Earth and Planetary Science Letters 152 1997. 203216
A multi-parameter rock magnetic record of the last glacialinterglacial paleoclimate from south-central Illinois, USA
Christoph E. Geiss ) , Subir K. Banerjee
Uniersity of Minnesota, Institute for Rock Magnetism, 283 Shepherd Laboratories, 100 Union St. S.E., Minneapolis MN 55455, USA Received 12 April 1997; revised 24 July 1997; accepted 24 July 1997
Abstract Pittsburg Basin, a small kettle lake in south-central Illinois, was formed after the retreat of the Illinoian ice sheet ; 130 ka B.P., and the sedimentary record includes the last interglacialglacial cycle. Curie temperature measurements, XRD analyses and thermal demagnetization of a low-temperature Jrs indicate that the magnetic fraction is characterized by oxidized magnetite and magnetic iron sulfides. The grain size of the magnetic fraction was estimated from ARMrJrs ratios, frequency-dependent susceptibility, and hysteresis parameters. These measurements show that glacial sediments contain coarse-grained, presumably detrital, material while interglacial samples have an additional fine-grained magnetic component, which cannot be explained by a single catchment or climatic process. Early interglacial samples show magnetic properties characteristic of authigenic biogenically produced magnetite, while late interglacial samples are probably influenced by pedogenesis around the basin. q 1997 Elsevier Science B.V.
Keywords: paleoclimatology; Sangamonian; magnetic susceptibility; magnetic hysteresis; pedogenesis
1. General background Recent reviews w1,2x have pointed out the usefulness of the magnetic properties of sediments as proxies of past climate change. The influence of Milankovitch cycles on oceanic sediment records has been frequently demonstrated w3x, including changes in the relative power of these periodicities w4x. Rock-magnetic properties have aided both in the detection of abrupt climate variations within the last interglacial period, as preserved in continental lake sediments w5x, and in the reconstruction of changes in east Asian summer monsoon intensity and rate of
Corresponding author. Tel.: q624 5274. Fax: q625 7502. E-mail: irm@geolab.geo.umn.edu
soil formation over the last 130,000 yr, as recorded in Chinese loess w68x. Because of the ease with which magnetic parameters can be measured from long cores, the low cost and high speed of the operation, and the ubiquity of iron-bearing minerals, such studies have become a popular tool for the reconstruction of paleoclimatic conditions from sedimentary archives. A time series of selected magnetic parameters from a sediment archive, however, does not a priori constitute a proxy of climatic change. In the true sense of the term, a proxy should be an accurate measure of a particular climate variable; for example, temperature or precipitation. In actuality, however, all magnetic proxies are strongly model-dependent. An increase of one parameter, for example, magnetic susceptibility, could represent warmer as
0012-821Xr97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S - 1 X 9 7. 3 - 7
C.E. Geiss, S.K. Banerjeer Earth and Planetary Science Letters 152 (1997) 203216
in Chinese loess w6x., or colder conditions as in French lake sediments w5x. The reconstruction of continental paleoclimate is often based on sedimentary archives from relatively small lakes. The processes that influence the sedimentary and magnetic properties of these records can change dramatically over time, therefore these records may be difficult to interpret. One simple model that links rock-magnetic parameters to paleoclimate might not hold true throughout the entire record. Variations in continental climate are important because it is on the continents where most people live and most of our food is produced. It is therefore crucial to determine how continental climate changed in the past in order to make predictions about future climate changes. High-resolution comparisons between oceanic and continental climate records may give us a better understanding of the physical nature and origin of the instability of climate systems. This paper presents the results of a

case study of lake deposits from an Illinois USA. lake, where multiple magnetic parameters were measured to arrive at a self-consistent history of past climate changes.
2. Site description Pittsburg Basin, located in south-central Illinois, is a small, now drained, kettle lake, which formed in till and glacial outwash of Illinoian age. It is situated between elongated, glaciofluvially deposited, ridges and knolls consisting mainly of a sand and gravel core capped by Wisconsinan loess w9x. It is located about 60 km south of the maximum extent of the Wisconsinan ice sheet Fig. 1. Due to the small drainage basin and low erosion rates, sedimentation rates in the lake were very low, producing a sedimentary record that extends from the late Illinoian into the Holocene w10x.
Fig. 1. Map showing the topography of Pittsburg Basin and the position of the two coring sites within the basin. Insert shows a map of Illinois and the location of Pittsburg Basin with respect to the maximum extent of the Wisconsinan glacial deposits stippled area.
In 1994 we obtained several cores from the center of Pittsburg Basin site 94-5. with a hollow stem augur and a Livingstone piston corer. The cores were correlated by lithology and magnetic-susceptibility profiles. One core 94-5a. was sampled for several magnetic and non-magnetic investigations. In 1996 two additional cores sites 96-1 and 96-2. were retrieved with a piston corer, at a location approximately 70 m away from the original site. These cores are approximately 50 cm shorter than the cores from site 94-5a but they contain the same lithological sequence. Depth correlation between sites 94-5a and 96-2 Fig. 2., based on core lithology and magnetic susceptibility, is good and all depths given in this paper are based on depths from site 94-5a. The cored
sequence at both sites consists of clay and silt-sized sediments with varying content of organic matter 060%. Clay and silt were most likely deposited from slope wash or from eolian processes. The base of the cores is sandy or gravelly glacial outwash. Drainage of the lake in the 1920s resulted in a fluctuating water table. The upper 100150 cm show abundant redoximorphic features indicating alteration of the iron minerals, they were therefore excluded from this study. The first paleoclimatic studies from Pittsburg Basin were conducted by Gruger w11x, who analyzed the pollen content of the sediments. According to his study, which is confirmed by recent analyses w12x, the base of the Pittsburg Basin sediments ) 755 cm
Fig. 2. Lithology and magnetic susceptibility x in m3 rm3. for cores 94-5a and 96-2. Lithological horizons and susceptibility peaks used for correlation are indicated by dashed lines. Correlation is good for most of the core, except for the depth interval between 490 cm and 530 cm, where the two susceptibility records disagree.

in core 94-5a. contains pollen assemblages characteristic of boreal forest, reflecting cold climatic conditions at the end of the Illinoian glaciation. The onset of interglacial conditions is marked by a deciduous forest period 755690 cm., which indicates a warm moist climate. A drier but equally warm prairie period 690570 cm. led to the deposition of a silty sequence abundant in ostracode shells. The ostracode d 13 C values suggest a relatively saline and eutrophic lake during this period w12x. The end of the interglacial is characterized by a second deciduous forest period 570485 cm., which gradually shifts into colder, glacial conditions of the Wisconsinan stage. This record differs from oceanic records in which the last interglacial marine oxygen-isotope stage 5. contains three warm substages 5e, 5c, and 5a. These warm substages do not appear to be recorded in Pittsburg Basin. There are two possible explanations for this difference: the first assumes that the Pittsburg Basin record does not extend to the beginning of the last interglacial and that the two deciduous forest periods correspond to substages 5a and 5c of the oceanic record w13x. The second explanation suggests that the smaller climatic variations of substages 5a through 5d did not severely affect the continental interior and are therefore not recorded at this midcontinental site, and that the double forest period corresponds to substage 5e. Dating of the core is still in progress and, once completed, should answer some of the above questions. Our current age estimates are based on the correlation of the pollen analyses with other dated sites. These will be supplemented by optically stimulated luminescence dating and by correlation of an ostracode oxygen-isotope record with oxygen-isotope records obtained from independently dated coeval speleothems from the same area work in progress. 3. Experimental procedures The magnetic properties of sediments can reflect the influence of a variety of different processes e.g. sedimentary, diagenetic or climatic. that affected the sediment during and after its deposition w1,2x. Different genetic histories can ultimately result in similar magnetic properties. We favor a multi-proxy approach in which many aspects of the magnetic signal

are examined in order to interpret results with greater confidence. Our approach concentrates on three variables: magnetic mineralogy, concentration of magnetic grains, and the grain-size distribution of the magnetic fraction. We have conducted a suite of experiments to characterize the magnetic fraction in terms of these parameters. We used a combination of magnetic and nonmagnetic techniques to characterize the magnetic mineralogy. Saturation magnetization Js. between 308C and 7008C was measured with a Princeton Measurement Corporation Micro-Vibrating Sample Magnetometer mVSM. The temperature dependence of magnetic susceptibility x. was determined with a Geofyzika Kappabridge, model KLY-2, for the same temperature range. From the resulting JsT. and x T. curves we estimated the Curie temperature Tc. of the samples. To eliminate the influence of paramagnetic minerals e.g., clays. we measured hysteresis loops as a function of temperature for several clay-rich samples on the mVSM. By fitting a line through the linear high field part of the loop we estimated Js values from the intersections of these lines with the y axis at zero field. The JsT. curves calculated from these loops also aided in the interpretation of the other high-temperature measurements. All heating experiments were carried out in an inert gas atmosphere He for the mVSM, Ar for the Kappabridge. We further identified the presence of various magnetic minerals by observing the presence or absence of magnetic phase transitions at low 10300 K. temperatures e.g., pyrrhotite at 3035 K w14x or magnetite between 90 K and 120 K w15x. Using a Quantum Design MPMS 2 magnetic-properties measurement system, we exposed samples to a 2.5 T field at a temperature of 10 K and observed changes in saturation remanence Jrs. as the samples were heated back to room temperature in zero field. In addition, magnetic measurements were complemented by X-ray diffraction XRD. analyses of magnetic separates. Changes in concentration of magnetic material were estimated by measuring x and saturation isothermal remanent magnetization Jrs. at room temperature for all samples. A Bartington Instruments M.S. 1 susceptibility bridge was used for the x measurements, and a cryogenic magnetometer 2G
Corp. SRM, model 760-R. was used to measure Jrs. The saturating field was 1.5 T. For intensely magnetized samples, Jrs was determined with a Schoenstedt SSM 1A spinner magnetometer. We characterized the magnetic grain-size distribution by measuring frequency-dependent susceptibility x FD s 100 x400 Hz y x4000 Hz.rx400 Hz. and anhysteretic remanent magnetization ARM. Most samples were too weakly magnetized to be measured with a Bartington Instruments dual-frequency sensor. We therefore used a Lakeshore 7000 susceptometer that allowed us to obtain x values for 20 frequencies between 40 Hz and 4000 Hz. The values corresponding to x400 Hz and x4000 Hz were obtained from a best fit through all of the data points see also Fig. 7. ARMs were acquired in a bias field of 50 mT and a peak alternating field of 99 mT. In addition we measured hysteresis loops with a maximum field of 0.8 T on a VSM constructed in our laboratory. XRD spectra were obtained from magnetic separates with quartz as an internal standard. Lattice parameters were calculated from a minimum of four identified peaks.

4. Results 4.1. Magnetic mineralogy Identification of the magnetic minerals present in the samples is an important step in the interpretation of magnetic data. The presence or absence of certain minerals can be a good indicator of certain diagenetic processes or environments. Furthermore, most of the parameters discussed later depend on magnetic mineralogy. However, identification of the main magnetic minerals present in the cores proved to be difficult because of their low concentrations - 150 ppm. 1. Many of the iron minerals are kinetically unstable when heated and convert into more stable phases during the heating process, even in an argon
1 The most intensely magnetized samples displayed Js values in the order of ;1.2=10y2 Am2 rkg. Assuming maghemite as the main carrier of magnetic properties yields Js rJsmaghemite. f1.5 =10y4 , which corresponds to a concentration of approximately 150 ppm.
or helium atmosphere. It was often impossible to obtain reliable mineral identifications from thermomagnetic curves alone. Our Curie temperature estimates were therefore complemented by XRD analyses from magnetic separates and low-temperature magnetic remanence measurements. On the basis of x T. curves it is possible to divide the samples into two groups, with their representative curves shown in Fig. 3. Group A, which consists of samples from the base of the core below 755 cm., produces JsT. and x T. curves that are dominated by a broad peak between 3608C and 5408C Fig. 3a. Most cooling curves show higher x and Js values after the sample is heated to 7008C, which suggests reducing furnace conditions at high temperatures. Repeated cycling of the samples to increasingly higher temperatures shows that neither the increase in x nor the following decrease is reversible. For some of the samples a strong smell of sulfur was noticeable during the heating experiments. To interpret these measurements we conducted a series of x T. measurements on natural iron sulfide samples of known composition: x T. curves measured for natural pyrrhotite, pyrite, and marcasite samples show a strong non-reversible increase in x between 4008C and 5008C, when the weakly magnetic or paramagnetic iron sulfides convert into a more magnetic phase. The exact conversion temperatures vary and depend on the heating environment; for example, whether the sample was heated by itself or was mixed into a clay matrix to simulate sediment conditions. In all cases conversion occurred between 4008C and 4608C. Based on these measurements, the increase in susceptibility found in samples from group A reflects the conversion of paramagnetic iron sulfides into a more ferrimagnetic phase, presumably maghemite or magnetite. This phase itself is thermally unstable and converts to weakly magnetic hematite on further heating. XRD analyses indicate the presence of marcasite in one of the magnetic extracts from group A, but the main component of all magnetic extracts is slightly oxidized magnetite. This is consistent with a weak Verwey transition, which is representative of oxidized magnetite w15x observed in the low-temperature measurements made on bulk samples Fig. 5a,c. Group B consists of samples from above 690 cm.

They display a non-reversible increase in x between 1508C and 2508C, which is followed by a gradual, only partially reversible, decrease in x until the sample becomes paramagnetic near 5708C Fig. 3b. In some samples this initial increase is absent or only weakly expressed. The decrease in x starts between 2508C and 3008C and coincides with the observed burning of the organic fraction of the sediment. x steadily decreases until a more magnetic phase is displayed above 4008C. Because there is a variety of minerals that could explain the high temperature behavior of these samples, we base our interpretation mainly on the XRD and low tempera-
ture studies. XRD analyses of these samples show the presence of magnetite or maghemite results for several samples range from 8.37 A to 8.40 " 0.02 A. as the only magnetic mineral. Low-T measurements Fig. 4a. show evidence of a weak Verwey transition, which is characteristic for partially oxidized magnetite w15x. Some 4 samples from the base of the core, 5 samples between 600 cm and 630 cm. low-T measurements display very weak indications of a possible 34 K transition Fig. 4c. According to our interpretations, magnetite that has been oxidized to varying degrees is the main magnetic mineral in the Pittsburg Basin sediments.
Fig. 3. Measurements of magnetic susceptibility x vs. temperature were used to estimate the Curie temperatures of the samples. The samples fall into two categories: a. Samples from group A show a distinct peak between 3608C and 5408C, which is attributed to the presence of paramagnetic or weakly ferrimagnetic iron sulfides. b. Samples from group B show an irreversible increase in x between 1508C and 2508C, which is followed by a decrease in x until a final blocking temperature is reached near 5605808C.
Fig. 4. Low temperature Jrs demagnetization curves for 3 characteristic samples. a. Sample of partially oxidized magnetite. Both field-cooled and zero-field-cooled curves are nearly indistinguishable. The Verwey transition at ; 120 K. is weak but can be seen by plotting the derivative of the curves. b. Sample likely to contain unoxidized biogenic magnetite. Field-cooled and zero-field-cooled curves differ distinctly, and the Verwey transition is clearly visible. c. Sample from the base of the core showing indications for a phase transition between 30 and 40 K. Both the transition at ; 35 K and a slight Verwey transition can be seen in the derivative of the demagnetization curve.
Iron sulfides, such as pyrite, marcasite, or pyrrhotite, may be present, especially in the lowest section of the core, but their contribution to the bulk magnetic signal appears to remain small. Greigite is another possible magnetic iron sulfide that might be present in lake sediments. Several authors w1618x point out that thermomagnetic experiments conducted in air might not detect greigites characteristic phase transitions. We therefore measured x T. curves for two characteristic samples in both air and argon Fig. 5a,c. Sample 96-2, 780 cm, displays a peak between 440 and 5508C, which is characteristic for the iron sulfide-bearing samples of group A. Sample 96-2, 560 cm, is an interglacial sample from group B. Fig.

5b,d shows the change in susceptibility at room temperature after stepwise heating in air. From Fig. 5 it becomes apparent that heating in air results in very similar heating curves and that major changes in mineralogy, as expressed by changes in x , do not occur until a temperature of ; 3408C is reached. The thermal stability of greigite is still poorly known, but many authors w1921x conclude that greigite becomes unstable at temperatures between 2508C and 3008C. The susceptibility curves shown in Fig. 5b,d do not indicate any significant changes in this temperature range, and we conclude that greigite is not a major magnetic mineral in these samples. This interpretation holds true for most of the core,
Fig. 5. a. and d. Measurements of magnetic susceptibility x vs. temperature in air solid lines. and argon dashed lines. For both samples the heating curves do not depend on the heating environment. b. and c. Normalized susceptibility, measured at room temperature, after stepwise heating. Changes in susceptibility above 3408C indicate mineralogical phase changes.
C.E. Geiss, S.K. Banerjeer Earth and Planetary Science Letters 152 (1997) 203216 Fig. 6. Several rock-magnetic parameters measured for Pittsburg Basin, site 94-5a. a. Simplified lithology of site 94-5a. b. Magnetic susceptibility is an indicator for the concentration of magnetic minerals in the sediments. c. ARMrJrs ratios are an indicator of fine-grained SD particles, with higher ratios indicating more SD grains. d. Frequency-dependent susceptibility, x FD , for selected samples. High values of x FD indicate the presence of ultra-fine-grained SP particles. e. Paleoclimatic interpretation based on pollen analyses w12x. 211
Fig. 7. Frequency dependence of magnetic susceptibility x for various glacial ` and I. and interglacial samples v and B. Sample 94-5a, 480 cm q. is from a late interglacial horizon. x values have been normalized to correct for concentration dependence of magnetic susceptibility.
Fig. 8. Plot of Jrs rJs vs. B crrB c w26x for samples from site 94-5a SD s single-domain, PSD s pseudo-single-domain, MD s multi-domain. Most samples plot in the PSD field but interglacial samples with high ARMrJrs ratios plot consistently closer to the SD field, supporting the interpretation of fine-grained magnetic enhancement in these samples.
except for the segment between 755 cm and 690 cm. These samples are too rich in organic carbon to yield reliable information when subjected to high-temperature analysis, but low-T measurements reveal the presence of a well developed Verwey transition in these samples, which is characteristic of stoichiometric magnetite. Furthermore, the decrease in a lowtemperature Jrs at the Verwey transition depends critically on whether the sample was cooled in a field or in zero field Fig. 4b. Such differences in demagnetization behavior are known to indicate the presence of biogenic. magnetite crystals that are aligned in chains w22x. 4.2. Magnetic granulometry We estimated the relative concentration of the fine-grained magnetic component by measuring ARM for all samples and x FD for 74 selected samples. ARM values, which depend heavily on the presence of single-domain SD. grains w23x, were normalized by either magnetic susceptibility x or Jrs to correct for any changes in concentration. Because x depends on both ferrimagnetic and paramagnetic minerals, while Jrs does not, we prefer Jrs as a normalizing parameter for ferrimagnetic mineral content. However, the difference between the ARMrx and ARMrJrs curves is small. ARMrJrs Fig. 6c. is low at the base ) 755 cm. and in the upper part of the core - 550 cm., which indicates that the magnetic component of these grains is dominated by coarsegrained particles. Between 550 cm and 755 cm ARMrJrs is relatively high, with three distinct peaks at 575 cm, 625 cm, and 720 cm. For the same section x FD , which is a measure of very fine superparamagnetic SP. content w24,25x, has high values, between 10% and 13% Fig. 6d., while the samples from above and below have moderate to low values - 6%. Fig. 7 shows the frequency dependence of x for various glacial and interglacial samples. The combination of high ARMrJrs ratios and high x FD values suggests the presence of a fine-grained component containing both SP and SD grains between 550 cm and 755 cm. Magnetic-hysteresis measurements results confirm the presence of fine-grained material between 755 cm and 550 cm. In a plot of M rs rM s vs. B crrB c w26x most samples fall into the pseudo-single domain

PSD. range of the graph but samples with high ARMrJrs ratios plot consistently nearer to the SD field than samples with low ARMrJrs values Fig. 8. These results are discussed further below.
5. Discussion A comparison with a pollen study w12x performed on cores from the same site reveals that samples between 550 cm and 755 cm, which represent the last interglacial period Sangamon., have high ARMrJrs ratios Fig. 6c,e. This increase in ARMrJrs may have resulted from an input of finegrained magnetic material during warm interglacial periods, probably related to erosion of magnetically enhanced interglacial soils in the catchment. Many modern soils show high ARM values for the upper soil horizons e.g., w24x., and both the modern soil and the Sangamon geosol that have been studied within the watershed of Pittsburg Basin w27x are enhanced in fine-grained magnetic material in comparison with the relatively unaltered parent material. It is reasonable to assume that erosion of strongly developed soils within the watershed of Pittsburg Basin produced the high ARMrJrs values in the lake sediments. These soils developed under a prairie environment, which, in general, is not very susceptible to erosion if it has continuous herb cover. Pollen studies, however, show a high abundance of Ambrosia, a weed that thrives on open ground, which might explain the postulated high erosion rates during this period. The high x FD values measured for these samples indicate an increase in SP grains, giving further support to this hypothesis w25x. However, a more detailed investigation shows that one single process is unlikely to explain the three maxima observed in Fig. 6c. The upper two peaks 570 cm and 625 cm. in ARMrJrs coincide with an increase in x and indicate the presence of a strongly magnetic and somewhat finer-grained component, while the main peak 700 cm. has low x values. This interval is rich in organic matter up to 60%., and all samples are weakly magnetized but contain mostly fine-grained magnetic material. Lowtemperature experiments reveal a striking difference between these samples and the rest of the core. Only the samples between 690 cm and 755 cm display a
marked Verwey transition Fig. 4b. Furthermore, the decrease in the low-T Jrs depends critically on whether the sample was cooled through the Verwey transition in zero field solid curves in Fig. 4. or in a field of 2.5 T dashed curves in Fig. 4. The separation of the two curves below the Verwey transition is an indicator for magnetite grains that are aligned in chains w22x. Such magnetite chains can be produced most easily by magnetotactic bacteria that live in lake sediments w28x. Samples from other depths do not show this separation of the two curves Fig. 4a,c.: the Verwey transition is barely visible but can still be detected by plotting the first derivative of the demagnetization curve. This difference in low-T behavior makes it unlikely that the lowermost peak in the ARMrJrs record was caused by the same pedogenic processes that influenced the rest of the interglacial sediments. Therefore the relative variations in ARMrJrs in the interglacial sediments may not be directly used as a proxy of climate variations between the apparent substages. A plot of Jrs rJs vs. B crrB c Fig. 8. indicates that interglacial samples with high ARMrJrs ratios plot closer towards the SD field than glacial samples, which have low ARMrJrs values. Thus the hysteresis measurements in general confirm the changes in magnetic grain size deduced from the previous parameters. Some samples have intermediate values of Jrs rJs , but high B crrB c ratios and plot above the MD field in Fig. 8, most of which give raise to wasp-waisted hysteresis loops. According to pollen studies these samples correspond to the end of interglacial conditions. Wasp-waisted loops can be produced by a mixture of two populations with different coercivities w29,30x, and we consider two different scenarios below that could explain these hysteresis loops. A mixture of low and high coercivity minerals could produce wasp-waisted loops. Our exploratory laboratory experiments with synthetic magnetite, goethite, and hematite samples indicate that high ) 75%. concentrations of goethite or hematite are required to achieve the desired effect as pointed out by Roberts et al. w29x. This is due to low values of Js for goethite - 1 Am2rkg. and hematite 0.08 Am2rkg. compared to magnetite or maghemite Am2rkg. Furthermore, the partial oxidation of magnetite to maghemite was considered; however,

when we oxidized a natural sample over a period of 2 months Jrs rJs decreased, but B crrB c remained more or less constant. We also considered a combination of magnetic grain sizes. In a series of experiments we used synthetic magnetite of known grain size ferrofluid mixed with quartz for the SP fraction, 0.05 mm for the SDrPSD fraction, 3540 mm for the MD fraction. and measured hysteresis loops for various mixtures. Some of the SPMD mixtures plot in the appropriate region of Fig. 8 but none of the loops is wasp-waisted. Mixtures of SD and SP are wasp-waisted, and variations in relative concentrations can achieve almost any ratio of B cr rB c. A closer inspection of the data from natural samples shows that the large variations in Jrs rJs and B crrB c are mostly caused by changes in Js and B c , while the remanence parameters Jrs and B cr are slightly above the average values of the core but remain fairly constant through the section of interest. This is consistent with the presence of both a super. paramagnetic phase and probably a slightly more coercive phase. Our magnetic results for these anomalous waspwaisted loops remain ambiguous. Low ARMrJrs ratios and x FD values indicate the absence of finegrained SD and SP grains but hysteresis loops are best explained by a mixture of SD and SP grains. It is possible that the wasp-waisted loops are caused by SP grains that lie outside the narrow sensitivity window of the x FD measurements 0.0180.02 mm. Another possibility is that our choice of SD and MD grain sizes in the hysteresis experiments was an unfortunate one and that wasp-waisted loops are not restricted to mixtures of SD and SP grains. Throughout the above discussion we have neglected the possible influence of magnetic iron sulfides. Our mineralogical studies show possible evidence for the presence of pyrrhotite at the base of the core and in a section between 600 cm and 650 cm. We consider the influence of iron sulfides on the overall magnetic properties to be weak also. Our main magnetic signal is a variability in ARMrJrs , which we interpret as a change in magnetic grain size. Pyrrhotite, with its high coercivity, should lead to higher ARM values but the part of the core where it might be most significant the base below 755 cm. shows very low ARMrJrs ratios. In contrast, the other zone where we found possible traces of
pyrrhotite 600640 cm. shows high ARMrJrs values, exactly the opposite of what we observe at the base. We therefore conclude that the magnetic signal is mainly carried by magnetite and its low-temperature oxidized form found throughout the core.
6. Conclusions No single climate or catchment process can account for the magnetic record of Pittsburg Basin. It is clear that a combination of several magnetic and non-magnetic measurements is necessary to obtain a reliable characterization of the sediments before any environmental or climatic interpretation can be made. Comparing our work with pollen studies reveals that various climatic processes affected the magnetic properties of Pittsburg Basin sediments. They can be summarized as follows: 1. Both Illinoian ) 755 cm. and Wisconsinan 485 cm. glacial sediments are characterized by relatively coarse-grained magnetite, which is interpreted as representing relatively unaltered sediment eroded from the slopes of the basin. 2. The onset of Sangamon interglacial conditions is marked by deciduous forest vegetation around the lake, which reduced the erosional input of detrital grains. Biogenic magnetite is most likely an important carrier of the magnetic properties in this part of the core 755690 cm. It is characterized by a combination of high ARMrJrs ratios and intermediate x FD values. Low-T measurements and recent transmission electron microscopy observations of the magnetic extract H. Petermann, pers. commun. confirm this interpretation. 3. The subsequent prairie period 690570 cm. contains more fine-grained magnetite, which spans the SDSP boundary. This grain size distribution leads to both high ARMrJrs ratios and high x FD values. We interpret this fine-grained component as representing extensive soil formation within the watershed of Pittsburg Basin, leading to a pedogenic enhancement of the parent material and resulting in erosional input of fine magnetic grains into the lake. 4. The nature of the magnetic carriers at the end of interglacial conditions 570485 cm. remains unclear. ARMrJrs ratios and x FD values point to a

relatively coarse-grained component, but hysteresis measurements are best explained by a mixture of SD and SP grains. To further characterize the influence of climatic change on magnetic records of small lakes it is necessary to combine magnetic studies with other climate proxies, such as information obtained from ostracode studies, stable-isotope analyses d 13 C and d 18 O., and pollen work. These investigations are currently under way, and we hope that a combination of all these techniques will yield a more reliable, self-consistent climatic model that can explain the changes observed in the Pittsburg Basin sediments. It will then be possible to compare our results with modern lakes situated under various climatic conditions and to test the validity of our model. Since the processes that affect sediments of small lakes are numerous and highly variable, it is not possible to interpret changes in the magnetic record directly in terms of climatic change, or use one single model throughout the core. It is necessary to come up with separate models that link climatic or sedimentary processes to the combination of magnetic parameters found in the record. In sections where the main climatic variable did not change drastically, the fine-scale properties of the sediment might be used to construct a high-resolution record of climatic change. In this case a combination of several magnetic parameters supported by non-magnetic data is most likely to yield a self-consistent and unique interpretation.
Acknowledgements We would like to thank Dr. H.E. Wright Jr. and Dr. E. Ito for their cooperation and willingness to share samples with us. Dr. B. Curry, J. Dorale, D. Gauss, M. Shapley and R. Teed were of great assistance during the coring operations and their help is gratefully acknowledged. Our work benefitted from fruitful discussions with Dr. D. Williamson and we thank Dr. E. Oches for reviewing the manuscript. This is contribution number 9706 of the Institute for Rock Magnetism IRM. The IRM is funded by the W.M. Keck Foundation, the National Science Foundation and the University of Minnesota. This project was supported by a grant from the National Science

C.E. Geiss, S.K. Banerjeer Earth and Planetary Science Letters 152 (1997) 203216 transition at 3034 Kelvin in pyrrhotite: insight into a widespread occurrence of this mineral in rocks, Earth Planet. Sci. Lett. 98 1990. 319328. O. Ozdemir, D.J. Dunlop, B.M. Moskowitz, The effect of oxidation on the Verwey transition in magnetite, Geophys. Res. Lett. 20 1993. 16711674. I. Snowball, R. Thompson, A stable chemical remanence in Holocene sediments, J. Geophys. Res. 95 1990. 44714479. A. Roberts, G.M. Turner, Diagenetic formation of ferrimagnetic iron sulphide minerals in rapidly deposited marine sediments, South Island, New Zealand, Earth Planet. Sci. Lett. 115 1993. 257273. R.L. Reynolds, M.L. Tuttle, C.A. Rice, N.S. Fishman, J.A. Karachewski, D.M. Sherman, Magnetization and geochemistry of greigite-bearing Cretaceous strata, North Slope basin, Alaska, Am. J. Sci. 294 1994. 485528. M. Krs, F. Novak, M. Krsova, P. Pruner, L. Kouklkova, J. Jansa, Magnetic properties and metastability of greigite smythite mineralization in brown-coal basins of the Krusne hory Piedmont, Bohemia, Phys. Earth Planet. Inter. 70 1992. 273287. A.P. Roberts, Magnetic properties of sedimentary greigite Fe 3 S 4., Earth Planet. Sci. Lett. 134 1995. 227236. J.W.E. Fassbinder, H. Staniek, Magnetic properties of biogenic soil greigite, Geophys. Res. Lett. 21 1994. 23492352. B.M. Moskowitz, R.B. Frankel, D.A. Bazylinski, Rock-magnetic criteria for the detection of biogenic magnetite, Earth Planet. Sci. Lett. 120 1993. 283300. C.P. Hunt, B.M. Moskowitz, S.K. Banerjee, Magnetic properties of rocks and minerals, in: Rock Physics and Phase Relations. A Handbook of Physical Constants, vol. 3, AGU Reference Shelf, 1995, pp. 189204. F. Oldfield, Towards the discrimination of fine-grained ferrimagnets by magnetic measurements in lake and near-shore marine sediments, J. Geophys. Res. 99 1994. 90459050. J.A. Dearing, R.J.L. Dann, K. Hay, J.A. Lees, P.J. Loveland, B.A. Maher, K. OGrady, Frequency dependent susceptibility measurements of environmental materials, Geophys. J. Int. 124 1996. 228240. R. Day, M. Fuller, V.A. Schmidt, Hysteresis properties of titanomagnetites: Grain-size and compositional dependence, Phys. Earth Planet. Inter. 13 1977. 260267. C.E. Geiss, S.K. Banerjee, Magnetic properties of the Sangamon paleosol and its implications for paleoclimate reconstruction, Ann. Geophys. Suppl. 14, C 129 1996. H. Vali, O. Forster, G. Amarantidis, N. Petersen, Magneto tactic bacteria and their magnetofossils in sediments, Earth Planet. Sci. Lett. 86 1987. 389400. A.P. Roberts, Y.-L. Cui, K.L. Verosub, Wasp-waisted hysteresis loops: mineral magnetic characteristics and discrimination of components in mixed magnetic systems, J. Geophys. Res. 100 1995. 1790917924. L. Tauxe, T.A.T. Mullender, T. Pick, Potbellies, wasp-waists, and superparamagnetism in magnetic hysteresis, J. Geophys. Res. 101 1996. 571583.
Foundation. Additional support from the University of Minnesota RTG. for coring and other field-related expenses was very much appreciated. We thank Drs. Weiming Zhou, Mike Singer and Andrew Roberts for their helpful comments on the paper. [RV] References

w1x R.L. Reynolds, J.W. King, Magnetic records of climate change, Rev. Geophys. 33 suppl. IUGG Report. 1995., 101110. w2x K.L. Verosub, A.P. Roberts, Environmental magnetism: Past, present, and future, J. Geophys. Res. 100 1995. 21752192. w3x J. Bloemendal, J.W. King, F.R. Hall, S.-H. Doh, Rock magnetism of Late Neogene and Pleistocene deep-sea sediments: Relationship to sediment source, diagenetic processes, and sediment lithology, J. Geophys. Res. 97 1992. 4361 4375. w4x J. Bloemendal, P. de Menocal, Evidence for a change in the periodicity of tropical climate cycles at 2.4 Myr from wholecore magnetic susceptibility measurements, Nature 342 1989. 897900. w5x N. Thouveny, J.-L. de Beaulieu, E. Bonifay, K.M. Creer, J. Guiot, M. Icole, S. Johnsen, J. Jouzel, M. Reille, T. Williams, D. Williamson, Climate variations in Europe over the past 140 kyr deduced from rock magnetism, Nature 371 1994. 503506. w6x B.A. Maher, R. Thompson, Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in the Chinese loess and paleosols, Quat. Res. 44 1995. 383391. w7x K.L. Verosub, P. Fine, M.J. Singer, J. TenPas, Pedogenesis and paleoclimate: Interpretation of the magnetic susceptibility record of Chinese loess-paleosol sequences, Geology 21 1993. 10111014. w8x C.P. Hunt, M.J. Singer, G. Kletetschka, J. Ten Pas, K.L. Verosub, Effect of citratebicarbonatedithionite treatment on fine-grained magnetite and maghemite, Earth Planet. Sci. Lett. 130 1995. 8794. w9x A.M. Jacobs, J.A. Lineback, Glacial geology of the Vandalia, Illinois region, Illinois State Geol. Surv. Circ. 442 1969. w10x E. Gruger, Late Quaternary vegetation development in South-Central Illinois, Quat. Res. 2 1972. 217231. w11x E. Gruger, Pollen and seed studies of Wisconsinan vegetation in Illinois, Geol. Soc. Am. Bull. 83 1972. 27152734. w12x R. Teed, B.B. Curry, E. Ito, A long )130,000 yrs. pollen and ostracode record from Pittsburg Basin, Illinois, AMQUA Program and Abstracts of the 14th Biennial Meeting, 1996, p. 133. w13x D.P. Adam, Correlations of the Clear Lake, California core CL-73-4 pollen sequence with other long records, Geol. Soc. Am. Spec. Pap. 214 1988. 8195. w14x P. Rochette, G. Fillion, J.-L. Mattei, M.J. Dekkers, Magnetic

Earth and Planetary Science Letters 228 (2004) www.elsevier.com/locate/epsl
Signature of magnetic enhancement in a loessic soil in Nebraska, United States of America
Christoph E. Geissa,*, C. William Zannerb, Subir K. Banerjeec, Minott Joannad
Department of Physics, Trinity College, McCook Hall 105, 300 Summit Street, Hartford, CT 06106, United States b School of Natural Sciences, University of Nebraska, 133 Keim Hall, Lincoln, NE 68583-0915, United States c University of Minnesota, Newton Horace Winchell School of Earth Sciences, Pillsbury Drive 310, Minneapolis, MN 55455, United States d Mt. Holyoke College, Mount Holyoke College, South Hadley, MA 01075-6419, United States Received 3 May 2004; received in revised form 7 October 2004; accepted 12 October 2004 Available online 21 November 2004 Editor: V. Courtillot
Abstract Our study of a loessic soil profile from east-central Nebraska shows that the A horizons of the modern soil are characterized by higher concentrations of fine-grained (b0.1 Am) magnetic minerals. This pedogenic magnetic component leads to higher values of concentration-dependent parameters, such as magnetic susceptibility (v), isothermal remanent magnetization (IRM) and anhysteretic remanent magnetization (ARM), combined with increases in frequency-dependent susceptibility (v fd) and ARM/IRM ratios. Hysteresis properties are relatively insensitive towards the presence of this pedogenic magnetic component. The magnetic properties of the soil profile are dominated by ferrimagnetic magnetite or maghemite. Analyses of bsoftQ IRM (sIRM) and bhardQ IRM (hIRM), however, do show that approximately 8090% of the remanence carrying magnetic component exists in the form of hematite or goethite and that the magnetically enhanced horizons are enriched in both ferri- and antiferromagnetic minerals. The pedogenic magnetic component is most likely caused by the conversion of paramagnetic, iron-bearing minerals to ferriand antiferromagnetic minerals. Soil compaction, lessivage or decalcification cannot explain the observed magnetic soil properties. Magnetic analyses of loess-paleosol sequences from the midwestern United States may yield valuable information about regional variability of paleoclimate. Based on the fine-grained nature of the pedogenic magnetic component, we expect grain-size-dependent proxies (ARM, ARM/IRM, v fd) to yield better paleoclimatic information than low-field magnetic susceptibility used in previous analyses. D 2004 Elsevier B.V. All rights reserved.
Keywords: environmental magnetism; loess; soil; Holocene; Nebraska; climate reconstruction
* Corresponding author. Tel.: +297 4191; fax: +987 6239. E-mail addresses: christoph.geiss@trincoll.edu (C.E. Geiss)8 czanner@unlnotes.unl.edu (C.W. Zanner)8 banerjee@umn.edu (S.K. Banerjee)8 jminott@hotmail.com (M. Joanna). 0012-821X/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2004.10.011
C.E. Geiss et al. / Earth and Planetary Science Letters 228 (2004) 355367
1. Introduction Over the last decades, rock-magnetic analyses of loess-paleosol sequences have provided valuable tools for the reconstruction of paleoclimatic change in mid-continental regions, where other recorders of climate are rare or absent. Numerous studies of the Chinese loess plateau show that modern and buried soils are characterized by higher concentrations of magnetic minerals in the A and B horizons. The degree of this magnetic enhancement is often quantified by changes in magnetic susceptibility between the enhanced horizons and the pedogenetically unaltered parent material (e.g., [1]). Several studies have shown a good correlation between the magnitude of magnetic enhancement in modern soils and present climatic conditions, especially rainfall (e.g., [24]), even though the processes that lead to the observed magnetic signal remain under discussion [5]. In situ formation of (ultra)fine-grained (b0.1 Am) ferrimagnetic minerals through a variety of processes is likely to explain the magnetic enhancement found in many loess-paleosol sequences in China, Europe and the North America. Potential pathways include the effects of repeated fires [68], changes in redox conditions due to plant decomposition [6,9] or repeated wetting and drying cycles in otherwise well-drained soils [10], and the effects of bacterial

activity [11]. Soil compaction and loss of weakly magnetic minerals through decalcification [12] or lessivage [13] have also been suggested as potential pathways of magnetic enhancement. Paleosols in Alaska and Siberia, however, are often characterized by low concentrations of magnetic minerals, compared to their loessic parent material. Such low concentrations might be due to the loss of iron oxide minerals during periods of gleying [14] or changes in wind vigor [15,16], which can result in changed sediment compositions. Singer et al. [17], Maher [18] and Tang et al. [5] provide additional information on the pathways of magnetic enhancement in soils and the relationship between soil magnetic properties and climate. While the influence of climate on soil magnetic properties is well established, it is less clear over which time periods the magnetic signature of soils continues to evolve. Based on evidence from modern or recently disturbed soils [19], as well as loesspaleosol sequences from the western edge of the Chinese loess plateau [20], Maher et al. argue that the magnetic enhancement signature is acquired rapidly within a few centuries. Analyses of a chronosequence from California by Singer et al. [21] and a loesspaleosol sequence from Jiaodao, China by Vidic et al. [22], however, suggest that the magnetic signal of soils continues to evolve, probably over several 100,000 years.
Fig. 1. Location of soil site 4G-99a in east-central Nebraska. Map insert shows location within continental United States. The approximate extent of Pleistocene Peoria loess is indicated by a light gray line (modified after [28]).
Unfortunately, successful application of magnetic proxies to the Chinese loess for paleoclimate reconstruction is not automatically extended to other loess regions. In Europe and Russia, many loessic modern and buried soils [4,18] show magnetically enhanced soil horizons, but some paleosols that formed during marine isotope stage 5 show a depletion of pedogenic ferrimagnetic magnetic minerals, possibly due to intense weathering [23,24]. As mentioned above, some paleosols in Alaska and Siberia contain low abundances of ferrimagnetic minerals, though the modern soils are still enhanced in magnetic minerals [25]. It is therefore necessary to establish the nature of the soil-magnetic signal for a given region before any attempts of paleoclimatic reconstructions can be made. Relatively few studies [26,27] are concerned with the magnetic properties of loess-paleosol sequences in the midwestern United States. The central Great Plains is an ecologically and agriculturally important semiTable 1 Soil description for site 4G-99a

Depth 020 Horizon Ap Boundary clear Color very dark gray (10YR 3/1) very dark gray (7.5YR 3/1) brown (10YR 4/3) Texture heavy silt loam light silty clay loam silty clay loam
arid region. Twenty-five percent of the worlds total production of wheat, corn, barley, oats, rye and sorghum are grown here, at the threshold limits of needed rainfall, and dependence on artificial irrigation from rivers and aquifers is high. Despite the regions dependence on natural rainfall, little is known about the variability of past climatic conditions, which is mostly due to the absence of good recorders of paleoclimate, such as lake basins or cave deposits. A magnetic analysis of soils and paleosols offers the opportunity to reconstruct paleoclimatic conditions for certain time slices. However, the absence of an universally applicable link between pedogenic magnetic enhancement and climate makes it necessary to carefully investigate the magnetic signature of modern soils before any climate reconstruction can be attempted. In this study, we use a magnetic multi-proxy approach to carefully characterize the magnetic properties of a modern loessic soil from east-central Nebraska, to constrain possible pathways of magnetic enhancement, and to
Structure weak fine to medium granular moderate fine granular moderate very fine and fine subangular blocky moderate fine subangular blocky moderate medium subangular blocky weak medium to coarse subangular blocky weak medium subangular blocky weak coarse subangular blocky

Effervescence none

2033 3344

clear clear

none none many thin continuous clay coatings (10YR 3/1) on ped faces many thick, slightly patchy (2.5Y 4/2) clay coatings on ped faces many thin patchy (2.5Y 4/2) clay coatings on ped faces very few fine (2.5Y 6/1) Fe depletions, very few fine 10YR 5/6 Fe concentrations common fine and medium CaCO3 concentrations common fine CaCO3 concentrations, less than in Ck1, faint sedimentary laminations (15 mm thick) few fine 10YR 5/6 Fe concentrations

gradual

light olive brown (2.5Y 5/3) light olive brown (2.5Y 5/3) light olive brown (2.5Y 5/3)
silty clay loam heavy silt loam silt loam

120135

light olive brown (2.5Y 5/3) light olive brown (2.5Y 5/3)

silt loam

135200
silt loam, high in very fine sand

none in matrix

200235
light olive brown (2.5Y 5/3)

massive

Description kindly provided by Joe Mason, University of Wisconsin.
suggest magnetic parameters that might yield reliable proxies for paleoclimate reconstructions.
2. Site description The site was sampled with a 7.6-cm (3-in.)diameter hydraulic soil probe and described by Dr. J. Mason in 1999. The dried cores were subsampled in 2000, and samples were packed tightly into weakly diamagnetic plastic cubes for magnetic analyses. Core 4G-99a was collected from a broad (~130 m) ridge top in Boone County, Nebraska (41.51918N, 98.21278W; Fig. 1). Uplands in this region of Nebraska are blanketed by approximately 120 m of Wisconsinan Peoria loess underlain by Illinoian Loveland loess. Some areas have been influenced by minor Holocene loess deposition. Presettlement vegetation was prairie. For the period between 1961 and 1990, mean precipitation was 710 mm/year and mean annual temperature was 8.78C (minimum=1.58C, maximum=15.98C) [29]. The area around the core is mapped as Hord silt loam, and classified at the time of the survey as a fine-silty, mixed, mesic Pachic Haplustoll [30]. The profile horizination is Ap-A-Bt1-Bt2-Bt3-BCCk1-Ck2-C (see also Fig. 4). Textures are silt loam in the Ap and Bt3-BC-Ck1-Ck2-C horizons with A, Bt1 and Bt2 horizons being silty clay loam. Ap and A horizons have granular structure and very dark gray (10YR3/1) colors. The C horizon is massive; remaining horizons have subangular blocky structure that increases in size with depth. Bt horizons are light olive brown (2.5Y5/3) with clay coatings. There are very few b2 mm distinct (2.5 Y6/1) and prominent (10YR5/6) mottles in the BC (90120 cm) and C (N200 cm) horizons, which are light olive brown (2.5Y5/3). The Ck horizons are light olive brown (2.5 Y5/3) and have b5 mm carbonate masses. A complete description of site 4G-99a is given in Table 1.

3. Methods We characterized the magnetic mineralogy and abundance of ultrafine (b0.01 Am) (super)paramagnetic particles through thermal demagnetization of isothermal remanent magnetization (IRM), acquired
Fig. 2. (a,b) Thermal demagnetization curves of low-temperature IRM, acquired at 268.158C in 2.5 T after cooling the samples from room temperature in a magnetic field of 2.5 T (dashed gray) and after cooling in zero field (solid black). Shown are the results from 10 and 200 cm depth. The drop in magnetic remanence at low temperatures is due to the thermal demagnetization of (super)paramagnetic minerals. (b) Zero-field cooled demagnetization curves, normalized by the magnetic moment at room temperature after demagnetization. Both curves are rather similar and show a marked drop in magnetic moment at low temperatures due to thermal unblocking of (super)paramagnetic particles. The Verwey transition, diagnostic of magnetite, is barely visible near 1558C.
at 5 K (268.158C) for several samples [31]. These analyses were performed on a Quantum Design Magnetic Measurement System (MPMS 5s). Curie temperature determinations aided in the characterization of the magnetic mineralogy [32] and were performed in an inert gas atmosphere on a AGICO Kappabridge KLY-2 with modified furnace apparatus. To estimate the concentrations of remanence carrying magnetic minerals, we acquired IRM at DC fields of 100 and 1300 mT and at backfields of 300 and 600 mT (Table 2). We distinguished between highand low-coercivity minerals by calculating bhardQ IRM (hIRM=0.5 (IRM1300 mT+IRMbackfield)) and bsoftQ IRM (sIRM=0.5 (IRM mT IRM b ack fiel d )) [33,34]. Measurements for hIRM and sIRM were repeated five times per sample and the resulting standard deviation is shown in Fig. 4g. Anhysteretic remanent magnetization (ARM) was acquired in a peak AF-field of 100 mT and a bias field of 50 AT, using a DTech 100 mT 2000 AF demagnetizer. The ratio of ARM/IRM100 mT was used as a proxy for the relative
abundance of single domain (SD) magnetic particles [35,36]. The relative abundance of low- and highcoercivity minerals was estimated from S-ratios (S 300=IRM300 mT/IRM1300 mT) [37]. All remanence parameters were measured in a cryogenic magnetometer (2G, Model 760-R) with a nominal sensitivity of 21011 A m2. Low-field susceptibility (v) was measured using a Kappabridge KLY-2 susceptibility meter with a nominal sensitivity of 4108 SI. Frequency-dependent susceptibility (v fd), a proxy for the presence of superparamagnetic (SP) particles [38], was measured at 20 frequencies between 40 and 4000 Hz. Frequency values for 400 and 4000 Hz were obtained from a best fit through the entire data set to calculate v fd=(v 400 Hzv 4000 Hz)/v 400 Hz. The errors associated with the measurement and shown in Fig. 4c were calculated from the error of the best fit line and the resulting error for v 400 Hz and v 4000 Hz. Hysteresis loops were measured for selected samples in a maximum field of 1300 mT using a vibrating sample magnetometer (Princeton Applied Research, modified by the Institute for Rock

Fig. 3. Curie temperature (T C) measurement for a sample from 20 cm depth. A Curie temperature of 5708C suggests a ferrimagnetic component close to magnetite as the principal magnetic mineral. The increase in susceptibility upon heating is likely due to the reduction of weakly magnetic antiferromagnetic minerals, such as goethite or hematite, to magnetite.
Table 2 Magnetic properties for samples from site 4G-99a Depth (cm) v if (m3/kg106) 0.81 0.84 0.83 0.76 0.76 0.75 0.56 0.55 0.50 0.54 0.50 0.46 0.46 0.48 0.43 0.45 0.46 0.50 0.48 0.43 0.48 0.46 0.47 0.47 0.39 0.46 0.54 0.48 0.43 0.47 0.48 0.43 0.48 0.42 0.44 0.42 0.53 0.49 ARM (mA m2/kg) 0.18 0.17 0.19 0.17 0.16 0.16 0.11 0.11 0.09 0.09 0.08 0.07 0.07 0.07 0.06 0.06 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.05 0.06 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.06 IRM100 mT (mA m2/kg) 6.10 5.93 4.99 4.78 5.20 5.21 3.91 3.86 3.54 3.86 3.65 3.35 3.46 3.64 3.23 3.47 3.51 3.96 3.50 3.52 3.33 3.73 3.77 3.81 3.19 3.71 4.11 4.08 3.35 3.61 3.76 3.45 3.72 3.29 3.35 3.31 4.15 3.78 ARM/ IRM100 (%) 2.97 2.88 3.72 3.61 3.13 3.03 2.90 2.85 2.54 2.42 2.29 2.07 2.02 1.93 1.95 1.77 1.85 1.71 1.86 1.62 1.94 1.56 1.60 1.59 1.60 1.63 1.76 1.64 1.65 1.72 1.61 1.60 1.65 1.72 1.68 1.69 1.58 1.61
v fd (%) 4.0F1.2 4.0F1.2 4.4F1.2 4.0F1.8 3.8F1.5 3.4F1.2 3.2F1.5 2.1F2.1 1.2F1.9 1.0F1.8
Ms (mA m2/kg) 58.70 61.90 55.20 52.00 48.90 29.00 41.80 37.70 34.70 32.20 33.00 34.70 31.90 32.10 32.10 33.50 33.00 35.70 36.80 34.50
M rs (mA m2/kg) 6.01 6.53 5.95 5.40 5.30 2.92 4.22 3.84 3.66 3.39 3.40 3.59 3.32 3.62 3.47 3.48 3.51 3.68 3.86 3.66
Hc (mT) 8.4 8.21 8.51 8.72 8.26 8.36 8.32 8.61 8.61 8.66 8.69 8.77 7.82 8.74 8.81 8.78 8.74 8.66 8.69

0.5F1.9

1.1F1.3

0.4F1.9

Magnetism) to estimate bulk magnetic particle size in a plot of saturation remanent magnetization/ saturation magnetization vs. coercivity of remanence/coercive force ( J rs/J s vs. H cr/H c) [39,40], and to correct for the v hf presence of para- or diamagnetic minerals using the high-field slope v hf of the hysteresis loop [32]. Concentrations of major and trace elements were measured for three magnetically enhanced horizons and two samples from the unaltered parent material using an Inductively

Coupled Plasma Mass Spectrometer (ICP-MS, Perkin-Elmer/Sciex Elan 5000).
4. Results Thermal demagnetization curves of low-temperature IRM show a distinct drop at temperatures below 2508C for all samples (Fig. 2), which indicates that para- and superparamagnetic minerals are abundant
H cr (mT) 25.85 26.79 26.12 45.09 26.22 27.16 26.86 27.05 26.3 27.07 27.39 27.16 26.4 28.31 28.51 28.26 28.49 28.28 28.17
v hf (m3/kg) 3.76E08 3.90E08 3.83E08 4.77E08 4.69E08 5.61E08 5.74E08 5.73E08 5.33E08 5.41E08 5.60E08 5.26E08 5.44E08 5.31E08 5.16E08 5.06E08 4.70E08 4.92E08 5.01E08

J rs/J s

H cr/H c
hIRM300 mT (mA m2/kg) 0.27F0.01 0.27F0.01 0.26F0.01 0.23F0.02 0.20F0.02 0.20F0.01 0.18F0.01 0.17F0.01 0.16F0.01
hIRM600 mT (mA m2/kg) 0.13F0.03 0.15F0.03 0.13F0.01 0.13F0.01 0.13F0.02 0.12F0.03 0.12F0.02 0.10F0.01 0.11F0.005 0.10F0.01 0.11F0.005 0.11F0.01 0.10F0.01 0.10F0.01 0.11F0.02 0.11F0.005 0.11F0.005 0.11F0.01 0.11F0.005 0.11F0.02
sIRM300 mT (mA m2/kg) 7.30F0.02 7.54F0.02 6.92F0.04 6.05F0.01 5.23F0.01 4.96F0.01 4.20F0.01 4.15F0.01 4.12F0.01
0.102 0.105 0.108 0.104 0.108 0.101 0.101 0.102 0.105 0.105 0.103 0.103 0.104 0.113 0.108 0.104 0.106 0.103 0.105 0.106
3.077 3.263 3.069 5.171 3.174 3.249 3.228 3.142 3.055 3.126 3.152 3.097 3.376 3.239 3.236 3.219 3.260 3.266 3.242
0.927F0.003 0.934F0.003 0.925F0.003 0.923F0.006 0.922F0.006 0.016F0.005 0.919F0.003 0.919F0.003 0.917F0.006

0.917F0.007

0.18F0.02

4.37F0.02

0.911F0.009

0.19F0.2

4.54F0.01

0.924F0.006

0.16F0.01

4.22F0.01

0.920F0.011

0.15F0.03

3.92F0.03

0.919F0.018

0.17F0.05

4.84F0.05

throughout the soil profile. In absolute terms, the loss in remanence at low temperatures is greatest in the magnetically enhanced horizons (Fig. 2a and b), consistent with a slight increase in SP particles deduced from an increase in frequency-dependent susceptibility (Fig. 4c). In relative terms, however, the concentration of (super)paramagnetic particles remains fairly constant (Fig. 2c), which is due to the simultaneous increase of remanence carrying SD particles in the magnetically enhanced horizons and a

loss of paramagnetic minerals suggested by a decrease in v hf (Fig. 4i). Curie temperatures between 550 and 5808C (Fig. 3) suggest that a ferrimagnetic mineral close to magnetite with probably minor Ti substitution (xb0.1) dominates the magnetic properties of the soil at ambient and high temperatures. Weak Verwey transitions confirm this interpretation (Fig. 2). Increases in concentration-dependent parameters (v, IRM) show that the A horizon is enhanced in magnetic minerals (Fig. 4a and b). The additional
362 C.E. Geiss et al. / Earth and Planetary Science Letters 228 (2004) 355367 Fig. 4. Summary of magnetic properties for site 4G-99a. (a) Magnetic susceptibility (v) and (b) isothermal remanent magnetization (IRM) are proxies for the abundance of magnetic minerals. (c) Frequency-dependent susceptibility (v fd) is a proxy for the relative abundance of ultrafine (b0.01 Am) superparamagnetic particles. (d) ARM/IRM is a proxy for the abundance of fine (0.010.1 Am) single-domain particles. (e) Hcr /Hc characterizes bulk magnetic grain size. (f) S-ratios quantify the relative abundance of low-coercivity ferrimagnetic minerals, while (g) bhardQ IRM (hIRM) quantifies the absolute abundance of high-coercivity antiferromagnetic minerals, such as hematite or goethite. Solid symbols: hIRM300 mT, open symbols: hIRM600 mT.d (h) bSoftQ IRM (sIRM) is used to estimate the presence of low-coercivity ferrimagnetic minerals. (i) High-field susceptibility tracks the abundance of paramagnetic Fe-bearing minerals.
C.E. Geiss et al. / Earth and Planetary Science Letters 228 (2004) 355367 Table 3 Concentration of iron and selected heavy metals for selected samples from site 4G-99a
Depth (cm) 220 Fe2O3 (%) 3.01F0.05 3.48F0.02 4.15F0.05 3.63F0.06 3.68F0.08 Cu (ppm) 22.18F0.44 24.39F0.80 28.50F1.16 26.51F0.85 28.97F0.89 Zn (ppm) 83.22F1.87 89.21F2.42 97.36F3.10 92.99F1.87 92.91F1.44 Pb (ppm) 17.80F0.42 18.96F0.46 19.44F0.65 17.21F0.30 17.47F0.14
pedogenic material is likely fine-grained (SD and smaller) leading to higher values of v fd and ARM/ IRM100 mT in the upper soil horizons (Fig. 4c and d). This shift to smaller, remanence carrying ferrimagnetic grains is quite small as maximum ARM/IRM100 mT ratios reach only slightly above 0.03. Hysteresis parameters (only H cr/H c shown, Fig. 4e), which are a measure of the average magnetic grain size, confirm this interpretation, showing only slight changes towards finer grains in the magnetically enhanced horizon. S-ratios between 0.9 and 0.95 (Fig. 4f) agree with the results of high- and low-temperature analyses, indicating that the magnetic component is magnetically dominated by low-coercivity (titano) magnetite or maghemite. S-ratios increase slightly in the A horizon, which suggests that pedogenetically produced magnetic signal is caused predominantly by magnetite or maghemite. The abundance of highcoercivity hematite or goethite has been estimated from hIRM measurements (Fig. 4g). hIRM increases in the upper soil horizon, which is either due to the direct generation of hematite or goethite, or the relatively rapid oxidation of pedogenic ferrimagnets (magnetite, maghemite) to antiferromagnetic minerals (hematite, goethite). We estimated the abundance of ferri- and antiferromagnetic minerals from measurements of hIRM and sIRM. Hysteresis measurements show that all samples have ratios of J rs/J sc0.1. Assuming that hysteresis properties are dominated by a component close to magnetite, we use a saturation remanence of J mt=0.1 J sc0.190 A m2/kg to estimate the amount of ferrimagnetic material based on observed values of sIRM. We further assume that the antiferromagnetic component consists of hematite or goethite with a saturation magnetization of J sc0.1 A m2/kg [41].

Since this component is likely finer-grained, which leads to higher ratios of J rs/J s, we approximate its saturation remanence as J hemc0.5 J s=0.05 A m2/kg. Using the values for hIRM and sIRM (both measured for backfields of 300 mT) shown in Fig. 4g, we obtain concentrations ranging between 1 and 0.6 mg/cm3 for ferrimagnetic minerals and concentrations ranging between 7 and 4 mg/cm3 for antiferromagnetic minerals. High-field susceptibility (v hf) measurements (Fig. 4i) show that the concentration of paramagnetic minerals remains relatively constant throughout the B horizon, but drops significantly for the magnetically enhanced upper part of the profile. ICP-MS analyses of iron and heavy metal concentrations are shown in Table 3. Total iron concentrations as well as the concentrations of copper, zinc and lead all show a decrease in the magnetically enhanced soil horizons.
5. Discussion Our study of a modern loessic soil profile from Nebraska shows a distinct magnetic enhancement in the upper soil horizon, similar to profiles in loess elsewhere in temperate regions. Preliminary, unpublished data further confirm that site 4G-99a is fairly typical for loessic soils of the midwestern United States. The magnetic enhancement is limited to the upper (Ap and A) soil horizons, though changes in grain size proxies extend farther down into the underlying Bt horizon to a maximum depth of approximately 50 cm. The added pedogenic magnetic component consists of fine-grained (SP and SD) ferri- and antiferromagnetic minerals, leading to increases in all concentration-dependent parameters (v, ARM, IRM, J s), as well as increases in grainsize-dependent magnetic parameters diagnostic of fine-grained magnetic material, such as v fd and ARM/IRM100 mT. Numerous coal-fired power plants exist in the western United States providing a potential source of magnetic fly-ash to our site. Site 4G-99a is approximately 100 km northeast of the nearest coal-fired power plant (Gothenburg NE) and more than 450 km downwind of the coal-fired plants located in eastern Wyoming. We did not observe any opaque spherules
in the uppermost soil horizon and concentrations of heavy metals, such as Cu, Zn or Pb do not increase towards the top of the profile, indicating the absence of significant levels of atmospheric pollution [42]. Furthermore, significant contamination of the soil profile by magnetic fly-ash seems unlikely since the magnetic enhancement is expressed in a shift towards finer SD and SP particles rather than coarse (MD) flyash spherules. The site has not been plowed for several decades and any recent anthropogenic contamination would be expected in the upper soil horizons, while the magnetic enhancement signal is found to a depth of approximately 40 cm. The site is well drained and located in a stable upland position, making it ideally suited to study pedogenic effects on soil magnetic properties. The few redoximorphic features in the BC (90120 cm) and C (N200 cm) horizons do not significantly affect the overall magnetic properties. 5.1. Mineralogy of the remanence carrying component Even though most magnetic parameters are dominated by low-coercivity ferrimagnets (magnetite or maghemite), antiferromagnetic minerals (hematite or goethite) are the dominant remanence carrying component by mass. Our rough estimate of both components makes several assumptions regarding the grain size and resulting saturation remanence of the two components. Our assumption of an SD antiferromagnetic component with a J rs/J r=0.5 may underestimate its concentration. However, since hematite and goethite are likely produced in situ or during (eolian) transport, their grain size is likely rather small and below the lower grain size limit of 0.1 mm for MDhematite [43]. We are likely to underestimate particles near the SPSD boundary (c0.003 Am for hematite and magnetite [32]) and all (super)paramagnetic components, which might represent up to 70% of the iron-bearing fraction, based the observed loss of remanence between 275 and 2008C in low-T analyses (Fig. 2). Mfssbauer analysis might allow further quantification of the ferri- and antiferromagnetic components. The identification of nanocrystalline, ionically substituted iron-components, however, is difficult and might not improve significantly our current estimates. Our simple estimates show that at

least 90% of the remanence carrying component exists in the form of hematite or goethite, rather than magnetite or maghemite. 5.2. Constraints on pathways of magnetic enhancement Our analyses can be used to shed some light on possible pathways of magnetic enhancement. Compaction of the top soil horizons can be ruled out since the enhancement is clearly expressed in massnormalized magnetic susceptibility and IRM, which are independent of changes in density and hence compaction. The concentration of magnetic minerals through lessivage or removal of carbonates [17] is unlikely as well. Our soil descriptions show that clay increase in the Bt horizon is slight, and carbonates are leached completely from the profile to a depth of 130 cm, approximately 90 cm below the highest values of v and IRM. The addition of Holocene Bignell loess and hence changes in the magnetic properties of the parent material are likely to influence the studied profile to some extent. Bignell loess is not distinguished as a separate sedimentary unit in our site as pedogenesis was able to keep up with loess deposition rates, creating a welded soil profile. The good correlation of magnetic properties with soil horizination, however, argues in favor of a pedogenic origin for the magnetic enhancement signal rather than a simple change in parent material. Magnetic enhancement is also not due to conversion of weakly magnetic antiferromagnetic minerals into strongly magnetic ferrimagnetic minerals, as both increase in concentration in the upper soil horizons. It is possible, however, that paramagnetic minerals are converted into ferri- and antiferromagnetic minerals, which is suggested by the v hf distinct drop in P in the Ap and A horizons (Fig. 4i).
6. Suggestions for paleoclimate reconstructions Many studies (e.g., [4446]) agree that the magnetic properties of loess-paleosol sequences are dominated by coarse ferrimagnets in the PSD to MD (N1 Am) grain-size range, but that the pedogenetic magnetic component, which is responsible for the observed
magnetic enhancement, is composed of fine- to ultrafine (b0.1 Am) magnetite or maghemite (e.g., [9,10,47,48]). Antiferrimagnetic minerals, such as goethite or hematite, may comprise the bulk of the iron-bearing component of the sediment, and changes in their mineralogy may be related climatic conditions during soil formation [49], but, due to their low saturation magnetization, these minerals contribute little to the bulk magnetic signal. For these reasons, we think that the magnetic proxies most useful for transfer function between climate and magnetic soil properties should be the ones that selectively respond to this change in magnetic grain size. Low-field magnetic susceptibility (v), which is commonly used as a paleoprecipitation proxy, has been shown to work well in well drained, buffered and unpolluted loessic soils that contain sufficient amounts of iron to allow for the formation of pedogenic magnetic minerals (e.g., [4,18,48,50,51]). However, v is relatively insensitive to SD particles, while it depends on a variety of other components, such as the abundance and mineralogy of paramagnetic minerals, or grain size and mineralogy changes in the parent material. For these reasons, we expect susceptibilitybased paleoclimate reconstructions to perform poorly in areas like the midwestern United States, where loessic parent material is derived from different sources [52]. Enhancement parameters based on changes in ARM, ARM/IRM ratios, and frequency-dependent susceptibility (v fd) focus on this (ultra)fine-grained magnetic component and therefore offer some promise as a useful paleoprecipitation proxy. ARM, or ARM its equivalent field-normalized ARM (v ARM), has been measured for several soil sites in the United States [53], Asia [4] and on the Chinese loess plateau (e.g., [31,54]), and Liu et al. [55] use it extensively to estimate grain size variations between loess and paleosol units. Basing a paleoclimate proxy on frequency-dependent susceptibility might have several disadvantages. The SP signal derived from v fd measurements is weak and the nanocrystalline particles can be easily altered by weathering processes. Furthermore, accurate determinations of v fd are time-consuming, and the results should ideally be corrected for paramagnetic contributions. It may therefore not be feasible to apply v fd measurements to large sample sets as it will be necessary to reconstruct regional climate

change. Measurements of time-dependent IRM [56] might allow for the rapid quantification of grains near the SPSD boundary. Time-dependent IRM is diagnostic of grain sizes slightly larger than these quantified by v fd, and both techniques yield comparable results for a wide range of well-characterized (natural and synthetic) samples. Presently, ARM or ARM/IRM might be a better foundation for paleoclimate reconstructions based on soil magnetic properties. They are easily and rapidly measured and reflect (slightly) larger SD particles, which may better withstand long-term weathering. IRM acquisition curves, S-ratios, or measurements of hIRM and sIRM may also yield useful information regarding the relative abundance and possibly mineralogy [34,57] of the high-coercivity antiferromagnetic component goethite and hematite. Combined with measurements of soil color, these magnetic parameters may yield another parameter to quantify soil moisture and temperature [49].
Acknowledgments We would like to thank Joe Mason for his help in the field and for providing the samples and soil description for this study. Insightful reviews by B. Maher and F. Lagroix helped to enhance the quality of the manuscript. Part of the analyses were performed at the Institute for Rock Magnetism at the University of Minnesota which is funded by the W.M. Keck foundation, the National Science Foundations Earth Science Divisions Instrumentation and Facilities Program and the University of Minnesota. This is IRM publication number 0409. The study was supported by NSF grant 9909523 to SKB and CEG. During her stay at the University of Minnesota, JM was supported by an NSF REU fellowship.

References

[1] G. Kukla, The mystery of the Chinese magnetic dust, LamontDoherty Geol. Obs. Yearbook, 1988. [2] B.A. Maher, R. Thompson, Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in the Chinese loess and paleosols, Quat. Res. 44 (1995) 383 391. [3] F. Heller, C.D. Shen, J. Beer, T.S. Liu, A. Bronger, M. Suter, G. Bonani, Quantitative estimates of pedogenic ferromagnetic
C.E. Geiss et al. / Earth and Planetary Science Letters 228 (2004) 355367 mineral formation in Chinese loess and paleoclimatic implications, Earth Planet. Sci. Lett. 114 (1993) 385 390. B.A. Maher, A. Alekseev, T. Alekseeva, Variations of soil magnetism across the Russian steppe: its significance for use of soil magnetism as a paleorainfall proxy, Quat. Sci. Rev. 21 (2002) 1571 1576. Y. Tang, J. Jia, X. Xie, Records of magnetic properties in Quaternary loess and its paleoclimatic significance: a brief review, Quat. Int. 108 (2003) 33 50. E. Le Borgne, Influence du feu sur les proprietes magnetiques du sol et sur celles du schiste et du granite, Ann. Geophys. 16 (1960) 159 195. G. Kletetschka, S.K. Banerjee, Magnetic stratigraphy of Chinese loess as a record of natural fires, Geophys. Res. Lett. 22 (11) (1995) 1241 1343. H. Lu, T. Liu, Z. Gu, B. Liu, L. Zhou, J. Han, N. Wu, Effect of burning C3 and C4 plants on the magnetic susceptibility signal in soils, Geophys. Res. Lett. 27 (2000) 2013 2016. X.M. Meng, E. Derbyshire, R.A. Kemp, Origin of the magnetic susceptibility signal in Chinese loess, Quat. Sci. Rev. 16 (1997) 833 839. B.A. Maher, R.M. Taylor, Formation of ultrafine-grained magnetite in soils, Nature 336 (1988) 368 370. D.R. Lovley, Organic matter mineralization with the reduction of ferric iron: a review, Geomicrobiology 5 (1987) 375 399. F. Heller, T.S. Liu, Magnetism of Chinese loess deposits, Geophys. J. R. Astron. Soc. 77 (1984) 125 141. M.J. Singer, P. Fine, Pedogenic factors affecting magnetic susceptibility of northern California soils, Soil Sci. Soc. Am. J. 53 (1989) 1119 1127. X.M. Liu, P. Hesse, T. Rolph, J.E. Beget, Properties of magnetic mineralogy of Alaskan loess: evidence for pedogenesis, Quat. Int. 62 (1999) 93 102. N. Bigelow, J.E. Beget, R. Powers, Latest Pleistocene increase in wind intensity recorded in eolian sediments from central Alaska, Quat. Res. 34 (1990) 160 168. M.E. Evans, N.W. Rutter, N. Catto, J. Chlachula, D. Nyvlt, Magnetoclimatology: teleconnection between the Siberian loess record and North Atlantic Heinrich events, Geology 31 (2003) 537 540. M.J. Singer, K.L. Verosub, P.T. Fine, J. Tenpas, A conceptual model for the enhancement of magnetic susceptibility in soils, Quat. Int. 3436 (1996) 243 248. B.A. Maher, Magnetic properties of modern soils and Quaternary loessic paleosols: paleoclimatic implications, Palaeogeogr. Palaeoclimatol. Palaeoecol. 137 (1998) 25 54. B.A. Maher, R. Thompson, L.P. Zhou, Spatial and temporal reconstructions of changes in the Asian paleomonsoon: a new mineral-magnetic approach, Earth Planet. Sci. Lett. 125 (1994) 461 471. B.A. Maher, H.-M. Yu, H.M. Roberts, A.G. Wintle, Holocene loess accumulation and soil development at the western edge of the Chinese Loess Plateau: implications for magnetic proxies of paleorainfall, Quat. Sci. Rev. 22 (2003) 445 451. M.J. Singer, P. Fine, K.L. Verosub, O.A. Chadwick, Time dependence of magnetic susceptibility of soil chronosequences on the California coast, Quat. Res. 37 (1992) 323 332. N.J. Vidic, M.J. Singer, K.L. Verosub, Duration dependence of magnetic susceptibility enhancement in the Chinese loesspaleosols of the past 620 ky, Palaeogeogr. Palaeoclimatol. Palaeoecol. 211 (2004) 271 288. E.A. Oches, S.K. Banerjee, Rock-magnetic proxies of climate change from loess-paleosol sediments of the Czech Republic, Stud. Geophys. Geod. 40 (1996) 287 300. J. Nawrocki, Magnetic susceptibility of Polish loesses and loess-like sediments, Geol. Carpath. 43 (1992) 179 180. P.A. Vlag, P.A. Solheid, E.A. Oches, S.K. Banerjee, The paleoenvironmental-magnetic record of the Gold Hill Steps loess section in central Alaska, Phys. Chem. Earth 24 (1999) 779 783. W.C. Johnson, K.L. Willey, Isotopic and rock magnetic expression of environmental change at the Pleistocene Holocene transition in the central Great Plains, Quat. Int. 67 (2000) 89 106. D.A. Grimley, L.R. Follmer, E.D. McKay, Magnetic susceptibility and mineral zonations controlled by provenance in loess along the Illinois and Central Mississippi river valleys, Quat. Res. 49 (1998) 24 36. D.R. Muhs, E.A. Bettis III, J. Been, J.P. Geehin, Impact of climate and parent material on chemical weathering in loessderived soils of the Mississippi river valley, Soil Sci. Soc. Am. J. 65 (2001) 1761 1777. HPRCC, High Plains Regional Climate Center: Historical data summaries, http://www.hprcc.unl.edu/products/historical.htm, pp. Developed by the Western Regional Climate Center, Accessed January 16, 2004, Lincoln, NE, 2004. C.L. Hammond, C.F. Mahnke, L. Brown, R. Schulte, W. Russell, Soil survey of Boone County, Nebraska, Unites States Department of Agriculture, Soil Conservation Service, Washington, DC, 1972. C.P. Hunt, S.K. Banerjee, J. Han, P.A. Solheid, E. Oches, W. Sun, T. Liu, Rock-magnetic proxies of climate change in the loess-paleosol sequences of the western loess plateau of China, Geophys. J. Int. 123 (1995) 232 244. D.J. Dunlop, O. Ozdemir, Rock Magnetism, Fundamentals and Frontiers, Cambridge University Press, Cambridge, 1997 (573 pp.). R.L. Reynolds, E. Callender, A. Goldin, J.G. Rosenbaum, P. Van Metre, M. Tuttle, Greigite (Fe3S4) as an indicator of droughtthe 19121994 sediment magnetic record from White Rock Lake, Dallas, Texas, USA, J. Paleolimnol. 21 (1999) 193 206. R. Egli, Characterization of individual rock magnetic components by analysis of remanence curves: 1. Unmixing natural sediments, Stud. Geophys. Geod. 48 (2004) 391 446. J. King, S.K. Banerjee, J. Marvin, O. O zdemir, A comparison of different magnetic methods for determining the relative grain size of magnetite in natural materials: some results from lake sediments, Earth Planet. Sci. Lett. 59 (1982) 404 419. C.P. Hunt, B.M. Moskowitz, S.K. Banerjee, Magnetic properties of rocks and minerals, Rock Physics and Phase Relations.

[24] [25]

[10] [11] [12] [13]
C.E. Geiss et al. / Earth and Planetary Science Letters 228 (2004) 355367 A Handbook of Physical Constants, AGU Reference Shelf, vol. 3, 1995, pp. 189 204. J. Bloemendal, J.W. King, F.R. Hall, S.-H. Doh, Rock magnetism of Late Neogene and Pleistocene deep-sea sediments: relationship to sediment source, diagenetic processes, and sediment lithology, J. Geophys. Res. 97 (1992) 4361 4375. H.-U. Worm, On the superparamagnetic-stable single domain transition for magnetite, and frequency dependence of susceptibility, Geophys. J. Int. 133 (1998) 201 206. R. Day, M. Fuller, V.A. Schmidt, Hysteresis properties of titanomagnetites: grain-size and compositional dependence, Phys. Earth Planet. Inter. 13 (1977) 260 267. D.J. Dunlop, Hysteresis properties and their dependence on particle size: a test of pseudo-single-domain remanence models, J. Geophys. Res. 91 (B9) (1986) 9569 9584. H.C. Soffel, Palaomagnetismus und ArchuEomagnetismus, Springer, Berlin, 1991, 276 pp. J.A. Dearing, Holocene environmental change from magnetic proxies in lake sediments, in: B.A. Maher, R. Thompson (Eds.), Quaternary Climates, Environments and Magnetism, Cambridge University Press, Cambridge, 1999, pp. 231 278. G. Kletetschka, P.J. Wasilewski, Grain size limit for SD hematite, Phys. Earth Planet. Inter. 129 (2002) 173 179. M.E. Evans, F. Heller, Magnetic enhancement and paleoclimate: study of a loess/paleosol couplet across the loess plateau of China, Geophys. J. Int. 117 (1994) 257 264. F. Florindo, R. Zhu, B. Guo, L. Yue, Y. Pan, F. Speranza, Magnetic proxy climate results from the Duanjiapo loess section, southernmost extremity of the Chinese loess plateau, J. Geophys. Res. 104 (1999) 645 659. X. Li, T. Liu, M. Chen, B.A. Maher, S.K. Banerjee, Origin of magnetic minerals and magnetic susceptibility variations in Chinese loess, in: T. Liu (Ed.), Loess Environment and Global Changes, Science Press, China, 1991.

[41] [42]

[43] [44]
[47] L.P. Zhou, F. Oldfield, A.G. Wintle, S.G. Robinson, J.T. Wang, Partly pedogenic origin of magnetic variations in Chine loess, Nature 346 (1990) 737 739. [48] F. Heller, X. Liu, T. Liu, T. Xz, Magnetic susceptibility of loess in China, Earth Planet. Sci. Lett. 103 (1991) 301 310. [49] U. Schwertmann, Occurrence and formation of iron oxides in various pedoenvironments, in: J.W. Stucki, B.A. Goodman, U. Schwertmann (Eds.), Iron in Soils and Clay Minerals, Nato ASI Series, Reidel Publishing, Dordrecht, 1988, pp. 276 308. [50] G. Kukla, F. Heller, X.M. Liu, T.C. Xu, T.S. Liu, Z.S. An, Pleistocene climates in China dated by magnetic susceptibility, Geology 16 (1988) 811 814. [51] B.A. Maher, R. Thompson, Paleoclimatic significance of the mineral magnetic record of the Chinese loess and paleosols, Quat. Res. 37 (1992) 155 170. [52] E.A. Bettis III, D.R. Muhs, H.M. Roberts, A.G. Wintle, Last glacial loess in the conterminous USA, Quat. Sci. Rev. 22 (2003) 1907 1946. [53] O. Ozdemir, S.K. Banerjee, A preliminary magnetic study of soil samples from west-central Minnesota, Earth Planet. Sci. Lett. 59 (1982) 393 403. [54] B.A. Maher, R. Thompson, Mineral magnetic record of the Chinese loess and paleosols, Geology 19 (1991) 3 6. [55] Q.S. Liu, S.K. Banerjee, M.J. Jackson, B.A. Maher, Y.X. Pan, R.X. Zhu, C.L. Deng, F.H. Chen, Grain sizes of susceptibility and anhysteretic remanent magnetization carrieres in Chinese loess/paleosol sequences, J. Geophys. Res. 109 (2004) B03101. [56] H.U. Worm, Time dependent IRM: a new technique for magnetic granulometry, Geophys. Res. Lett. 26 (1999) 2557 2560. [57] D. Heslop, M.J. Dekkers, P.P. Kruiver, I.H.M. van Oorschot, Analysis of isothermal remanent magnetization acquisition curves using the expectation-maximization algorithm, Geophys. J. Int. 148 (2002) 58 64.