SSSAJ
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 May 2008
Published in Soil Sci Soc Am J 72:650-659 (2008)
DOI: 10.2136/sssaj2007.0112
© 2008 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stiles, C. A.
Right arrow Articles by Stensvold, K. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Stiles, C. A.
Right arrow Articles by Stensvold, K. A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Stiles, C. A.
Right arrow Articles by Stensvold, K. A.
Related Collections
Right arrow Soil Geomorphology and Geography
Right arrow Pedology
Right arrow Soil Analysis

PEDOLOGY

Loess Contribution to Soils Forming on Dolostone in the Driftless Area of Wisconsin

Cynthia A. Stilesa,* and Krista A. Stensvoldb

a National Soil Survey Center, 100 Centennial Mall, Lincoln, NE 68508-3886
b Dep. of Soil Science, Univ. of Wisconsin, 1525 Observatory Dr., Madison, WI 53706-1299

* Corresponding author (Cynthia.Stiles{at}lin.usda.gov).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soils of the southern Driftless Area of Wisconsin are derived from loess mantles and dissolving carbonate from dolostone bedrock found in formations of the Sinnippi Group. Contributions of the two parent components can be estimated by utilizing volumetric compositions of the relatively immobile elements Ti and Zr. The dolostone bedrock is depleted in both elements (3.22 and 0.11 µmol cm–3, respectively) relative to loess derived from sediments of the St. Croix and Upper Mississippi river systems, which contains 184.54 and 7.44 µmol cm–3, respectively. This strong contrast between materials allows contribution indices based on Ti and Zr to be determined using a simple algebraic relationship based on the end-member concentrations. Contribution indices were determined from Ti and Zr concentrations of horizons from 11 pedons in Iowa County in southwestern Wisconsin. These pedons showed typical morphology for the southern Driftless Area, with loess-derived silt loam to silty clay loam Ap and Bt horizons and clay-enriched 2Bt subsoils overlying dolomite residuum (3BC/3C). Mean contribution estimates for the pedons show maximum loess contribution indices (0.802–0.854 kg kgTi basis–1) in the argillic horizons, with lower contribution indices (0.718 kg kgTi basis–1) in the surface horizons due to organic matter dilution and increased porosity (lower bulk density). Mean contribution indices for loess in the horizons just above the dolomitic bedrock are predictably low (3.82 and 1.62 g kgZr basis–1) and not significantly different (P < 0.0005). Contribution indices based on the two elements showed differences in trend with depth due to the individual behavior of the elements, i.e., the affinity of Ti for Fe oxides and clay minerals, and physical transport of small durable zircons moving into the solum through pores. The value of these indices is that they can be utilized to discern not only how much material the weathering loess actually contributed to soil formed on the bedrock surface, but also the relative intensity of argillic horizon development.

Abbreviations: DA, Driftless Area • GD, Galena Formation dolostone • IPL, Ipswich Prairie loess


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Carbonate (limestone and dolostone) enriched bedrock comprises approximately 1647 x 103 km2, or 20%, of the total land area of the contiguous United States. In the states east of the Mississippi River, where rainfall and abundant plant growth accelerate carbonate weathering, 27% of the land area is underlain by carbonate bedrock (~881 x 103 km2; USGS, 2002). Because of their relatively high base and nutrient status, soils formed from carbonate-rich parent materials are often considered prime agricultural and forest production land, and are thus significant ecological components in landscapes where they occur. In humid climatic zones, weathering carbonate is often capped with clay-rich soils characterized by high-chroma red hues (7.5YR, 5YR) and strong angular blocky to prismatic structure, with thicknesses often a function of both the duration of exposure and landscape stability (Olson et al., 1980; Benac and Durn, 1997). These soils, called terra rossa in the Mediterranean regions where they were originally described, have a characteristic redness associated with accumulations of Fe oxides (Glazovskaya and Parfenova, 1974; Jackson and Frolking, 1982; Bech et al., 1997) and are found in temperate to tropical climate zones throughout the world (Stace, 1956; Atalay, 1997).

Once considered to form exclusively on stable landscape positions as the result of long-term residual weathering (Verheye, 1974), terra rossa soils have been found in a variety of geomorphic settings that suggest that erosion and deposition play a role in their formation (Yassoglou et al., 1997), thus countering in situ residual weathering as the primary soil-forming mechanism. It is now commonly accepted that, along with dissolution of the carbonate parent material, external material additions contribute to the formation of these soils (Olson et al., 1980; Rapp, 1984; Moresi and Mongelli, 1988). These mechanisms act in a cooperative fashion, the first being the dissolution of the carbonate rock and the second being the addition and pedogenic alteration of aeolian or colluvial materials.

Congruent dissolution of carbonate minerals, most commonly limestone (CaCO3) and dolomite [(Ca,Mg)CO3], yields a noncarbonate fraction, including elements such as Si, Al, and Fe, which are necessary materials for pedogenic clay formation, and provides increased pore space for secondary soil mineral formation. It is widely accepted that the noncarbonate fraction alone cannot account for the total thickness of many of these soils and that residual weathering as the only source of a soil may only be appropriate in some settings (Olson et al., 1980; Durn et al., 1999). Typical carbonates have fairly low noncarbonate impurity content (global mean weight percentage of noncarbonate components in continental carbonate formations = 20.02%, insoluble mean = 17.17 %; Ronov and Yaroshevskiy, 1976; Ronov, 1982). This suggests that for every meter of soil formed, 5 to 6 m of limestone must be dissolved. Given the thicknesses of some terra rossa soils (>2 m) and the low percentage of impurities inherent in carbonates, particularly in high-purity phases such as dolomite where the impurity content drops to ≤5%, rock thicknesses necessary for forming 1 m of soil substantially increases. MacLeod (1980) estimated that nearly 130 m of high-purity limestone would have to dissolve and the residual materials stay in place to have formed only 40 cm of soil at one location in Greece. Thus, additions of materials on top of the bedrock in the form of loess, colluvium, or alluvium provide materials conducive to the formation of soils, i.e., quartz, feldspars, and micas (Frolking et al., 1983).

The addition of external materials as eolian, colluvial, or alluvial deposits is essential to the second mechanism of terra rossa formation, and related to the first mechanism through the replacement of the dissolved carbonate by components derived from those materials accumulated on top of the existing bedrock saprolite. The presence of carbonate dissolution products (base cations and HCO3–) tends to stabilize pedogenic clay formation and movement (Schaetzl and Anderson, 2005, p. 189), and the dissolution itself provides spatial accommodation porosity for secondary phases including, but not limited to, clay minerals, Fe and Mn oxyhydroxides (Ehrlich et al., 1955; Frolking et al., 1983; White, 1995), and stabilized humates (Levine et al., 1989). Thus, soil formation on carbonate bedrock is the result of two inputs, the degradation of authigenic carbonate and the transformation of surficially added allogenic materials.

The degree to which the two primary processes dominate the formation of terra rossa soils is difficult to ascertain. One tool that holds promise is the use of "immobile" elements, as defined in mass balance studies (sensu Brimhall et al., 1991) to evaluate elemental gains and losses in weathering profiles. In theory, immobile elements remain within pedogenically altered parent materials after the more labile elements have been released from soluble phases in leached zones in the profiles. Under humid-climate soil-forming conditions, both Ti and Zr are commonly considered immobile due to the relatively insoluble nature of the minerals in which they occur, such as zircon (ZrSiO4) and rutile or anatase (TiO2; Marshall and Haseman, 1942; Jackson and Sherman, 1953). In many studies, a test of parent material uniformity is the evaluation of the Ti/Zr ratio, which should remain constant throughout the profile depth if they are both immobile and derived from the same source (Reheis, 1990; Birkeland, 1999, p. 164). Zircon is an extremely durable mineral that can chemically survive numerous geological recycling events and is often added to soil surfaces by aeolian processes (Brimhall et al., 1991). Titanium is more labile and may replace Fe in hydrous Fe oxyhydroxide phases (Kaup and Carter, 1987; Fitzpatrick and Chittleborough, 2002), particularly in finer textured soils. Despite the chemical reactivity of some Ti-bearing detrital phases compared with zircon, Ti is still generally conserved in the bulk soil matrix, particularly in fine-textured, Fe-enriched matrices (Driese et al., 2000; Stiles et al., 2003). This complementary behavior, i.e., the sensitivity of Zr to physical transport and the conservative hydrochemical reactivity of Ti in clays, allows these elements to be particularly useful in determining the contribution of different parent materials to a soil and assessing the intensity of clay genesis and dynamics in this pedogenic setting (Stiles et al., 2003).

The Driftless Area (DA) of southwest Wisconsin (Fig. 1 ) is an ancient landscape of Cambrian through Silurian formations comprised of dolostone and carbonate-intercalated sandstone formed in a shallow tropical marine environment and then exposed since the early Devonian (360 million yr; Choi, 1998). During the intervening time, the bedrock has been mantled by a variety of terrestrial deposits including alluvium, colluviums, and loess (Dott and Attig, 2004). During the Quaternary, the landscape of the DA escaped direct glaciation (Mickelson et al., 1982) but was no doubt strongly influenced by periglacial processes during glacial advances, which left relict ice wedge casts and deposition and erosion features such as talus cones, solifluction rubble at the base of hillslopes, block streams, and deep erosional gullies (Smith, 1949; Johnson, 1990; Clayton et al., 2001). Since that time, the exposed rock has been weathering and mantled with thick layers of windblown silt-sized loess deposits, the most recent deposition associated with the last full glacial period when periglacial conditions favored landscape instability and heavier sediment delivery to drainageways (Leigh and Knox, 1994; Mason and Knox, 1997; Muhs et al., 2001; Bettis et al., 2003). Once loess deposits were stabilized, warmer and wetter conditions accelerated pedogenesis and carbonate dissolution, with the formation of terra-rossa-type soils, especially in interfluve positions of relatively high landscape stability. The red-clay subsoils are formally recognized in Wisconsin as the Rountree Formation, named after a locale in southwestern Wisconsin on the Galena Formation of the Sinnippi Group (Knox et al., 1990). It is generally accepted that the Rountree Formation may have had its origin in pre-Wisconsinan loess deposits that pre-date the Peoria and Roxana Formations found in other regional loess sequences (e.g., Lindbo et al., 1997; Muhs et al., 2001), and may even be a relict from Cretaceous times owing to its association with the Windrow Formation (Andrews, 1958; Knox et al., 1990). The time-transgressive pedogenic nature of this formation makes it nearly impossible to date by any conventional means (Knox et al., 1990). The upper solum of the landscape is formed from loess derived from Mississippi and St. Croix river sediments following retreat of the last glacial advance (Peoria Formation equivalent, Leigh and Knox, 1994). Because of the duration of DA bedrock exposure (throughout the Pleistocene glacials and interglacials, possibly longer), there were several episodes of loess deposition and erosional removal that contributed to forming the upper solum in this landscape. The loess presumably comes from the same source, and thus the mineralogy and elemental composition of the loess has not varied significantly through time. This setting is ideal to evaluate the quantitative contribution of loess deposition to soil formation on carbonate parent material.


Figure 1
View larger version (48K):
[in this window]
[in a new window]

  		Fig. 1. Map of location: (a) planform contour depiction of sampling locations; (b) digital elevation model, oblique view, showing landscape features and sampling locations.

 
The main hypothesis of this investigation is that terra-rossa-like soils formed on the Galena Formation dolostone are primarily the product of external additions of loess with some geochemical contribution of the underlying dolostone. To test this hypothesis, evaluations were made of several pedons in a southern DA landscape formed in this two-component parent material setting, with the objectives being to: (i) quantify Ti and Zr contents in soils and contributing parent materials; and (ii) propose and determine contribution indices from the compositions of these materials and compare the differences to elucidate pedogenic evolution in relatively simple two-component parent material settings.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Geographic Setting
The watershed from which samples were taken for this study is located in Iowa County, Wisconsin, in the southwestern corner of the state. The location is approximately 70 km west of the terminal moraine of the Wisconsinan glaciation and is 68 km east of the Mississippi River and 30 km south of the Wisconsin River (Fig. 1). The present regional soil climatic regime is classified (Soil Survey Staff, 2006) as mesic and udic, with a mean annual air temperature of 10°C and mean annual precipitation of 840 mm (Wisconsin State Climatology Office, www.aos.wisc.edu/~sco/clim-history/index.html, accessed September 2004). Most of the annual precipitation is in the form of rain, which falls during the growing season from May through September (Sartz, 1966). Winter precipitation is typically in the form of snow. The parent material is reworked loess over Galena Formation dolostone. The topography has rounded slope morphology with slopes ranging from 2 to 40%, with steeper slopes along steeply dissected valleys or coulees. Although much of the land of the southern DA is heavily farmed, the area in which the soil pits were excavated had been in the Conservation Reserve Program (CRP) for 5 yr before the investigation, with vegetation dominated by bromegrass (Bromus spp.) with localized stands of mixed perennial forbs. Before CRP enrollment, the land had been in a standard corn (Zea mays L.)–soybean [Glycine max (L.) Merr.]–alfalfa (Medicago sativa L.) rotation. A total of 12 soil pits were excavated, described, and sampled following standard methods described by Schoeneberger et al. (2002). One pedon was excluded in this study because it was deemed to be a cut-and-fill feature in a swale and hence did not fit the standard soil types expected for each landscape position. The geomorphic characteristics, horizon sequences, and taxonomic designations for all the pedons are listed in Table 1 . For all pedons, coherent bedrock was assumed to lie beneath the bottom horizon.


View this table:
[in this window]
[in a new window]

  		Table 1. Landscape characteristics, profile horizonation, and classification of pedons from Iowa County study area.

 
Field descriptions were completed by NRCS staff using standard descriptive procedures for soil properties and geomorphic relationships (Soil Survey Division Staff, 1993). Clod and bulk samples were taken of each horizon down to and including the bedrock and returned to the Soil Characterization Laboratory at the University of Wisconsin-Madison. Four samples of the Galena Formation dolostone (GD) were collected from both the base of a profile in Iowa County and from nearby construction sites where dolostone from the formation was being excavated and disaggregated. The ground GD samples were composited and a subsample taken from the composited powder. Samples of unweathered Upper Mississippi River Valley Peoria loess were obtained from a Giddings core recovered from an uncultivated remnant of Ipswich Prairie, located to the south and west of the Iowa County site (from the collection of J.C. Knox, Univ. of Wisconsin, Geography Dep.), referred to herein as the Ipswich Prairie loess (IPL). Three depth intervals were sampled from this core to account for possible elemental variability of the loess deposits through time: 176 to 180, 330 to 334, and 406 to 410 cm. Because it is impossible to obtain samples of unaltered loess from earlier depositional episodes, the IPL composite was used in this study as the analog parent material for any loess contributions to the landscape through time, presuming that the source did not change throughout the Pleistocene. All samples were air dried and the bulk samples were split for a portion to be ground to pass a 125-µm (120 mesh) sieve for pressed powder pellet preparation. In the case of the GD, the sample was initially pulverized in a shatterbox and no further grinding was required before pressing a portion into a pellet. All samples were dried at 105°C before being made into pellets. Total elemental concentrations were determined using x-ray fluorescence (XRF) on a Bruker AXS Model 3400 spectrometer (Bruker AXS Corp., Fitchburg, WI) (Singer and Janitsky, 1986; Karathanasis and Hajek, 1996). Loss-on-ignition at 1025°C was determined on a subsample of each ground sample for entry into the XRF sample input system to recalculate total elemental contents and account for compounds or elements not evaluated in the scans. Bulk densities ({rho}b) of three clod samples per horizon were determined by the paraffin-coated air-dried clod method (Blake and Hartge, 1986). Statistical analyses of data used the Student's one-tailed t-test of two sample populations with unequal variances statistical subroutine in Quattro Pro (Version X3, Corel Corp., Ottawa, ON) to determine levels of significance. Correlation of parameters was evaluated by using the linear regression subroutine in SigmaPlot (Version 10, Systat Software, San Jose, CA).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Titanium and Zirconium Trends
Trends of decreasing Ti and Zr contents with depth were noted in all profiles examined (Table 2 ). To facilitate comparisons across multiple pedons, Table 2 is presented such that each row represents an individual pedon and the columns are the horizons found in that pedon. Note that not all horizons were found in all pedons. The profile descriptions indicate that these soils are polygenetic in nature, with three lithologic units (topsoil and two lithologic discontinuities). A lithologic discontinuity describes a distinct change in parent material, particle-size distribution, grain morphology, or color. It is presumed that the three lithologic units in this landscape, from the top down, are comprised of soil materials derived from (i) the most recent Wisconsinan (Peorian) loess accumulation, (ii) pedogenically altered loess deposits that accumulated during earlier glacial intervals, and (iii) GD residuum. The values given for Ti and Zr contents of the IPL are means of the three depth intervals that were sampled. Although there were slight differences in compositions with depth in the core, a trend originally noted in an earlier study by Mason and Jacobs (1998), the variation is relatively small and thus is not considered as a separate issue in the interpretations of the data.


View this table:
[in this window]
[in a new window]

  		Table 2. Concentrations of Ti and Zr in pedon horizons from the Iowa County study area.

 
Titanium/zirconium concentration ratios support the differentiation of three lithologic discontinuities found in these pedons (Fig. 2 ). The upper horizons (Ap and Bt) have Ti/Zr ratios that fall generally within a narrow range of values and are statistically similar (P < 0.01), confirming that these horizons are derived from one lithologic unit. The 2Bt horizon Ti/Zr ratios are statistically different from the overlying horizons (P < 0.0005), confirming that these argillic horizons comprise a second lithologic unit. A third lithologic unit is indicated by the Ti/Zr ratios of the 3Bt upper horizon values, ranging from 19.37 to 30.89, which are again statistically different from both of the overlying lithologic groupings (P < 0.0005). At depth in the 3Bt (lower)–3BC–3C horizons, however, the Ti/Zr ratios have a wide range of values (18.00–38.18), some of which are similar to values found in the overlying upper 3Bt horizon values and approaching the mean Ti/Zr value of the GD. Most significantly, the parent material Ti/Zr values are similar to each other (13.02 and 15.00 for IPL and GD, respectively) and to the upper horizons of the sola. Although there are definitely differentiating trends in the Ti/Zr ratios that indicate lithologic discontinuities possibly attributable to different parent materials or depositional episodes, the findings to support this are inconclusive due to the ambiguous Ti/Zr values of the two parent materials. Thus, evaluation of Ti/Zr ratios by depth and horizon is not helpful in discerning the contributions of parent materials identified in this landscape.


Figure 2
View larger version (13K):
[in this window]
[in a new window]

  		Fig. 2. Titanium/zirconium ratios of horizon groupings from the Iowa County locations. Parent materials are indicated by their abbreviations: GD = Galena Fm. dolostone; IPL = Ipswich Prairie loess.

 
Contributions of different parent materials can be more clearly discerned by plotting the relationships of Ti/Zr ratios in soil as functions of the concentrations of component elements (Fig. 3 ; Stiles et al., 2003). The Ti/Zr ratio values are dependent on the concentrations of those components, which do not covary consistently in the pedons of this study. In Fig. 3A, horizon groupings occur in vertical trends, suggesting that concentrations of Ti are similar within each horizon, which is also evident from the data presented in Table 2. In contrast, the linearity of the distributions in the upper three horizon groups (Fig. 3B) indicates that shifts in the Ti/Zr ratios are dictated primarily by changes in Zr concentrations, with concentration generally decreasing with depth. Horizons deeper in the profiles, however, have low concentrations of both elements and thus the linear relationship shown in Fig. 3B breaks down and there is wide scatter in the ratios within a relatively narrow range of concentrations. At low concentrations, small changes in concentration yield great variability in the ratio product. Some of the Ti/Zr values at depth again approach those of the upper solum ratios and the usefulness of ratios for discerning lithologic discontinuities is compromised.


Figure 3
View larger version (14K):
[in this window]
[in a new window]

  		Fig. 3. Titanium/zirconium ratios as functions of elemental concentrations: (A) Ti-based relationship; (B) Zr-based relationship.

 
One notable feature in Fig. 3 is the significant difference of the Ti and Zr concentrations between the two parent materials, loess and dolostone. The loess has concentrations nearly two orders of magnitude higher than dolostone (Table 2, loess: Ti = 5290 mg kg–1 and Zr = 406 mg kg–1; dolostone: Ti = 60 mg kg–1 and Zr = 4 mg kg–1). Combined with the obvious decrease in elemental compositions with depth, this substantial difference lends itself to evaluating the contribution percentages of both parent materials by evaluating Ti and Zr inputs as functions of horizon and parent material compositions as well as also evaluating the intensity of pedogenic processes as a function of the difference between the index values with depth.

Contribution Indices
To do a meaningful evaluation of contribution trends by index element, differences in {rho}b need to be taken into account, as well as the difference in molecular weights between the two elements. Differences attributable to variations in parent material contributions can be enhanced by converting mass concentrations to micromoles and then multiplying by the bulk densities to obtain a unit of measure—micromoles per cubic centimeter—that clarifies the true compositional content of these two elements in the pedons. By representing the values used in the calculation of parent material contributions as mass per volume, a unit of measure used in mass balance assessments (Brimhall and Dietrich, 1987), any change in {rho}b within the profiles is taken into account. Conversion of concentrations to molar values allows for one-to-one comparison of the contribution factors without distortion due to differences in the atomic weights of the two elements.

Table 3 shows the volumetric molar contents of the pedons, in similar fashion to Table 2. The strong contrast between loess and dolostone compositions is attenuated by the differences in {rho}b of the loess (1.67 g cm–1) and the dolostone (2.60 g cm–1). Means and ranges of the {rho}b values for the horizons are shown in Table 4 . The highest mean value for {rho}b is found within the subsoils of the second lithologic discontinuity. This is not unexpected, considering that the overlying horizons have increased porosity due to bioturbation and roots. The explanation for lower {rho}b in the third lithologic discontinuity 3Bt and 3BC horizons is due primarily to the increasing content of residuum coarse fragments and related fracturing induced by the presence of resistant coarse fragments (chert) within smectite-enriched zones in these horizons. Episodic drying that occurs both at the height of the growing season, when evapotranspiration is greatest, and again in the coldest months of winter, when water infiltration is limited, probably promotes shrink–swell processes that cause fracturing. Roots growing into these fractures have been observed in numerous such profiles examined in the field (unpublished data, 2003). The highest {rho}b is understandably within the 3C materials, where in situ residuum is dominant. This material is generally sandy in texture and can be carved with a knife to show lithologic structures.


View this table:
[in this window]
[in a new window]

  		Table 3. Volumetric molar concentrations of Ti and Zr based on gravimetric concentrations and bulk densities of soil materials from the Iowa County study area.

 

View this table:
[in this window]
[in a new window]

  		Table 4. Bulk density means and ranges for horizons.

 
The contribution indices are calculated as a binary function of the two contributing sources: IPL and GD. Given that the molar volumetric values for IPL are 184.5 and 7.4 µmol cm–3 for Ti and Zr, respectively, and these same values for GD are 3.3 and 0.1 µmol cm–3, a simple arithmetic function can be set up to calculate the degrees of contribution to the soil materials made by the loess:

Formula 1[1]
where ICi,jL is the contribution index of the loess (L) based on element j (either Ti or Zr) for horizon interval i; (mol {rho}b)i,j is the molar volumetric concentration of element j within horizon interval i; and (mol {rho}b)j,GD and (mol {rho}b)j,IPL are molar volumetric concentrations of element j in the dolomitic parent material and loess parent material, respectively. Because two of the three variables are fixed [(mol {rho}b)j,GD and (mol {rho}b)j,IPL] with set values listed above, the equation can be simplified to

Formula 2[2]

Formula 3[3]
Thus, we are then able to determine, using geochemical and physical parameters, the degree to which the loess mantle has contributed to the composition and pedogenic processes in southern DA soils. Although it was not performed in our study, we recognize that inclusion of variances of parent material molar concentrations to the equations would account for analytical uncertainty and strengthen the interpretive power of observed trends.

Loess Contribution Trends
Figure 4 shows the range of values for loess contribution trends based on Ti and Zr as functions of depth. In this figure, points represent mean contribution index values as a function of mean depths of the horizon groups and the bars are ranges of those variables. The two graphs are not identical, which confirms that the two elements, although both geochemically conservative in soils, do not behave consistently with each other and indicates the differences in pedogenic processes that dictate their behavior in the lithologic units.


Figure 4
View larger version (14K):
[in this window]
[in a new window]

  		Fig. 4. Loess contribution indices based on (A) Ti concentrations and (B) Zr concentrations. Range bars are given for bottom depth of horizons in the profiles (horizontal bars) and contribution index values (vertical bars). The centroid indicates the mean value for all like horizons within the category. The dashed line indicates decreasing slope between the two argillic horizons in the upper solum; the solid line indicates the best-fit linear correlation of contribution indices with depth in the second lithologic unit.

 
As expected, the largest loess contribution values occur within the Ap and Bt horizons of the upper solum. Contribution indices calculated on a Ti basis show this trend particularly well. For both elements, the maximum index values did not occur in the Ap but rather in the subsoil Bt1 and Bt2 horizons. This is not surprising, considering that the Ap horizons are highly enriched in organic matter, have lower overall {rho}b (Table 4), and have probably been amended in the past when the area was under standard crop rotation. The translocation of materials out of the uppermost portion of the solum through lessivage probably accounts for Ti removal as well. In Fig. 4A, the horizontal groupings of contribution indices by lithologic unit suggest that Ti is conserved within the argillic horizons of each sequence, as maximum clay content (range 50–80%, data not shown) occurs in these horizons. Also, the highest concentration of Fe (9.5–20.1% [w/w] Fe2O3, data not shown) occurs in association with the clays in the lower Bt horizons of the second lithologic unit and the top of the third lithologic unit. Studies of Ti behavior in soils strongly suggests that Ti will remain in association within the structure of Fe (oxy)hydroxides and clay minerals (Fitzpatrick and Chittleborough, 2002), hence conserving the element within Fe-enriched horizons.

The ability of Ti to act as a sole reference for parent material contribution is compromised by its secondary association with pedogenic Fe and clay phases. The groupings and plot orientation of the Ti data (narrow contribution factor ranges), however, suggest that there are differences in the primary formation mechanisms. The upper solum and the second lithologic unit fall into a statistically similar range (P < 0.206) and the third lithologic unit has a statistically different range from the other two (P < 0.0005) (Fig. 4A). These groupings support the idea that the upper portion of the solum and the second, older loess are formed primarily through loess weathering and organic enrichment, and the third unit formed (and possibly continues to evolve) from lessivage of loess-derived materials into the dolomitic residuum, as suggested by Frolking et al. (1983).

The contribution index trend demonstrated by Zr (Fig. 4B) is significantly different than that of the Ti-based values. Zirconium is far less likely to be hydrogeochemically altered in these soils and also is a distinctive indicator of aeolian inputs (Stiles et al., 2003). As was the case in the Ti-based contribution indices, the highest value occurred not in the uppermost surface horizon (Ap), but in the Bt horizons of the upper solum. Again, this has to do with biological activities decreasing the {rho}b and possible inputs from past farming operations. Mean values from the second lithologic unit are negatively linear as depth increases, losing roughly 0.005 of a unit for every centimeter of depth (indicated by the solid line in Fig. 4B described by the regression equation ICZr,L = 0.908 – 0.005(depth in cm), R2 = 0.986, {sigma}2 = 0.024. This also supports the postulation that the upper portion of these soil profiles is derived primarily from the loess, but the lower portion is derived from the secondary illuviated products of loess weathering into the dolostone bedrock.

Zirconium-based contribution indices in the lowermost horizons have a nearly horizontal distribution, with values significantly different from the second lithologic unit (P < 0.0005). Since the third lithologic unit is most closely associated with the dolomitic residuum and largely isolated from direct aeolian inputs (low Zr content), the contribution indices are low, ranging between just under 0.15 to nearly 0. At this depth, the soils have formed almost solely from the dissolution of carbonate and precipitation of secondary clays and Fe and Al (oxyhydr)oxides formed from the weathering of primary minerals in the loess.

The Galena dolostone is comprised of primarily fine-sand-sized baroque to saddle dolomite crystals cemented together with calcite (fine sand/total sand mean = 0.60). Exposure to acidic pore fluids charged with dissolved organic acids from vegetation decomposition degrades the intergranular cement, leaving the more resistant dolomite grains, which is often why the 3BC and 3C horizons have coarser textures. This dissolution process also promotes clay intercalation into the intergranular spaces once the flocculative effects of the carbonate have been overcome. Thus, the clay enrichment at depth in the third lithologic unit has been enhanced by the presence of the carbonate dissolution products, and soil genesis is ongoing at the soil–bedrock interface as long as weatherable materials exist in the overlying horizons.

Interpretations of Contribution Trends
The issue of terra rossa pedogenesis contends that these clay-rich soils formed from three possible mechanisms: (i) isovolumetric replacement of dolomite with clay by migrating pore fluids; (ii) supergene enrichment of the relatively insoluble components contained in the carbonate bedrock; or (iii) evolution from external inputs onto the surface of the bedrock. It is most probable that varying combinations of these mechanisms are responsible for the pedogenesis of terra rossa soils, a postulation supported by utilizing Ti and Zr, relatively immobile elements that occur in greatly differing concentrations in the two-component parent material setting evaluated in this study. Ratios of Ti to Zr, although indicative of lithologic discontinuities, are not helpful in discerning the degree of contribution of aeolian inputs to the carbonate bedrock simply because the ratios for the two parent materials are very similar.

The use of arithmetically derived contribution indices, specifically focusing on the loess as the additive material to the carbonate, is much more demonstrative in confirming that both parent materials contribute to the formation of soils in these systems. Contribution indices calculated on the two elements were not identical, nor did their trends behave correspondingly with depth. Figure 5 illustrates the mean contribution index values based on the two different elements as a function of depth. The bar extending between the symbols shows the differences between the values, attributable to differential geochemical behaviors of these elements. In the shallowest horizons (Ap, Bt1, and 2Bt1), the contribution of loess determined by Zr is higher than the Ti indices because Ti was conservatively translocated out of these horizons into the deeper argillic horizons and Zr was residually enriched. In the 2Bt2, 3, and 4 horizons, the difference between contribution index values calculated on the basis of the two elements is maximized. Because Ti was retained within the pedogenic clays and secondary oxyhydroxides, the Ti contribution index values are consequently higher than Zr values, which more accurately reflect true levels of loess inputs into the developing profiles. The differences lessen considerably as the parent dolostone is contacted in the third lithologic unit (3BC and 3C horizons), which strongly suggests that within this unit, the pedogenic formation process has shifted from loess weathering and argilluviation to carbonate dissolution.


Figure 5
View larger version (10K):
[in this window]
[in a new window]

  		Fig. 5. Comparison of contribution indices determined from Ti and Zr, with differences indicated as a bar between the symbols.

 

   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
In situations where soils are forming from two parent materials, it is challenging to discern the degree of contribution from each component and the dominant processes that lead to formation of the existing soils. If the parent materials have a strongly contrasting geochemical composition, particularly in concentrations of the relatively immobile elements Ti and Zr, pedogenic processes can be elucidated by determining contribution indices based on these elements. Calculation on a mole per volume basis allows direct comparisons between the contribution trends that indicate differences and similarities in those processes.

In our setting of loess over carbonate bedrock, the trend noted in the Zr-basis index values indicates that direct loess contributions diminish with depth until they reach nearly nil in the 3C horizons. The differences throughout the second lithologic unit indicate that argilluviation (i.e., Ti-basis trends) is very prevalent and conserves pedogenic materials derived from loess mineral component weathering in the solum. Finally, the diminishment of the difference between the two contribution index trends in the third lithologic unit reveals that carbonate dissolution is the dominant mechanism of soil genesis at depth.

In regions underlain by carbonate bedrock that received regular influxes of loess during glaciations, soils formed from both the dissolution of the bedrock and the addition of fresh weatherable materials from the loess. The use of contribution indices allows us to evaluate not only the veracity of this process, but also the intensity of the transformation and translocation in the argillic horizons. From the contribution indices, we can see that the clay-rich subsoils of the loess-blanketed hills of the southern Driftless Area of Wisconsin are derived from additions of loess through time, which weathered and then stabilized in the presence of the underlying dolostone. And thus we defend that both parent materials play significant roles in the formation of the terra-rossa-like soils of this area. Contribution indices will prove to be useful tools to elucidate pedogenic processes in any area where soils are forming in geochemically contrasting parent material systems.


   ACKNOWLEDGMENTS
 
This study was funded by USDA-CREES Project no. WIS04753. We wish to thank Duane Simonson, Chanc Vogel, and Phil Meyer of the Wisconsin USDA-NRCS for choosing and excavating the site and providing valuable descriptive information. Our thanks to Dave Kromm of Mineral Point, who graciously allowed us to use his land. We also thank Kathleen Arrington, UW-Soil Science, for creating the ArcGIS figure of our study area.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication March 21, 2007.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stiles, C. A.
Right arrow Articles by Stensvold, K. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Stiles, C. A.
Right arrow Articles by Stensvold, K. A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Stiles, C. A.
Right arrow Articles by Stensvold, K. A.
Related Collections
Right arrow Soil Geomorphology and Geography
Right arrow Pedology
Right arrow Soil Analysis

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome