1B

Arthur Casagrande, Harvard Uni-versity;.P. C. Rutiedge, Moran, Proctor,Meuser & Rutledge; and K. B. Woods,Purdue University, are the investigationalconsultants. Acknowledgment is alsomade to T. William Lambe of Massachu-setts Institute of Technology, who wasengaged as a consultant on studies of ef-fect of mineral composition of soil finesand investigations of admixtures to pre-vent or minimize frost action in soils.

REFERENCES

1. Frost Effects Laboratory, NewEngland Division, Corps of Engineers,U. S. Army, Boston, Massachusetts,"Report on Frost Investigation 1944-1945", dated April 1947.

2. Frost Effects Laboratory, . NewEngland Division, Corps of Engineers,U. S. Army, Boston, Massachusetts,"Frost Investigation ß4!L947 . AddendumNo. 1 to Report on Frost Investigation1944-1945", dated June 1948.

3. Haley, James F. andKaplar, Ches-

ter rff. , "CoId Room Studies of FrostAction in SoiI', Highway Research BoardSpecial Report No. 2, 1952, "Frost Actionin Soils, A Symposium", pp. 246-267,

4, Bouyoucos, G. J, , "Soil Tem-peratures" Michigan Agricultural College,Experimental Station, Technical Bulietin,No. 26, Jan. 1916, 133 pp.

5. Kerr, P. F. , Kulp, J. L,. , andHamiton, P. K. , "Differential ThermalAnalyses of Reference ClayMinerat Speci-mensrr, American Petroleum InstituteReport of Project 49, New York, 1951.

6. DyaI, R. S. and Hendricks, S. 8.,"Total Surface of Clays in Polar Liquidsas a Characteristics Index", Soil Science,Vol. 69, June 1950.

7, Grim, Ralph E, , "Relationof FrostAction to the Clay Mineral Composition ofSoil Materialsr', Highway Research BoardSpecial ReportNo. 2,1952, "Frost Actionin Soils, A Symposium", pp. 167-L72,

8. Michaels, Alan S. , "Altering SoiI-Water Relationships", Proceedings ofConference on Soil Stabilization, Massa-chusetts Institute of Technoloey, June 1952.

Frost Design Criteria for Pavements

KENNETH A. LINELL, Chief, Frost Effects Laboratory,New England Division, Corps of Engineers

THE increases in traffic andwheel loadingson airfieldand highwaypavements inthe past 10 to 15 years; the rising costs of pavement construction, maintenance,and repair; the greater need formaintaining pavements in fully serviceable con-dition at ali times; and theincreasing of operatingspeedshave madeit necessaryto consider frost action in greater detail in pavement design. This paper de-spribes criteria formulated by the Corps of Engineers to meet the needs of itsconstruction in areas of seasonal frost.

The variation of subgrade strength through the seasons is illustrated, and itis indicated that the frost-meltinþ þeriod is critical when conditions are con-ducive to active frost action. Methods forrecognitionof conditions of soil, tem-' perature, and moistune which result in detrimentat frost action are described.Base composition requirements are given. Load design charts for airfield andhighway flexible pavements for various types of loadings are presented. Load-design criteria for rigid pavements are also given. The application of thesemethods is illustrated by means of design examples. Needed studies to furtherimprove the present design criteria are discussed.

O fUn detrimentaleffectsof frost action during the winter and by ioss of strengthin subsurfacematerials aremanifested by of affected soils rvith a corresponding re-heave of pavements or other structures duction in ioad-supportingcapacity during

the period of weakening which ensues. Inpavements, these effects mayresult in un_satisfactory riding qualities, excessivemaintenance, hazardous operational con_ditions, or pavement breakup.

In highways, the great increases intraffic and wheel loads in the past decademay cause pavements which formerlyappeared adequate to deteriorate in thespring. The interruption or slowing oftraffic and damage to equipment which mayresult from frost action involves muchmore money value under the increasedtraffic conditions than in the past. Thecorresponding cost for maintenance, re_pair, and rebuitding of frost_damagedpavements iS also increased.

On airfields, the great increases inwheel loadings have cr.eated problemsnot encountered on highways, because gfdeeper and more intense s[ressing of iliesubgrade. There mustalso be conãideredthe possibility of damage or hazard toexpensive present-day þlanes and theircrews. This involves the necessity formaintaining smooth surfaces on runwayswhgre speeds may exceed 100 mph.

Thus, the detrimental effects of frostaction must be taken intoaceount in pave_ment design. At the same time, we muststrive to avoid over design, since everyextra inch of pavement structure will addenormously to the cost of pavements whenmultiplied over all highway and airfieldpavements constructed over soils subjectto the frost action.

Thepresent paper describes frost designcriteria developed by the Corps of Engi_neers to meet design needs in areas-ofseasonal frost action. The criteria arebased on the assumption that permanentmilitary pavements should be dósigned sothat there will be no interruption õf traf_fic at any time of the year ãue to differ_ential heave, reduction in load-supportingcapacity, or deterioration'of the pãvementresulting from frost action.

The criteriahave beendeveloped in theFrost Effects Laboratory, New England?luigi9l, Boston, Massachusetts, fo-r theAirfields^Branch, Engineering óivision,Military Construction, Office ót ttre Ctrieiof Engineers, U. S. Army. The studiesare continuing, and it is anticipated thatimprovements will be made in the cri_teria from time to time in the future.

19

DEFINITIONS

The following specialized terms areused in this paper:

Degree-day is each degree in any oneday-Iñãt--Iñõ average daity ãir te-p""ät,r""varies from 32 F. The difference be-tween the average daily temperature and32 F. equals the degree daysfor ihat day.The degree days are minus when the aver-age daily temperature is below 82 F. a¡dplus when above. A cumulative degree-days-time curve is obtained by ptotttngcumulative degree-days against time asillustrated in Figure 2.

Freezing index is the number of degree_daysTffi eêñTñã-h i gh e s r and ro *"li pTint son the cumulative degree-days_time curvefor one freezing season (sðe Fig. 2). Itis used asa measureof thecombined'dur_ation and magnitude of below-freezingtemperatures occurring during any givenfreezing season. The index ãetei"rñmeofor air temperatures at 4. b ft. above theground is commonly designated as the airfreezing index, while thai determined fortemperatures immediately below the sur_face is known as surface freezing inàex.. ,Mea! free?ing. index is the fieezingrnoex determined on the basis of meantemperatures.

Frost action is a general term used inrefõFenõffi-Tieezing and thawing of mois_ture in materials and the resultait effectson thèse materials and the structures ofwhic.h they are a part or with which theyare in contact.

- Ice segregation in soils is theot ice as distinct lenses, Iayers,and masses; commonly, ¡ut noioriented normal to the directionIoss.

growthveins,

always,of heat

Frost boil is the breaking of a highwayor ffifiõIìlGîrrface under trJtic

"nãäu.-tion-of subgrade soils in a soft and sðupycondition caused by the melting of thesegregated ice formed by frosi action.

Frost heave is the raising of a surfacedue-ïõTñ-eErm-ation of ice i-n ttre unãer_lying soil.,, Frost:.me.Iting period is an interval ofrne year during which the ice in the foun_dation materials is returning to a liquidstate. It ends when all thã ice in itreground has melted or when freezing isresumed. Although in the generaliäed

'J¿-

'lÌl.4..:ì"

::..l:Ì.

..ù,;1rlìllr

\tj

'..:1¿;

Y,i.

20

zotsoJ 1,4

o

3 l.e&0@e t.o

oz; o.Bo<i

ã 0.6ôr<lolJl o.4ololô

frlÉ ",elstslcilH o.o=l--l<'ø l=Pl6clzoÉÉ

:\t

t.'

..''.,: I

i(;r,'r.ã!r

¿'.

..::,.';1.

Figure I. Hesults of stabic-Ioad tesls \¡/rth 30-in.-diameter plateon lB-in, bituminous-surface-treated gravel base course for Lime-stone, ltlaine, frost test section' average of Positions 3 and4'

case there is visualized only one frost-metting period, beginning during the gen-eral rise of air temperatures in the spring,one or more significant f rost-meltingintervals may occur during the winterseason.

Frost-susceptible soils are those inwh@gation willoc-cur when the requisite moisture and freez-ing conditions are present.

Non-frost-susce4iqs_jrnateria]s arema gravel,sand, slag, cinders, or any other cohe-sionless material in which ice segregationdoes not occur under natural freezingconditions.

Pavement pumping is the ejection ofsoiffim jointsand cracksof rigid pavements under the action oftraffic.

A coverage is one application of thewheeT-Iõ-over each point in the mostheavily travelled area of the traffic lane.Four and 40 coverages are assumed hereto correspondapproximately to30 and 300Ianding and take-off cycles, respectively,on an airfield.

Frost capacity design is a pavementdes@uateundermaxi-mum traffic usage throughout the frost-

melting period. For airfields, this usagecorresponds to approximately 40 totalcoverages per day during the period ofweakening due to frost, by the designwheel ioad and assembly. Greater fre-quencies of operation can be toleratedwith loadings lighter than the designloading.

Frost timited design is design intendedto b@ef initelY r est rictednumber of coverages per day during theperiod of weakening due to frost, by thedesign wheelload andassembly. For air-fieids, this design corresponds to approxi -mately four coverages per day. Greaterfrequencies of operation can be toleratedwith loadings lighter than the designloading.

ICE SEGREGATION AND ITS EFFECTS

Evidence indicates that ice segregationwitt always occur in a frost-susceptiblesoil when it is frozen gradually, under con-ditions similar to those experienced innature, with ample water available. Across section through such a frozen soilwill commonly show the segregated ice aslenses, Iayers, veins, and irregular mas-ses. However, under someconditions and

with some soils, the segregation may oc-cur so uniformly through the soil massthat it maybe difficult to determine visual-ly the extent to which ice formation mayhave occurred.

In soils where appreciable water isdrawn from below into the zone of freez-ing, pavement heave is the most obvioussignof icesegregation. However, in claysconsiderable ice segregation may occurwithout appreciable heave, the water toform ice lenses being obtained by pullingit òut of the di'rectly adjacent soil, whichtends to be consolidated in the process.In this case, the oniy heave is that repre-sented by the relatively small expansionresulting from change of a portion of thesoil water from the liquid to the solidstate. Spring breakup of pavements onfine-grained soils may sometimes beattributed erroneously to other phenomenathan frost action simply because no heavehas been noticed. Also, it is a normaltendency to dig an exploratory test pitonly afterthe failure hasbecome apparent;such an investigation may find no signs ofice segregation, the ice having, by thattime, already melted.

Atthough pavement heave is definitelya problem in design, the subsequentweak-ening in the frost-melting period is moreserious. As frost melting penetrates afrost-susceptibie subgrade underlying apavement, the melting of segregated icereleases an excess of water which must(1) escape upward to the surface of the

21

subgrade or (2) beabsorbed by the thawedsoil. In either case, the tharved soil isloose, with a disrupted structure and anexcess of moisture. Its shearing strengthis therefore low. Even though ice segre-gation may have been weak, marked weak-ening of the load-supporting capacity mayresult if the soil conditions are such,ti¡ata small increase iri moisture content re-sulting from the frost action will causeappreciable loss of strength. The mostcritical period is usually from a few daysto 3 weekfs duration, and as the excesswater drainsfrom thesubgrade, the pave-ment gradually regains strength. Thetime duringwhich strengthcontinues to beregained varies from a few weeks to sev-eral months, depending upon the intensityof ice segregation, depth of frost pene-tration, rate of thawing, permeability ofthe soil, drainage, and traffic conditions.

Figure 1 shows the variation of subgradestrength under a bituminous-surf ace-treat -ed, l8-in. gravel base course at Lime-stone Air Force Base in Maine during1950 to 1952, as measured by plate-bear-ing tests. A plate bearingtest on a frost-softened subgrade is not considered a goodmeasure of magnitude, of reduction inbearing capacity under traffic. However,it does provide a means of picturing (1)the relative change of strength throughthe season and (2) theduration of weaken-ing. Conditions representedby the seriesof tests shownin Figure 1 arenot atypicalof many pavements on frost-susceptiblesoils, in areas subject to frost action.The regain of bearing capacity after thesudden drop during the spring frost-melt-ing period is seen to be rapid at first,then somewhat slower. It is interesting tonote that the loss of strengthwas apparent-ly considerably greatér in the spring of1952 than in the spring of 1951. The testarea did not receive traffic during theseries of tests. If it had, the wheel load-ings might havehastened the reconsolida-tion, but aiso the remolding effect of thetraffic might have increased the degree ofweakening.

It is obvious from Figure 1 that underseasonal frost conditions the bearing ca-pacity of thepavement isnot a fixed thing,but is something which is constantlychanging. If our design is based on fieldsoil tests and natural conditions, we ar'e

l

Ìlov.

Figure 2.

OEO. JAI¿. FEB.

Determination of

MAR. APR.

freezing index.

ø

ôúrUG0UoU:ts

22

ù%

-.ì

rù%o

aFigure 3. lvlean freezing-index

liliety to obtain different designs dependingupon the particular time of year when we

oLtain oui field data.' The strength in thefrost-melting period is seen to be thecritical value which must be considered indesign. This spring weakening has longbeen recognized in the practice of limitingwheel loads to arbitrarily reduced valuesin the spring. On some airfields a com-parable result has been obtained by limit-ing number of dailylandings and take-offsduiing the weakened period. IJVhile suchapproaches are practieal in many instan-

Volueg of eonlours oreFreezing lndcx cxPressed inDogree Doys belou 32oE

il,r,finil APQTOZhA/Csaa//zer/V /¿ht/alare¿ ràere,/zaslma4 áe et2ec/cdla peze/ra/e 4are'mczl I áase /¿ edep/2 afe//e¿s//Z¡zcâe.s az.29edlerd?e a//Yean'/þ ./O.

data influencrng trost actlon.

ces, theypresent anenforcement problemand are impractical for high-traffic-ca-pacity roads or airfields.

Less conspicuous than either heave orweakening due to frost melt is the generalroughening of an iriitially level pavementwith time as a resqlt of either differentialchanges in subgradbdensity as a result offrost action, or of overstressing duringthe frost-meltingperiod, causing deterio-ration at a f aster ràte than otherwise wor¡ldappty and which Ìogically must be con-sidered as an ecoq'omic loss.

RE COGNITION OF FROST -SUSC EPTIBLECONDITIONS

Direct evidence of the frost activity ofthe subgrade may not be una'¡ailable orinadequate, or it may not be feasible toperform necess¿ry field invesitgationsduring the freezing and frost-melting per-iods. Therfore, determination as towhether or not frost-evaluation criteriaare applicable is usually based on thegradations of the base course and sub-grade soils, and on ground-water condi-tions and air-temperature records. How-

Is

õ¡

E

!i

t

Figure 4.and base

rR€cz$3 rnotx (oroi¿É o^Yg)

Combined ¿hicknees ofreguired to prevent

of subgrade.

pavementfreezing

Irt_-'T'-'t''.\

\\

-.-. -{-.-.

I

---\.,-.

I

tit

'\i

tl::

l

I

ever, all reliable information on pastperformance of comparable pavements inthe area during the freezing season andf rost -melting period should be considered.Maintenance and traff ic records may assistin confirming whether or not frost-sus-ceptible conditions exist. Visibie surfaceeffects which maycontribute to the pictureand which areassociated with frost actionare pavement heave and surfaee crackingduring the winter season, and alligatorcracking, frost boils, noticeable weaken-ing, shoving or deflection, and pumpingin cracks and joints during the frost-melting period. When the presence of suchdefects, or the absence of them, is usedas aguide inthe design, the freezingindexand moisture-condition data for the loca-tion during the actual years of recordmust be examined carefully to determinewhether the severity of the frost conditionsduring the period of observation can beconsidered representative.

In order for ice segregation to occur,three conditions must exist simultaneously:the soii must be f rost susceptible; freezingtemperatures must penetrate the frost-susceptible soil, and a sufficient supply.of water must be availabie.

sgr

The potential intensity of ice segrega-tion in the soil is dependent to a largedegree on its void sizes and may be ex-pressed as an empirical function of grainsize as follows:

Inorganic soils containing 3 percent ormore of grains finer than 0.02 mm. indiameter byweight are considered gener-ally frost susceptible, Certain sandy soilsmay have as high as three or four timesthis percentage of grains finer than 0.02mm. without being frostsusceptible; how-ever, because of their tendency to occurinterbedded with other soils, it has beenconsidered generally impractical to con-sider them separately. Inorganic soilscontaining less than 3 percent of grainsfiner than 0.02 mm. in diameterbyweightare generally not frost susceptible.

In borderline cases the Frost EffectsLaboratory performs freezing tests on the.soils to measure the relative frost sus-ceptibility to ice segregation. ThiS servicehas beenperformed onsoils from govern-

23

ment construction projects covering a widerange of geographical locations.

Frost - susceptible soils have been clas -sified into four groups, listed in order ofincreasing susceptibility. Group F4 soilshave particuiarly high frost susceptibiiity.SoiI names correspond with those definedin the Department of the Army UniformSoil Classification.

Group

F1

F2

Description

Gravelly soils containing be-tween 3 and 20 percent finerthan 0.02 mm. by weight.Sands containing between 3 andL5 percent finer than 0.02 mm.by weight.

F3 (a) Gravelly soils, containingmore than 20 percent finerthan

0. 0.02 mm. by weight, andsands,except fine silty sands, contain-ing more than 15 percent finerthan 0.02 mm. by weight.(b) Clays withplasticity indicesof more than 12 except varvedclays.

F4 (a) All silts including sandysilts.(b) Fine silty sands contain-ing more than 15 percent finerthan 0.02 mm. by weight.(c) Lean clays with plasticityindices of less than 12.(d) Varved clays.

Temperature

The depth to which frost will penetratebelow the surface of the pavement keptclear of snow depends principally on themagnitude and duration of below-freezingair temperatures, of which the freezingindex provides a measure, ând on theamount of waterwhich isfrozeninthe sub-grade. Figure 2 illustrates the computa-tion of thefreezingindex. An approximatevalue of the mean freezing index may beobtained from Figure 3, on which areplotted isograms of mean freezing indexfor the continental United States. How-ever, it is prefereble to compute the freez-ing index from actualdaily mean air tem-peratures measured at aweather observa-tion station in the particu,lar locality, basedon a long period of record.

À9È

---t-+-øí.lô

g¿ r t, o¿-';-Ê{v" 1 22n3Þ- EÈ P-56õ i

lg :'. â-_- 0--9

*åå å3. sã

e:IÉ-Eã:;-l-J . al<tøl

c Coldcal yaor in SOyaorr Ayarog. of 3 coldctl ytor¡ in 30 taor p.r¡od

c Aycrogo of 5 cofdcrt t6or3 in 30 t.or per'od

l{ota: Mc6n lrcazing ¡ndicos rarr comPutod

Szzoosocã zooooL- 18000ão< t600

l¿¡oo

r200

tooo

800

60c,

400

2o0

o27æ 2€(þ2900400 500 600 700 9ol0 1000 lloo 1400 1500 1600 l?oo 1800 ¡9æ 2000 2þo Øo æo 2400

Maon Frcaziog lndèr

freezing index and freezing indexes during colder years for 50 consecutive years'Figure 5. Relationship between mean

An empirical chart showing the rela-tionship between air freezing index anddepth of frost penetration for tlte case ofa drained, granular, non-frost-suscep-tible base course beneath a paved areakept free of snow, is shown on Figure 4.The relationship shown on this chart is anapproximátion, since it averages togetherthe effects of such variables as pavementtype and base-course density and moisturecontent. More comprehensive methods,which take these variables into account,have been studied in the Frost EffectsLaboratory (1, 2) and the Perma-frostDivision, St.- Pãul District, Corps ofEngineers. Carlson has described thePerma.frost Division studies (3), in whichemphasis is on calculation õf depth ofthaw. Improvement of these methods isstill in progress. These methods shouldbe used where detailed computations arerequired.

Fluctuations in the severity of winterfreezing and in the rateof spring thawing,combined with fluctuations in seasonalground-water conditions, cause wide vari-ations in frost action from year to yearand alsobetween localities in anyone year.Experience in the continental United Statesindicates that in 1 year out of perhapsevery 5 to 10 yearsfrostaction conditionsat any given locality are considerably moresevere than the aveïage. The results of acomparison between the mean freezingindex and the freezing indices during thecolder years, for a B0-year record fromapproximatety 35 Weather BureauStationswith a wide geographical distribution, isshown on Figure 5. It will be observedthat inareas of relatively lowmean freez-ing index the relative variation from meantemperature conditions maybe very large.This condition also exists for some dis-tance south of the zero mean freezingindex isogram (see Fig. B).

The design freezing index is, at pres-ent, usually based on the mean freezingind9x, However, thedata onFigureS sug-gest that a more significant freezing indéxvalue would be obtàined by setecting thevalue which occurs, Iet us say, aboui oneyear in ten, particularly in areas of lowmea-n freezing index. On Figure b thiswould correspond to the aveiage of thetnree highest freezing indices ìn the 30years of record. In general, the magni_

25

tude of the freezing index value is not ameasure of the intensity of ice segregation,but rather of the depthof frostpenetration.Use of the one-year-in-ten freezing indexin place of the mean freezing index wouldbringunder thefrost designcriteria pave-ments and regions to which they would nototherwise appty. The greater depth offrost penetration would also correspondwith longer critical weakened periods un-der traJfic. The selection of design Lreez-ing index involves abalancingof the cumu-lative loss from frost damage over the lifeof theconstructionagainst thecost of pro-tectivemeasures. Theproposalforuse ofa one-year-in-ten freezing index ratherthan the mean is tentative and is still un-der study.

Water

Moisture required for formation ofsegregated ice within the soil may be de-rived from an underlying ground-watertable, from infiltration through the pave-ment orat the shoulders, froman aquifer,or from the water held within the voids offine-grainedsoils adjacent to the freezingpiane. A potentially troublesome watersuppiyfor ice segregation is considered tobe present if the highest ground water atany time of the year is within 5 ft. of theproposed subgrade surface or of the top ofany frost susceptible subbase materialsused. When the depth to the uppermostwater table is in excesé of 10 ft. through-out the year, a condition of troubiesomewater supply is considered usually notpresent. However, these water-tablecriteria are not necessarily applicable forimpervious clay soils if the water contentis sufficiently high, sinceit hasbeenfoundthat ice segregation will occur in homo-geneous clay soils at moisture contentsdown to approximately shrinkage timit ofthe clay, without any other water beingavailable for freezing than that originallycontained in the soil voids.

MAGNITUDE OF SUBGRADE WEAKENINGDUE TO FROST ACTION

The degree to which the soil losesstrength during the frost-melting periodand the length of the period during whichthe strength of the soii is reduced depend

26

on the type andcondition of the soil, depthof frost penetration, temperature condi-tions during freezingand thawing periods,the amount and type of traffic during thefrost-melting period, the availability ofwater during the freezing period, and

drainage conditions. The application oftraffic may cause remolding or develophydrostatic pressures within the pores ofthe soil during the period of weakening,resulling in subgrade strength reducedappreciably betow that measured by statictests. Traffic tests performed by theCorpsof Engineers (4) haveshownthat thewheel-load-supporting capacity of a ftexi-ble pavement during the frost-rheltingperiod may be of the order of one third ofthe normal period wheel-load evaluation.Rigid pavements, on the other hand, havebeen found to retain about three quartersof their normal period wheel-Ioad-sup-porting capacities. The smaller reductionin strength of rigidpavements due to frostaction is attributed to the fact that thesupporting capacities of rigid pavementsare not influenced to as great a degree asflexible pavements by changes in the sub-grade strength, and in addition, there isIess loss in strength due to shearing de-formation and remolding during the criti-cal spring frost-melting Period.

BASE COMPOSITION REQUIREMENTS

All base-course materials specified bythe design criteria are required to be notfrost susceptible, except for any portionswhich may extendbelowthe predicted depthof frost penetration. Where the combinedthickness of pavement and base over afrost-susceptibte material is lessthan thepredicted depth of frost penetration, thefoltowing additional design requirementsapply:

1. For both flexible and rigid pave-ments, the bottom 4 in. of base course isrequired to be composed of any non-frost-susceptible gravel, sand, or crushed stoneand is required to be designed as a filterbetween the subgrade soil and overlyingbase course, inorder to preventmixing ofa frost-susceptible subgrade with the baseduring, and immediaiely following, thefrost-melting period. The giadation ofthis filter material is determined in ac-cordance with the filter criteria used in

subsurface drainage design, withthe added

overriding limitation that the filter ma-terial shall, in no case, have more than3 percent by weight finer than 0.02 mm'

2, For rigid Pavements, the B5-Per-cent size of the fiiter or regular base-course material placed directly beneaththe pavement is required to be equal to orgreater than a given diameter in order toprevent loss of support by pumping soilthrough the joints of the rigid pavement'This diameter is presently specified as

'/, ín. , but study has indicated that this isprobably too conservative under modernconstruction practices and considerationis nowbeing givento reducingthis 85-per-cent-size value.

LOAD DESIGN CRITERIA

Where the investigations of soil, tem-perature, and moisture conditions indicatethat a frost-weakening problem does notexist, the pavement designis made in ac-cordance with the standard methods forflexible or rigid pavements. However,if the investigations showthata frostprob-lem does exist, then twoalternate methodsof design are available which will assuresafe carrying of the design wheel load.First, enough thickness of pavement andbase can be provided so that frost doesnot enter the susceptible soil. Second,we can base our design on the reducedstrength of thefrost-susceptible soil dur-ing the frost-meiting period and providesufficient thickness of pavement and non-frost-susceptible base above it to carry,during the most critical days of the frost-melting period, the anticipated rate ofcoverages of the design load.

Pref¡enting Freezing _of Subgrade

In this method the combined thicknessof rigid or fiexible pavement and base ischosen to be not less than the depth offrost penetration determined from Figure4, using the design freezing index deter-mined for the particular locality. Thismethod is required (1) over the extremelyfrost-susceptible, type F4 soils or (2)

wherever significant diff erential pavementheave will be detrimental to high-speedtraffic. However, the method is not re-quired in the case of flexible-pavement

27

areas where appreciable differential heavemay be tolerated, such as parking areas.Exception is aiso permitted when thecauses of nonuniform heaving can be sat-isfactorily corrected by the removal ofisolated pockets of highly frost -susceptiblesoils for the full depth offrostpenetrationor by providing gradual transitions atabrupt changes in subgrade conditioils.In these cases, frost may be permitted toenter the subgrade and design is then basedon reduction in subgrade strength.

The full-protection-design method pro-duces pavement+upporting characterist ic swhich are fairly uniform through the yearbut which may be very conservative withrespect to strength. The cost of this meth-od rises with increase in depth of annualfrost penetration. In regions of deepseasonal frost penetration the method isnot feasible, even in cases where heavywheel loadings would require heavy basecourses in any event, and here arbitrarylimits must be set upon the maximumthickness of base course. In extremenortherly areas the situation is reversed,and thère full protection against thaw ofthe subgrade may be possible. The full-protection - design methods describedshould be considered limited to conditionscomparable to those encountered withinthe continental limits of the United States.

Reduction in Subgrade Strength

Design based on the reduction in strengthof the subgrade during the spring frost-melting periodwill frequently permit Iessdepthof pâvement and base than is requiredfor prevention of freezingof the subgrade.This method is appticable for both flexibleand rigid pavements on subgrade soils ofGroups FL, F2, and F3, when subgradeconditions are sufficiently uniform to as-sure that obj ectionable differential heavingwiII not occur or where subgrade variationsare correctible to thiscondition. As pre-viously noted, the method may also beused where appreciable nonuniform heavecan be tolerated, in flexible pavements oflesser importance not subject to highspeed traffic. The design procedure pro-vides a pavementwhichis justbarely ade-quate during the frost-melting period butwhich necessarily has excess strength dur-tng the remainder 'of the year.

* Wàan ,roal il 9arñ¡llad lo panal.ota lro0g F¡l toll¡ru¡a romr daa¡ln curva oa lorl.ouD F3 lo¡ll, Froal t¡ovld ba ,a.ñ¡ttad lo 9aÀalrola F4 ¡oilr oâlt uñdaa lla¡iòla Þov¡dorao3 ol hararim9ortoôca ¡nara og9rac¡oòta mn-ùnitorn toËmañt ùaova ñoy ùa totarotad.

POiJNDS ONroo fo ?oo Psl

LL B€ NTOUCEO IO PER CEII

.*i:' ;:"ï "ii;i " -;i:;"ï' ;" ";"' î,- "for taxiways, etc., for frost action in

subgrade soil.

For airfield usage, two design cate-gories have been chosen for design basedon reduction in subgrade strength. Frostlimited and frost capacity are used to de-note approximately 4 and 40 total cover-ages per day, respectively, during theperiod of weakening due to frost, by air-plane with weights equal to the designioadings. The frost-Iimited informationrepresented here is tentative only. It issubject to revision and is presented onlyto illustrate the concepts involved in cur-rentstudies. The frost-capacity categoryis normally used for design. The frost-limited category will normally be used onlyfor evaluation of existingpavements to de-termine what loadings may be tolerated inthe springunder a relatively small numberof traffic operations. For flexible pave-ments, separate design curves have beenprepared for the two operational condi-tions, the frost limited curves being ten-tative, as noted. No separate frost-limiteddesign curves have been prepared for rigidpavements. However, where this condi-

UIo4Io

@

ôz

FzUEU

À

oøUzv9Its

oUzo¡o ôo3. 3. 3.oo

SINOLÊ W8EELfrnE Pn€ssuR€

:r

tìRoup NFSCRIPTIONGRAVELLY SOILS CONIAIMNC SETW€€il 3 ANO 20 PER C€NI F¡il€R IHAN

O32ßñ 8Y WÊlcaf

F2 SANOS CoNTÂlNlro ðElwEEl 3 AHO 15 PER C€Nf FINeR TNA{ O.O26m.8Y W€rGXt

F3{oI GRAVELLY SOILS CONTAININc üOR€ IAAI 20 P€R CEilf FIIER THAfl

0.O2ññ 0Y w€lGXl Âflo SANoS,€xC€Pt Flile SILÍY SÂNDS,COflTÂtNtNGMoR€ faÂfl 15 P€R CEflT F|NER faAN 0.02ñm. 8y WErGhf. (bl CLAYS W|TXPLASI¡CIIY INOIC€S OF MORE ÍHAN 12,EXCEPI VARV€O CLAYS,

(o) aLL srLts rNcluortc satoy srLts. (ôl ftN€ stLty saflos ooNratNrNcMORE IHAN 15 P€R CEflÎ FIilER IHAfl O,O2 MO. gY W€I6HI. (C) LEAN CLAYSwrTH PLASTTCTTy rNorôÊS OF LESS THAI t2. (d) VARVÉo CLAYS.

¡o

60

50

ao

p

20

LOAO IN

28

dÍo=ooóêztszU¿u

(o

uz9fFôu¡3o=oo

70

60

50

30

25

20

r5

Toral Loao 't*,99',l3'"", rr'6 p'€ssu.€

Figure ?.

tion must be considered, 10 percent lesspavement thickness than that reguired forfrost-capacity design is considered ade-quate for the frost-Iimited traffic.

1. Reduction in Strength Criteria forFlexibt@rade, the combinedthickness of flexible pavement and non-frost-susceptible base required for thedesign assembly and loading is determinedfrom the applicable curve among Figures6 through 10, Figure 10 being the highwaydesign chart. The designs produeed bythese curves have previously been com-pared with the results of a.n extensiveseries of traffic tests in the frost-meitingperiod, in a paper presented before thettigtrway Research Board in 1951 g).Ho-wever, since the referenced paper wãspresented, the curves have been adjustedto give approximateiy 10 percent less re-quired pavement thickness and the frost-limited concept has been added.

The curves reflect the reduction instrength of the soilduring the frost-melt-ing period. It is considered that the re-duction in strength of subgrades tends,generally, to be greater in cuts than infills. If field data and experience defi-nitely indicate that the reduction in strengthin fitl areas may be expected to be less,because of such factor as the greater depthto water tablq, a reduction in combinedthiclmess of base andpavement forthe fillarea may be pèrmitted. In no case is acombined thictmess of pavement and non-frost-suseeptible baselessthan I in. per-mitted in design, where frost action is a

factor, atthough smaller combined thick-ness may have to be considered in evalu-ation of an existing pavement. Curves on

Figures 6, ? and 10. are therefore showndotted below f-in. thickness.

2, Reduction in Strength Criteria forRigidt io-ñs,-tnon:Î ro s f - sus c eptiblebas e c ou r s e

equal in thiclicress to the thicli¡ress of theconcrete slab is requiredwhere frost pene-tration is permitted into a f rost susceptiblesubgrade beneath a rigid pavement. Thespecifie exceptions to this requirementare as follows:

A. lVhere soils of Groups Fl, F2, andF3 occur under very uniform conditionsof subgrade and the freezing index is lessthan 500, the thickness of the non-frost-susceptible base under a rigid pavementmay be reduced to 4 in. ; it is designed tomeet filter requiremehts outlined pre-viously under "Base Composition Require-ments. t'

B. rffhere soils of Groups Fl, F2, andF3 occur underuniform conditionsand thedepth to the uppermost water table isgreater than 10 ft., the thickness of thenon-frost-susceptible base under a rigidpavement may bereduced to4 in., and thebase is designed as a filter.

C. Over Group F4 soils the combinedthickness of rigid pavement and base isdetermined according to ihe criteria forprevention of freezing of the subgrade.

The thiclmess of concrete pavement isdetermined on basis of anticipated flexuralstrengths and subgrade modulus using

Loao lN poul¡o3 oN ouaL WHEELS 37,6'c.0.267 SO. N. COñrACr Ai€A EAC¡ V¡ÉÉl

Figure 8.

øUEozuøoêz

F2UTE

fo

øUzIIfÞôu2óIo0

60

50

ao

35

50

25

20

r5

øU¡?ID

oôz

ts2U.u

o

Uz9¡Fouz@

=oo

standard Corps of Engineers rigid-pave-ment-design methods. Subgrade modulusvalues for use in these computations aredetermined from Figure 11, which allowsfor the reduced strength of the subgrade.Should actual field test subgrade modulusvalues prove to be lower than those ob-tained from Figure 11, the field test val-ues govern the design.

roraL Loao 'n

toll?tfï å""ti:l IåïD5^T"T"TMBLY

(3r"x6o"c'c''

Figure 9.

EXAMPLES OF PAVEMENT DESIGN

Example 1

Design an access-road pavement for aflexible pavement to withstand a 12,000-lb, wheel load (24,000-tb. axte load) un-der maximum traffic conditions during thefrost-melting period, for the followingconditions:

Design freezing index, 700Subgrade, silty sand 20 percent finer

than 0.02 mm.Highest ground water, 1 ft. below top

of subgradeBase CBR, B0 percent1. Preventiònof Freezingofsubgrade.

Frompavement and base to prevent freezing efthe subgradefor a designfreezingindex of700 is 38 in.

- 2, Reduction in Subgrade Strength.From Fof pavement and base is required over thetype f3 subgradesoil toprovide sufficientsupporting capacity during the weakenedperiod in the spring. Since this thicknesswiII still allowsonle frostpenetration into

29

thè subgrade, some heave should be ex-pected, but it should not be detrimental inthis application provided it occurs uniform-ly, i. e. , there are no abrupt changes insubgrade conditions.Example 2

Design an airfield taxiway for bothflexible and rigid pavernents to with-stand a 25,000-lb. , single wheel loadwith 200-psi. tire pressure undqr maxi-mum traffic e.onditions during the frost-melting period, for the following con-ditions.

Design freezing index, 300 degree-daysSubgrade, uniform lean clay and plas-

ticity index 14.Highest ground water, 2 ft. below sur-

face of subgradeSubgrade CB& B percent (rormalperiod)Base CBR - B0 percentSubgrade modulus, k - 100 lb.

in. per in.Concrete fiexural strength,

per sq. in.

per sq.

650 rb.

1. Flexible Pavement. (a) Preventingfreezi@m Figure 4, thecombined thickness of pavement and basetopreventfreezingofthe subgrade is 25 in.

(b) Reduction in subgrade strength:Soil is type F3. From Figure 6, the re-quired total thickness of pavement and basefor frostcapacityoperation is 28 in. Thisis greater than the depthof frost penetra-tion; therforg design on bases of reductionin subgrade strength is not applicable.Analysis by standard California BearingRatio procedures shows that a thicknessof 22 in. is required forthe normalperibdsubgrade strength. Since the thicknessof 25 in. required to prevent freezing ofthe subgrade is less than the value fromFigure 6 and greater than the 22 ín, re-quired for thenormal period, 25 in. wouldbe selected as the combined thickness ofpavement and base.

2, Rigid Pavement. (a) Preventingfreezin{iFGõ[iãiêl-Trom Figure 4 theminimum thickness of pavement and baserequired to protect the subgrade fromfrost action is 25 in. The required siabthickness from standard rigid-pavement-design chartg with no subgrade weakening,is 11.4 in. The Corps of Engineers' de-sign manual specifies that when the thick-ness from the design curves indicates a

ooôooooooooodoÔooÔ

9

.r:'1,

30

fractional value greater than 1/4 in., thenext full-inch thickness is used for con-struction. The adopted slab thicknessshould therefore be 12 in. , resultingin abase-course thickness of 25-12=13 in.

(b) Reduction in subgrade strength:Since subgrade conditions are uniform andfreezing .index is less than 500, exceptionto the ri gid-pavement-base-course-designcriterion is applicable. A minimum basecourse of 4 in. is required to protectagainst loss of support by pumping. Thesubgrade modulus duringthe frost-meltingperiod is 25 lb. per sq. in. per in. asdetermined in Figure 11. The slab thick-ness required, from standard rigid-pave-ment-design charts, is 13 in. Cost com-parison then indicates whether this designor the one obtainedin the preceding para-graph should be used.

El3ryls_3.

Design an airfield taxiway for bothfiexibie and rigid pavements to withstanda 25,000-1b. single wheel ioad with 200-psi. tire pressure, under capacity opera-tion during the frost-meiting period, forthe foliowing conditions:

Design freezing index, 2000 degree-days

Subgrade, uniform lean clay and plas-ticity index 14

Highest ground water, 3 ft. below sur-face of subgrade

Subgrade CBR, B Percent (normalperiod)

Base CBR, B0 PercentSubgrade modulus, k is 100 Ib. Per

sq.

sq.

irt. per in.Concrete flexural strength, 650 ib. perin.

1. Flexible Pavement. (a) Preventing

required to protect subgrade from freez-ing is 62 in. The slab thickness accordingto standard rigid-pavement-design curveswould be 11.4 in. A 12-in. slab thicknessshould be used in construction, therebyresulting in a base course of 50 in.

(b) Reduction in subgrade strength:Assuming a 12-in. base thickness, thesubgrade modulus, as determined in Fig-ure 11, is 65 lb. per sq. in. per in. Usingthis value, the required siab thickness,from standard rigid - pavement - designcurves, is 12 in., which requires a 12-in. base thiclness in accordance with thepreviously stated general criterion forbase courses under rigid pavements.This confirms the original assumption ofbase thickness. Fortheuniform subgradeconditions, this would be the adopteddesign.

If, in this case, the ground-water-table depth should be in excess of 10 ft. ,

with atl other conditions the same' a 4-in. -minimum-thickness base would bepermitted in accordance with exceptionpreviousty outlined. The design subgrademodulus in accordance with Figure 11

would then be 25 Ib. per sg. in. per in.Using this value of subgrade modulus, therequired slab thickness, from standardrigid-pavement-design charts, would be13.

NEEDED IMPROVEMENTS IN CRITERTA

Actual application of traffic on frost-susceptible areas is the only effectivemeans we have at present of evaluatingsimultaneously all factors rvhich influenceweakeningof frost-susceptible soils. Thetraffic method gives us, particularly, ânevaluation of the remolding action of traf -fic and ofany subtlechanges insoil struc-ture resulting fromthe frost action on thesoil strength. Therefore, we should at-tempt to obtain additional fuliy correiatedrecords of traffic experience on border-line designs during the frost - meltingperiod. These should extend coverageover a fullrangeof soil, temperature, andmoisture conditions. Frost traffic teststo date have covered a fair variety of con-ditions, but much remains to be learned.It is essential that the surface obierva-tions be carefully correlated with data on

the coverages, wheel loadings, soil mois-

freezin@m Figure 4 thecombined thickness of pavement and baseto prevent freezing of the subgrade, forthe design freezing index, is 62 in.

(b) Reduction in subgrade strength:When allowing a reduction in strength ofthe subgrade due to frost action and al-lowing uniform heave of the pavement, atotal thickness of 28 in. is required, ac'cording to Figure 6.

2. Rigid Pavement. (a) PreventingfreezirfõIffi[iãõ:-Trom Figure 4 theminimum thickness of pavement and base

WË"1): r!

&lt

a0

t!

t5oz

ftot'''> raÀ¡lloolto5aflo9¡Þo6Uz_¡¡8e

Figure I0.ture, depth of freezing, soil density, andother basic soil information.

A thin base under a rigid pavement isbelieved to contribute relatively litüestructural effect, since the layer is sothin relative to the depth to which the ma-terial is stressed. However, a thin basecourse will definitely have some drainagefunction in distributing and'removing in-filtration through the pavement or excessmelt water which emerges from the sub-grade during the frost-melting period.This function wiII, of course, vary withthe permeability of the subgrade, amountof ice segregation which has occurred,Iateral drainage distance, rate of thaw.AIso, it is assumed the thin base coursewill reduce or prevent pumping, if it issatisfactorily designed as a filter. pos-sibly conventional filter criteria may betoo eonservative for this special apptica-tion. Much work needs to be done tocrystallize design criteria in this field.

It may be that we should give greaterattention to the difference in spreadingeffects of different base courses underPavements, as for example, betweenrounded natural gravels and angularcrushed rocks. These effects possiblyare of iittle consequence in relativelythin base courses but may be of distinctimportance when base courses reachthicknesses of the order of 2 ft. or more.

The question of how much heave ispermissible under a rigid pavement isalso a thorny problem, nôt süsceptible oftneoretical analyses.

31

We need better methods of analyzingthose combinations of precipitation, freez-ing and rate of thaw which are conduciveto especially severe spring pavementweakening.

We need to obtain better information onthe duration of the maximum weakeningperiod in various soils. In relativetypervious frost-susceptible soils, the meltwater may escape nearly as rapidly as itbecomes available. In soils of lowerpermeability the effective time of drainagemay be slower and the duration of theweakening may be greatly extended. Wealso need more information on the effect-ive permeabiiity of soilsduring the frost-melting period, whichmaynot be thesameas the permeability of a homogeneoussample, due to fissured structure.

- -F.osl 6holl not ba pô.n¡ltcd to poncl.ote group F4 ro¡t8 bonooth.¡g¡d povomenlg.

Figure 11. Rigid-pavement-subgrade mod-ulus curves for frost action in subgrade

Study is needed to determine whetherthe present rule, which states that a soilcontaining 3 percent or more of grains byweight finer than 0.02 mm. is generallyfrost susceptible, can be improved. Var-ious possible methods of making use ofborderline frost-susceptible base-coursematerials should be explored. More dataare needed onthe effectsof depth to watertable, degree of saturation, soil struc-

?

=o

lioJ

ooÉu4035uIFøoÊLYaÞJÞoo:E

uoÉIoÞ

w32

ture, surcharge, depth of cover and other.variabies. James F. Haley described inthe previous paper cold-room studiespresently being made in the Frost EffectsLaboratory to investigate some of theseeffects with the objective of deriving im-proved design criteria (B). Actual fieldperformance data and celtain basic theo-ietical studies are needed in addition tothe cold-room studies.

England Division, Corps of Engineers,U. S. ArmY, BostÖn, Mass', "FrostInvestigation 1945-1947. Addendum No' 1

to Repórt on Frost Investigation 1944-

1945", June 1948.3. Carlson, HarrY, "Calculation of

Depth of Thaw in Frozen Ground", High-*ay ResearchBoard SpecialReport No' 2,

1952, ("Frost Action In Soils, A Sympos-ium"), pp. t92-223,

4, Linell, K. A. and HaleY, J' F',"'Investigationof the Effect of Frost Actionon Pavement Supporting Capacity", High-way ResearchBoard SpecialReport No' 2,

1952 ("Frost Action in Soils, A Sympos-ium"), pp. 295-325.

5. Haley, J. F. "Progress RePort on

CoId Room Studies", presented at HighwayResearch Board Thirty - First Annual'

Meeting, JanuarY 1953.

i::

Soil'Temperature ComParisons

lJnder Varying Covers

GEORGE A. CRABB, JR. , Research Hydraulic Engineer'soil conservation service, u. s. Department of Agriculture, and

IAMES L. SMITH, Hydrotogic Research Assistant,Department of Forestry, Michigan State College

REFERENCES

1. Frost Effects LaboratorY, NewEngland Division, Corps of Engineers,U.' S. Army, Boston, Mass., "RePorton Frost Investigation 1944-1945", April1947.

2. Frost Effects LaboratorY, New

a THE Michigan Hydrologic ResearchStation was established at East Lansingin 1940 as a cooperative study betweenthe Soil Conservation Service of the U. S.

Department of Agriculture and the Michi-gan Agricultural Erperiment Station tostudy the effectof landuse on thehydrologyof farm landsunder varying types of snowcover and frozen soil. As additionalobjectives, it was planned for the stationto: (1) determine the manner in whichfreezing and thawing of soils on water-sheds with varying types of land usecontribute to runoff, erosion, and floodflow under northern' winter conditions,and (2) to determine the fundamêntalhydrotogic relationships of typical Mich-igan soils under varying types of landuse, witl esþecial emphasis upon themovement of water through the soil pro-file during the faII and winter months.

In order to accompiish these objectives,one of the most complete hydrologic in-strumentations in this country was de-vised and installed on lands of MichiganState College and the Rose Lake VfildlifeEx¡leriment Station, near East Lansing(14),

--The mqitiplicity of climatic ánd hydro-logic factors working together to causerun'off , erosion and flood flow, and con-trolling the hydrologic relationships ofsoils requires a broad program of basicreSêarch tci include investigatiöns in manylittle known fields of climatology that areof coirsiderable interest to highway engi-neers, agronómists, agricultural engi-neers, and many other specialists, aswell as hydrologists. One set of rela-tionships which interests both the hy-drologist and the highway engineer isthe air-soil temperature relationship.