| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Natural Resource Ecology Lab., Colorado State Univ., Fort Collins, CO 80523 USA
johan{at}nrel.colostate.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Abbreviations: CT, conventional tillage • NSI, normalized stability index • NT, no-tillage • NV, native vegetation • SOM, soil organic matter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Aggregate stability is often used as a measurement of soil structure. Aggregate stability has been shown to be a good indicator for erodibility (Chan and Mead, 1988; Coote et al., 1988). However, aggregate stability is often measured on a specific aggregate size class which is not a measurement of whole soil structure. The mean weight diameter (MWD), on the other hand is an index that characterizes the structure of the whole soil by integrating the aggregate size class distribution into one number. The MWD has often been used to indicate the effect of different management practices on soil structure. For example, 2 yr of moldboard plowing and chisel plowing significantly reduced the MWD of water-stable aggregates in comparison to no-tillage (Angers et al., 1993b). Haynes and Francis (1993) reported an increase in MWD after 32 mo in the order perennial ryegrass > annual ryegrass > perennial white clover = barley.
The use of MWD, however, is questionable if the aggregate distribution is skewed, that is, relatively nonsymmetrical (Stirk, 1958). In addition, there are often complications when different sites and/or management practices are compared for soil structural differences by means of the MWD. Three confounding factors have been identified: pretreatment of soil samples (Beare and Bruce, 1993; Gollany et al., 1991), antecedent water content (Angers et al., 1993a; Perfect et al., 1990), and sand content (Angers et al., 1993b; Caron et al., 1992; Elliott, 1986; Gollany et al., 1991; Perfect et al., 1990).
Beare and Bruce (1993) compared four pretreatment effects (air-dried, capillary wetted; air-dried, tension wetted; air-dried, slaked; field moist, capillary wetted) on water stable aggregation. They found that the field moist, capillary wetted treatment had the least variability. However, this pretreatment causes aggregation to be a function of antecedent water content (Perfect et al., 1990; Gollany et al., 1991; Angers et al., 1993a). A negative correlation was found between the antecedent water content and aggregate stability when soils were prewetted to field capacity, but antecedent water content and aggregate stability were positively correlated when soils were analyzed at field moisture (Gollany et al., 1991). Caron et al. (1992) also observed a decrease in aggregate stability with increased water content when measured on field moist samples rewetted to field capacity. The positive correlation between aggregation and antecedent water content is due to a decreased pressure build-up and aggregate disruption (slaking) upon fast wetting when the soil is at high antecedent moisture content. In contrast, when the soil is capillary wetted to field capacity from field moist conditions, aggregation is related to the length of antecedent drying time, because inorganic cementing agents precipitate at points of particle contact under dry conditions (Kemper and Rosenau, 1984). Consequently, aggregate stability measurements on field moist or field moist, capillary-wetted samples show a seasonal variability in aggregation which is partly induced by pure physical processes related to seasonal differences in water content. Since this physical induced variability is not related to the soil quality it is not desirable in assessments of soil quality.
The confounding effect of sand size distribution is a result of preferential accumulation of sand within certain size fractions during sieving. Sand of the same size as the aggregate is most likely not within the aggregates but is retained on the sieve together with the aggregates and therefore weighed as an aggregate. As a result, aggregation is expected to be higher (if not corrected) for a sample with a high proportion of coarse sand compared to a sample with a high proportion of fine sand. Consequently, differences in sand size distribution confound the measurement of aggregate distribution and structural stability of the soil.
In 2:1 clay-dominated soils, SOM is a major binding agent because polyvalent metal–organic matter complexes form bridges between the negatively charged 2:1 clay platelets. However, SOM is not the only major binding agent in oxide and 1:1 clay mineral dominated soils. Part of the soil stability in oxide and 1:1 clay dominated soils is induced by the binding capacity of oxides and 1:1 minerals (Tisdall and Oades, 1982; Oades and Waters, 1991). Consequently, the mineralogical characteristics of a soil can influence the potential soil stability and the relationship between SOM content and soil stability.
Since each pretreatment has its own specific advantages and disadvantages, Beare and Bruce (1993) suggested comparing the responses of soils to different pretreatments to study the effects of environmental factors on soil structure. The objectives of this study were: (i) to develop a single index for soil stability based on two different pretreatments, which takes into consideration the confounding effects of sand size distribution, antecedent water content, and sample pretreatment, (ii) to test the index for applicability of detecting effects of agricultural practices on soil structure, and (iii) to study the effect of mineralogy on soil stability.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
|
Normalized Stability Index
Rationale for Normalized Stability Index
The normalized stability index (NSI) is an index improved from the conceptually defined aggregation index (AI) developed by van Steenbergen et al. (1991). The NSI measures the stability of the soil by comparing the aggregate distribution before and after disruption. We chose the rewetted and slaked aggregate distribution as the initial distribution and the distribution after disruption, respectively. The rewetted aggregate distribution is considered the initial aggregate distribution because maximum aggregate yield is achieved after rewetting the soil to a moisture content of field capacity plus 5% (g/kg) (Hofman and De Leenheer, 1975). The slaked aggregate distribution is taken as the aggregation level after disruption.
By air drying the soil, the effect of antecedent water content, as observed by Perfect et al. (1990), Gollany et al. (1991), Angers et al. (1993a), and Caron et al. (1992), is minimized. However, precipitation of inorganic binding agents is favored upon drying and increases with time of storage (Kemper and Rosenau, 1984). It has been suggested that an increase in surface acidity upon drying also increases binding between organics and particles (Caron et al., 1992). By subtracting the slaked distribution from the rewetted distribution, the increased aggregation due to precipitation of inorganic binding agents and increased adsorption of organics onto particles is nullified. It has been shown that 24 h soaking of soil in distilled water does not influence the effect of air drying on aggregation (Caron et al., 1992; Kemper and Koch, 1966). Therefore, there is no interaction between the air drying and rewetting of the soil and subtracting the slaked from the rewetted aggregate distribution does indeed nullify the effect of air drying.
Since sand of the same size as the aggregate size class (= aggregate-sized sand) is unlikely to be a part of an aggregate, whereas sand with a smaller size than the aggregate cut-off size is certainly part of the aggregate, it is necessary to correct for the aggregate-sized sand content. This is in contrast to the AI calculated by van Steenbergen et al. (1991) which corrects for the whole sand content of the aggregates. The NSI also normalizes the instability of the aggregates for the maximum disruption level possible because a sandy soil texture has an inherently lower maximum disruption level than a clay soil.
Calculation of Normalized Stability Index
The formula for calculation of disruption level of a size class upon slaking (DLSi) is
 | (1) |
,
,
,
. The factor before the multiplication sign in Eq. [1] calculates the disruption level caused by slaking and corrects the proportions of aggregate size classes for aggregate-sized sand content. This factor also ensures that only weight losses are used in the calculation of the index, that is, this factor is 0 if there is a weight gain. The factor after the multiplication sign normalizes the weight loss to the maximum weight loss for that size class. This normalization is done because a loss of 5% from an aggregate size class with an initial proportion of 10% indicates a lower stability than a 5% loss from a fraction with an initial proportion of 50%. In contrast to the disruptive value calculated by van Steenbergen et al. (1991), the DLSi is based on weight losses and not on weight gains. We used weight losses because this enabled us to correct for aggregate-sized sand content. With weight gains, the correction for sand is only possible if the whole sand content of the aggregates is used. This is a result of the change in aggregate distribution upon disruption; weight gains occur in the smallest size classes. However, the sand is associated with the larger size classes. If the calculations are based on weight gains then the larger size classes are not used in the calculations and consequently neither is the sand distribution. Only when the whole sand content of the aggregates is used, would the sand correction come into the calculation by difference as a cumulative sand content of the larger size classes where a weight loss occurs. However, when weight losses are used in the calculations, the weight losses occur in the larger size classes with which sand is associated; the aggregate-sized sand content can be used for the correction.
The whole soil disruption level (DL) is then calculated as
 | (2) |
The maximum disruption [DLSi (max)] is calculated with the following formula:
 | (3) |
Pp = primary sand particle content with the same size as the aggregates size class after complete disruption of the whole soil. Whole soil DL (max) is calculated with Eq. [2], except DLSi is replaced by DLSi (max).
The normalized stability index (NSI) is then computed as
 | (4) |
The DL is divided by the DL (max) to normalize the DL for the maximum disruption possible based on the primary sand particle-size distribution.
We calculated the mean weight diameter (MWD) for all our soils to compare and indicate the differences between NSI and MWD in interpretation of cultivation effects on soil stability.
Mineralogical Analyses
A 50-g subsample was taken from the 8-mm-sieved soil from each replicate of the 0- to 5-cm CT samples and sieved through a 2-mm sieve. The 0- to 5-cm samples of CT were chosen because they represent the 0- to 20-cm soil layer due to mixing by plowing. The 2-mm-sieved soil was treated with 30% H2O2 at 60 to 70°C until there was no further reaction. For x-ray diffraction analyzes, the samples were rinsed with deionized water and were shaken for 18 h to disperse the soil. The <20 µm fraction was isolated by sieving and suspended in 250 mL deionized water. Oriented samples for x-ray diffraction were made by the millipore filter transfer method (Moore and Reynolds, 1997). A 10-mL suspension was deposited by vacuum filtration on a 0.20-µm Gelman GA Metricel filter, 47-mm diam. The filter was then inverted and laid down on a glass slide. The sample filter-glass slide was partially dried at 50°C and the filter stripped off. X-ray analyzes were done on air-dried and ethylene glycol treated samples with a SCINTAG XRD (CuK
radiation). Vermiculite was identified by a collapse of the 14 
spacing upon heating (1 h, 180°C). Samples were scanned from 5 to 45° 2
.
Noncrystalline components were determined with the citrate–ascorbate (CA) method described by Reyes and Torrent (1998). Most studies use the acid ammonium oxalate (AOD) method for determination of the noncrystalline components. We chose, however, this new method because of: (i) the similar amounts of Fe extracted by AOD and CA but higher selectivity of CA (Reyes and Torrent, 1998) (ii) the simplicity of the method and (iii) the use of nontoxic chemicals. The procedure used was: 250 mg H202 treated whole soil was weighed out in centrifuge tubes and 50 mL 0.2 M sodium citrate–0.05 M sodium ascorbate solution
was added. Samples were shaken for 16 h and centrifuged. The supernatant was analyzed for Fe, Al, and Si by atomic absorption spectrophotometry. The residue of the CA extraction was used for the determination of "free" sesquioxides with dithionite (Blakemore et al. (1987). One gram sodium dithionite and 50 mL of 0.75 M sodium citrate were added to the residue and shaken for 16 h. The suspension was centrifuged and Fe, Al, and Si concentrations in the supernatant were determined by atomic absorption spectrophotometry.
Statistical Analyses
Data were analyzed as a complete randomized block design using the SAS statistical package for analysis of variance (ANOVA-GLM, SAS Institute, 1990). Within depth, tillage treatment was the main factor in the model, with replicate as secondary factor. Separation of means was tested using Tukey's honestly significant difference with a 0.05 significance level. Since there was no replication for the NV treatment at Wooster, KBS and Lexington the NV data for these sites was not included in the statistical analysis.
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
|
|
|
|
The NSI was similar for both depths (0–5 and 5–20 cm) in the CT treatments at all sites (Fig. 3). This was a consequence of mixing during plowing which resulted in similar soil properties over the whole plow depth. Both rewetted MWD and slaked MWD, on the other hand, showed differences in stability between depths in CT at some of the sites.
The lower stability value (NSI) in the surface layer (0–5 cm) compared to the subsurface layer (5–20 cm) in NV and NT (Fig. 3) indicates that surface related processes have an influence on aggregate stability (Paustian et al., 1997). Wet–dry cycles, freeze–thaw cycles, and raindrop impact lead to aggregate destabilization and disruption (Kay, 1990). The soil surface is the most susceptible to wetting and drying, raindrop effect and freeze–thaw cycles. Therefore, it is not surprising that the aggregate stability tended to be higher in the subsurface layer compared to the surface layer in NV and NT. In addition, every plowing event in CT brings up depth protected aggregates to the surface layer where they are exposed to wet–dry and/or freeze–thaw cycles, and raindrop impact (Paustian et al., 1997). This repeated exposure of aggregates decreases the amount of stable aggregates in CT. In the soils dominated by 2:1 clays, the NSI differed generally in the order: NV > NT > CT (Fig. 3), which indicates that the NSI is a good indicator for detrimental effects of agricultural practices on soil structure.
In the surface layer, NSI was significantly different between NT and CT at Sidney and Wooster, but not at KBS and Lexington (Fig. 3). The similarity of NSI for NT and CT at KBS is probably a result of the young age of the experiment (9 yr old). Total carbon, particulate organic matter C and aggregate distribution were also not different between NT and CT at this site (Six et al., 1999, 2000). The highest soil stability for NT and CT was observed at Lexington. The similar NSI value for all management treatments at Lexington indicates that the soil stability was not strongly affected by cultivation. However, total C and particulate organic matter C were significantly lower in CT compared to NT at Lexington (Six et al., 1999). Therefore, the soil stability of this soil does not seem to be related to SOM content as is often observed in temperate soils. We suggest that the specific mineralogy of the soil in Lexington could be the cause of this higher stability (see below).
Before we can use the NSI as an effective indicator for soil quality across sites and ecosystems, however, we have to understand the effect of driving variables on NSI (Elliott, 1997). Driving variables are factors controlling soil quality that are themselves not influenced by soil quality. For example, soil organic matter is not a driving variable of soil quality because it affects aggregate stability or NSI, but it is also related to soil quality itself and is influenced by aggregate stability. Mean annual temperature, mean annual precipitation, clay content and clay type, on the other hand, are driving variables that influence soil structure independently from the quality of the soil and are not affected by aggregate stability. Therefore, the statistical relationships between these driving variables and NSI must be developed before we can account for the effect of driving variables on the potential NSI for a soil and use the NSI as an effective indicator. If a direct comparison between management treatments is not possible then a valid judgement of the soil quality for a specific soil can only be made in comparison to the potential soil structural quality for that specific soil.
Mineralogical Effect
The effect of mineralogy is illustrated in the comparison between the Lexington soil and the three other soils (Fig. 3). The Lexington soil is characterized by a clay mineralogy dominated by kaolinite and vermiculite whereas the other soils are dominated by chlorite and/or illite (Table 3)
. In addition, and of even more importance for soil stability, the Lexington soil contains 2 to 16 times more Fe and Al extracted by citrate–ascorbate (Feca, Alca) and dithionite (Fed, Ald) compared to the other soils (Table 4) . Compared to literature values of sesquioxide concentrations in Oxisols (Colombo and Torrent, 1991; Pinheiro-Dick and Schwertmann, 1996; Reyes and Torrent, 1998), the Feca and Alca concentrations are similar, whereas the Fed and Ald are rather low in the Lexington soil. This indicates that, even though the Lexington soil is not an Oxisol (Table 1), it contains a substantial amount of amorphous and poorly crystalline oxides.
|
|
Evidence for the flocculation capacity of kaolinite has been reported. Bühmann et al. (1996) found that 2:1 clay minerals were on average slightly more easily disaggregated than kaolinite. In addition, Seta and Karathanasis (1996) found a significant negative correlation between kaolinite content and water dispersibility of soil aggregates. The high flocculation capacity of kaolinite is caused by electrostatic interaction between the positive charges on the edges of the clay platelets and the negative charges in the body of the crystal (Dixon, 1989; El-Swaify, 1980; Schofield and Samson, 1954; Tama and El-Swaify, 1978); both charges co-exist at prevailing field pH (El-Swaify, 1980).
In addition to the stabilizing effects of Fe- and Al-oxides and kaolinite by themselves, interactions between the two components have been reported (Seta and Karathanasis, 1996; El-Swaify, 1980). Iron oxide deposits on kaolinite platelets have been observed by Kitagawa (1983) and Fordham and Norrish (1983). This adsorption of Fe- and Al-oxides on the few negative charged sites of kaolinite reduces the CEC of kaolinite and increases the positive charge property of the kaolinite (Dixon, 1989). Therefore, this interaction between Fe- and Al-oxides and kaolinite is synergetic and increases the aggregation potential of kaolinite. Schofield and Samson (1954) observed also electrostatic interactions between negatively charged 2:1 minerals and positively charged kaolinite edge faces by x-ray examination. Since vermiculite (which is present in the Lexington soil (Table 3) has the highest CEC (or negative charged sites) of all clay minerals (Alexiades and Jackson, 1965), it seems that such an interaction would occur between vermiculite and kaolinite.
The electrostatic interactions between kaolinite, oxides, and vermiculite seem to result in a soil stability not as dependent on SOM content as soils dominated by 2:1 clays. Due to the binding of particles by electrostatic interactions, SOM does not have to function as the critical binding agent. This is supported by the observation that C concentrations do not increase with increasing aggregate size in kaolinitic soils (Elliott et al., 1991; Feller et al., 1996; Six et al., 2000). Similar C concentrations across aggregate size classes is in contrast to soils dominated by 2:1 minerals where SOM forms bridges between negative charged clay minerals within aggregates and consequently leads to increased C concentrations with increasing aggregate size (Elliott, 1986; Six et al., 2000). The similar soil stability but different SOM levels among different management systems in the Lexington soil indicates a partly decoupling of aggregation from SOM. However, Six et al. (1999) presented data for the Lexington soil which suggests that increased aggregate turnover under CT increased SOM turnover due to the reduced protection of SOM by aggregates. Therefore, increased management intensity seems to increase aggregate turnover and decrease SOM levels without a concomitant decrease in soil stability in kaolinitic soils. In conclusion, soils dominated by 2:1 minerals show positive feedbacks between SOM and aggregation, but the feedback from SOM to aggregation is diminished by 1:1 minerals and oxides in kaolinitic soils.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
The higher soil stability at Lexington compared to the other sites and the smaller effect of management practices on soil stability at Lexington were a result of the presence of kaolinite and oxides. We conclude that the feedback of SOM on aggregation, as observed in soils dominated by 2:1 minerals is reduced in soils with oxides and 1:1 clay minerals due to the electrostatic interactions between these mineral components.SAS Institute 1990
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 5, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Paul, G. O. Martinson, E. Veldkamp, and H. Flessa Sample Pretreatment Affects the Distribution of Organic Carbon in Aggregates of Tropical Grassland Soils Soil Sci. Soc. Am. J., February 15, 2008; 72(2): 500 - 506. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rasmussen, M. S. Torn, and R. J. Southard Mineral Assemblage and Aggregates Control Carbon Dynamics in a California Conifer Forest Soil Sci. Soc. Am. J., September 29, 2005; 69(6): 1711 - 1721. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kubota, J. Bordon, K. Hoshiba, T. Horita, and K. Ogawa Change in Physical Properties of "Terra Rossa" Soils in Paraguay under No-tillage Soil Sci. Soc. Am. J., August 4, 2005; 69(5): 1448 - 1454. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Zotarelli, B. J. R. Alves, S. Urquiaga, E. Torres, H. P. dos Santos, K. Paustian, R. M. Boddey, and J. Six Impact of Tillage and Crop Rotation on Aggregate-Associated Carbon in Two Oxisols Soil Sci. Soc. Am. J., March 1, 2005; 69(2): 482 - 491. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Denef, J. Six, R. Merckx, and K. Paustian Carbon Sequestration in Microaggregates of No-Tillage Soils with Different Clay Mineralogy Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 1935 - 1944. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Eynard, T. E. Schumacher, M. J. Lindstrom, and D. D. Malo Aggregate Sizes and Stability in Cultivated South Dakota Prairie Ustolls and Usterts Soil Sci. Soc. Am. J., July 1, 2004; 68(4): 1360 - 1365. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. O. Marquez, V. J. Garcia, C. A. Cambardella, R. C. Schultz, and T. M. Isenhart Aggregate-Size Stability Distribution and Soil Stability Soil Sci. Soc. Am. J., May 1, 2004; 68(3): 725 - 735. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. I. C. Wakindiki and M. Ben-Hur Soil Mineralogy and Texture Effects on Crust Micromorphology, Infiltration, and Erosion Soil Sci. Soc. Am. J., May 1, 2002; 66(3): 897 - 905. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
