SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 27 October 2006
Published in Soil Sci Soc Am J 70:1983-1990 (2006)
DOI: 10.2136/sssaj2006.0054
© 2006 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 Vaught, R.
Right arrow Articles by Miller, D. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Vaught, R.
Right arrow Articles by Miller, D. M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Vaught, R.
Right arrow Articles by Miller, D. M.
Related Collections
Right arrow Structure and Properties
Right arrow Soil Analysis
Right arrow Soil Physics

Soil Physics

Relationships among Coefficient of Linear Extensibility and Clay Fractions in Expansive, Stoney Soils

R. Vaughta, Kristofor R. Bryeb,* and D. M. Millerb

a USDA-NRCS, 200 HWY 70 East Suite 2, Glenwood, AR 71943
b Dep. of Crop, Soil, and Environmental Sciences, Univ. of Arkansas, Fayetteville, AR 72701

* Corresponding author (kbrye{at}uark.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
An expansive soil is any soil that has a potential for shrinking and swelling under changing moisture conditions. Structural damage to homes (i.e., walls and foundations) due to expansive soils is costly to repair and may be somewhat avoidable if soil properties, such as clay content and the coefficient of linear extensibility (COLE), are investigated. The objectives of this study were to examine the relationships among COLE, clay fractions, and coarse-fragment content and determine the usefulness of COLE as an indicator of structural damage in expansive and stoney soils. Sixteen individual homeowner sites were selected for sampling in the Fayetteville area of Washington County, Arkansas based on a homeowner questionnaire and visual inspection. COLE was determined based on the rod method from the length change of a moist and oven-dry soil-paste rod (COLErod). Measured COLErod values were adjusted for coarse-fragment content (COLEadj) by multiplying COLErod by the volume fraction of fine-earth plus pores, yielding a 0.003 unit reduction (P < 0.001) in COLErod, but no change in hazard-class rating. COLErod was positively correlated (P < 0.001) with total (r = 0.88), coarse (r = 0.55), and medium plus fine (r = 0.79) clay fractions. However, COLErod adjustments for coarse-fragment content did not improve linear relationships between COLErod and clay fractions. The summation of soil profile COLErod values for a site was a reasonable predictor, while the average COLErod value for a site and COLEadj were poor predictors of home structural damage. Results demonstrate that COLErod is a reasonably reliable predictor of potential home structural damage.

Abbreviations: CEC, cation-exchange capacity • COLE, coefficient of linear extensibility • COLEadj, adjusted coefficient of linear extensibility • COLErod, COLE determined based on the rod method


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
EXPANSIVE SOILS are a global engineering problem (Al-Rawas et al., 2002) and cause more damage to buildings and roads in the USA each year than any natural disaster, including floods and earthquakes (Nelson and Miller, 1992). Structural damage to foundations, homes, and roads is costly to repair but may be avoided altogether provided sufficient knowledge is available about the soil on which structures are built.

An expansive soil is any soil that has a potential for shrinking and swelling under changing moisture conditions (Nelson and Miller, 1992). Expansive soils experience three-dimensional volume changes during wetting and drying cycles, increasing volume when wetting and decreasing volume when drying; hence expansive soils often have some shrink-swell potential as a result of wetting-drying cycles. Building on expansive soils becomes problematic when the soil's inherent tendency to shrink and swell results in structural damage (Azam et al., 2000; Erguler and Ulusay, 2003; Nusier and Alawneh, 2002; Al-Rawas et al., 2002). Buildings on expansive soils are most often damaged by the upward heaving (swelling) of the soil in a non-uniform manner, which can result in buckled floors, cracked walls, settling, and directional shifting. Expansive soils used as fill material can also be damaging to retaining walls due to corresponding cycles of increasing and decreasing earth pressure (Jennings, 1969) resulting in soil creep (Selby, 1982).

Soil shrink-swell behavior is primarily governed by the dominant clay mineralogy (Davidson and Page, 1956; Greene-Kelley, 1974; Nettleton and Brasher, 1983; Erguler and Ulusay, 2003; Kariuki and van der Meer, 2004) and arises from the movement of water into and out of interlayer spaces of the 2:1 phyllosilicate clay minerals (e.g., predominantly montmorillonite and vermiculite) that causes the mineral to expand and contract on a molecular level. Soil shrink-swell potential is also affected by numerous other factors and soil properties, such as soil particle- and pore-size distribution, texture, water content, the rate of moisture change due to natural and man-made drainage (Komornik, 1969), specific surface area, cation exchange capacity (CEC; Ross, 1978; Smith et al., 1985), organic matter, exchangeable cations, iron content (Davidson and Page, 1956; Azam et al., 2000), and parent material (Thomas et al., 2000). In general, soil shrink-swell potential generally increases as clay content increases regardless of mineralogy (Anderson et al., 1973; Schafer and Singer, 1976a; Ross, 1978; Smith et al., 1985). More specifically, Reid-Soukup and Ulery (2002) documented a positive correlation between soil creep, as a result of shrink-swell behavior, and the amount of expandable clay (i.e., smectite) present in soil. The extent to which these and other factors interact make it difficult to predict a soil's shrink-swell potential. As a result of the numerous properties that affect soil shrink-swell potential, various classification schemes have been developed to categorize soil shrink-swell potential based on both soil physical (Schafer and Singer, 1976b; Simon et al., 1987; USDA-NRCS, 2006) and chemical properties, namely CEC (Soil Survey Staff–Natural Recource Conservation Service–United States Department of Agriculture [SSS–NRCS–USDA], 1999).

The COLE, which is the one-dimensional length change of a natural soil ped (i.e., <2-mm fine-earth fraction) between two different moisture conditions, is commonly used to determine soil shrink-swell capacity (Grossman et al.,1968) and to categorize the potential level of associated risk (Schafer and Singer, 1976b). When it is not possible to extract intact soil samples, the COLE rod method (COLErod), which has been shown to be highly correlated with COLE (Schafer and Singer, 1976b; Simon et al., 1987), is often used. Recently, Kariuki and van der Meer (2004) established a unified swelling potential index across a wide range of properties for expansive soils by incorporating numerous commonly used swelling potential indices including COLE.

Aside from characteristics of the fine-earth fraction, the presence of coarse fragments is a confounding factor in predicting overall soil shrink-swell potential. For example, if a home is built on a soil that has an overall high capacity to shrink-swell, as may be stated in a county soil survey report, but the soil also contains a large percentage of coarse fragments, which do not possess shrink-swell capacity, one would expect for that soil to have a much lower potential for foundation damage than the same soil without coarse fragments. There has been little investigation into possible adjustments to COLE values to account for the reduced shrink-swell risk and potential for structural damage that would be expected in soils with a high percentage of coarse fragments. Holmgren (1968) presented a series of nomographs prepared from calculations of the fine-earth volume fraction to determine a coarse-fragment correction factor, but the applicability of this approach was not tested against actual field observations.

Expanding urbanization has forced development into areas previously considered poorly suited for urban use because of the shrink-swell potential of the soils; thus rendering the potential for structural damage high in new homes (Williams, 2003). Therefore, reliable prediction of the potential for home structural damage would be of great benefit in many rapidly expanding communities. Recent attempts have been made to characterize expansive soil behavior and predict damage to masonry structures using numeric simulation models (Masia et al., 2002; Masia et al., 2004), but with mixed results. However, based on our survey of the literature, there have been no evaluations of COLErod and clay fractions as potential predictors of home structural damage.

Washington and neighboring Benton County in northwest Arkansas are among the regions of highest population growth and urbanization in the USA in recent years. In addition, many of the expansive soils in northwest Arkansas contain a substantial amount of coarse fragments (>2.0 mm). Therefore, the objectives of this study were to examine the relationships among COLErod, clay fractions, and coarse-fragment content and determine the usefulness of COLErod as an indicator of home structural damage in expansive, stoney soils. It was hypothesized that COLErod would be more highly correlated with medium plus fine clay than coarse- or total clay because most expansive clay minerals are found in the medium plus fine clay size classes (i.e., <0.2 µm) and that COLErod values adjusted for the presence of coarse fragments will be a better indicator of home structural damage than unadjusted COLErod values because of accounting for the actual volume of soil with shrink-swell capability. It was also hypothesized that COLErod and/or medium plus fine clay fractions could be used to differentiate between little to no home structural damage and severe damage.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
Site Selection and Assessment of Home Structural Damage
Based on previous information regarding problems with expansive soils in several areas, 16 individual homeowners within the city limits of Fayetteville, Washington County, AR were asked and agreed to be included in this study. Homes younger than 4 yr of age were excluded from the study because they were less likely to have visible structural damage. Rental homes were also excluded from the study due to the likelihood that current renters would not sufficiently know the history of their home's foundation. This study was set up as a survey and 16 sites was determined to be a reasonable compromise between number of sites and the amount of post-sample processing that was going to take place. The 16 sample sites were comprised of three soil orders (13 Ultisols, 2 Alfisols, and 1 Inceptisol) and nine soil series (Table 1; Harper et al., 1969).


View this table:
[in this window]
[in a new window]
  		Table 1. Summary of soil series sampled, number of sites sampled within a soil series, and number of sites within a soil series sampled that had severe home structural damage.

 
Once sites were determined, a questionnaire was administered to homeowners to gather information regarding home age, foundation type, homeowner opinion as to cause of any structural damage, pest problems, and whether corrective measures were ever taken if structural damage had occurred. Homes included in this study varied in age from 4 to 45 yr old. Seven of the homes had crawl-space-type foundations, six had slab foundations, and three of the homes had a combination of slab and crawl-space foundations. Based on the questionnaire information and visual inspection of all sites before any soil examination, sites were subjectively categorized as either homes with severely damaged foundations (i.e., many and/or large cracks to foundations and/or walls) or those with little or no damage (i.e., few and/or small cracks to foundations and/or walls). Nine of the sixteen sites were categorized as having little or no damage and seven sites were categorized as having severe damage.

Sample Collection
Samples from the soil profile were collected from two spots at 13 sites and from one spot at the other three sites that were too rocky to provide acceptable soil samples from two sample spots. To avoid buried objects near the foundation, actual sample sites were located within a relatively undisturbed area in the vicinity of the house that was judged to have the same soil properties as the soil under the house. Soil was collected with a 6.6-cm diameter bucket auger in approximately 17.5-cm increments to a depth of 123 cm or until collection was inhibited by coarse fragments. Sampling depths ranged from 35 to 123 cm. A total of 127 individual soil samples were collected from the 16 sites.

Sample Processing
Field-moist samples were air dried for 72 h and large pieces of organic material (i.e., roots) were removed by hand. Soil samples without coarse fragments were mechanically ground and sieved to pass through a 2-mm mesh screen. For stony samples, coarse fragments (i.e., >2-mm diameter) were removed by wet sieving. Once > 2-mm diameter coarse fragments were completely separated from the <2-mm fine-earth, coarse fragments were air-dried and weighed. Free water in the fine-earth slurry was allowed to evaporate for 6 d at 25 to 28°C. Once air dry, the soil sample was mechanically ground to pass a 2-mm mesh screen and weighed.

Sample Volume Determination
Bucket-auger volumes were determined after all soil samples were collected based on the method described by Brye et al. (2004) for determining bulk density of stony alluvium. Ten soil cores were extracted from the 0- to 17.5-cm depth of a Captina silt loam using the same bucket auger that was used to extract all soil samples to mimic the actual sampling interval. The resulting auger holes were filled with Great Stuff expanding foam sealant (Dow Polyurethane Systems, Marietta, GA). After curing for 24 h, the foam cores were marked at the soil surface and extracted. Extracted foam cores were allowed to dry for an additional 24 h, after which soil and organic matter were removed from the outside of the cores with a soft brush. A band saw was used to remove the tops of the cores, so that only the filled void from the auger hole remained. The cores were then cut in half to fit into a 2-L graduated cylinder. The cut faces of the foam cylinders were coated with urethane so that they would not imbibe water. Foam cores were submerged in a 2-L graduated cylinder and the volume determined based on the amount of water displacement. The mean of the 10 foam-core volumes (i.e., 726 [standard deviation = 10] cm3) was used as the estimated soil-core volume extracted with the bucket auger to express coarse fragment content on a volume basis.

COLE Measurement
COLErod was determined in quadruplicate for each sample according to the method of Schafer and Singer (1976b). Previously air-dried, ground, and sieved (<2 mm) soil samples were used for COLErod measurement. For each sample, approximately 100 g of soil were added to small plastic cups, and enough deionized water was added to make a saturated paste. The soil paste was covered and left to equilibrate for 24 h. Twenty-five cubic centimeter disposable plastic syringes were modified using a drill press so their orifices were smooth and a uniform 1.0-cm diameter. Four 60- to 100-mm long soil rods were slowly extruded onto a Teflon-coated baking sheet. Samples with higher clay contents were extruded to a length of 60 to 70 mm to decrease the likelihood of cracking on drying. A spatula was used to trim the ends of the rods for measurement with a digital caliper. The rods were dried in an oven for 2.5 h at 105°C. COLErod was determined by the equation

Formula 1[1]
where lm is the moist-rod length and ld is the dry-rod length (Schafer and Singer, 1976b); thus COLE values are reported unitless. The mean COLErod value from the four replicate measurements of each sample was used for subsequent data analysis.

COLE Adjustment for Coarse Fragments
Since coarse fragments do not possess shrink-swell capability, COLErod values were adjusted to reflect only the soil volume that had the capability to shrink and swell (i.e., the volume fraction of fine earth and pores). A procedure similar to that described by Holmgren (1968) was used. The volume fraction of the fine earth and pores was estimated using the following formula:

Formula 2[2]
where R was the measured coarse-fragment mass (g), 2.65 g cm–3 was the assumed particle density, B was the estimated bucket auger volume of each collected sample (726 cm3) based on the previously described method of Brye et al. (2004), and Vfp was the resulting estimate of the volume fraction of fine-earth plus pores. The adjusted COLErod value (COLEadj) was then calculated as

Formula 3[3]
where COLErod was the average of the four COLErod measurements for each sample.

Clay Fraction Determination
Total Clay
Total clay content (<2 µm) was determined in triplicate using the micropipette method of Miller and Miller (1987). Four grams of dried, ground, and sieved soil were mixed with 40 mL of dispersing solution, created from 10 mL each of 5% (w/v) sodium hexametaphosphate and 1.0 M sodium hydroxide diluted to 1 L, and shaken on an end-over-end shaker for 8 h at approximately 60 rev min–1 to disperse the soil. Based on Stoke's law, a settling time of 1 h and 50 min and a sampling depth of 2.5 cm were used for particles <2.0 µm (Miller and Miller, 1987). A 5-mL adjustable pipettor was used to slowly withdraw a 2.5-mL sample of the soil suspension. Samples were dispensed into preweighed aluminum dishes, dried at 105°C for 2 h, and weighed. Final clay percentages were corrected for the solids content of the dispersing solution. The mean value of three total clay measurements per sample was used for subsequent data analysis.

Medium Plus Fine Clay
The medium plus fine clay fractions (<0.2 µm) were isolated together. To shorten the extensive period of time required to isolate these soil fractions, a centrifuge was used to increase the force of gravity. Stoke's law was integrated by Jackson (1969) to account for centrifugal force as:

Formula 4[4]
where tmin is the time for sedimentation in minutes, {eta} is the solution viscosity in poises (i.e., a dyne s cm–2 or 10–5 N s cm–2), R is the radius (cm) of rotation of the top of the sediment in the tube, S is the radius (cm) of rotation of the surface of the suspension in the tube, Nm is the rotations per minute, Du is the particle diameter in microns (µm), and SGDs is the difference in specific gravity between the solvated particle and the suspension liquid.

Four milligrams of ground, sieved soil were transferred in triplicate to round-bottom centrifuge tubes. To ensure a uniform distance from the center of rotation to the surface of the suspension, tubes were brought to the 10-cm depth mark with an aqueous solution prepared by diluting 10 mL each of 5% (w/v) sodium hexametaphosphate and 1.0 M sodium hydroxide to 1 L. Capped tubes were shaken vigorously by hand for 1 min and placed on an end-over-end shaker for 8 h at approximately 60 rev min–1. After shaking, tubes were again shaken vigorously by hand for 60 s to make certain that no particles had settled to the bottom and then centrifuged for 56 min at 1700 rpm.

After centrifugation, the supernatant suspension was decanted into a 500-mL volumetric flask until clear. Each flask was then brought to volume with deionized water and was mixed vigorously by hand for 50 s. Following mixing, approximately 40 mL of the suspension was poured into a 50-mL centrifuge tube, and 5.0 mL of suspension was slowly withdrawn. Samples were dispensed into pre-weighed aluminum dishes, dried for 2 h at 105°C, and weighed. Final medium plus fine clay percentages were corrected for the salt content of the dispersing solution. The mean value of three medium plus fine clay measurements per sample was used for subsequent data analysis.

Statistical Analyses
Though prediction was not the goal, regression analyses were performed to describe soil property trends with depth using Minitab version 13.31 (Minitab, Inc., State College, PA). Paired-sample (i.e., corrected and uncorrected) t tests were performed to evaluate the overall effect of the COLErod adjustment for coarse fragments. Linear correlations were performed to evaluate the relationships among measured and adjusted COLErod values and clay fractions, and analysis of covariance techniques were used to ascertain any differences in linear relationships between COLErod and COLEadj and clay fractions (SAS version 9.1, SAS Institute, Inc., Cary, NC).

To account for potential risk differences due to depth to bedrock, single site COLErod values were determined by two methods, averaging and summing across all replicates and depths at a site. Only mean site values were calculated for clay fractions. Two-sample t tests with equal variances were performed to ascertain whether COLErod or clay fractions could significantly differentiate between little to no structural damage and severe structural damage. Since there were a small number of degrees of freedom that resulted after the data set was reduced from n = 127 (i.e., the total number of individual measurement values) to n = 16 (one value per site), P < 0.1 was used to judge statistical significance for this test.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
SUMMARY AND CONCLUSIONS
 REFERENCES
 
The Alfisols and Ultisols of the Ozark Highlands that encompass the Fayetteville area generally have a series of argillic (Bt) horizons in the subsoil (Harper et al., 1969); thus it would be expected that clay contents increase with depth in the soil profile. Since it has been documented that COLE tends to increase as clay content increases (Anderson et al., 1973; Schafer and Singer, 1976a; Ross, 1978; Smith et al., 1985; Reid-Soukup and Ulery, 2002), it was also expected that COLErod values would increase with depth. With the exception of coarse clay, medium plus fine clay (P < 0.001), total clay (P < 0.001), and COLErod (P < 0.001) increased with depth across all sample sites (Table 2). Though not significant, coarse fragment content tended to increase to about 50 cm then decrease to 105 cm (Table 2). COLErod ranged from a minimum of 0 in several samples in the 35- to 70-cm depth interval to a maximum of 0.145 in one sample from the 70- to 105-cm depth interval.


View this table:
[in this window]
[in a new window]
  		Table 2. Summary of mean (± standard error) soil properties measured by depth.

 
Hazard Classes
To qualitatively evaluate the risk of or potential for structural damage, several sets of shrink-swell hazard classes have been proposed based on COLE values. Although the COLE and COLErod methods for determining the shrink-swell capacity of a soil are highly correlated (Schafer and Singer, 1976b; Simon et al., 1987), the specific shrink-swell hazard classes for the two methods often do not possess the same numeric ranges (Table 3; Schafer and Singer, 1976b). Since the standard COLE method was not used, shrink-swell hazard classes based on COLErod measurements proposed by Simon et al. (1987) were used to define hazard classes for the samples in this study because the soils analyzed here were most comparable with the Alfisols and Ultisols analyzed by Simon et al. (1987).


View this table:
[in this window]
[in a new window]
  		Table 3. Hazard classes for COLE (USDA-NRCS, 2006) and COLErod. COLErod classes are taken from those previously reported by Simon et al. (1987) and Schafer and Singer (1976a). The Schafer and Singer (1976a) classification is included for comparison purposes.

 
Based on Simon et al. (1987), all samples analyzed in this study fell into either the low or moderate hazard classes. Though some sites were sampled where foundation damage was subjectively categorized as severe, only one of 127 soil samples fell into the high or very high hazard classes (i.e., COLErod > 0.14).

The method for classifying COLErod values for shrink-swell hazard as proposed by Simon et al. (1987) and the USDA classification (USDA-NRCS, 2006), included for comparison purposes, as listed in the county soil survey report (Harper et al., 1969) resulted in only 55% agreement. Thirteen percent of the time the COLErod system rated the soil as having a higher shrink-swell hazard than the USDA system. Thirty-two percent of the time COLErod values resided in a lower shrink-swell risk hazard than in the USDA system. These data suggest that potentially major discrepancies exist between USDA-estimated shrink-swell potential as listed in county soil survey reports and site-specific hazard-class ratings based on direct measurement. Though often used to make site-specific decisions, county soil survey reports are based on few actual observations that are generalized to much larger areas that may or may not well-represent the original site of observation. To make the most well-informed decisions possible, home builders and home owners should utilize direct measurements from on-site evaluations of soil profile properties and characteristics to assess structural damage risk in areas with expansive soils.

The clay mineral activity class, as part of the family-level taxonomic soil description, can give an indication as to the shrink-swell potential of a given soil (SSS–NRCS–USDA, 1999). Clay mineral activity classes are based on soil properties in the activity class control section of the soil profile, which is similar to the control sections for the particle-size and mineralogy classes, generally the upper part of the argillic (Bt) horizon. Soils in the superactive or active activity classes have a CEC to clay percent by weight ratio of >0.4 and will tend to exhibit greater shrink-swell potential than those of the semiactive or subactive activity classes, which have a CEC to clay percentage by weight ratio of <0.4 (SSS–NRCS–USDA, 1999). Of the nine soil series sampled in this study, five had active and four had semiactive or subactive activity classes (Table 1). None of the sites sampled with soils having semiactive or subactive activity classes (five sites) were subjectively categorized as having severe structural damage, while all seven of the sites sampled with severe damage had soils with an active activity class.

Despite only one of 127 soil samples in the high or very high hazard class for potential structural damage, increasing shrink-swell potential with soil depth, as demonstrated by increasing COLErod values with soil depth (Table 2), has ramifications for the construction industry. Topsoil is often stripped away and foundations and stem walls are laid up to 1 m deeper than the original soil surface in the region when possible. It is commonplace to haul soil into a construction site to place below a slab foundation. With these practices in use, the knowledge that soils from greater depths might have a greater tendency to shrink-swell are critical in constructing a sound structure and saving future owners or investors money.

Adjusted COLErod
Based on a simple mathematical adjustment (Eq. [2] and [3]) that corrected COLErod values to reflect only the volume fraction of fine earth and pores capable of responding to wet-dry cycles by shrinking and swelling, the average COLErod value decreased significantly (P < 0.001) by 0.003 units, ranging from a minimum of 0 units for samples with no coarse fragments to a maximum of –0.038 units for a sample with 75.9% coarse fragments by weight. Though the COLErod adjustment for coarse fragments was significant, the –0.003-unit difference was too small to have altered the hazard class; thus all COLEadj values had the same hazard class as the non-adjusted COLErod values. However, the hand-auger sampling method used in this study may have limited the coarse fragment sample size by excluding coarse fragment that physically did not fit through the auger (ASTM, 2004). Therefore, the coarse fragment measurements and COLErod adjustments in this study represent minimal values. If coarse fragments larger than what could have physically fit thought the auger occurred in the soils sampled, the total coarse fragment content was likely greater than that measured and the COLErod adjustment would also have been larger.

Correlations among Clay Fractions
Similar to previous reports (Schafer and Singer, 1976a; McCormack and Wilding, 1975), COLErod was positively correlated (P < 0.001) with total, coarse, and medium plus fine clay fractions. Awareness of the significant correlation between COLE and clay percentage regardless of clay mineralogy has practical field importance because, as one might expect, simply estimating the clay content by hand texturing could give much insight into potential problems that may arise in expansive soils. As hypothesized, COLErod was more highly correlated with medium plus fine clay (r = 0.79) than coarse clay (r = 0.55). However, contrary to that hypothesized, COLErod was more highly correlated with total clay (r = 0.88) than medium plus fine clay. This result is somewhat counterintuitive, since the majority of expansive clay minerals are found in the fine-clay fraction. However, Schafer and Singer (1976a) reported that total clay was a better predictor of COLErod than medium plus fine clay because the ratio of expandable- to total clay varied little. Since COLEadj was derived from COLErod, it is not surprising that COLEadj was similarly positively correlated (P < 0.001) with total- (r = 0.86), coarse- (r = 0.51), and medium plus fine clay fractions (r = 0.80).

Though both COLE rod and COLEadj were highly correlated with several clay fractions, their linear relationships with clay fractions differed somewhat (Fig. 1 , Table 4). The slopes of the linear relationships between COLE rod, COLEadj, and clay fractions were all significant (P < 0.001). The intercepts differed slightly, though significantly, from 0 (P < 0.001) for the relationships between both sets of COLE values and medium plus fine clay and coarse clay. The intercept for the relationship between both sets of COLE values and total clay did not differ from 0. For the relationships between both sets of COLE values and medium plus fine clay, coarse clay, and total clay, intercepts differed significantly (P < 0.001), while all slopes were similar (Fig. 1, Table 4). These results indicate that adjustment of COLErod values for coarse-fragment content did not improve the correlations with clay fractions and that there were no practical differences between COLErod or COLEadj and any clay fraction.


Figure 1
View larger version (28K):
[in this window]
[in a new window]
  		Fig. 1. Linear regression relationships among coefficient of linear extensibility using the rod method (COLErod), the adjusted COLE (COLEadj), and clay fractions. Regression equations and associated statistical analyses are summarized in Table 4.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Summary of linear regression relationships among COLErod, COLEadj, and clay fractions. Formal comparison among slopes and intercepts of COLErod and COLEadj equations are also included.

 
COLErod, Clay Fractions, and Home Structural Damage
To test the hypothesis that COLErod and/or medium plus fine clay fractions could be used to differentiate between little to no structural damage and severe damage, it was necessary to calculate single overall values for each site. It is reasonable to assume that a site that possesses COLE values that fall into a high or very high shrink-swell category, but are shallow to bedrock, pose a lesser threat for structural damage than does a site with COLE values that fall into a moderate shrink-swell category that are much deeper to bedrock; there is simply more expansive soil in the later to potentially damage a foundation.

Neither total-, medium plus fine, nor coarse-clay fractions differed among structural damage categories. As is evident in Fig. 1, there is a relatively large range of clay values regardless of clay fraction for any give COLE value. Therefore, simply knowing a particular clay fraction for a given soil or site does not appear to be a good predictor of potential home structural damage.

The two methods used to obtain a single site COLErod value differed greatly in their ability to differentiate COLErod values for the two structural damage categories (Table 5). Neither mean site COLErod nor COLEadj differed among structural damage categories (P > 0.1) indicating that averaging to obtain a single site COLErod value unified the COLErod data too much (i.e., reduced variability). However, both summed site COLErod (P = 0.058) and COLEadj (P = 0.074) significantly differentiated the two structural damage categories indicating that summing COLErod values across depths, rather than averaging them, was able to account for sites with varying soil thicknesses or depths to bedrock. It also appears that adjusting COLErod values to account for coarse fragments yielded no improvement in the ability to differentiate among structural damage categories. Based on these data, the threshold summed site COLErod value that became a critical indicator of home structural damage was 0.373 (Table 5).


View this table:
[in this window]
[in a new window]
  		Table 5. Summary of mean and site summed COLE values by sample site and their associated home structural damage assessment.

 
Though the data set was somewhat small, these results also suggested that, since it is common practice for builders to set home foundations from 0.6 to 1.0 m deep in the Fayetteville area where depth to bedrock is a limiting factor, the overall sum of the COLErod values found within a soil profile can be used to predict the potential home foundation damage in expansive soils.


   SUMMARY AND CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
REFERENCES
 
COLErod adjustments for coarse-fragment content were small, but significant. However, no differences in hazard class rating resulted because of the small adjustment that was made to COLErod values. COLErod was positively correlated (P < 0.001) with total, coarse, and medium plus fine clay fractions. As hypothesized, COLErod was more highly correlated with medium plus fine clay (r = 0.79) than coarse clay (r = 0.55). However, contrary to that hypothesized, COLErod was more highly correlated with total clay (r = 0.88) than medium plus fine clay. COLErod adjustments for coarse-fragment content did not improve the linear relationships among COLErod and clay fractions. Summed site COLErod was a significant predictor of structural damage, whereas mean site COLErod was not. Results of this study demonstrated that it is likely not essential to account for coarse fragment contents less than 15% when using COLE to make interpretations regarding land-use risks associated with expansive soils and that it may be possible to predict potential home structural damage in expansive soils using knowledge of COLErod values, but likely not clay fractions alone.

Received for publication February 8, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND 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 Vaught, R.
Right arrow Articles by Miller, D. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Vaught, R.
Right arrow Articles by Miller, D. M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Vaught, R.
Right arrow Articles by Miller, D. M.
Related Collections
Right arrow Structure and Properties
Right arrow Soil Analysis
Right arrow Soil Physics

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