freezing effects on aggregate stability of soils amended with lime

CATENA SUPPLEMENT 24 p. 115-127 Cremlingen 1993

Freezing Effects on Aggregate Stabilityof Soils Amended with Lime and Gypsum

G.A. Lehrsch, R.E. Sojka & P.M. Jolley

Summary

Aggregate stability, which influences soilresponse to raindrop impact and othererosive forces, is affected by freezing. Wehypothesized that lime or gypsum addedat agriculturally feasible rates may actas bonding agents to mitigate the ef-fects of freezing on the stability of aggre-gates from different soils. Thus, the ob-jectives of the laboratory study were todetermine the effects of freeze-thaw cy-cles, water content at freezing, and limeand gypsum additions on the aggregatestability of six soils differing in texture,mineralogy, and organic matter. To allsamples but the control, either CaCO 3or CaSO4 • 2H2 0 was added at a rateof either 0.2 or 1.0% by weight of oven-dry soil. Field-moist aggregates werevapor-wetted to either a low water con-tent (5-10% w/w) or a high water con-tent (25-30%), then frozen and thawedeither 0, 1, or 3 times in brass cylin-ders. After thawing, they were vapor-wetted to 0.30 g g' and their stabilitywas characterized by wet sieving. For allsoils, aggregate stability decreased sig-nificantly with increasing water contentat freezing. When averaged over bond-ing agents and addition rates, soils corn-

ISSN 0722-0723ISBN 3-923381-33-6©1993 by CATENA VERLAG,38162 Cremlingen-Destedt, Germany3-923381-33-6/93/5011851/US$ 2.00 + 0.25

monly increased in stability, at timessignificantly, after one freeze-thaw cy-cle. Soils with clay contents of 20% ormore and organic matter contents over3% were the most stable after freez-ing. When averaged over soils and freezethaw cycles, aggregate stability was notaffected by bonding agents at low wa-ter contents. At high water contents,however, stability increased significantly(P<0.05) when CaSO 4 at a rate of 0.2%was added. Though the differences fromagent to agent at high water contentswere not significant at the 0.05 level,CaSO4 appeared to increase aggregatestability more than CaCO 3 . Surpris-ingly, stability tended to be greater atthe lower rather than higher applicationrates. In conclusion, aggregate stability

1. decreased with increasing watercontent at freezing, and

2. increased after CaSO 4-amended soilsamples at high water contents werefrozen.

1 Introduction

Aggregate stability, a measure of a soilaggregate's resistance to breakdown, in-fluences many soil physical and hy-draulic characteristics, such as surfacesealing, infiltration, and hydraulic con-ductivity. Processes that affect aggre-gate stability are not well understood

J.W.A. Poesen & M.A. Nearing (Eds.): Soil Sealing and Crusting

116 Lehrsch, Sojka & Jolley

and should be studied further.Freezing is one process that has been

shown to affect aggregate stability. Sev-eral studies have indicated that aggre-gate stability decreases with increasingsoil water content at the time of freez-ing (Bullock et al. 1988, Bryan 1971,Lehrsch et al. 1991, Logsdail & Web-ber 1959). Aggregates from poorly ag-gregated soils, however, are more sta-ble when frozen at intermediate watercontents (Mostaghimi et al. 1988, Per-fect et al. 1990b, Sillanpaa & Webber1961). The number of freeze-thaw cyclesto which aggregates have been subjectedis important. The stability of air-driedaggregates generally decreases with in-creasing number of freeze-thaw cycles(Logsdail & Webber 1959, Mostaghimiet al. 1988) but may increase for somesoils (Mostaghimi et al. 1988, Richard-son 1976). For aggregates maintainedat field moisture levels until analyzed,Lehrsch et al. (1991) found stability toincrease with the first few freeze-thawcycles.

No consensus exists in the literatureregarding the effects of lime and/or gyp-sum amendments on soil aggregationand aggregate stabilization in non-sodicsoils. For sodic soils, however, gypsumis commonly used to improve drainage.Gypsum amendments enhance the floc-culation of clays and thus facilitate boththe formation and stabilization of soilaggregates. The effects of gypsum ad-ditions to non-sodic soils, however, areless clear. Rimmer & Greenland (1976)noted that CaCO 3 can act as a bondingor cementing agent between soil parti-cles and, since it is slightly soluble, asa source of Ca ions which reduce dif-fuse double layer thicknesses of clay do-mains. Kemper et al. (1974) suspectedboth CaCO 3 and CaSO 4 to be bonding

agents, whereas Uehara & Jones (1974)discounted CaCO3 and CaSO4 as ce-menting agents. Uehara & Jones (1974)stressed that silica was most importantin arid and semiarid regions. They alsonoted the importance of the bondingagent's solubility in water (affecting dif-fusion to contact points) as well as theequilibrium attained (if any) betweenthe agent's solid phase and dissolvedphase.

The dissolution of lime or gypsumdoes affect soil structure. Exchangeablepolyvalent cations (e.g., Ca) near clayparticle surfaces reduce the thickness ofthe diffuse double layer and thus the re-pulsive forces acting between clay par-ticles (Emerson 1983) and cause strongflocculation of clays and increased resis-tance to dispersion (the Schultze-Hardyrule). Calcium rather than Mg or Kon the exchange complex was associatedwith stable aggregates in Australiansubsoils (Emerson & Bakker 1973). Po-jasok & Kay (1990) found aggregate sta-bility to increase with Ca concentrationsin soil solution.

Because of the lack of consensus inthe literature, the influence of inor-ganic bonding agents on the aggregatestability of non-sodic soils has beenwidely studied. For example, Rimmer& Greenland (1976) presented evidencethat CaCO 3 cemented clay tactoids 0.1to 0.3 microns apart. Al-Ani & Du-das (1988) found that aggregate meanweight diameter increased with addi-tions of calcium carbonate from 0 to 4%by weight but thereafter decreased withCaCO3 additions from 4 to 32%. In con-trast, Chepil (1954) reported that theaddition of 3% calcium carbonate to asoil had no effect on aggregate stabilitybut that 10% increased it. Others (Kern-per Si Koch 1966, Shanmuganathan

CATENA SUPPLEMENT 24

Freezing effects on aggregate stability 117

Oades 1983, Toogood 1978) found nosignificant effect due to lime content.

Some of the above inconsistencieswere related to the nature and prop-erties of the soils studied. This studywas conducted under a uniform set ofconditions with six different soils hav-ing paired soil properties. Since soils re-spond differently to the freezing processand to amendments, it was hypothesizedthat aggregate stability, after freezing,of soils differing in texture, mineralogy,and organic matter would be affectedby lime and gypsum additions. A lab-oratory experiment was designed to testthis hypothesis. The objective was to de-termine the effects of number of freeze-thaw cycles, water content at freezingand the addition of two common inor-ganic amendments at two agriculturallyfeasible rates on the aggregate stabilityof six continental U.S. soils differing intexture, mineralogy, and organic matter.

2 Materials and methods

Samples of six agriculturally impor-tant soils (AP horizons only) were ob-tained from across the United States.The soils differing primarily in tex-ture were a Barnes loam (fine-loamy,mixed Udic Haploboroll) from Morris,MN, and a Sharpsburg silty clay (fine,montmorillonitic, mesic Typic Argiu-doll) from Lincoln,- NE. The soils dif-fering in mineralogy were a Cecil sandyloam (clayey, kaolinitic, thermic TypicKanhapludult) from Watkinsville, GA,and a Sverdrup sandy loam (sandy,mixed Udic Haploboroll) from ElbowLake, MN. The soils differing primarilyin organic matter content were a Palousesilt loam (fine-silty, mixed, mesic PachicUltic Haploxeroll) from Pullman, WA,and a Portneuf silt loam (coarse-silty,

mixed, mesic Durixerollic Calciorthid)from Kimberly, ID. Properties of thesix soils (tab. 1) were determined (SoilConservation Service Staff 1984) by thepersonnel of the National Soil Surveyand Soil Mechanics Laboratories, Lin-coln, NE.

These six soils also differed in proper-ties other than the contrasted primaryproperty (clay content, mineralogy, ororganic matter). For any given soilpair, however, the difference betweenthe soils in their primary property wasmuch greater than any difference in asecondary property. For example, theorganic matter of the Palouse is 144%greater than that of the Portneuf whilethe clay content differs by less than halfthat amount, the clay in the Palouse be-ing only 67% greater than that in thePortneuf. Even so, lesser effects on ag-gregate stability may be related to dif-ferences in other properties among thesoil pair.

All soils had a similar cropping andtillage history. In the growing seasonprior to sample collection, a monocotcrop species, either corn, sorghum, or asmall grain, was grown. After harvest,crop residues were removed from thesoil surface and the sampling sites weremoldboard plowed, then lightly disked.Thereafter, the sites were fallowed forthree to twelve months to allow rootsand other organic residue to decomposebefore sampling.

This study which focused on the ef-fects of lime and gypsum amendmentson aggregate stability was similar inexperimental approach to but separatefrom an earlier study (Lehrsch et al.1991). For this study, a randomizedcomplete block design with three repli-cations and a factorial arrangement oftreatments was used. Sources of vari-



Particle size distribution Pre-Soil type Very dominant Organic

fine Total Base Exch. mineral pH mattersand sand Silt Clay sat. Ca CEC t type (in content s

% % % cmolc kg-1 CaC1 2 )

Barnes loam 11 49 34 17 100t -t 19.5 2:1 7.1 3.41

Sharpsburg silty clay 3 3 56 41 94 19.4 29.4 2:1 5.4 3.19

Cecil sandy loam 6 67 16 17 87 2.2 3.5 1:1 4.6 1.24

Sverdrup sandy loam 4 76 15 9 100 9.5 11.8 2:1 6.0 2.21

Palouse silt loam 9 10 70 20 82 12.7 19.6 2:1 4.5 3.03

Portneuf silt loam 20 22 66 12 100t -t 12.6 2:1 7.8 1.24

t Acid-equivalent lime was present.t Determined using NH 4 OAc at pH 7.§ As estimated from the organic C content using the Van Bemmelen 1.724 factor.

Tab. 1: Soil properties.

ation were soils, number of freeze-thawcycles, water content, and potential in-organic bonding agents added at tworates. Freeze-thaw cycles were either 0,1, or 3 with the 0 level signifying nofreezing. All samples were constrainedin brass cylinders 5 cm high with an in-side diameter of 2.75 cm. Before freez-ing, the water content (qualitatively ei-ther low or high) was adjusted to ei-ther 0.05 or 0.25 g/g for the coarse-textured Cecil and Sverdrup soil or 0.10or 0.30 g/g for the remaining soils. Foreach soil, water contents at matric po-tentials of -1500 and -33 kPa are givenin Elliot et al. (1989). Statistical anal-yses were performed on arcsin (x)°•5-transformed data (to stabilize the vari-ance) using an analysis of variance (SASInstitute, Inc., 1985) 1 . For presentation,the treatment means have been back-transformed to the original scale of mea-surement. The subsequent results anddiscussion have focused mainly on three

'Trade names are included for the benefit ofthe reader and do not imply endorsement of orpreference for the product by the USDA.

statistically significant interactions thatdescribe the variation in aggregate sta-bility caused by all four factors. Meanswere separated using confidence inter-vals constructed to be equivalent to testsof significance at the 5% level. Ad-ditional pre-planned, single degree-of-freedom comparisons of selected treat-ments were also made.

Samples were prepared by sievingfield-moist soil (gravimetric water con-tents at sieving ranged from 7 to 22%and averaged 13%) through a 4-mmsieve. After the soil was sampled inthe field, it was not dried but ratherstored at field water content in air-tight containers at +6°C until ana-lyzed in the laboratory. The moist,freshly sieved soil was coated with pow-dered reagent-grade CaCO 3 or CaSO4 .2H2 0 (as a potential inorganic bondingagent) at a rate of either 0.2 or 1.0%by weight of oven-dry soil. To coatthe soil, the desired mass of CaCO 3 orCaSO 4 was distributed over all surfacesof a known mass of sieved soil on plas-tic sheeting. The soil was then thor-



oughly mixed. The water content of theamended soil was then adjusted to ei-ther the low or high level. The soils'water content was lowered, if necessary,by evaporating slowly in air or raisedusing a non-heating vaporizer (Humid-ifier Model No. 240, Hankscraft, Reeds-burg, WI) 1 . Each soil sample wassubsequently packed by tapping into abrass cylinder to a dry bulk density of1.15 1g/cm3 . Each packed cylinder wassealed in a polyethylene bag to both in-hibit water loss and prevent water up-take, inserted into a styrofoam tray,and stored at +6°C until the remainingcylinders were packed. The styrofoam, aminimum of 7 cm underneath and 2 cmaround each cylinder, served as insula-tion so that freezing occurred primarilydownward from the surface (confirmedby subsequent thermocouple measure-ments).

All prepared samples were subjectedto either 0, 1, or 3 freeze-thaw cy-cles. One cycle was completed when pre-pared soil samples were frozen convec-tively (without access to water) at -14°Cfor 24 hours, then thawed at +6°C for48 hours. Little, if any, freezing-inducedvertical expansion of the soil in the cylin-ders occurred. For all samples, a datalogger within each enclosure recordedambient air temperatures. The 0 cyclesamples were not frozen but were storedat +6°C for a minimum of 48 hours. Be-fore the aggregate stability analysis, allsamples were brought to room tempera-ture on a lab bench for 2 hours. Aggre-gate stability was determined using theprocedure of Kemper Si Rosenau (1986)modified by Lehrsch & Jolley (1989) sothat field-moist 1- to 4-mm aggregateswere vapor-wetted to 0.30 g/g prior towet sieving.

3 Results and discussion

3.1 Interaction between soils,freeze-thaw cycles, and watercontent

In the statistical analysis, freeze-thawcycle was considered a discrete ratherthan continuous variable. When freeze-thaw cycle was initially modeled as con-tinuous, a trend whenever statisticallyidentified, was seldom a type that aidedinterpretation or summarization.

Freeze-thaw cycle effects on aggre-gate stability (averaged over amend-ments and addition rates) are illustratedin fig. 1. Large differences in aggregatestability existed from soil to soil evenat the same water content. For exam-ple, Palouse and Portneuf, two silt loamsfrom the Pacific Northwest, differed instability by over 30 percentage points af-ter one freeze-thaw cycle at a 10% wa-ter content. When frozen at least once,the most stable soils were the Sharps-burg and, to a lesser degree, the Palouse,each having 20% or more clay and over3% organic matter.

From 0 to 1 freeze-thaw cycle (one oc-currence of freezing), aggregate stabil-ity generally increased (fig. 1). Others(Mostaghimi et al., 1988, Perfect et al.1990a and 1990b) have also found ag-gregate stability to increase with freeze-thaw cycles. This increase in stabilitywas statistically significant (P<0.05) atwater contents of 0.10 and 0.30 g/g forthe Palouse silt loam and at 0.10 g/gfor the Barnes loam, two of the threesoils highest in organic matter (tab. 1).In contrast, stability decreased signif-icantly only for the low organic mat-ter Portneuf silt loam at 10% with onefreeze-thaw cycle.

Several factors may have caused some


100

80

60

40

20

WATER CONTENT

• HIGH 3 LOW


0 1 3 0 1 3 0 1 3 0 1 3 0 1 3 0 1 3

FREEZE-THAW CYCLE

BARNES SHARPSBURG CECIL SVERDRUP PALOUSE PORTNEUF

I- TEXTURE -1 I-.- MINERALOGY --I 1- ORGANIC MAT. -I

Fig. 1: Aggregate stability as a function of freeze-thaw cycle for each soil at eachwater content. The water contents were 5 and 25% (w/w) for the Cecil and Sverdrupsoils and 10 and 30% (w/w) for the remaining soils. The first pair of soils differsprimarily in texture, the second in mineralogy, and the third in organic matter.

or all of this increase in stability withfreeze-thaw cycle(s). First, particlescould have been brought into contactwith one another due to ice lens growth.Second, water moving toward a devel-oping ice lens from the soil matrix sur-rounding and below the lens could haveencouraged soil drying (Hoekstra 1966,Miller 1980). This drying could have po-sitioned polysaccharides on soil particlesurfaces (Myers 1937; Reid & Goss 1982)and/or precipitated bonding agents atpoints of contact between soil particles(Kemper et al. 1987; Lehrsch et al.1991). Third, water flowing to an icelens would have increased the concentra-tion of ions (notably Ca) in the soil so-

lution near the lens. Once there, the Caas a counter ion would diffuse into thedouble layer between clay platelets andhelp to bond the platelets to each otherdue to electrostatic attraction (Kemperet al., 1987). Thus, upon thawing, ag-gregates could have re-formed as well asincreased in strength.

As portions of each soil sample driedduring the first few freezes, bondingagents from the soil solution likely pre-cipitated at particle-to-particle contactpoints. This precipitation would havetaken place because, as water was re-moved from the soil below an ice lens,the meniscus at the air-water interfaceapproached a contact point between soil



particles. There, due to the agent's con-centration exceeding its solubility limitin the unfrozen water film, precipitationcould occur, thus bonding the soil parti-cles to one another (Kemper et al. 1974,Wada & Nagasato 1983). Since thisbonding probably involved irreversibleor slowly reversible reactions (Kemperet al. 1987), these newly precipitatedbonding agents did not go back into so-lution during the subsequent thawingperiod. As freeze-thaw cycles continuedto accrue with additional aggregate re-orientation and reshaping, more of thebonding agents that had remained in so-lution in the thin unfrozen water filmssurrounding the soil particles (Miller1980) during the first cycle likely precip-itated from the soil solution. This pre-cipitation mechanism and/or Ca's elec-trostatic bonding of clay platelets men-tioned above could account for the of-ten observed gradual increase in stabil-ity with freeze-thaw cycles (fig. 1).

The aggregate stability of every soilwas greatest at the low water content(0.05 or 0.10 g/g) at every level of freeze-thaw cycle (fig. 1). Moreover, aggregatestability decreased significantly from lowto high water contents for all soils atall levels of freeze-thaw cycle. Pack-ing would have weakened the wet aggre-gates more so than dry aggregates (For-manek et al. 1984). This would ac-count for the observed decrease in theaggregate stability of the samples thatwere not frozen (freeze-thaw cycle 0).This same weakening may also have ac-counted for some of the decrease in ag-gregate stability for the frozen samples(freeze-thaw cycles 1 & 3). However,Lehrsch et al. (1991) presented evidencein their fig. 1 that demonstrated that,even for soil samples that were not sub-jected to stress by packing, after freez-

ing aggregate stability decreased signifi-cantly (i.e., a statistically significant lin-ear trend) with increasing water content.In the frozen samples at the high watercontent, pressure exerted by expansionof ice crystals within the constrainedsamples could have developed planes ofweakness in the aggregates whose hori-zontal displacement was limited by thebrass cylinder walls. During subsequentwet sieving, these aggregates likely frac-tured along these planes. The decreasein aggregate stability as water contentincreased varied from soil to soil. In gen-eral, stability decreased most for coarse-textured soils and least for fine-texturedsoils.

One can compare the two soils dif-fering primarily in organic matter con-tent. Aggregates of a Palouse silt loam(two and one half times as much or-ganic matter as Portneuf, tab. 1) weremore stable than aggregates of Portneuf(fig. 1). Elasticity provided by organicmatter may have enabled Palouse aggre-gates at low water contents to withstandice lens expansion pressures before frac-turing. High molecular weight organicmolecules, while remaining attached tomineral particles, may be able to un-fold or flex to a degree before break-ing. In contrast, at high water contents,there may have been insufficient elas-ticity present to prevent fracture. In-deed, stability differences among thesetwo soils were least at the high watercontent level.

As mentioned earlier, the high organicmatter soils, Barnes and Palouse, usu-ally increased significantly in stabilitywith freeze-thaw cycles. In contrast,the Sverdrup and Portneuf soils (organicmatter contents of 2.21% or less and claycontents of 12% or less) in general ei-ther decreased in stability or did not

J.W.A. Poe•en de M.A. Nearing (Eds.): Soil Sealing and Crusting

60

40

BONDING AGENT & RATE

n NONE CaCO3, 0.2 CaCO3,M 1.0

CaS0 4, 0.2 El CaS0 4, 1.00

20


100

80

BARNES SHARPSBURG CECIL SVERDRUP PALOUSE PORTNEUF

I- TEXTURE I— MINERALOGY -I I- ORGANIC MAT.

Fig. 2: Aggregate stability of each soil as a function of bonding agent and rate ofaddition.

change significantly with freeze-thaw cy-cles. The Barnes with its high organicmatter content likely had more solu-ble organic matter (a potential bond-ing agent) in its soil solution than didthe other soils. The quite stable Sharps-burg also exhibited properties (high or-ganic matter content and relatively highbase saturation, tab. 1) that could ac-count for relatively high concentrationsof bonding agents in solution. Also,the Sharpsburg's high clay content hintsthat clay platelets, suspended and mov-ing in the soil solution, could form claybridges between soil particles.

The effects of clay content can alsobe seen in fig. 1. The soil highest inclay content (Sharpsburg, 41%, tab. 1)had the highest aggregate stability at

both water contents. Mostaghimi et al.(1988) found a similar relationship.

3.2 Interaction between soils,bonding agents, and rates ofaddition

Of particular interest was the influenceof CaCO 3 and CaSO 4 on each soil's ag-gregate stability (fig. 2). Statisticallysignificant effects of a bonding agent onaggregate stability averaged over freeze-thaw cycles and water contents wereidentified only for the Sharpsburg. Ag-gregate stability, when compared to thecontrol, increased significantly (P<0.05)when CaSO 4 at a rate of 0.2% was addedand also when CaSO 4 at a rate of 1.0%was added (P<0.06). The Sharpsburg


100

80

60

40

20


None CaCO 3, 0.2 CaSO 4, 0.2CaCO 3, 1.0 CaSO 4 , 1.0

BONDING AGENT AND RATE (%, w/w)

Fig. 3: Aggregate stability (averaged over soils and freeze-thaw cycles) as a functionof bonding agent and rate of addition at each water content. The water contentswere 5 and 25% (w/w) for the Cecil and Sverdrup soils and 10 and 30% (w/w) forthe remaining soils.

was fine-textured with less than 100%of its exchange capacity saturated withbases, tab. 1. For five of the six soils,calcium sulfate increased aggregate sta-bility the most, though at a probabil-ity level much higher than 0.05. Therelative ranking of the two agents andrates was similar only for the two siltlooms, Palouse and Portneuf. For theother soil pairs, the rankings were notconsistent. On the whole, fig. 2 revealsthat, on an individual soil basis, neitherof the amendments at either of the ap-plied rates exerted a marked influenceon aggregate stability.

Both lime and gypsum additionstended to increase aggregate stability(i.e., strengthen bonds between soil par-ticles) when compared to a control(fig. 2). Though the increases were not

significant at the standard 0.05 level,the trend for five of the six soils inthis experiment was clearly for stabil-ity to increase when either amendmentwas added. Only for the weakly sta-ble Portneuf silt loam did the additionof potential bonding agents tend to de-crease aggregate stability. The Portneufprobably responded differently becauseit was Ca-saturated initially, tab. 1. Assuch, it was among the least affected bythe addition of the Ca-containing bond-ing agents. Complete base saturation ofthe exchange complex does not entirelyexplain this phenomenon, however, be-cause the Barnes and Sverdrup also hadbase saturations of 100% yet showedimprovements in stability with bondingagent additions, fig. 2.

J.W.A. Poesen do M.A. Nearing (Eds.): Soil Sealing and Crusting


3.3 Interaction between watercontents, bonding agents,and rates of addition

When averaged over soils and freeze-thaw cycles, at the low water content,bonding agents and addition rates didnot affect aggregate stability (fig. 3). Atthe high water content, however, the ad-dition of CaSO 4 increased stability overthat of the control. When gypsum wasadded at a rate of 1.0%, the increase wassignificant at P<0.072 and when addedat a rate of 0.2%, the increase was sig-nificant at P<0.05. Shanmuganathan& Oades (1983) found gypsum applica-tion rates of 0.2% (w/w) to coagulateclay and thus encourage aggregate for-mation. For soil management, they rec-ommended annual gypsum additions of0.2% rather than higher rates.

In this experiment, the aggregate sta-bility (averaged over freeze-thaw cycles)of soil samples at relatively high watercontents having received either lime orgypsum, was increased more by calciumsulfate than calcium carbonate (fig. 3).This could well be attributed to CaSO 4being more soluble (well over 150 timesas soluble in water) than CaCO 3 . Moregypsum could have dissolved in the soilsolution where the Ca could have movedby mass flow to the enlarging ice lens.Once there, due to solute exclusion uponwater freezing (Miller 1980), it couldhave precipitated as CaSO 4 or someother Ca-containing molecule at pointsof contact between soil particles or claydomains (Rimmer Si Greenland 1976),thus strengthening bonds within aggre-gates. Perfect et al. (1990b) detectedboth water and solute (Ca) movementto freezing zones of a medium-texturedsoil. They found as well that Ca min-imized dispersion and was thus impli-

cated as compressing the diffuse doublelayer surrounding clay particles. Thisdouble layer compression could permitelectrostatic bonding by Ca of adjacentclay platelets, as mentioned above.

Other studies provide additional ev-idence of gypsum being more effectivethan lime as a soil additive. Shan-muganathan Oades (1983) found ex-changeable Ca to be increased more byCaSO4 than by CaCO 3 . Uehara & Jones(1974) expected gypsum to be more ef-fective than lime as a soil amendmentbecause CaSO 4 raised neither soil pHnor surface charge density.

Though the differences in stability be-tween addition rates within a bondingagent averaged over all six soils were notsignificant at the 0.05 level, aggregatestability decreased consistently with in-creasing bonding agent addition rates.One may speculate why this would oc-cur. Three explanations can be of-fered. First, supersaturation seldom oc-curs when a compound is present in asolution at a concentration greatly ex-ceeding its solubility limit. At the lowerrate (slightly over the solubility limit),the soil solution may have been super-saturated with Ca. As a consequence,more Ca moved in the water stream tothe ice lens, there to be precipitated. Incontrast, at the higher rate (far abovethe solubility limit), no supersaturationoccurred and less Ca was transportedvia mass flow. Second, at the higherrate a much greater proportion of thebonding agent would have been presentas undissociated molecules or discreteparticles. It was likely much more dif-ficult for these relatively large particlesto pass through the unfrozen water filmsand come to rest at the particle con-tact points. Third, with increasing addi-tion rates, precipitated lime or gypsum



molecules rather than Ca ions alone mayhave occurred between clay platelets orclay domains. The resultant weak Ca-lime or lime-lime van der Waals bondsso formed were thus broken more easilythan the stronger Ca-clay electrostaticbonds formed at the low addition rates.In other words, at the higher rates, thebonding forces per unit of separationdistance were weaker than at the lowerrates.

3.4 Additional singledegree-of-freedomcomparisons

Statistical comparisons of the two soilsin each pair were made using singledegree-of-freedom contrasts. Withoutexception, the soils differing in texture(Barnes and Sharpsburg) and the soilsdiffering in organic matter (Palouse andPortneuf) responded differently to bothfreeze-thaw cycles at each level of watercontent and also to bonding agents andrates of addition. In contrast, the miner-alogically dissimilar Cecil and Sverdrupresponded alike. Thus, as figs. 1 and 2imply, texture and organic matter sig-nificantly affected aggregate stability inthis experiment. Mineralogical differ-ences, however, had little influence.

3.5 Results in relation to others'findings

The results of this investigation are sim-ilar in many but not all respects to theresults obtained by other investigators.The finding of this study that aggre-gate stability increases with at least thefirst freeze-thaw cycle (the Barnes andPalouse data of fig. 1) is supported bythe results of Mostaghimi et al. (1988)for a Crofton silt loam (fine-silty, mixed,

mesic Typic Ustorthent) with 0.5% or-ganic matter. On the other hand,Mostaghimi et al. (1988) reported de-creases in stability with the first freeze-thaw cycle for two other soils. Thisdiscrepancy may have occurred becausetheir soils were air-dried before theirexperiment was conducted. Air-dryingcould have strengthened bonds withinthe aggregates, thus effectively maskingthe increase in aggregate stability oftendetected in this study for the first fewfreeze-thaw cycles (fig. 1). When, as inthis study, soils are not air-dried prior toanalysis, this increase in stability withfreeze-thaw cycles may commonly occur(Lehrsch et al., 1991). The finding thatCaCO 3 added at either 0.2 or 1.0% byweight did not significantly affect aggre-gate stability agrees with the results ofChepil (1954), Kemper & Koch (1966),and Toogood (1978) but disagrees withthose of Al-Ani Dudas (1988). Therewere, however, differences in both soilsand procedures between the study of Al-Ani Dudas (1988) and this study. Thesoil samples used by Al-Ani Dudas(1988) were from the BC horizons of twosoils with clay contents of 32 or 42%whereas the soil samples used in thisstudy were from the A p horizons of soilsgenerally with much less clay (tab. 1).The pH values and/or ions present onthe clays were also surely different. Pro-cedurally, air-dry soil and 4- to 8-mmaggregates were used by Al-Ani & Du-das (1988) while field-moist soil and I-to 4-mm aggregates were used here. Anumber of the earlier studies that foundCaCO3 to affect aggregate stability usedvery high calcium carbonate applicationrates. The application rates consideredin this investigation were more feasiblefor agricultural production.

J.W.A. Poesen & M.A. Nearing (Ed..): Soil Sealing and Grunting


4 Conclusions

Soil texture and organic matter, but notmineralogy, affected aggregate stabilityafter freezing. The stability of field-moist aggregates usually increased withat least the first freeze-thaw cycle. Theaddition of either lime or gypsum tendedto increase aggregate stability, but at aprobability level much higher than 0.05.At low water contents, neither poten-tial bonding agents nor addition ratesaffected aggregate stability after freez-ing. In contrast, at high water contents,CaSO4 added at a rate of 0.2% by weightsignificantly increased aggregate stabil-ity over that of a control. The one re-sponse most consistent throughout theexperiment was that aggregate stabil-ity decreased significantly with increas-ing water content at the time of freezing.

Acknowledgement

The authors gratefully acknowledge thecontributions of Dr. Russ Bruce, Dr.John Gilley, Dr. Hans Kok, and Dr.Bob Young for initial soil sampling andof Miss Michelle Garrison, Miss TonyaPearson, Mr. Ron Peckenpaugh, andMiss Deb Watkins for sample prepara-tion and laboratory analyses.

ReferencesAL-ANI, A.N. & DUDAS, M.J. (1988):Influence of calcium carbonate on meanweight diameter of soil. Soil Tillage Research11, 19-26.

BRYAN, R.B. (1971): The influenceof frost action on soil aggregate stability.Transactions of the Institute of British Ge-ographers 50, 71-88.

BULLOCK, M.S., KEMPER, W.D. &NELSON, S.D. (1988): Soil cohesion asaffected by freezing, water content, time andtillage. Soil Science Society of America Jour-nal 52(3), 770-776.

CHEPIL, W.S. (1954): Factors that in-fluence clod structure and erodibility of soilby wind. III. calcium carbonate and decom-posed organic matter. Soil Science 77(6),473-480.

ELLIOT, W.J., LIEBENOW, A.M.,LAFLEN, J.M. & KOHL, K.D. (1989):A compendium of soil erodibility data fromWEPP cropland soil field erodibility exper-iments 1987 & 88. NSERL Report No. 3.USDA-ARS National Soil Erosion ResearchLaboratory, W. Lafayette IN.

EMERSON, W.W. (1983): Inter-particlebonding. In: Division of Soils, CSIRO, Mel-bourne, Australia. Soils: an australian view-point. Academic Press, London, 477-498.

EMERSON, W.W., & BAKKER, A.C.(1973): The comparative effects of ex-changeable calcium, magnesium, and sodiumon some physical properties of red-brownearth subsoils. II. the spontaneous dispersionof aggregates in water. Australian Journal ofSoil Research 11, 151-157.

FORMANEK, G.E., McCOOL, D.K., &PAPENDICK, R.I. (1984): Freeze-thawand consolidation effects on strength of a wetsilt loam. Transactions of the American So-ciety of Agricultural Engineers 27(6), 1749-1752.

HOEKSTRA, P. (1966): Moisture move-ment in soils under temperature gradientswith the cold-side temperature below freez-ing. Water Resources Research 2, 241-250.

KEMPER, W.D., EVANS, D.D. &HOUGH, H.W. (1974): Crust strengthand cracking: Part I. strength. In: J. W.Cary & D. D. Evans (eds.), Soil crusts. Ari-zona Agricultural Experiment Station Tech-nical Bulletin 214, 31-38.

KEMPER, W.D. & KOCH, E.J. (1966):Aggregate stability of soils from westernUnited States and Canada. U.S. Depart-ment of Agriculture Technical Bulletin 1355.U.S. Government Printing Office, Washing-ton DC, 52 pp.

KEMPER, W.D. & ROSENAU, R.C.(1986): Aggregate stability and size distri-bution. In: A. Klute (ed.), Methods of soilanalysis, Part 1. Agronomy Monograph, No.9, 425-442.

KEMPER, W.D., ROSENAU, R.C. &DEXTER, A.R. (1987): Cohesion devel-opment in disrupted soils as affected by clay



and organic matter content and temperature.Soil Science Society of America Journal 51,860-867.

LEHRSCH, G.A., & JOLLEY, P.M.(1989): Temporal variation in the aggre-gate stability of 14 U.S. soils. In: Agronomyabstracts. American Society of Agronomy,Madison WI, 284.

LEHRSCH, G.A., SOJKA, R.E., CAR-TER, D.L. & JOLLEY, P.M. (1991):Freezing effects on aggregate stability af-fected by texture, mineralogy, and organicmatter. Soil Science Society of AmericaJournal 55(5), 1401-1406.

LOGSDAIL, D.E. & WEBBER, L.R.(1959): Effect of frost action on structureof Haldimand clay. Canadian Journal of SoilScience 39, 103-106.

MILLER, R.D. (1980): Freezing phenom-ena in soils. In: D. Hillel (ed.), Applicationsof soil physics. Academic Press, New YorkNY, 254-299.

MOSTAGHIMI, S., YOUNG, R.A.,WILTS, A.R. & KENIMER, A.L.(1988): Effects of frost action on soil ag-gregate stability. Transactions of the Ameri-can Society of Agricultural Engineers 31(2),435-439.

MYERS, H.E. (1937): Physicochemical re-actions between organic and inorganic soilcolloids as related to aggregate formation.Soil Science 44, 331-359.

PERFECT, E., KAY, B.D., VANLOON, W.K.P., SHEARD, R.W. &POJASOK, T. (1990a): Rates of changein soil structural stability under forages andcorn. Soil Science Society of America Jour-nal 54, 179-186.

PERFECT, E., VAN LOON, W.K.P.,KAY, B.D. & GROENEVELT, P.H.(1990b): Influence of ice segregation andsolutes on soil structural stability. CanadianJournal of Soil Science 70, 571-581.

POJASOK, T. & KAY, B.D. (1990):Effect of root exudates from corn andbromegrass on soil structural stability.Canadian Journal of Soil Science 70, 351-362.

REID, J.B. & GOSS, M.J. (1982): In-teractions between soil drying due to plantwater use and decreases in aggregate stabil-ity caused by maize roots. Journal of SoilScience 33, 47-53.

RICHARDSON, S.J. (1976): Effect of ar-tificial weathering cycles on the structuralstability of a dispersed silt soil. Journal ofSoil Science 27(3), 287-294.

RIMMER, D.L. & GREENLAND, D.J.(1976): Effects of calcium carbonate on theswelling behavior of a soil clay. Journal ofSoil Science 27, 129-139.

SAS INSTITUTE, INC. (1985): SAS/STAT guide for personal computers. 6th ed.SAS Institute, Inc., Cary NC, 378 pp.

SHANMUGANATHAN, R.T. &OADES, J.M. (1983): Modification of soilphysical properties by addition of calciumcompounds. Australian Journal of Soil Re-search 21, 285-300.

SILLANPAA, M. & WEBBER, L.R.(1961): The effect of freezing-thawing andwetting-drying cycles on soil aggregation.Canadian Journal of Soil Science 41, 182-187.

SOIL CONSERVATION SERVICESTAFF (1984): Procedures for collectingsoil samples and methods of analysis for soilsurvey. USDA-SCS Soil Survey Invest. Rep.No. 1. U.S. Government Printing Office,Washington DC.

TOOGOOD, J.A. (1978): Relation of ag-gregat e stability to properties of Albertasoils. In: W.W. Emerson, R.D. Bond & A.R.Dexter (eds.), Modification of soil structure.John Wiley and Sons, Inc., Chichester, 211-214.

UEHARA, G. & JONES, R.C. (1974):Bonding mechanisms for soil crusts: PartI. particle surfaces and cementing agents.In: J.W. Cary & D.D. Evans (eds.), Soilcrusts. Arizona Agricultural ExperimentStation Technical Bulletin 214, 17-28.

WADA, S.A. & NAGASATO, A.(1983): Formation of silica microplates byfreezing dilute silica acid solutions. Soil Sci-ence and Plant Nutrition 29, 93- 95.

Address of authors:Dr. G.A. LehrschDr. R.E. SojkaMs. P.M. JolleyUSDA-Agricultural Research ServiceSoil and Water Management Research Unit3793 N 3600 EKimberly, ID 83341U.S.A.

J.W.A. Poesen dr M.A. Nearing (Eds.): Soil Sealing and Crusting

freezing effects on aggregate stability of soils amended with lime

Documents