late quaternary climate change, loess sedimentation, and ...€¦ · ondary carbonate appears to be...

For permission to copy, contact [email protected]© 2007 Geological Society of America

ABSTRACT

Research on soil genesis often assumes a “top-down” model, in which the soil profi le develops downward from a stable land sur-face. This model is inapplicable to upland landscapes affected by frequent dust deposi-tion, where soils grow upward as they develop. On the central Great Plains, late Quaternary loess sections proximal to immediate source areas contain the Brady Soil, a prominent marker separating late Pleistocene Peoria Loess from Holocene Bignell Loess. Farther from immediate dust source areas, the Brady Soil and Bignell Loess are not recognizable in the fi eld. On loess tablelands in these dis-tal regions, surface soils typically contain a prominent, clay-rich B horizon below a thick silty A horizon. Assuming top-down pedogen-esis, this could be interpreted as a postglacial soil profi le formed in Peoria Loess, with the B horizon produced by weathering and clay illuviation. We propose a strikingly different interpretation, in which the upper B horizon at distal sites is the Brady Soil A horizon that has been transformed by burial, organic mat-ter loss, and modern subsoil structure forma-tion processes. The overlying modern A hori-zon represents Bignell Loess. Properties of the Brady Soil at proximal sites (a distinctive bur-rowed zone, high clay content and TiO2/ZrO2, and low volcanic glass content) can be traced to the B horizon in distal soils. A decrease in smectite abundance above the Brady Soil at proximal sites is identifi able at the top of the clay-rich B horizon in distal soils. The spatial variation of clay content in loess and soil hori-zons is best explained by eolian sedimenta-tion patterns. The higher clay content in the Brady Soil and distal B horizons defi nes a fi ne-grained zone that represents a late phase

of Peoria Loess accumulation. Evidence of chemical weathering is minimal, and illuvial clay is rare to absent in the clay-rich B hori-zons. Illuvial clay does often occur deep in the solum and is related to the depth below the top of the Brady Soil. The depth of occur-rence of illuvial clay is not related to modern climate parameters, although depth to sec-ondary carbonate appears to be in equilib-rium with modern climate. Upland soils in the central Great Plains are composite soils; their properties are the result of a pedosedi-mentary history linked to regional climate change that has infl uenced sedimentation and pedogenesis since the late Pleistocene.

Keywords: Great Plains, soils, stratigraphy, loess, Holocene, paleosols.

INTRODUCTION

Research on soil genesis often implicitly or explicitly assumes a “top-down” model, in which the soil profi le develops downward from a stable land surface after the parent material is deposited or exposed by erosion. In fact, the time factor in Hans Jenny’s well-known model of soil formation is defi ned relative to a specifi c t

0, the

time when sedimentary or erosional processes were replaced by pedogenesis (Jenny, 1941). This conceptual model is inapplicable to upland landscapes affected by frequent dust deposition, where soils grow upward as they develop, under the competing infl uences of pedogenesis and sedimentation (McDonald and Busacca, 1990; Kemp et al., 1995, 2001; Kemp and Derbyshire, 1998; Almond and Tonkin, 1999). This soil development process has also been described as cumulization (Schaetzl and Anderson, 2005).

In this paper, we use evidence from 122 soil profi les, sampled at 41 sites, to develop a model of aggradational pedogenesis on the loess-mantled uplands of the central Great Plains. This model leads to an interpretation of

the soil profi les that is strikingly different from interpretations based on an assumption of top-down pedogenesis in initially uniform loess. We propose that the upward transition from largely unaltered C horizon loess to a clay-rich B hori-zon, overlain by a thick A horizon with lower clay content, is primarily a refl ection of changes in the accumulation rate and grain size of atmo-spheric dust over the late Pleistocene and Holo-cene. Postdepositional pedogenesis does have a signifi cant effect on soil morphology in this setting, but the overall soil horizon sequence is more a stratigraphic record of climatically driven dust infl ux than a product of pedogenesis as traditionally defi ned.

Multiple reconstructions of paleoenviron-ments indicate that the climate of the Great Plains has varied throughout the late Quater-nary and has included episodes of persistent and severe drought, dune activity, and dust transport (Feggestad et al., 2004; Forman et al., 2001; Muhs et al., 1999). In particular, the central Great Plains in Nebraska has a high-resolution record of dune mobilization, loess deposition, and soil formation driven by climate change (Goble et al., 2004; Johnson and Willey, 2000; Mason et al., 2003; Miao et al., 2005). The loess-mantled uplands that are the focus of this paper have largely served as sinks for dust derived from less stable areas upwind. Given this history of recurrent dust infl ux, soils on these uplands must be interpreted in light of the cumulization model and the spatial and temporal interactions between loess sedimentation and pedogenesis.

Loess thins and fi nes systematically with distance from its source, which infl uences the nature and properties of soils developed in the loess deposit, as well as the stratigraphy of loess-paleosol sequences. Studies of loess-derived soils indicate stronger morphological expression, greater clay content and measures of B horizon clay accumulation, increased depth of carbonate leaching, and increased mineral weathering rates with increasing distance from

Late Quaternary climate change, loess sedimentation, and soil profi le development in the central Great Plains: A pedosedimentary model

Peter M. Jacobs†

Department of Geography and Geology, University of Wisconsin–Whitewater, 800 W. Main St., Whitewater, Wisconsin 53190, USA

Joseph A. MasonDepartment of Geography, University of Wisconsin, 550 N. Park St., Madison, Wisconsin 53706, USA

†E-mail: [email protected].

GSA Bulletin; March/April 2007; v. 119; no. 3/4; p. 462–475; doi: 10.1130/B25868.1; 12 fi gures.

462

Climate change, loess sedimentation, and soil profi le development

Geological Society of America Bulletin, March/April 2007 463

the loess source (Caldwell and White, 1956; Hutton, 1951; Mason and Nater, 1994; Muhs et al., 2004; Putman et al., 1988; Ruhe, 1969; Smith, 1942). Most of these effects are attrib-uted to the greater fi ne silt and clay content of distal loess deposits, and more uniform par-ticle-size distribution throughout the profi le, in contrast to proximal loess, which is coarser and often displays more grain-size variation within a single section. Finer textures and greater unifor-mity with depth enhance moisture retention and uniformity of leaching, which enhances struc-ture development along with weathering and clay accumulation (Mason and Nater, 1994). A contributing effect to fi ner loess textures is the lower rate of loess deposition in distal sedi-mentation sites, which may allow at least some pedogenesis to occur as the loess accumulates (Almond and Tonkin, 1999; Huang et al., 2003; McDonald and Busacca, 1990; Muhs et al., 2004; Smith, 1942).

Loess sedimentation also interacts with pedo-genesis when a thin loess deposit is superimposed over a landscape that is undergoing active pedo-genesis. This process will raise the zone of active pedogenesis and produce a new set of horizons refl ecting the modern pedogenic environment while burying horizons developed during for-mer episodes of soil development (Almond and Tonkin, 1999; Kemp et al., 1995, 2001; Kemp and Derbyshire, 1998). Follmer (1982) and Thorp et al. (1951) found that thin increments of middle Wisconsin loess accumulation in a cooler climate produced thick upper sola (namely A horizons) that inherited the B horizon of the last intergla-cial Sangamon Soil and that had weathering and morphological characteristics not in equilib-rium with middle Wisconsin climate conditions. McDonald and Busacca (1990) concluded that morphologic characteristics, namely extensively burrowed soil horizons that acted to reduce sur-face area, greatly enhanced rates of petrocalcic horizon development as the landscape aggraded and the zone of carbonate accumulation moved upward in the profi le.

Finally, an entire loess-paleosol sequence can display trends of soil stratigraphy related to the proximal to distal decrease in sedimenta-tion rate. Often, proximal stratigraphic sections contain multiple buried soils that merge into one pedostratigraphic unit at distal sedimen-tation sites (Huang et al., 2002; Kemp et al., 1997; McDonald and Busacca, 1990; Muhs et al., 2004). Discriminating multiple loess units and soils after they have been incorporated into a single soil profi le can be diffi cult because pedogenesis may obscure original lithological characteristics and stratigraphic contacts. Most studies have overcome this diffi culty by using ratios of particle-size fractions (Follmer, 1982)

or chemical or mineralogical analytical tech-niques such as ratios of chemically resistant Ti and Zr or rare earth elements (Aide and Smith-Aide, 2003; Follmer, 1982; Haseman and Mar-shall, 1945). Micromorphological evidence has also been used to identify evidence for multiple increments of loess, or at least temporal variation in sedimentation rate, within superfi cially simple soils at distal sites in the loess system (Kemp et al., 1995, 1997; Woida and Thompson, 1993).

In the loess deposits of the central Great Plains, formal late Quaternary stratigraphic units include the Peoria Loess, Brady Soil, and Bignell Loess. The Brady Soil formed in late Wisconsin Peoria Loess during the Pleistocene-Holocene transition and was the original basis for recognizing postglacial Bignell Loess in the Great Plains (Schultz and Stout, 1948). Bignell Loess is a widespread Holocene loess unit that began accumulating ca. 9–10 ka, when it bur-ied the Brady Soil (Johnson and Willey, 2000; Miao et al., 2005). Bignell Loess is thickest at the downwind margins of dune fi elds in west-ern and central Nebraska, where it is up to 6 m thick, contains up to four buried soils, and can be divided into regionally correlative stratigraphic zones (Mason et al., 2003; Jacobs and Mason, 2004). The ultimate source of the silt and clay in Bignell Loess is not known, but we believe the particles were transported through the sand sheets and dune fi elds, and were last entrained there by the wind. This conclusion is based on the coarse grain size and great thickness of Big-nell Loess immediately downwind of the dune fi elds (Mason et al., 2003; Miao et al., 2005), and the rapid thinning of this unit with greater distance to the southeast. Jacobs and Mason (2005) concluded that distal Bignell Loess has largely been transformed into the modern A horizon. Kuzila (1995) found that particle-size distribution, chemistry, volcanic ash content, and clay mineralogy indicated a lithologically distinct increment of loess within the sola of modern soils of southeast Nebraska, with a lower boundary near the base of the modern argillic horizon. He hypothesized that the loess increment is correlative with Bignell Loess.

Soils formed in loess on stable upland sum-mits of the central Great Plains typically dis-play a distinct clay content peak in the upper B horizon, and maximum clay content generally increases southeastward from the sources of both Peoria and Bignell Loess (Fig. 1). In most upland soils, U.S. Department of Agriculture (USDA) county soil surveys designate the clay-rich subsurface horizons as “Bt,” even though clay fi lms are often not specifi cally described in the series description. Apparently, the relative increase in clay content is the basis for using the subordinate distinction “t.” For an example, see

the Hastings series description in Ragon (1974). The origin of these texture-contrast soils has been contentious. Bronger (1978, 1991) con-cluded that these soils are largely the result of a lithologic discontinuity in loess deposits, and that the clay-rich subsoil horizons mapped as Bt horizons show no evidence of illuvial clay coatings in thin section. Bronger described soil B horizon microfabrics in Nebraska as rich in aggregates due to biological activity that resulted from paleopedogenesis. Ruhe (1984) found that soils from the central Great Plains show little evidence of weathering, based on clay mineralogy and extractable base cations, yet have the greatest clay content in B horizons of all mid-continent Mollisols (Fig. 1 illustrates the spatial extent of clay content in B horizons). Ruhe assumed that the soils in central Nebraska were formed entirely in Bignell Loess and were therefore much younger than loess-derived Mol-lisols further east in Iowa and Illinois, but the work of Kuzila (1995) has suggested that most soils Ruhe sampled contained both loess units. Ruhe (1984) suggested that extraneous dust deposition may explain the complex character of these soils. Ransom et al. (1997) investigated the micromorphology of some loess soils in Kansas, determining that the B horizons show no evidence of illuvial clay coatings. They did, however, note illuvial clay coatings in buried soils that occurred beneath the modern solum of some soils. Gunal (2001) found illuvial clay coatings only deep in the solum of Kansas loess soils, coincident with crystalline smectite as the dominant clay mineral. Higher in the solum, clay mineralogy is dominated by what is likely interstratifi ed mica-smectite, and the only ori-ented clay occurred around grains as stress fea-tures. Ransom et al. (1997) and Gunal (2001) suggested in situ weathering of biotite as the source of clay in the B horizons. To review, all across the mid-continent loess belt, B horizon clay content is greatest in loess-derived soils of the Great Plains, yet they do not show evidence of weathering or clay illuviation. This research applies methods and knowledge of the interac-tions between loess sedimentation and pedo-genesis to understand the complex character of central Great Plain soils.

This paper is based on a much larger set of soil profi les than previous studies of pedogen-esis in Great Plains loess deposits, and the pro-fi les are distributed across a wide range of loess deposition rate and grain size. Furthermore, our interpretation of these profi les is informed by a much better understanding of the distribution and stratigraphy of Bignell Loess, as well as detailed observations of Brady Soil morphology at many localities (Mason et al., 2003; Jacobs and Mason, 2004; Miao et al., 2005). We focus

Jacobs and Mason

464 Geological Society of America Bulletin, March/April 2007

fi rst on regional variation of soil morphology across the study area, in light of new informa-tion on Bignell Loess sources and dispersal pat-terns. The Brady Soil is then traced downwind from near-source sites, where it is clearly recog-nizable beneath thick Bignell Loess, to locations where it is subsumed within the modern surface soil. This allows identifi cation of the horizons at the distal sites that correspond to the original

Brady Soil profi le, as well as horizons formed in Bignell Loess. We then use the results of this analysis to develop a broader model of aggrada-tional pedogenesis.

STUDY AREA AND METHODS

The study area includes most of southern Nebraska, extending >200 km eastward from

the boundary between loess and the Nebraska Sand Hills and other dune fi elds (Fig. 1). The topography of the region is fl at to highly dis-sected, and it is underlain by Pleistocene loess. The upland landforms in the region are referred to as “tables.” The areal extent of very fl at tables is greatest in the east, and the size decreases while relief and degree of dissection generally increase to the west.

Nebraska

Kansas

Colorado

-20 -10

1 2

34

5

6

7

8

Eolian sand

Loess-derived soils withshallow buried A horizon

Approximate peak B horizon clay contentin loess-derived soilson upland summits

40°N

104°W

15-25

0

10

0

-10-20

NebraskaSand Hills

%25-35

35-4545-55

Figure 1. Map of central Great Plains and study site locations (white dots). The extensive eolian sand sheets and dunes northwest of the study area are inferred to be the immediate source areas of both Peoria and Bignell Loess (Mason, 2001; Mason et al., 2003). Gray shades illustrate general southeastward increase of peak B horizon clay content in soils formed from loess on level to gently sloping summits across southern Nebraska and Kansas (interpreted from STATSGO soil survey data, Natural Resource Conservation Service). Unshaded areas lack thick loess cover or are too dissected to have extensive stable summits. Dashed lines are contours of annual precipitation minus annual potential evapotranspiration (P – PET) in cm yr–1, based on 1961–1990 means (data from High Plains Climate Center). Sites specifi cally mentioned in text or fi gures are labeled: (1) Wauneta, (2) 15G01, (3) 6G01, (4) 14G01, (5) 8G00, (6) 5G00, (7) 3G00, and (8) 3G98.



Modern annual precipitation in the study area varies from ~725 mm in the southeast to <450 mm in the west. Most precipitation falls during the April through September growing season. Mean annual temperatures vary from ~11 °C in the southeast to 9 °C in the north-west. Native vegetation is tall grass prairie in the east, changing to mixed and short grass prai-rie in the west. Most of the region is presently or historically cultivated, but many of our soil sampling sites were located in pasture or native range. Only one of our sites was infl uenced by irrigated agriculture, and it was not used for detailed analysis.

Sites were located along transects perpendic-ular to loess thickness contours, extending from thick loess near the Sand Hills–loess border to thin loess in southeast Nebraska (see thickness trends in Mason, 2001). The soils studied are all well drained, located on level sites, and are clas-sifi ed as Typic and Aridic subgroups of Argius-tolls and Haplustolls.

Soils were sampled on fl at upland landscape positions using a hydraulic soil-coring rig with a 7.6-cm-diameter core tube. At most sites, fi ve replicate soil profi les were extracted every 20 m along a 100 m transect. Depth of sam-pling varied with loess thickness and in nearly all instances exceeded 3 m, often 6 m. Some thick stratigraphic sections containing multiple buried soils were sampled from outcrops, typi-cally road cuts. Soils were described in the fi eld in pedologic detail using Natural Resources Conservation Service terminology (Soil Survey Staff, 1993). Samples were collected from each genetic horizon described, and horizons thicker than 20 cm were subdivided to maintain close-interval sampling.

Analytical data presented here include high-resolution particle-size analysis by laser diffrac-tion using a Coulter LS100Q instrument, along with sand (>63 µm) or clay content determined by wet sieving and pipette (Kilmer and Alex-ander, 1949). Pretreatments included carbonate and organic matter removal with Na acetate (or 10% HCl in some cases) and 30% H

2O

2, and

dispersal with a 5% Na metaphosphate solution. Oriented thin sections, 25 × 45 mm and 30 µm thick, were prepared by Spectrum Petrograph-ics, Inc., Winston, Oregon; terminology fol-lows that in Stoops (2003). Selected samples were also examined using scanning electron microscopy (SEM). Other physical and chemi-cal analyses include bulk density by the clod or core method (Blake and Hartge, 1986), organic carbon (OC) by loss on ignition (LOI) (Konen et al., 2002), whole-rock elemental chemistry using a lithium meta- or tetraborate fusion and X-ray fl orescence (XRF) at a commercial labo-ratory (ALS-Chemex Labs, Sparks, Nevada).

Samples used for mineralogical analy-sis were given pretreatments similar to those used in particle-size analysis. Clay mineral-ogy of the <2 μm fraction was analyzed using smears on glass slides. Dispersed clay samples were Mg2+ saturated; subsequent treatments included: (1) air drying, (2) ethylene glycol solvation, (3) heating at 450 °C. Scans from 2° to 70° 2θ, with a step of 0.1°, were run on a Rigaku Minifl ex X-ray diffractometer using Cu kα radiation. Coarse silt and sand fractions were separated from samples by elutriation after pretreatment and dispersion and were used to prepare grain mounts. Volcanic glass content was quantifi ed using line counts under a petrographic microscope.

RESULTS AND DISCUSSION

Stratigraphy and Soil Morphology

Variation of Soil Morphology with Distance from Immediate Loess Source Areas

Thick stratigraphic sections proximal to sand sheet and dune fi eld areas contain Peoria Loess, the buried Brady Soil, and typically four buried soils within Bignell Loess (Jacobs and Mason, 2004; Mason et al., 2003). Detailed morphologi-cal descriptions and interpretations are presented in those papers and are only summarized here.

In thick stratigraphic sections, the Brady Soil is isolated from the modern surface by several meters of loess and is morphologically well expressed, typically with A/AB/Bk/C horizo-nation. Distinctive properties of the Brady Soil include higher clay content than overlying or

underlying loess, the thick dark A horizon, and a ubiquitously burrowed zone. Clay content in isolated Brady Soil profi les is higher than over-lying Bignell Loess and underlying C horizon Peoria Loess, based on fi eld textures and pipette data from every section we observed or analyzed (e.g., site 1, Fig. 2). The infl uence of clay con-tent on soil development (e.g., structure devel-opment and organic carbon retention) is well documented, and the clay enrichment of the Brady Soil has likely contributed to the strength of its morphological expression.

The thick dark A horizon(s) of the Brady Soil has been the basis for recognizing this soil stratigraphic unit where it is isolated from the modern surface by thick Bignell Loess. These A horizons are a very dark gray or black (10YR 3/1 or 2/1 moist) silt loam with moderate fi ne subangular blocky structure. Multiple A hori-zons up to 50 cm thickness occur in some pro-fi les. Most Brady Soil A horizons have common to many pedogenic carbonate features, typically threads or soft masses. Examination of these horizons by SEM indicates that the carbonates bridge between pedogenic aggregates; thus, they formed after burial when the A horizon was no longer undergoing active bioturbation. The Brady Soil A horizon is commonly darker than most intra-Holocene soils in thick sections of Bignell Loess (Jacobs and Mason, 2004), but this dark color is likely the result of recalcitrant clay–soil organic carbon (SOC) complexes (Jacobs and Mason, 2005). Most isolated Brady Soil profi les have transitional AB horizons with lighter color and coarser subangular blocky structure relative to A horizons.

A1ACAbABbBtb2Ab22Akb22Bkb2

2BCkb2

2Cb2

% clay

BradySoil C horizon

Peoria Loess

BignellLoess

10 20 30 40 10 20 30 40 10 20 30 40 10 20 30 40

100

200

300

400

500

600

700

100

200

300

A1A2A1b2Ab22ABb22Btb2

2BCb2

2BCkb2

2Ckb2

A

AB

A1bA2bABbBtb

Bkb

BCkb

Cb

Ab

Ab

Ab

2Akb2ABkb2Bkb2Cb

Site 1 Site 2 Site 5 Site 6

Dep

th (

cm)

Figure 2. Stratigraphy, soil horizonation, and pipette clay percentage depicted for an array of study sites with increasing subsurface clay content. Study sites are located in Figure 1. Site 1 is Wauneta old road cut (495 mm precipitation, uncultivated), a proximal strati-graphic setting with an isolated Brady Soil; site 2 is 519 mm, site 5 is 577 mm, uncultivated, and site 6 is 664 mm, uncultivated. Sites 2, 5, and 6 are profi les that have a buried soil with characteristics associated with the Brady Soil. The stippled pattern indicates the location of the burrowed zone. Lines mark the top of the Brady Soil and the top of C horizon Peoria Loess (base of the solum). Note change in vertical scale from site 1 to sites 2 through 6.

Jacobs and Mason


The single most distinctive characteristic of Brady Soil profi les is the ubiquitous occur-rence of a zone that has been thoroughly bur-rowed by insects, perhaps cicada nymphs (Jacobs and Mason, 2004; Mason et al., 2003). Burrows of similar size and morphology are essentially absent from the Bignell Loess and Holocene soils, and this burrowed horizon is a reliable marker of the Brady Soil. The bur-rowed zone typically is most evident in AB and underlying upper B horizons because of color contrasts between soil matrix (10YR 5/4 moist) and burrow fi lls (10YR 3/1 moist). B horizons are brown (10YR 5/3 moist) with silt loam or very fi ne sandy loam textures. Mod-erate medium or fi ne prismatic structure that parts to subangular blocks is typically coated with pedogenic carbonate as threads or soft masses. Transitional BC horizons are typical in most profi les; they are often 1 m thick and have prismatic structure, pedogenic carbonate, and slightly redder colors (10YR 5/3 moist) than C horizon Peoria Loess (2.5Y 5/3 moist). The base of the BC horizon is typically the start of an upward increase in clay content, culminat-ing in the clay peak in the A horizon.

In thick stratigraphic sections, soils formed within Bignell Loess primarily have A/C, A/AB/C, or A/Bk/C horizonation. Morphologi-cally, they are identifi ed by very dark gray to dark grayish brown A horizons, typically with weak subangular or granular structure. Subja-cent AB or Bk horizons are recognized on the basis of weak subangular blocky or prismatic structure and colors slightly darker than C hori-zon Bignell Loess.

Bignell Loess thins rapidly to the southeast (Mason et al., 2003), and soil morphology changes also. Overall morphology changes from profi les with obvious complex stratigraphic and genetic history to sites with what generally fi rst appear to be simple A/B/C profi les with a silty A horizon(s) overlying dark, clay-rich Bt hori-zons. With few exceptions, sites within 10 km of the sand-loess border have soil profi les with multiple horizons, or they have genetic profi les formed within ≥1 m of Bignell Loess, which overlies a clearly discernible Brady Soil with its characteristic burrowed zone (site 2, Fig. 2). Bignell Loess has visible evidence of pedogen-esis throughout the entire thickness, generally expressed as a pedocomplex of multiple A hori-zons and often with transitional AB, BA, or AC horizons or a B horizon. Pedogenic carbonate is common. The A horizons in this thinner Big-nell Loess probably correlate with the isolated buried soils in the thick sections, but specifi c correlations are not possible at this point. The B horizons are dark and prismatic structured, but will typically part to granular or fi ne subangular

blocky structure, indicating an origin as an aggradational A horizon that has lost organic matter and acquired B horizon structural charac-teristics after burial (Jacobs and Mason, 2005). When buried at suffi cient depth, subsoil struc-tural processes of shrink-swell predominate and prismatic structure develops.

With increasing distance (10–50 km) from the sand-loess border, the Brady Soil is no lon-ger consistently recognizable as a buried A hori-zon, except in the westernmost part of the study area. Surface soils contain thick A horizons, and a transitional AB horizon typically occurs over a clayey horizon typically described as “Bt” (Jacobs and Mason, 2005). The AB horizons are dark and show macro- and microscopic evi-dence of biologic activity such as granular struc-ture and channels and chambers, suggesting an origin as A horizons before the land surface aggraded. The actual depth of burial required to allow transformation of A to AB or Bw appears variable and probably depends on rate of aggra-dation, clay content, and rainfall to drive shrink-swell processes. The underlying “Bt” horizon in these profi les is dark and often meets mollic requirements (value and chroma ≤ 3). In fact, in some cases, the horizon with peak clay content clearly has the appearance of a buried A hori-zon, and we described it as such (e.g., sites 5 and 6, Fig. 2). A burrowed zone very similar to that observed in the lower Brady Soil occurs below the zone of peak clay content at some sites (e.g., sites 5 and 6, Fig. 2).

Finally, at far distal sites, profi le morphology consists of thick dark silty A horizon(s) over a very clay-rich, relatively dark-colored Bt horizon, where transitional AB horizons are less frequent. The profi le is similar to sites 5 and 6 (Fig. 2), but the clay peak is at a shallower depth.

Figure 2 allows comparison of clay content profi les across much of the range of distance from the sand-loess border, excluding only the most distal sites. Sites 1, 5, and 6 have never been cultivated, according to landowners, thus they provide the best possible record. Figure 2 suggests that the clay-rich B horizons at distal sites are correlative with the Brady Soil, and they represent former A horizons transformed by burial and organic matter loss. Localities such as site 6, where the clay peak occurs in a horizon that retains enough A horizon characteristics to be described as Ab rather than B, are consis-tent with that interpretation. Furthermore, as distance from immediate Bignell Loess source areas increases, the horizon(s) with peak clay content becomes thinner, and its clay content increases. This observation strongly suggests a sedimentological trend, since loess deposits typ-ically thin and increase in clay content with dis-tance from the source. Thus, the morphological

trends observed across our study area strongly suggest that the clay-rich zone at all sites is at least partly sedimentary in origin, regardless of whether it is recognized as the Brady Soil or described as a “Bt” horizon.

Of course, this interpretation cannot be accepted without testing other more conven-tional explanations for clay enrichment, such as in situ weathering or clay illuviation. In the fol-lowing sections, we consider macro- and micro-morphological, geochemical, and mineralogical evidence in order to distinguish between effects of loess sedimentation and in situ pedogenesis.

Quantitative Analysis of Macromorphological Trends: Effects of Climate versus Loess Sedimentation

We further explored the proximal-to-distal trend in peak clay content suggested by Fig-ure 2 using pipette clay measurements from one or more profi les at 21 of our sampling sites (Fig. 3). Horizons of maximum clay content were grouped according to their identifi cation in the fi eld as the Brady Soil or as a B horizon at distal sites; we also grouped samples repre-senting baseline clay content in underlying C horizon Peoria Loess. Clay content within each of these groups of samples varies systematically with distance downwind from the northwest-ern edge of Peoria Loess, as defi ned by Mason (2001), which effectively represents downwind transport distance for this loess unit. Clay con-tent trends are offset toward higher values in the Brady Soil and B horizon samples but are other-wise similar between the two groups. The trend observed for maximum clay content at our study sites is consistent with the general southeastward increase in B horizon clay content across the central plains as inferred from soil survey data (Fig. 1). The southeastward fi ning of C horizon Peoria Loess resembles similar trends in other loess systems (e.g., Smith, 1942; Hutton, 1951), and it is clearly sedimentary in origin. The sim-plest explanation for the trend observed in the Brady Soil and B horizon samples is that it is sedimentary as well.

Clay content in both groups of samples also increases systematically with modern mean annual precipitation (Fig. 3). For the C horizon loess, this is almost certainly an artifact of the fact that precipitation and distance from loess source(s) both generally increase eastward. We interpret the trend in the Brady Soil and B hori-zon samples in a similar fashion, rather than as the result of enhanced clay production by in situ weathering as precipitation increases. If the lat-ter interpretation were valid, we would expect a greater relative increase in clay content from the C horizon to the Brady Soil to B horizon clay peak in areas with higher rainfall. Instead, Brady



Soil and B horizon clay content is ~1.7 times C horizon content at drier proximal sites and ~1.6 times C horizon content at wetter distal sites. A few extreme values of B horizon clay content at distal sites that are well above the overall trend may refl ect some addition of clay through in situ weathering.

If the horizon of peak clay content at all sites is in fact the Brady Soil, the total thickness of horizons above that peak should decrease with distance southeast from the dune fi elds previ-ously identifi ed as immediate Bignell Loess source areas (Mason et al., 2003). Thickness data support this, and, while it does not provide large explanatory power, the trend is statistically signifi cant even if proximal locations with thick Bignell Loess are excluded (Fig. 4). In an earlier paper, we demonstrated a similar statistically signifi cant trend for total thickness of A hori-zons (Jacobs and Mason, 2005). The variable of distance from dune fi elds used in these analyses is similar to the distance from the northwestern edge of Peoria Loess for many sites, but differs in some cases because it accounts for proximity to small dune fi elds that occur within the gener-ally loess-mantled region.

It is unlikely that climatic effects on pedogen-esis can explain the observed trends in depth to the clay peak or A horizon thickness. Figure 4B is a plot of depth to the clay peak against pre-cipitation, stratifi ed to separate profi les with or without thick Bignell Loess. While the relation-ships are statistically signifi cant, the slope is negative. Depth to peak clay content should be positively correlated with precipitation, if clay enrichment results from either illuviation or weathering. Higher rates of precipitation should lead to deeper translocation of clay, and higher moisture content at greater depths, favoring deeper clay production by weathering.

Micromorphological Evidence on the Occurrence and Timing of Clay Illuviation

Pedogenic features observable in thin sec-tion provide evidence on two issues. The fi rst is whether the present B horizons at distal sites contain features indicative of initial formation as the Brady Soil, consistent with the correla-tion proposed herein. Jacobs and Mason (2005) demonstrated that this is the case; granular aggregates typical of those observed in mollic epipedons are common in B horizons, often preserved within larger blocky peds. Bronger (1991) made similar observations. Granular structure is also evident, in the fi eld as well as in thin section, in some BC horizons and even in C horizon Peoria Loess at sites far southeast of the major loess sources (Fig. 5A). This is consistent with slow loess accumulation on a vegetation-covered land surface.

All points includedy = 295.14 x -0.567

r 2 = 0.53, p < 0.001

Thick sections excludedy = -0.316x + 57.38r 2 = 0.16, p = 0.034

A

0

100

200

300

400

500

400 500 600 700 800

Dep

th to

cla

y m

axim

um (

cm)

Mean annual precipitation (mm)

All points includedy = 5E+13x -4.30

r 2 = 0.50, p < 0.001

Thick sections excludedy = -0.182x + 158.34r 2 = 0.40, p < 0.001

B

0

100

200

300

400

500

0 25 50 75 100 125

Dep

th to

cla

y m

axim

um (

cm)

Distance from dunes (km)

Figure 4. Plots showing statistical relationships between depth to the clay maxima or Brady Soil and predictor variables (A) distance from sand dune immediate source areas, or (B) precipitation. Statistics are stratifi ed according to exclusion or inclusion of thick profi les (•) proximal to sand dune source areas of Bignell Loess.

0

10

20

30

40

50

60

1 10 100 1000

% C

lay

Distance from NW edge of loess (km)

y = 4.821Ln(x) + 0.915r 2 = 0.70, p < 0.001

y = 6.081Ln(x) + 9.380r 2 = 0.74p < 0.001

A

0

10

20

30

40

50

60

400 500 600 700 800

y = 0.0598x – 15.86r 2 = 0.59, p < 0.001

y = 0.087x – 19.66r 2 = 0.71, p < 0.001

B

% C

lay

Mean annual precipitation (mm)

Bt maximum C horizon Peoria Loess

Figure 3. Statistical relationships between pipette clay percentage of horizons with clay maxima and C horizon Peoria Loess and predictor variables (A) distance from northwest edge of loess (km) or (B) precipitation (mm).

Jacobs and Mason


The second issue, which we focus on here, is whether clay illuviation has occurred in these horizons, potentially resulting in higher clay content. Examinations of thin sections from 14 profi les, plus data from the Mirdan Canal sec-tion (Mason and Kuzila, 2000), agree with the conclusion of previous studies (Bronger, 1978, 1991; Gunal, 2001; Jordan, 1967; Ransom et al., 1997) that clay illuviation is not a dominant pro-cess in the central Great Plains soils today, nor has it been at any time in the late Pleistocene or Holocene, and it is most likely not responsible for much of the B horizon clay content. Evi-dence of clay illuviation exists, but the percent-age of clearly illuvial clay is low and typically occurs deep in the solum; furthermore, most of this illuvial clay is probably relict from the period of Brady Soil formation.

The clay-rich upper B horizons typical of soils at distal sites do not contain illuvial clay, with one exception, the easternmost site sampled (site 7, Fig. 1). Most of the clay in the clay-rich B hori-zons occurs as grain coatings (granostriated b-fabric) with occasional striated appearance in ped interiors, especially in the more clay-rich eastern part of the study area (Fig. 5B). Nine of the 15 profi les do contain illuvial clay coatings, but they are well below the clay peak, except at site 7. The clay coatings are generally limpid, occasionally microlaminated, and they occur along the walls of voids or fi ll circular channels (Fig. 5C).

Plotting the depth of fi rst occurrence of illu-vial clay coatings versus depth to the top of the Brady Soil or the top of the horizon with peak clay content reveals a strong linear relation-ship (Fig. 6A). If we assume that the peak clay

content corresponds to the former Brady Soil A horizon, as discussed already, the explana-tion for the observed trend becomes evident. The illuvial clay coatings formed during active pedogenesis of the Brady Soil profi le, at a con-sistent depth of ~34 cm below the land surface at that time (intercept of regression equation in Fig. 6A), before the Brady Soil was buried by varying thicknesses of Bignell Loess.

The location of illuvial clay coatings relative to pedogenic carbonate in most profi les also indicates that most illuvial clay originated dur-ing Brady Soil formation. Pedogenic calcium carbonate typically coats or fi lls voids, but it also occurs as hypocoatings adjacent to voids or quasicoatings within the matrix (Fig. 5C). In most profi les, pedogenic carbonate occurs above illuvial clay, which indicates that clay illuviation

0.5 mm

2BAkb matrix2B2BAkbkb matrimatrix

burrow fill

1 mm

B

1 mm

0.5 mm

ca

icic

A

C D

Figure 5. Photomicrographs of: (A) 2C horizon (Peoria Loess) of profi le at site 2 showing open fabric interpreted as former A horizon that formed as the loess aggraded. (B) A2b horizon (80 cm, clay maximum) of profi le at site 6 with 38% clay but no evidence of illuvial clay in existing planar voids; only grains are coated with clay, which we attribute to origin by eolian sedimentation during a late phase of Peoria Loess deposition. (C) BCkb horizon (151 cm) of profi le at site 5 showing a microlaminated illuvial clay coating partially fi lling a channel (ic) and pedogenic calcium carbonate (ca) in an adjacent planar void. (D) 2BAkb horizon (143 cm) of Brady Soil at site 3 with burrow fi ll con-taining an aggregate from the underlying horizon that includes illuvial clay coatings, indicating that burrowing postdated clay illuviation.



is highly unlikely to be a modern pedogenic process. The occurrence of free calcium car-bonate in a horizon is believed to preclude the dispersion and migration of clay (Jenny, 1980; Schaetzl and Anderson, 2005). The illuvial clay therefore is likely the result of paleopedogenesis prior to Holocene climatic drying and accumula-tion of pedogenic carbonate. Evidence from site 3 indicates that clay illuviation occurred before formation of the burrowed zone characteristic of the Brady Soil. Figure 5D shows burrow fi ll from the burrowed zone that contains illuvial clay coatings, while the surrounding matrix lacks coatings. This suggests that the burrowed zone may have formed in a late phase of Brady Soil pedogenesis.

A plot of depth to the fi rst occurrence of illu-vial clay and/or pedogenic carbonate against modern P – PET (precipitation minus potential evapotranspiration) for each of the 15 profi les provides additional evidence that the illuvial clay coatings are relict features (Fig. 6B). Note

that the depth to illuvial clay decreases as pre-cipitation increases, and the depth to pedogenic carbonate increases with precipitation. Only the relationship between carbonate and climate fi ts with current scientifi c understanding of soil carbonates and, assuming the occurrence and location of pedogenic carbonates is in equilib-rium with modern climate conditions, provides a climofunction for the region (this assumption is supported by the occurrence of pedogenic car-bonate in Bignell Loess younger than 2000 yr old). Only the illuvial clay in profi les at points to the right of −3 (i.e., P – PET > –3), where pedogenic carbonates occur deeper in the pro-fi le than illuvial clay coatings, are possibly the result of late Holocene clay illuviation. In one profi le (3G00), illuvial clay does occur in the AB horizon above the clay maximum, in what our hypothesis indicates is Bignell Loess. Based on the data collected in this study, a value of P – PET ≥ –3 cm appears to be a threshold for modern clay illuviation.

The implications of illuvial clay associ-ated with the Brady Soil are signifi cant for soil formation rates and paleoclimate. The climate across most of Nebraska during Brady Soil formation evidently allowed leaching of carbon-ates from the entire profi le, along with disper-sion and migration of clay, perhaps in as little as 2000 yr (the mean duration of Brady Soil pedogenesis, based on radiocarbon dating of the upper and lower 5 cm of the A horizon; Johnson and Willey, 2000). Rapid leaching may be rea-sonable because the content of primary calcium carbonate in Peoria Loess in our study area is low, and the quantity of clay translocated was small. If we assume that a P – PET value of −3 is the modern threshold for clay illuviation, the illuvial clay associated with the Brady Soil in west-central Nebraska implies a 350 km west-ward shift of the P – PET = −3 isoline in the late Pleistocene and earliest Holocene.

Mineralogical and Geochemical Evidence

We investigated the clay mineralogy, volcanic glass content, and bulk (whole-rock) geochem-istry of selected samples to identify any prove-nance-related contrasts between Peoria and Big-nell Loess that could help us place the boundary between these units within surface soils at dis-tal sites. In most cases, however, we also have to consider effects of in situ weathering on the same geochemical or mineralogical properties. Furthermore, geochemistry and mineralogy may vary systematically within a loess unit because of particle-size sorting during loess transport (Muhs and Bettis, 2000; Eden et al., 1994).

Clay MineralogyIn thick proximal sections, the <2 µm clay

fractions of Bignell Loess and Peoria Loess below the Brady Soil are distinctly different (Fig. 7A; for data from another thick section, see Mason and Kuzila, 2000). The clay fraction of Peoria Loess below the Brady Soil is dominated by smectite, with lesser amounts of mica and kaolinite, based on diffractogram peak heights. In contrast, Bignell Loess is dominated by mica and kaolinite, with much smaller amounts of smectite. Detection of smectite in Bignell Loess is often problematic because of poor expres-sion of broad and choppy peaks. In the section studied by Mason and Kuzila (2000), the clay mineralogy of the lower part of the Brady Soil is similar to underlying Peoria Loess, but in the upper part of the Brady Soil, there is a transition toward the low-smectite mineralogy of Bignell Loess. Similarly, in the thick Wauneta section (site 1 in Fig. 7A), the clay mineralogy of the uppermost Brady Soil approaches that observed in overlying Bignell Loess.

0

100

200

300

400

500

600

0 100 200 300 400

Dep

th to

firs

t illu

vial

cla

y (c

m)

Depth to Brady A/clay maxima (cm)

y = 1.153x + 33.8r 2 = 0.97, p < 0.001

A

0

100

200

300

400

500

600

-25 -20 -15 -10 -5 0

Dep

th to

cla

y or

car

bona

tes

(cm

)

P-PET (cm)

Depth to secondary carbonatesy = 4.68x + 172.1r 2 = 0.55, p = 0.004

B Depth to first illuvial clayy = -15.241x + 84.0r 2 = 0.39, p = 0.053

5

Figure 6. Scatterplots portraying statistical relationships between (A) depth from modern land surface to the top of the Brady Soil or horizon with clay maximum versus the depth to fi rst occurrence of illuvial clay coatings; and (B) modern climate (precipitation – Thorn-thwaite potential evapotranspiration) and the depth to the fi rst occurrence of illuvial clay and pedogenic carbonate.

Jacobs and Mason


A very similar upward change in mineralogy occurs in the surface soil at distal sites, where the transition occurs within, or at the top of, the hori-zon with maximum clay content. This is true of distal sites where the clay peak displays A hori-zon morphology in some of the replicate cores, (e.g., site 6, Fig. 7B), and at sites where the clay

peak is in a horizon described as “Bt” (e.g., site 8, Fig. 7C). These observations are consistent with clay mineralogical data from many surface soils of the central Great Plains. Numerous authors have noted that the surface horizons of these soils are enriched with mica (illite) relative to smec-tite (Bronger, 1991; Gunal, 2001; Kuzila, 1995;

Ruhe, 1984). Gunal (2001, p. 49) found broad and choppy smectite peaks in diffractograms from surface soils of western Kansas, which he attributed to the occurrence of interstrati-fi ed mica-smectite, possibly due to weathering in near-surface horizons. These previous stud-ies also reported much more abundant smectite,

A

B

C

Figure 7. X-ray diffractograms from 3° to 18° 2θ to illustrate the clay mineralogy of: (A) Peoria Loess, Brady Soil, and Bignell Loess in a core collected at site 1, where thick Bignell Loess has isolated the Brady Soil (Mason et al., 2003; Miao et al., 2005); (B) selected horizons in a profi le at site 6, a distal sedimentation site where the horizon with peak clay content (A1b) retains some morphological characteristics of an A horizon; and (C) selected horizons in even more distal profi le at site 8, where the horizon with greatest clay content has been described as Bt1. Peaks corresponding to 001 refl ections of major clay minerals are labeled, including S—smectite, M—mica or illite, and K—kaolin-ite (diffractograms after heating to 500 °C indicate that little or no chlorite is present in these samples). All samples are Mg saturated and ethylene glycol solvated.



producing much more well-defi ned peaks, in Peoria Loess below the surface soil profi le.

We interpret the upward change in clay mineralogy at both proximal and distal sites as primarily the result of a change in provenance from Peoria to Bignell Loess. The coincidence of the change with the most clay-rich horizon at distal sites supports our correlation of that horizon with the Brady Soil. An alternative explanation must be considered, however. Singer (1989a) indicated that illitization of smectite or vermiculite is a common occur-rence in arid-region soils, where K+ is relatively abundant in the soil solution, and Joeckel and Ang Clement (2005) recently suggested that illitization of smectite occurs in saline wet-lands of western Nebraska. We doubt that in situ illitization can explain the upward transi-tion at our study sites, for two reasons. First, this transition occurs at sites where the present climate is actually subhumid rather than arid, and K+ is unlikely to have accumulated in the upper solum. Second, the transition occurs in thick proximal sections where Bignell Loess accumulated rapidly and illitization would have to have proceeded quite rapidly.

The possibility that illitization can occur in saline wetlands on the western Great Plains (Joeckel and Ang Clement, 2005) suggests an explanation for a change in clay mineralogy from Peoria to Bignell Loess. Climatic change in the early Holocene may have exposed exten-sive saline wetland and playa sediment to wind erosion in the loess source areas, increasing mica

(illite) content and decreasing smectite content in the loess that accumulated downwind.

An intriguing observation potentially consis-tent with this hypothesis is the identifi cation of palygorskite within thick Bignell Loess at site 1 (Fig. 8). Palygorskite is identifi ed on the basis of peaks at ~1.04 nm (8.5 two-theta), ~0.446 nm (19.8), and ~0.365 nm (24.4) in Mg-EG slides and also an increasing in intensity of the 0.365 nm peak with heat treatments (Singer, 1989b, p. 843); other possible refl ections are weak or obscured. While palygorskite can form in the soil environment, notably calcareous crusts, most authors attribute its occurrence to parent material inheritance (Singer, 1989b, p. 859). Palygorskite is known to form in alkaline lakes and playas with high Mg and Si activities (Singer, 1989b, p. 857). Alternatively, palygorskite has been identifi ed in the Ogallala Formation in New Mexico (Frye et al., 1974), which could indicate active Holocene erosion surfaces stripping areas of Ogallala Formation and supplying clay.

Volcanic Glass ContentVolcanic glass occurs in many Quaternary

sediments of the central Great Plains, prob-ably reworked from volcaniclastic rocks of the White River and Arikaree Groups (Oligocene to Lower Miocene; Swinehart et al., 1985). Mason and Kuzila (2000) proposed that volcanic glass contents may differ between Peoria and Bignell Loess as a result of changing provenance; there-fore, we determined glass content in fi ve pro-fi les representing proximal to distal locations.

In Figure 9A, coarse silt and sand fraction glass content of samples stratifi ed by stratigraphic position and/or pedologic horizon are plotted against distance from the northwestern edge of Peoria Loess. Volcanic glass content is low at distances >30 km, but downwind trends are not well defi ned. At study sites located more than ~175 km downwind from the edge of Peoria Loess, preliminary glass counts often produced values not signifi cantly different from zero, so we did not pursue more detailed analyses.

With one exception, samples of C horizon Peoria Loess consistently contain more glass than any shallower samples at a given site (Fig. 9A). An upward decrease in glass content begins well below the Brady Soil A horizon (at proximal sites) or the upper B horizon (at distal sites). Pat-terns of variation at shallower depths differ from site to site, but an increase in glass content in the uppermost one to three samples is common (Fig. 9B). Kuzila (1995) reported vertical profi les of glass content in surface soils at distal sites in south-central Nebraska that showed an upward decrease to the top of the most clay-rich B hori-zon, with a small increase above that depth.

These results can be interpreted as indicating a change in provenance, or as the result of in situ weathering of glass. One major problem with the latter interpretation is that it would require greater mass loss of glass by weathering at prox-imal sites, where the climate is drier. Also, rapid loess accumulation at proximal sites would have more quickly buried glass below the near-sur-face zone where rapid weathering is more likely. If the upward change in glass content refl ects a shift in loess provenance, this occurred during the latter part of Peoria Loess deposition, before deposition of the sediment that is now incorpo-rated in the Brady Soil at proximal sites, or the modern solum at distal locations.

Bulk GeochemistryThere is strong evidence that the bulk (or

whole-rock) content of several major and trace elements is affected by particle-size sorting dur-ing loess transport, and there is little evidence that this sedimentary effect is overprinted by signifi cant in situ weathering (Fig. 10). TiO

2

and Fe2O

3 increase with increasing distance

downwind from the northwestern edge of Peo-ria Loess and with increasing clay content, while Na

2O decreases with increasing distance

downwind and clay content. The TiO2 and Fe

2O

3

trends are similar to those in Iowa (Muhs and Bettis, 2000). The Na/TiO

2 ratio decreases with

increasing downwind distance and clay content, as expected given the trends of the two elements involved. Very similar trends emerge whether all samples are pooled or samples are stratifi ed by position relative to stratigraphic and pedologic

Site 1 Core

3 8 13 18 23 28

Degrees 2θFigure 8. X-ray diffractogram illustrating Bignell Loess sample (255 cm) from site 1. Peaks identifi ed as palygorskite are labeled P; Q is quartz. Sample is Mg saturated and ethylene glycol solvated.

Jacobs and Mason


0

5

10

15

20

11 0 100 1000

% g

lass

in s

ilt a

nd v

ery

fine

sand


A

Bignell Loess

Brady A/Bt max Deep Peoria Loess

Upper Peoria Loess

100

200

300

400

05 10 15 20

% glass in silt and very fine sand

Dep

th (

cm)

Site 4

Site 5

Site 6

B

0

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

11 0 100 1000

0

100

200

300

400

500

600

11 01 00 1000

Zr

(ppm

)


0

1

2

3

4

5

11 01 00 1000

Fe 2

O3

(%)


0

1

2

3

11 01 00 1000

Na 2

O/T

iO2


0.0

0.4

0.8

1.2

1.6

11 0 100 1000

Na 2

O (

%)


K2O

(%

)


Bignell Loess Brady A/Bt max Upper Peoria Loess Deep Peoria Loess

Pooled statistics:y = -0.005x + 2.50r 2 = 0.56, p < 0.001

Pooled statistics:y = -0.077x + 364.55r 2 = 0.01, p = 0.300

Pooled statistics:y = 0.0003x + 2.69r 2 = 0.02, p = 0.160

Pooled statistics:y = 0.005x + 2.83r 2 = 0.60, p < 0.001

Pooled statistics:y = -0.001x + 1.30r 2 = 0.36, p < 0.001

0

0.2

0.4

0.6

0.8

11 0 100 1000

TiO

2(%

)


Pooled statistics:y = 0.001x + 0.52r 2 = 0.68, p < 0.001

Figure 10. Scatterplots of loess transport distance versus selected elements. Data points are stratifi ed by position relative to stratigraphic and pedologic boundaries, with statistics for all data pooled.

Figure 9. (A) Scatterplot relating distance from northwest edge of loess to volcanic glass percentage in samples stratifi ed into deep lower Peoria Loess, upper Peoria Loess near pedogenic horizons, the Brady Soil/Bt maximum, and soil horizons formed in Bignell Loess. (B) Vol-canic glass content in three profi les: site 4 is a proximal loess section with an isolated Brady Soil and three buried soils formed in 250 cm of Bignell Loess, and sites 5 and 6 are located at increasing distance from immediate loess source areas. Arrows indicate the top of the Brady Soil or clay-rich horizon.



boundaries (Fig. 10). Stratifi cation of the sam-ples also reveals that, at proximal sites, Peoria Loess, the Brady Soil, and Bignell Loess overlap in their contents of TiO

2, Fe

2O

3, and Na

2O and

in Na2O/TiO

2. At distal sites, samples with simi-

lar clay content have similar contents of those elements, regardless of whether they are in A horizons or in largely unaltered Peoria Loess. Weathering typically depletes relatively mobile elements such as Na and enriches less mobile ele-ments such as Ti and Fe, and we used Na

2O/TiO

2

as a weathering index. Any weathering that has occurred at our sites apparently has been insuf-fi cient to create offsets in the content of these elements or in Na

2O/TiO

2 between unweathered

loess and surface or buried soil horizons.In contrast to the results obtained by Muhs

and Bettis (2000) from Peoria Loess in south-western Iowa, we found no evidence for sorting effects on Zr and K

2O, neither of which is sig-

nifi cantly related to distance downwind or clay content (Fig. 10). Contents of Nb and Y are also unrelated to those variables (data not shown).

When we group Bignell Loess samples at proximal sites with samples from above the horizon of peak clay content at distal sites, these samples are consistently distinguished by their high Zr content relative to other samples at the same site (Fig. 10). Because Zr is a rela-tively immobile element, this observation could indicate greater in situ weathering in both Big-nell Loess and upper solum horizons at distal locations, causing relative enrichment of Zr. We doubt this interpretation is valid, however. First, the ratios TiO

2/ZrO

2, Nb/Zr, and Y/Zr are

all distinctly lower in the same Bignell Loess/upper solum samples that have high Zr com-pared to underlying horizons (Fig. 11). Given that Ti, Nb, and Y are all considered relatively immobile during weathering, they should be enriched along with Zr if signifi cant in situ weathering has occurred. Second, the increase in Zr above the clay-rich B horizon at the more humid distal sites is not consistently larger than the increase in Zr above the Brady Soil at proximal sites, where the climate is drier and rapid Bignell Loess accumulation would prob-ably have limited the effectiveness of weather-ing. Thus, we favor an interpretation of the Zr increase above the Brady Soil or clay-rich B horizons as the result of a change in loess prov-enance, which marks the boundary between Peoria and Bignell Loess at both proximal and distal sites.

Pedosedimentary Model and Climatic Control on Aggradation

Based on the evidence described herein, we propose a pedosedimentary model applicable to

the distal study sites, and more broadly to upland loess-derived soils across a large part of the cen-tral Great Plains. These soils are the result of ongoing dust aggradation on a pedogenically active surface since the late Pleistocene. Modern soils of the central Great Plains are composite soils (sensu Morrison, 1967), and their proper-ties refl ect paleopedogenesis of the Brady Soil (Geosol) and episodic sedimentation and pedo-genesis of Bignell Loess throughout the Holo-cene. The pedosedimentary history of these soils is closely linked to regional climate change.

The clay-rich “Bt” horizons of the modern soil profi les are distal equivalents of the buried Brady Soil. This conclusion is supported by the change in clay mineralogy and increase in Zr within, or at the top of, both the Brady Soil at proximal sites and the “Bt” horizon in distal profi les, and by the upward decrease in volcanic glass content culminating in the Brady Soil or distal soil “Bt” horizon. More direct evidence is available at site 5G00, where we can even trace the “Bt” laterally to a buried A horizon that has a burrowed zone characteristic of the Brady Soil.

Furthermore, we propose that the clay-rich B horizons, including the regional pattern of increasing clay content southeastward (Fig. 1), originated almost entirely by sedimentation, and they are not the result of mineral weath-ering or clay illuviation. Illuvial clay coatings occur in modern soils, but where P – PET is < –3 cm, the illuvial clay is present in only the lower solum, beneath pedogenic carbonate and the clay maxima. This illuviation cannot explain clay enrichment in overlying horizons, and it is instead interpreted as having occurred when the Brady Soil (now the “Bt” horizon) was undergoing active pedogenesis at the land surface under a wetter climate than present. This relationship is visibly evident stratigraph-ically as the thickness of Bignell Loess exceeds 1 m. Clay illuviation likely occurred during the initial phases of landscape stabilization prior to thorough burrowing by insects.

In our conceptual model, the Brady Soil formed in a clay-enriched late phase of Peoria Loess, and its granular soil structure developed during ongoing slow aggradation (or up-build-ing) of the fi nal increments of Peoria Loess (Fig. 12). Although we cannot rule out some chemical weathering during Brady Soil devel-opment, bulk geochemical data provide no unequivocal evidence of signifi cant mass fl ux or immobile element enrichment by in situ weath-ering. The strong downwind trend of increasing clay content from the Brady Soil to distal soil B horizons also suggests a sedimentary rather than weathering origin for clay enrichment in both cases. We suggest that increasing clay content, decreasing volcanic glass content, and probably

decreasing sedimentation rates characterized the last phase of Peoria Loess accumulation because of a shift in the nature of loess sources and/or a change in transport distance. Climatically driven changes in source area vegetation or hydrology are clearly possible factors, but this is a problem that needs much further work.

Climatically driven vegetation change contin-ued throughout the deposition of the clay-rich late Peoria Loess and Brady Soil development. Stable isotopes of C contained in Peoria Loess in Colorado, Nebraska, and Kansas indicate that the proportion of soil organic matter derived from C

4 plants perhaps was increasing as the

latest phase of Peoria Loess was accumulating. The upward increase in C

4 inputs peaks in the

upper Brady Soil (Feggestad et al., 2004; John-son and Willey, 2000; Muhs et al., 1999). All of these authors have attributed the increase in C

4 species primarily to climatic warming, per-

haps as much as 10 °C by the earliest Holocene (Johnson and Willey, 2000).

More rapid loess sedimentation began to bury the Brady Soil between ca. 11 and 9 ka (John-son and Willey, 2000; Miao et al., 2005), and this period is marked by an upward decrease in pedogenic alteration (Jacobs and Mason, 2004). Bignell Loess accumulated in response to drier climate conditions that allowed devegetation of sand sheets followed by dune mobilization, dust defl ation, and Bignell Loess accumulation on upland landscapes (Mason et al., 2003). Bignell Loess accumulation has previously been recog-nized mainly at proximal locations where it is thick and the Brady Soil is clearly identifi able.

Throughout the Holocene, accumulation of Bignell Loess also occurred on uplands in dis-tal areas, however. The slowly accumulating loess progressively buried the clay-rich Brady Soil, and the loess increments were transformed into A and/or AB horizons as the land surface aggraded. At the same time, subsoil shrink-swell processes acted on the higher clay con-tent of the buried Brady Soil, transforming the subsurface into the blocky and prismatic “Bt” horizon(s) that still retain granular structure in ped interiors. Low levels of recalcitrant SOC have maintained dark colors in the B horizons, confounding recognition of pedologic discon-tinuities. Pedogenic calcium carbonate accu-mulated in Brady Soil horizons and in Bignell Loess in response to drier climate, and the upper limit of accumulation probably refl ects near-modern climate conditions.

CONCLUSIONS

Clay-rich B horizons are characteristic of soils formed in loess on upland summits of the central Great Plains. Using stratigraphic

Jacobs and Mason


observations, soil macro- and micromorphol-ogy, clay mineralogy, volcanic glass content, and bulk geochemistry, we interpret these B horizons as former buried A horizons, cor-relative with the Brady Soil identifi ed in thick loess sections at the sand-loess border. These B horizons have clearly been affected by pedogenic processes, but their high clay content is not the result of weathering or clay illuviation. Instead, it primarily represents the distal equivalent of a clay-rich upper zone in

the late Pleistocene Peoria Loess, in which the Brady Soil developed as slow dust accumu-lation continued. During the Holocene, espe-cially during periods of drier-than-present climate, thin Bignell Loess buried the Brady Soil, which then acquired B horizon charac-teristics. The modern A and AB horizons are formed in Bignell Loess. Upland soils in the central Great Plains with silty A horizons overlying clayey B horizons are composite soils that resulted from dust sedimentation

and pedogenesis controlled by changing cli-mate throughout the late Quaternary.

ACKNOWLEDGMENTS

This research was supported by the National Sci-ence Foundation, Geography and Regional Science and Geology and Paleontology Programs (grants BCS-0079252 and BCS-0079320). Thanks are due to Xia-odong Miao, Paul Hanson, and Aaron Feggestad, for fi eld and laboratory assistance, and to Arnt Bronger for encouragement of this research. Reviews by Dan Muhs and an anonymous reviewer improved the manuscript.

AC

C

A

Bt

CBk

A

C

Bw

A

C

Bw

Ab

C

Bw

AC AB

Figure 12. Successive schematic panels illustrating major climatically driven changes in loess aggradation and the evolution of upland soils of the central Great Plains. The solid line in each panel represents clay content, increasing to the right. Panel one: Cold climate accumula-tion of deep silty Peoria Loess. Panel two: Clayey late-phase Peoria Loess aggradation and development of Brady Soil horizonation includ-ing clay illuviation. Panel three: Insect burrowing in Brady Soil. Panel four: Early to middle Holocene accumulation of Bignell Loess buries Brady Soil; secondary calcium carbonate is emplaced. Panel fi ve: Accumulation of Bignell Loess continues and subsurface of former Brady Soil A horizon is transformed to blocky structured “Bt” horizon.

0

0.5

1

1.5

2

2.5

3

11 0 100 1000

TiO

2/ Z

rO2


0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

11 0 100 1000

Nb

/ Zr


0.00

0.02

0.04

0.06

0.08

0.10

0.12

11 0 100 1000

Y /

Zr


Bignell Loess Brady A/Bt max Upper Peoria Loess Deep Peoria Loess

Figure 11. Scatterplots of loess transport distance versus ratios of conservative index elements. Data points are stratifi ed by position relative to stratigraphic and pedologic boundaries and indicate provenance change rather than weathering of Bignell Loess.



REFERENCES CITED

Aide, M., and Smith-Aide, C., 2003, Assessing soil genesis by rare-earth elemental analysis: Soil Science Society of America Journal, v. 67, p. 1470–1476.

Almond, P.C., and Tonkin, P.J., 1999, Pedogenesis by upbuild-ing in an extreme leaching environment and slow loess accretion, south Westland, New Zealand: Geoderma, v. 92, p. 1–36, doi: 10.1016/S0016-7061(99)00016-6.

Blake, G.R., and Hartge, K.H., 1986, Bulk density, in Klute, A., ed., Methods of Soil Analysis, Part 1 (second edition): Madison, Wisconsin, American Society of Agronomy, p. 363–375.

Bronger, A., 1978, Climatic sequences of steppe soils from Eastern Europe and the USA with emphasis on the gen-esis of the “argillic horizon”: CATENA, v. 5, p. 33–51, doi: 10.1016/S0341-8162(78)80005-8.

Bronger, A., 1991, Argillic horizons in modern loess soils in an ustic soil moisture regime: Comparative studies in forest-steppe and steppe areas from Eastern Europe and the United States: Advances in Soil Science, v. 15, p. 41–90.

Caldwell, R.E., and White, J.L., 1956, A study of the ori-gin and distribution of loess in southern Indiana: Soil Science Society of America Proceedings, v. 20, p. 258–263.

Eden, D.N., Wen, Q., Hunt, J.L., and Whitton, J.S., 1994, Mineralogical and geochemical trends across the Loess Plateau: North China: Catena, v. 21, p. 73–90.

Feggestad, A.J., Jacobs, P.M., Miao, X., and Mason, J.A., 2004, Stable carbon isotope record of Holocene envi-ronmental change in the central Great Plains: Physical Geography, v. 25, p. 170–190.

Follmer, L.R., 1982, The geomorphology of the Sangamon surface: Its spatial and temporal attributes, in Thorn, C.E., ed., Space and Time in Geomorphology: London, Allen and Unwin, p. 117–146.

Forman, S.L., Oglesby, R., and Webb, R.S., 2001, Tempo-ral and spatial patterns of Holocene dune activity on the Great Plains of North America: Megadroughts and climate links: Global and Planetary Change, v. 29, p. 1–29, doi: 10.1016/S0921-8181(00)00092-8.

Frye, J.C., Glass, H.D., Leonard, A.B., and Coleman, D.D., 1974, Caliche and clay mineral zonation of Ogallala Formation, central-eastern New Mexico: New Mexico Bureau of Mines Mineral Resource Circular 144, 16 p.

Goble, R.J., Mason, J.A., Loope, D.B., and Swinehart, J.B., 2004, Optical and radiocarbon ages of stacked paleo-sols and dune sands in the Nebraska Sand Hills, USA: Quaternary Science Reviews, v. 23, p. 1173–1182, doi: 10.1016/j.quascirev.2003.09.009.

Gunal, H., 2001, Clay illuviation and calcium carbonate accu-mulation along a precipitation gradient in Kansas [Ph.D. thesis]: Manhattan, Kansas State University, 246 p.

Haseman, J.F., and Marshall, C.E., 1945, The use of heavy minerals in studies of the origin and development of soils: Missouri Agriculture Experiment Station Research Bulletin 387, 75 p.

Huang, C.C., Pang, J., Huang, P., Hou, C., and Han, Y., 2002, High-resolution studies of the oldest cultivated soils in the southern Loess Plateau of China: Catena, v. 47, p. 29–42, doi: 10.1016/S0341-8162(01)00169-2.

Huang, C.C., Pang, J., Chen, S., and Zhang, Z., 2003, Holo-cene dust accumulation and the formation of polycy-clic cinnamon soils (Luvisols) in the Chinese Loess Plateau: Earth Surface Processes and Landforms, v. 28, p. 1259–1270, doi: 10.1002/esp.512.

Hutton, C.E., 1951, Studies of the chemical and physical characteristics of a chrono-litho-sequence of loess-derived prairie soils of southwestern Iowa: Soil Science Society of America Proceedings, v. 15, p. 318–324.

Jacobs, P.M., and Mason, J.A., 2004, Paleopedology of soils in thick Holocene loess, Nebraska, USA: Revista Mex-icana de Ciencias Geologicas, v. 21, p. 54–70.

Jacobs, P.M., and Mason, J.A., 2005, Impact of Holocene dust aggradation on A horizon characteristics and car-bon storage in loess-derived Mollisols of the Great

Plains, USA: Geoderma, v. 125, p. 95–106, doi: 10.1016/j.geoderma.2004.07.002.

Jenny, H., 1941, Factors of Soil Formation: New York, McGraw-Hill, 281 p.

Jenny, H., 1980, The Soil Resource: New York, Springer-Verlag, 377 p.

Joeckel, R.M., and Ang Clement, B.J., 2005, Soils, surfi cial geology, and geomicrobiology of saline-sodic wetlands, North Platte River Valley, Nebraska, USA: Catena, v. 61, p. 63–101, doi: 10.1016/j.catena.2004.12.006.

Johnson, W.C., and Willey, K.L., 2000, Isotopic and rock magnetic expression of environmental change at the Pleistocene-Holocene transition in the central Great Plains: Quaternary International, v. 67, p. 89–106, doi: 10.1016/S1040-6182(00)00011-2.

Jordan, R.H., 1967, Formation and transfer of clay in Hast-ings silt loam, a Udic Argiustoll [M.S. thesis]: Lincoln, University of Nebraska, 99 p.

Kemp, R.A., and Derbyshire, E., 1998, The loess soils of China as records of climatic changes: European Jour-nal of Soil Science, v. 49, p. 525–539, doi: 10.1046/j.1365-2389.1998.4940525.x.

Kemp, R.A., Derbyshire, E., Meng, X.-M., Chen, F.H., and Pan, B.T., 1995, Pedosedimentary reconstruction of a thick loess-paleosol sequence near Lanzhou in north-central China: Quaternary Research, v. 43, p. 30–45, doi: 10.1006/qres.1995.1004.

Kemp, R.A., Derbyshire, E., and Meng, X.-M., 1997, Micro-morphological variation of the S1 paleosol across northwest China: Catena, v. 31, p. 77–90, doi: 10.1016/S0341-8162(97)00028-3.

Kemp, R.A., Derbyshire, E., and Meng, X.M., 2001, A high-resolution mircomorphological record of changing landscapes and climates on the western Loess Plateau of China during oxygen isotope stage 5: Palaeogeogra-phy, Palaeoclimatology, Palaeoecology, v. 170, p. 157–169, doi: 10.1016/S0031-0182(01)00236-X.

Kilmer, V.J., and Alexander, L.T., 1949, Methods of mak-ing mechanical analysis of soils: Soil Science, v. 68, p. 15–24.

Konen, M.E., Jacobs, P.M., Burras, C.L., Talaga, B.L., and Mason, J.A., 2002, Equations for predicting soil organic carbon using loss-on-ignition for north central U.S. soils: Soil Science Society of America Journal, v. 66, p. 1878–1881.

Kuzila, M.S., 1995, Identifi cation of multiple loess units within modern soils of Clay County, Nebraska: Geoderma, v. 65, p. 45–57, doi: 10.1016/0016-7061(94)00030-E.

Mason, J.A., 2001, Transport direction of Peoria Loess in Nebraska and implications for loess sources on the cen-tral Great Plains: Quaternary Research, v. 56, p. 79–86, doi: 10.1006/qres.2001.2250.

Mason, J.A., and Kuzila, M.S., 2000, Episodic Holocene loess deposition in central Nebraska: Quaternary International, v. 67, p. 119–131, doi: 10.1016/S1040-6182(00)00013-6.

Mason, J.A., and Nater, E.A., 1994, Soil morphology–Peo-ria Loess grain size relationships, southeastern Minne-sota: Soil Science Society of America Journal, v. 58, p. 432–439.

Mason, J.A., Jacobs, P.M., Hanson, P.R., Miao, X., and Goble, R.J., 2003, Sources and paleoclimatic signifi -cance of Holocene Bignell Loess, central Great Plains, USA: Quaternary Research, v. 60, p. 330–339, doi: 10.1016/j.yqres.2003.07.005.

McDonald, E.V., and Busacca, A.J., 1990, Interaction between aggrading geomorphic surfaces and the forma-tion of a late Pleistocene paleosol in the Palouse loess of eastern Washington State: Geomorphology, v. 3, p. 449–470, doi: 10.1016/0169-555X(90)90016-J.

Miao, X.-D., Mason, J.A., Goble, R.J., and Hanson, P.R., 2005, Loess record of dry climate and eolian activity in the early to mid-Holocene, central Great Plains, North America: The Holocene, v. 15, p. 339–346, doi: 10.1191/0959683605hl805rp.

Morrison, R.B., 1967, Principles of Quaternary soil stratigra-phy, in Morrison, R.B. and Wright, H.E., eds., Quater-nary Soils: Reno, Desert Research Institute, p. 1–69.

Muhs, D.R., and Bettis, E.A., 2000, Geochemical variations in Peoria Loess of western Iowa indicate paleowinds of midcontinental North America during last glaciation: Quaternary Research, v. 53, p. 49–61, doi: 10.1006/qres.1999.2090.

Muhs, D.R., Aleinikoff, J.N., Stafford, T.W., Jr., Kihl, R., Been, J., Mahan, S.A., and Cowherd, S., 1999, Late Qua-ternary loess in northeastern Colorado: Part I. Age and paleoclimatic signifi cance: Geological Society of Amer-ica Bulletin, v. 111, p. 1861–1875, doi: 10.1130/0016-7606(1999)111<1861:LQLINC>2.3.CO;2.

Muhs, D.R., McGeehin, J.P., Bean, J., and Fisher, E., 2004, Holocene loess deposition and soil formation as com-peting processes, Matanuska Valley, southern Alaska: Quaternary Research, v. 61, p. 265–276, doi: 10.1016/j.yqres.2004.02.003.

Putman, B.R., Jansen, I.J., and Follmer, L.R., 1988, Loessial soils: Their relationship to width of the source valley in Illinois: Soil Science, v. 146, p. 241–247.

Ragon, L.G., 1974, Soil Survey of Adams County, Nebraska: Washington, D.C., U.S. Government Printing Offi ce, U.S. Department of Agriculture Soil Conservation Society Publication, 78 p., 42 map sheets.

Ransom, M.D., Nettleton, W.D., Porter, D.A., and Fraser, S.H., 1997, Micromorphology of Ustolls and Ustalfs from Western Kansas, U.S.A., in Shoba, S., ed., Soil Micromorphology: Studies on Soil Diversity, Diagnos-tics, Dynamics: Wageningen, International Society of Soil Science, p. 139–149.

Ruhe, R.V., 1969, Application of pedology to Quaternary Research, in Pawluk, S., ed., Pedology and Quaternary Research: Edmonton, Alberta, Canada, University of Alberta, p. 1–23.

Ruhe, R.V., 1984, Soil climate systems across the prairies in midwestern U.S.A.: Geoderma, v. 34, p. 201–219, doi: 10.1016/0016-7061(84)90039-9.

Schaetzl, R.J., and Anderson, S., 2005, Soils: Genesis and Geomorphology: New York, Cambridge University Press, 817 p.

Schultz, C.B., and Stout, T.M., 1948, Pleistocene mammals and terraces in the Great Plains: Geological Society of America Bulletin, v. 59, p. 541–630.

Singer, A., 1989a, Illite in the hot-aridic soil environment: Soil Science, v. 147, p. 126–132.

Singer, A., 1989b, Palygorskite and sepiolite group miner-als, in Dixon, J.B., and Weed, S.B., eds., Minerals in Soil Environments: Madison, Soil Science Society of America, p. 829–872.

Smith, G.D., 1942, Illinois loess—Variations in its proper-ties and distribution: Illinois Agricultural Experiment Station Bulletin, v. 490, p. 139–184.

Soil Survey Staff, 1993, Soil Survey Manual: Washington, D.C., U.S. Government Printing Offi ce, U.S. Depart-ment of Agriculture–Soil Conservation Service, Agri-culture Handbook 18, 437 p.

Stoops, G., 2003, Guidelines for analysis and description of soil and regolith thin sections: Madison, Wisconsin, Soil Science Society of America, 184 p.

Swinehart, J.B., Souders, V.L., DeGraw, H.M., and Diffen-dal, R.F., Jr., 1985, Cenozoic paleogeography of west-ern Nebraska, in Flores, R.M., and Kaplan, S.S., eds., Cenozoic Paleogeography of the West-Central United States: Denver, Rocky Mountain Section, Society for Sedimentary Geology (SEPM), p. 209–229.

Thorp, J., Johnson, W.M., and Reed, E.C., 1951, Some post-Pliocene buried soils of the central United States: Jour-nal of Soil Science, v. 2, p. 1–19, doi: 10.1111/j.1365-2389.1951.tb00586.x.

Woida, K., and Thompson, M.L., 1993, Polygenesis of a Pleistocene paleosol in southern Iowa: Geological Society of America Bulletin, v. 105, p. 1445–1461, doi: 10.1130/0016-7606(1993)105<1445:POAPPI>2.3.CO;2.

MANUSCRIPT RECEIVED 6 JULY 2005REVISED MANUSCRIPT RECEIVED 17 MAY 2006MANUSCRIPT ACCEPTED 9 OCTOBER 2006

Printed in the USA

late quaternary climate change, loess sedimentation, and ...€¦ · ondary carbonate appears to be...

Documents