the dominant mineral. For the western sample, the major loss is between 300 and 500 °C, due to the dominant presence of maghemite. It has long been known that fine-grained magnetite oxidizes towards maghemite. The key point is that pedogenic enhancement does not result in the formation of maghemite directly. Rather, iron is reduced (by direct or indirect bacterial action as soils are wetted and dried) to form the mixed Fe 2+ /Fe 3+ iron oxide, magnetite. This authigenic magnetite is prone to surface oxidation to maghemite, postformationally, the degree of oxidation controlled by the moisture regime of the soil. In the dry western Loess Plateau, oxidation has proceeded more completely than in the more humid eastern area. Thus, in any magnetically enriched soil, the mineralogy will fall along the solid-solution series between magnetite and maghemite (Longworth et al., 1979). Mem- bers of this series have virtually indistinguishable magnetic rema- nence and susceptibility properties, so it is not obvious that precise differentiation will be of major paleoclimatic significance. 4. Finally, we agree that CBD extraction is a useful chemical Figure 1. Thermomagnetic curves for (A) paleosol S5, central Loess Plateau (x = 256.10" 8 m 3 /kg), and (B) paleosol SO, western Loess Plateau (X= 110.1(T a m 3 /kg). GEOLOGY, September 1994 857 on December 16, 2014 geology.gsapubs.org Downloaded from