._--,...

- - -

Elsc vier Seienee Puhlishing Compa ny, In e.

52 Vand erbilt Avenue. New York, NY 10017

Sole di stribu tors outside the U nited States an d Canada :

Elsevi er Seie nee Pu blishers B. V.
P.O. Box 2 11, 1000 AE Amsterdam , The Netherland

Thjs book has been registered with the Copyright Oearancc Cenler, lne.
For furth er information, please eontact the Copyright C learance Ccnter,
Salern , Massae husetts.

1986 by Elsev ie r Seienee Publishing Co., 1ne.

AH ngh ts reserved.

Library ofCongress CataJoging in Publication Data

Rah n. Pe rry H., 1936­
Engineeri ng geology.

Bibliography: p. 539

Inc\udes ind exo

1. Engin eering geol ogy. 1. Title.

TA705. R2 8 1985 624. 1' 51 85-10190
ISBN 0-444-00942-6

Current prinling (I ast digit ):

109 8 7654 32 1

Manufaetured in ¡he U nited States of America

"'-:"1'1 'JT:7 !. - '\.. '" ~: ~T ........ _~ . . ~'.-,'. ....,.1'". " ~

 3
Weathering and Soil-Forming
Processes

No e ngineering structu re is better than the mate rial of whi ch, a nd on whieh. il is buil!.
A. B. Brink, 1979

Lel us nOI be very hopeful aboul ur humar. co nquesl over na ture. For eaeh su eh viclOry,
natu fe manages 10 lake her re enge.
Friedrich Engels

3.1 Soil and Bedrock

Many engineering geo logists have encountered a situation where it is necessary
to differe ntiate between soil and bedrock. The con tact may b gradational or very
sharp. Contractor specifications which read ' excava te to bedro k" in a n area of
gradational contact may lcad to constructio n delays and litigation. An example
of this sort occurred during the excavation of th e Erie Ca nal when a con tractor
(Mr. Barrett) asked James Hall, the state geologist of N w York for his opinion
as to whether he should get paid for excavating soil or bedrock in a certain cal­
careous shale bed.
Lockport
.Tune 8th. 1839

Jas. Hall Esquir

State Geologist 4th

Dear Sir
1 am very desirous f obta ini ng your opin jon of lhe rock which occurs in the Excavation
opposite the prescnt Locks in this village upon the orth Side of the Ca nal. That is its
Geological Classificat io n and the constituent parts of the dif'ferent lasses as nearly as you
can j udge from lheir appearance as it presents its self to view Commencing with the G ray
lime stone on the surface the rock appcars 10 Cha nge by incensible degrees from the under
urface of the grey li me to the bOllom of our Exca a tion Containing a greater proportion
of AJumÍJla a we deseend the Strata. In Our Original Estimate th is Material was called
Slate rock & Shale we supposcd it would all Come under the head ofShale below the gray
lime Slo n . In observ tng the face from which the rock has becn recentl y blasted 1 See that
there are very heavy laycrs of this materia l with appare ntly li!tle or no Seams, but on Expo­
ure to the Atmosphere il s on yields and crumbl s to pieces in 5 111 a11 Cubcs, and by Con­
tinucd Exposure it becomes decomposed or disintegrates and for ms a very tenatiolls
clay . .. . by gtving your opin ion upon Ih is Subject yo u wilI confer a very great favour. . . .
 - - --- ---- ---

30 Chapler 3 Weathering and Soi l-Forming Proccsses

It is rair to sta te to you that this informa lion is desired lo Enablc Ihe Canal Commis­
sioner and my self tO decide a q uestion raised by Ihe Ca na l C onlraclors (un readable n a me)
in relatio n lo their C ontrael for lhe Excavati on 01' Ihis Material. In the Con tra et they have
a priee fo r "Sa lid rock" and a priec for "S la te rock & Shale."
They Cla im Ihal the whole is Solid rock. T herefore your p rofcssional opinion will be of
greal Service
1 am very respeclfully
Yours
A. Barretl
C hief Engi neer

Alfred Barretl Esq uir

C hief Engineer &c
Dear Sir,
1 ha ve recei ved your favor o f ycsterday and hasten to give a n answer lo your inq uiries
in relation to Ih e rock occurring in the excavation opposite Ih e Locks, 1 understan d you 10
rcquire Ihe names by wh..ich lhe several roeks are known geologically-
The fa ce of the cl ilr presenls tbe foU owing rocks in Ihe descen d ing order-l sl a bo ul 10
fcet of gra y, encri nal li mestone of inches to 2 feet thick. Below the Iim estone are abou l 6
feet composed o f la yers o f a few inches thickness a nd alternaling with seams ofsha le. T his
rock ma)' be lermed an a rgillo-siJi ceous !imestone and probably it contains magncsia or a
trace of iron a nd perha ps ma nganese. lt belongs to the variely wh ich are term ed hydrau lic
lim estones, though this term IS rather vague in its appli catio n, Below the hydraulic Iime­
stonc are nearly 80 feet in thickness of "ca lcareous shal e" - (T he same rock was termed
ealciferous slate" by Prof. Eaton) wilh occasio nal layers of si liceous Iimesto ne and from
onc to fo ur ineh es tbickness. In a ge neral description these would searcely be menlioned
as Ihe a m ounl is so small as not to affect the charac!er of the ma ss as a whole which com es
slrictly with in !he dc no mi natio n of shale, a nd no other name can properly be applied 10
jI. Where it has never bcen exposed 10 lhe weather il se para tes into salid blocks which
readily c1eave in to irregular lami nae. this 1S a character common lo all our sha le rocks. The
upper part of Ihis rock is more compac! and apparently eontai ns a larger porportion o f
carbonate of lime than t hal be low which has a more slaty strueture. O n exposure thi s shale
decomposes into a te naceous day which i5 in Ihc co ndilion of much of thal pOlt ion alo ng
the banks or sides ofthe clilrs below lhe locks. T he change from Ihe shalc 10 a perfect slate
of deco m positi o n is very gradual and ¡here arc so many intermediale stages that il may
no! be easy to d ecide the point when one begins or the other ends, yet for all practical
purposes the disti nc tion is suffic iently ob vious.
1 have here sta ted di stinc tl y my opin io ns ofthe characters ofthese roc ks, and Ihe names
are thosc by which they are known 10 a ll geologists. I make these sta!em ents impartial ly
withoul reference to indiv iduals or circums tances, regarding onl)' Iruth. Should therc be
aDy points which are not sati sfactorily cxpla ined 1 shall give any farther explanation wilh
pleasuTe. You a re aware tha! there can be noth ing assu med or arbitrary on my part as rock
terms are suited and in com mO n usage among geologists of Europe and America. 1 cannot
give an opinion of wha ! constitutes "solid rock " in you r contracts. It is evidenl that you
have certal n specificati ons attached 10 il in contradistinction to shale-if Ihe term solld
rock is applied lo t his l hcn therc is no furt her use of the term " sLa te roek & shale" exee pt
as app licable to the weat hercd edges of such strata which are partially decomposed . Th
circumslances 01' the mass cru mbl.ing in to cubical or angular fragments denotes the pres­
ence of sorne sca lene ma ller, which in this case is probabl y sulphate o f iron and sulphate
or magnesia a rising from lhe deco m posi ti on o firon pyrites, thc sulphuric acid u niting bot h
witb !he iron of the pyriles a nd wi th Ih e magnesia 01' the rock.
 3. 2 Wealhering 31

Wilh regard lo hardness this rock IS far inferior 10 the lim estone a nd is one oflhe sofieSl
rocks which occur in any series.

1 am ve ry rcspectfully
yours
James Hal l
Geologisl 4th Dist.

Ja mes Hall did not hedge the issue by avoid ing a direct response, but simply
tricd to explain th at lhe con tact bctween fum u nweathered shale bedrock and
loase soi l wa s gradationaI.
Real and semanti cal probl ems involving sim ilar problems ofsoi l/bedrock con­
lacts have happened over and over during lhe past cent ury. To a large degree, the
argument of what const itutes soil and bedrock originates with people who do nOl
understand geology.

3.2 Weathering
In arder to understand soil/bedrock contacts, it is necessary to study wealhering
and soil-forming processes.
T he grad ational cha ng o f bedrock to soil is due t o w athering (Reiche, 1950'
Oll ier, 1969; CarroIl, 1970; Hu nt, 1972; Birkeland, 1974, 1984). Weathering is the
physical and chemica l d isintegration or deco m posit ion of geologic deposits. AH
ro ks exposed at th e earth's surface have bcen buried to sorne degree in lhe geo­
logic past and subject d to greater heat and pressure. At 01' near the earth's sur­
face, the envi ro n rnellt is d ifferenl physically and chemically from condiitons al
depth, and the rocks cha nge to accomodate these new co nditio ns.
Fo r di scussion purposes, it is possible to classify wealhering pr cesses into two
categories, 1) physical weath ring, and 2) chernical wealhering.

3. 2.1 Physical Weathering

Physical weathering is the mechanica l breakdown ofroc k. and inc\udes processes
such as pressure and ternperature changes which tend to break up rocks. These
ínclude the following.

Frost. If water is avai lable and l he te mperat ure d rops be low DOC, ice forms
a nd the resul ting 9% increase in volume creates a pressure that can produce tre­
mendou compressive forces. Frost is probably the ehief reasan for the break up
of jointed bed rock outcrops in high mountains with cool clirnates. Thus frost
action works best in join ted rock or where other agencies of weathering initiate
eracks in rocks.
T he spal ling of concrete has been found to be due LO thc formation of ice in
the po re of the aggrcgate (Iyer el al., 1975). High l porous and absorptive partí­
eles of aggrega te, such as sorne cherts, argi llaceous limestones and dolomitcs, and
shales, cause "popouts" or general surface scali ng of concrete due to freezi ng and
thawing (Mielenz. 1962 ).
32 Chapter 3 Weathering and Soil-Forming Processes

Figure 3. 1 Unloading joints (sheeti ng) in gra nite, Sierra Nevada Range, Oroville, CA.

Unloading. Du ring erosion, underlying rocks experience reduced pressure as

overburden is removed. The rocks expand as pressure is relea sed, a proeess called
unloading. Lithostatic stress release and accom pan ying strain (dilation) occurs
slowly, during geologic time, but it is possible for cracks to sudd enJy propagate in
quarries, or fo r wall rocks in mines to burst witb explosj ve violence (Bajn, 1931).
In certain rocks, particularly massive pl utonic igneous rocks (e.g., granite) and
massive sandstone, the process of un)oading causes large joints (sheeting) to
develop (Jahns, 1943). These joints tend to be oriented paraIl el to the genera)
slope of the terrain. A granitic mountain may appcar to be spalling off like layers
ofa giant onion. The magn ificent domes ofYosemite Park, CA, and Stone Mou n­
ta in, GA, are manifestati ons of the process of unloading (Gilbert, 1904; Mathes,
1930; Wahrhaftig, 1965).
Sheeting may ha ve considerable engineering significance. For instance, in a
vertical cut into a granite mountai nside. joints caused by unload ing that dip at
an angle paraUel to the slope of the moun tai nside form perfeet avenues along
whieh failure can oceur (Fig. 3.1). Figure 3.2 shows a cross section of a spillway
cut for a dam in New Hampshire. Widely spaced sheeting joints allowed a small
roekslide to occur in 1943, 3 yr after construction. This was followed by a major
rockslide. Rehabilitation costs totalled 70% of the initial cost of the entire dam .
The engineering geologist should be aware tbat any excavation can trigger
unloading to occur. Natural erosion of overlying rocks ha ve already induced
 3.2 Weatheríng 33

Second S I i de

20 m
,,
,, 10m
S heeting
,
[
Om
Figure 3.2 Rockslidc t
Surry Mt. Dam, NH (after
Om 20m Pariseau and Voight, 1979).

unload ing stresses in an y exposed racks. Further remaval of material by man can
crea te rapid strai n. Railroad cuts slawl heave together over the years, and quar­
ries may shrin k in size due to unload ing. Rack quarry operations have been
knawn to produce rapid strain release, which may be due to li thostatic stress but
also include tectonjc stresses (Block et aL, 1979). (Chapter 4 contains quan titative
examples of unload ing phenomena. )
D uri ng the construction of lunnels unloading may cause spalling rock or 'pop­
ping rocks." For exam ple, during the 1906 construction ofthe Catskill Aqueduet
(for tbe New York City water supply), frequent oeeurren es of unloading phe­
nomena were observed in tunnels in granite gneiss and the Shawa ngun k Con­
glomerate. When a deep shaft was put down in the Storm King granite on !he
west side of the Hudson River, the popping action was so pronou nced that walls
had to be repeatedly scaled to remove loosening spal1s which caused dange r to
workmen below (Sanbom , 1950).
Rack bursts are da ngerous manifestations of unloading. The walls of de p
mines, pru1icularly in q uartz-rich rock su h as gr nite or quartzíte, ca n Li terally
explade, kil1 ing the miners. In the quartzite host rack areas of the lead-silver
mines in Coeur d'Alene, ID microseismographs are installed to monitor for pre­
cursors of da maging rack bursts. Rock cores from deep drilli ng areas have blown
up due to rapid unloading. "Pressure ridges" are areas of raek in q uarries whieh
suddenly pop up on floor exeavatians. In 1981 , a magnitude 2. 5 earthq ua ke was
generated by displacement along a 0.6-km fau lt in a California diatom ite quarry.
Tbe fault on the quarry floor showed a maximum displacemen t of 23 cm dip-slip
and 23 cm strike-slip. The rupture is believed due 10 unloading following the
excavation of 44 m of diatomite (Yerkes et al., 1983).
Sorne shales exhibit a form of unloading. Shale has the ca pacity for rebound
during an excavation or due to natural erosional processes. For example. the
removal of overlyi ng glacial ice 00 the Bearpaw Formation (a Cretaceous shale)
in western Ca nada is believed to be r sponsible for the unloading and attendant
gradual weakening of t be rock. resulting in land lides along the oll th Saskatch­
ewan River (Peterson, 1958). Erosional unloading may ca use anomalous fl uid
pressure in shale (Neuzil and Pollack, ! 983).

Salts. The combinatton of maisture and salts (halite, gyp um, etc.) has beeo
observed to cause decay of building stones (Winkler, 1978). Tbe gr wth of salt
crystals can cause stresses in the rock wbich physically break it apart. Thermal
34 Chapter 3 Weathering and Soil-Forming Processcs

ex pansion of salt crystals which bave grown in former voids roay also work to
disintegrate rock.

M iscellaneous physical weathering processes. Rocks expand when lhey are

heated . In lhe desert sun, rocks reach temperat ures of 60 e (Kerr e l al., 1984).
0

Typical coefficients of therma\ expansion are 10 6 m/m per oc.Temperature

changes which result in sma!l expansion and coo tractioo of mineral grains were
long thought to be a major cau se of the disi ntegrat ion of rocks. This theory prob­
ably has little credence, however, because Griggs ( 1936) experimental1y heated
and cooIcd rocks an d found no disi ntegrat ion. He exposed dry granite to the
equivalent of 244 yr of diurna! temperature fluctuation between OT and 61 °C.
but found visible cracks formed ooly wb en water was added. Wb ile a campfire or
forest fire may pro vide sufficient local heat a nd stress for the splitting of rocks,
most geologists believc that daily and seasonal te mperature changes are not of
s ufficient magnitude to weather rocks, even 011 descrt outcrops (Blackwel der,
1933).
On the moon, temperature changes fro m + 130 e during 1unar day lO - 150 e
0 0

during the lunar nigh t have been observed. During a lunar ecl ipse, changcs of
more than l40 e in less than 1 hr have been determ ined. These tem perature
0

changes probably crea te sufficient stress lO brea k up lunar rocks. Wea thering by
temperature-induced stress, aJong with m icro meteorite impact and solar radia­
tion, helps form the ubiq uitous ma nlle of lunar dust.
Other physical weathering processes, such as colloid p1 ucking by soil coLloid
form ation (Reiche, 1950), evaporation of salt water acco mpanying wave splash
(Mustoe, 1982), and organic forces such as tree roots, lichcns, etc. , also exert
mechanical forces which he1p weatber rocks. Prim itive plants such as llchens may
also cause weathering of rocks by form ing chelate compounds wh ich re move sil­
ica and metal ions.

3. 2.2 Chemical Wealhering

Chemical weathering reactions are exotbermic and produce mineral s of increased
volume. The reactions cause a decomposition (chemical change within the m in­
eraIs constituting the rack), and the expansion results in the physical disintegra­
ti on or break up of lhe rack. Dissolved mineral ions (sol utes) found in ground
water are largeIy the resu lt of chemical wealhering of rock through which water
has passed.
As a sim plification, let us cODsider the following fo ur ki nds of chemical weat h­
ering reactions:

Hydration. Hydration is the process whereby a mineral combines with water

to form a hydrated mi neral. An exampIe is the hydration of anh yd rite to gypsum :
CaS04 + 2H 20 = CaS04· 2H20. (3.1)
anhyd rite gypsum

Hydrolysis. Hydrolysis is the reachon of a m ineral with water to prod uce a

new mineral or minerals. An exampIe is the weatheri ng of feldspa r by reacting
 3.2 Wcalhcring 3S

Figure 3.3 Exfolialio n of basal! and formation of grus, near Weiser, ID.

with water to form day and sorne sal utes. For example, consider the hyd rolysis
of orthoclase:
4KAISi J O g + 22 H 20 = AI 4 Si4 0 lO(OHh + 4K-t- + 90H - + 8H 4Si0 4 • (3 .2)
m icrocline kaolinlle
Because ofthe expansion of feldspa úlic minerals undergoing hydrolysis, the rock
increases in volume. T he gra nular decay of coarse-grained igneous rocks such as
granite is largely attributed to the hydrolysis of the feldspars (Leopold et aL,
1964). Other grani tic minerals such as biotite al50 contribute to weathering in
granite.
Equatíon (3.2) and other representative hydroly is reactions are oversimplified
because these reactions undoubtedly require several sequential steps before th e
prod ucts are formed. Also, they are very s)ow processes and may require years 10
reach eq uilibriu m . Ma ny reacti on s have been studied in the laboratory but almost
always al elevated tem peratures, circa 20De C (Faust and Aly, 1981).
Because of di lational stresses due to chel1Úc I weathering, many massive
igneous rocks tend 10 deve)op spherical sheU and sp heroidal boulders. T his pro­
cess is called exfoliation . The spheroidal bou lders may be buried in the decayed
prod ucts which in clude shell fragments as well as disaggregated particles called
grus (Flg. 3.3).

Solution. Limestone (CaCO) dissolves due to its reaction with perco lating
36 Chapter 3 Wcatheri ng and Soil-Forming Processes

water which con lains dissolved carbon dioxide. Tbe reaetion ean be wriuen sim­
plyas:

H:O + CO2 + CaCO] =: Ca + > + 2HC0 3 • (3.3)

Beeause earbonie acid, biearbona te, a nd carbonate can exist in water, dependjng
on pH (Freeze and Cherry, 1979), it is often more meani ngful to express the solu­
lÍon or preeipitation of caJcite in lerm s of three reversible rcaetions:

H lO + CO2 = HC0 3 + H+ = H 2CO], (3.4)

B CO] = H~ + CO] , (3.5)
CaC0 3 + H' = Ca -+ + HCOj. (3.6)

Meteoric water (rain and snow) picks up earboo dioxide in the atmosphere and
in the soil zone. Tbe atmospberie CO, content o f 0.03% is increased up to 3% in
the soil zone due to root respiration ~nd orga nic decay. T he hyd rogen ions thus
prod uced (Eqs . (3.4) and (3.5» dissolve the lim estone (Eq. (3.6» .
Whe n loss of dissolved COl occurs, such as near a spring or pu mped well, loss
oft he gas ca uses the equations to work in reverse, and calcium carbonate precip­
ita tes, ca using incrustatioD around the spring or on well screens (Baron, 1982).
The effectiv eness of chemicaJ weathcring of Ii mestone and its relationship to
ground wa ter quality has beeo stud ied in t he high plaiDs (Back et al. , 1983) and
in F lorida and the Yucatan Pennisula of Mex ico (Back a nd Hanshaw, 19 70). Jt
was fou nd that, in general, ground water is undersat urated with respect to calcite
in the recharge areas and progressively reaehes eq uilibri um a nd supersaturation
downgradien t.
Chemieal weathering by soJution is responsible for the rapid weatbering O
Iimestone and marble dimension ston es a nd tombstones in hu m id regions. Com­
pare, fo r ex am ple, the rates of weathering of the granite and m arble tombstones
hown in Figs. 3.4 and 3.5. Jen ni ngs (1983) reports tbat Ii mestone tombstones
weather al a rate of at least 5 mm/lOO yr in England; a nd, based on weathering
of glaciated bedrock surfaces, lirnestone ou tcrops have weathered at a rate of
about 4 m m/lOO yr over the past 12,000 yr.
The rapid deterioration of buildi ng stones in urba n areas has been linked to
the inerease in pollutants, partieuJarly "acid rain" assoeiated with tbe release of
su lfurie aeid fro rn coal-fired power pla nts (Win kler, 1966). The rate of chemical
weat hering depends 00 lhe bydrogen ion concentration (pH ) as well as moisture
a vailabi li ty. Delicately carved faces on the ma rble statues on the Areh of Con­
stantine, Lhe offieial eotrance to Rome bui lt in 31 5 A. D., ha ve been almost com­
pletely destroyed in the past 20 years due to automobile exhausts. Many Italian
rn onumen ts a re now hidden bebind cloth cages to help protect them.

Oxidation. Oxidation involves th e ionie combination of an elernen t with

oxygen. By defin ition, a substance is oxidized when it loses electrons. For iron,
the reaetion is:

Fe+2 - e = Fe+3 (3 .7)

Mi neralogi cal exarnples of iron oxidation include lbe weathering of pyrite to

 37

Figure 3.4 Weathering of 110­

ycar-old marble tom bstone,

Connecticut. The letteri ng is

almost obliterated.

Figure 3. 5 Graoite tomestone from Connecticut cemetery shows no wcathc ring despitc
exposure for 113 years in an area hav ing 11 4 cm/ year precipitation.
38 Chapter 3 Weatheri ng and Soil-Forming Processcs

limonite, or the weatbering of siderite to hematite:

4FeCO) + O 2 + 4H 20 = 2Fe20 J + 4H 2CO). (3.8)
Siderite Hematite

3. 2.3 Rales oI Weathering

M inerals that crystallize at high tem peralures are generally the easiest to weather
(Goldich, 19 38). Bowen's reacfÍon series (Bowen, 1922) shows the general order
of crystallization of minerals in igneo us rocks as a m agma cools:
H ottcr (- lOOO°C) O li vine
Pyroxene
Horn blende Ca Plagiocl ase
Biotite
O rthoclase Na P lagioclase
M uscovite
Cooler ( - 300°C) Q uartz
From the very nature ofweathering, it follows that the mi nerals which have high
crystallizalion temperatures are fa rthest [rom equilibrium with atmospheric con­
diti ons, and hence a re the first to weather. Q uartz, with a low crystal lization tem­
perature, is virtually im mune from chem ical weathering; th is, and lhe abu ndancc
of silicon and oxygen , explain why quartz is the mosl common mi nera l in rocks
exposed al the earth 's surface.
Climale controls lhe rate of chemical reactÍons. T h us, in a hum id area li me­
stone wealhers rapidly and forms a vaJley, bu t in arid regio ns limestone is a resis­
tant rock and in many desert locales fo rm s the cap-rock fo r buttcs, mesas, and
hogbacks. Ancient li mestone builcüngs in th e a rid middle cast are remarkably
unweathered.
A general ranking of the rates of chem ical weatheri ng of common rocks is given
in Table 3.1. (Freeze an d Cherry (19 79) present a quantitative anal ysis of the sol­
ubility of individual minerals based o n the "equilibri um constan!" (K"Q ) for
reversible reactions.) Table 3. 1 gives an indication, at least in an engi neeri ng
sense, why foundations should not be bui lt on certai n rocks such as gypsu m . Gyp­
sum is nol only weak in terms of structural strength, but it dissolves readily. Th is
problem is compounded by houses where roof drains empty near tbe house,
allo wing for rapid solution and differen tial settlement.
Dam construclion problems associated with soluble founda tions are discussed
by Sherard et al. (1963). A noteworthy example is the Buena Vista dam in Cal i­
fornia , where considerable seepage thro ugh the foundation had been detected for
many years. In 1938 a chemicaJ analysjs of reserv oir water was com pared to seep­
age water, and it was estimated that seepage was responsible for dissol ving as
much as 3 m 3 of solids (mainly gypsum ) fro m the foundation each day .
G ranite is a particularly resistant rock, and bui lding stones show virt ually no
evi dence ofweathering even in moist clim ates (see Fig. 3.5). Rodgers and Holland
(1979) report only min ute cha nges of orth oclase to ka olinite along hai rl ine cracks
in Wisconsinan glacial bouJders (-10 ,000 yr old). However, older (Ill inoian
-100,000 yr) glacial gra nite boulders are decomposed.
 3.2 Weathcring 39

Table 3. 1 Ralc of Weatheri ng of Common Rocks

Catcgory Rock Rate

Gypsum (halite) Soluble, unwcathcrcd or as

alcite (arragonite, dolomite) ../ sccondary d eposi ts
P rimary minerals OJivine-hornblende (djo pside)
Biotite (chlorite, glauconi tc )
J' Easily and rapidl y wcathcrcd
{
Albitc (microcline, a n nhite )

Quartz }
I1lite (muscovitc)
Slowly weathcred
Hydrous mica intermediates
Montmorillonite
Secondary mincrals
Kaolimte }
G ibbslte
Wcathcred extremely slowly
HematIlc (gocthIlc, iLmonite)
Anatase (rutIle, llmen ite, corundum)
Adaptcd from Lcopold et al. 1964.

Birkland et al. ( 1979) diseuss tile use of weathering rinds and degree of grus­
sifieation of gra nite stones as a mea ns ofrelalive dating ofglaeial deposits. Birke­
land ( 1982) noles that weathering rind tluckness, comer angularity, surfaee oxi­
dalion, a nd pi tti ng of c1asts in Holoc ne glacial deposi ts allow for the subdivision
of five d ifferent ages of glacial and/or rock-glacier deposits.
Weatb ring rea tions are also important for understanding lhe properties of
dimensÍon stone. Stone has long been one of man's primary means of construe­
lio n. The Roma ns built enormous aq ueducts a nd eoliseums with slo ne. Today
stone quarri es are a major min ing industry. Modem buildings rarely use stone as
structural members as in the past. Rather, d imension stone is used as an orna­
mental fac ing on outside waIls.
Tombstones provide vivid evidence oftlle rates and kinds ofwealh ring whieh
oecur in d ifferent roeks (Figs. 3.4 and 3.5). Even in humid areas such as ew
England, lhe rates of weatbering are slow. The diffi rent weathering rales of var­
ious tomb tone lithologies are relatcd to the general topography of New England
in Ihat rocks !hat weather fast fo rm low topographíc areas (Rahll, 197Ia).
PracticalJy speaking, most comm n rocks used in cngineering struetures, eit her
as d imensio n (faciog) stooe or as a Lructural wa lI , wilI retain their integrity over
the Life of the structure. For exam ple, exposed para pet capings of J urassie-aged
Portla nd Limestone on St. Pa ul's Cathedra l in London were fou nd to have weath­
ered only 13 m m in 250 years (Schaffer, 1933).
The importa nce of wea theri ng 10 engineeri ng geo logy lies mainJy in the rec­
ogni tion of soi l/bedroek conlacts and properties of sound rock, weathered rock
or so il. Because almo st al! engineering excava tiolls are near the surface, lhey
involve rocks weathered to s me degree.

3. 2.4 Concrete Aggregale

Another enginccri ng application of rock weath ring relates to the use of rock as
aggregate fo r concrete. For exa mpl crushed li mestone was fo und to be inferi or
40 Chapter 3 Weathering and Soil-Forming Processes

to quartzite as concrete aggregate for interstate highways, not only beca use of the
rel ative hard ness of quartzite, bul because its lower porosity precl uded water
entrance in to the aggregate, and bence freezi ng and breaking up ofthe rock (lyer
et al., 1975).
Concrete is com posed of coarse aggregate, fi ne aggregate, cernent, and water.
The ratio of the above (by volume) is approx imately 4: 2: 1: l . Wben hardened, me
cement (normally Portland variety) paste ineludes hydraÜon products and a min­
ute pore system. Hardened concrete has sufficiently low permeability and absorp­
tivity to prevent rigid ingress and destructive agents, th e most important ofwhich
are water, air, acid s, an d sulfates. The increase in hydrostatic pressure within the
cement or the particles of aggregate as a result of the crystalluation of ice a nd
consequ ent increase in volume occupied by ice and water is noted to be particu­
larly destructive. Highly porous aggregate, sucb as sorne limestones. cause
popouts and surface sealing.
The reaction of alkalies (sodium and potassium) with certain forms of silica
can ca use the expansion and disintegration of concrete (McCon nell et al., 1950;
Erlin, 1969). Aggregate containing concentrations of about 0.5% alkalies with
chalcedony, opaline chert, rhyolite, or cristobalite should be considered suspect.
Chert is a hard rock, but chert and other cryptocrystall ine silicates are sorne of
tbe worst reactors in concrete. The hardening of concrete (a hydrati on reaction)
prod uces heat and releases alkalies, particularly the hydroxides of calciu m,
sodium, and potassiurn. Alkalies can be derived from 1) release from aggregate
particles sucb as water-solu ble salts of soruum or potassiu m, includin g zeoli tes or
clays, 2) alkaline mixing water such as sea water, or 3) alkaline water s that pen­
etrate the concrete after hardening (Mielenz, 1962). T hese al kalies react with cer­
tain cryptocrystalline silicates to form silica gels which absorb water from lhe
co ncrete paste and exert osmolic pressure, creating tensio nal cracks in the hard­
eni ng concrete. The small cracks can act as aven ues of subseq uent water pcnetra­
tion and destruction of concrete by freeze-tbaw. Chert is a com mon componen !
in strea m gravel, and presents a serious problem which can, fort unatcly. be
reduced by lhe use of low alkal i cement o
Ordi na ry Portland cement can be chemically attacked by graund water con­
taini ng high sulfate. In addition, fragments of shale fou nd in coarse aggregate may
disintegra te in time. Shale fragments can be remo ved by separa ti on prior to use
as aggregate in a high specific gravity liquido

3.2. 5 Weathering and Landscape Reduction

Over geologic time, the amo unt of weathering and subsequent removal of weath­
ered products (erosion) in a drainage basin are equal. The two processes can be
considered to be in "dynamic equilibriurn," and the earlh's surface is reduced
accordingly.
Chemical weatheri ng is most important in moist climates and in areas under­
la jn by soluble rock (J udson a nd Rüter, 1964). Because limestone, dolom ite, or
evapori tes dissolve, leaving essentia lly no residual products, the rate of weather­
ing and landsca pe reduction can be deterrnined by measuring the rale of remo val
of dissolved co nstiluents in sunace or ground water. For example, the total dis­
solved solids (TDS) of large springs draining a Paleozoic limestone terrain (a ver­
 3.2 Weat hering 41

.... 500
>.
--.... Col or ado
N 400
E
.><.
--.... 300
<f>

-
"O
e
o
200
Gulf
o
O
-.J 100 N. Atlantic S. Atl anl
t · c an d
"O
O Colum io East Gu lf
O
(J)
O 10 20 30 40 50 60 70
Dissolved Lo ad (tons/km 2 /yr.

Figure 3.6 Relatíonship of dissolved and solid load in streams in various parts of the
United States (after Judson and Ritter, 1964).

age an nual precipitation = 43 cmjyr) eireUng the Black Hills in South Dakota
ranges from 150 to 2280 mgj l. The solutes are maioly Ca +.J., Mg++ , and HCO.l.
Based 0 0 tbe product of discharge times the TDS of eac h of 23 springs a total of
143 X lOó kgjyr of rock is weat hered and removed via springs drai o ing into trib­
utaries of the Cheyenne River. Assurning a rock specific gravity of 2.7 gj ec, the
volume of limestone rem oved is:
6
143 X 10 kgjyr = 53 X 103 ro] j r.
2.7 X 103 kgj m 3 y
Because the area oflimeston e exposed and drained by these spri ngs is 3900 krn2,
then the rate of landscape reduction by cheruical weathering is:
53 X 103 m .1 yr
39 09 2 = 14 X 10- 6 ml yr = 14 mmj lOOO yr.
. X I m
Th is chem ical erosion rate is sim ilar to olher measurements reported by Walli ng
(19 77) and Jennings (1983). The total rate oC landscape reduction o f Black Hi lls
limestone terrain wou ld actually be somewhat larger than th is, becaus occasion­
ally a fla sh flood rem oves so rne clastie debris (Ii mestone fragments and residual
so i1 formed from clay and sand particles in the limestone). This debri s constit utes
the suspended and bedload of floods in the ephemeral streams draini ng the lime­
stone area. Due to the karst topography, however these floods are very rare, and
are probably not as effective in reducíng th e Iimesto ne landscape as chemical
weathering.
Figure 3.6 illustrates the amount of dissolved and solid (det ri tal) load in U .S.
strea ms. Note that tbe Colorado River has an eno rmous sol id load, deri ved from
overland ru noff and stream erosion in semia rid areas conta ining poorl vegetated
slopes. In co ntrast, trea ms in the southeastern U.S. derive a much greater pereent
of their total load from dissol ed constituents. U e ofthese data to calcula te long­
term landscape evolution may be suspect becaus oft he increased erosion accom­
pa nying man' s act ivities (Douglas, 1967).
Ma n rocks, such as granite, do not dissolve, but che m icall y cha nge to fo rm a
resid ual soil upon weathering. Bartb (1961 ) proposed a for m ula for the amount
42 Chapter 3 Wealhering and Soil-Forming Processes

of rock weathering if weathering produces insoluble products as well as solutes:

W=~
C, - SI
, (3.9)

where:
W = amou ot of rock weathered,
DI = amount of an element i removed in solution,
CI = fractio nal concentration of ¡ in lhe original rocte,
SI = fracti onal concentration of i in the residual soil.
For a hilly graniti c terrain in Rhodesia (average an nual precipitation = 122 cm),
Owens an d Watson (19 79) used calciurn a nal yses in streams to determine the rate
of granite weathering:
2.0 kg/ba/ l OOO yr

W = 0.0005 _ 0.0000 = 400 kg/ha/ l 000 yr.

This corresponds to a depth of chemical weathering (and by inference landscape

reduction) of about 15 mm/ IODO yr. Because granite minerals such as quartz are
insoluble, lhe 15 mm of cbemical weathering was cstimated to produce a residual
soil of 11 m m thickness in 1000 yr. Presumably this residual soil is removed by
stream erosion (suspended and bedload) in order 10 mai ntain a dynamic equilib­
rium of weathering and erosiono
Gibbs ( 1967) studied the salinity of Ihe Amazon Rivcr, lhe world's largest
ri ver, which contributes 18% ofthe world's total runo tfreaching the occan. Ofilie
dissolved salts reaehing the oecan (36 rng/ I), approximately 86% are su pplied by
only 12% oflhe total area ofthe Amazon Basin wh ich ls the mou ntainous region
of the Andes. The reason for trus is tha t ehemical weathering proceeds rapid ly in
the mountain ous areas where fresh rocks are exposed by crosion, whereas in the
lowlands an enormous tb ickness of saprolite (mostly gibbsite and kaol init.e min­
eraIs which are weatbering end products) effectively isolates unweathered bedrock
from the vast amou nts of rain that cha racterizes the ju ngle area.
General surface lowering of the earth's landscape at 20-80 m m/! 000 yr for
gentle to moderate slopes have been reported by You ng ( 1969). Judson (1 968)
reports that, based on average sediment loads fo r rivers in the U nited States, the
landscape is being eroded at arate of 60 mm /IODO yr. l n a study of 70 rivers in
the United States, approximately 20% of the total measured load is carried in
solution (Leopold et al., 1964). Trus percen tage ra nges from l % in tbe semiarid
basin of the Little Colorado River in Arizona to 64% in the humid basi n of tbe
Juniata River in Pcnnsylvania. The worldwide ratio of mechanical to chemical
denudation is about 5: 1 (Birkeland, j 984). (Additio nal information of Ou via l ero­
sion and sedime ntatjon is presented in Chapter 8.)

3.3 Soil and Bedrock Contacts

T he origin and properties of soiJ are technically com plex subjects. The word
"soil" is typically used by engineers to indlcate fra gmented material below cobble
size that can be excavated without blasting. Bed rock is thc couo terpart. The top­
soil of agricultural soil scientists (pedologists) is more restricted, and ineludes soil
 3.3 Soil and Bedrock Contacts 43

horizon development near the surface which is capable of supporting plan t life.
The Glossary of Geology (American Geological Institute, 1972) defines bedrock
as: "A ge neral term fo r the rock, u uaUy solid, that underlies soil or other u ncon ­
solidated surticial material. ' Soil (somctimes 100sely refen-ed to as saprol ite, rego­
lith, or mantle) is defined as "unconsoLidated material aboye the bedrock...."
The defi nition of soil and bedrock is an excrcise in semantics; the definition
largely depends on a person's area of interest. To a civil engineer, soil is simply
un onsolidated material which typically disintegrates in water. Agricu ltural engi­
neers and most geologists consider soil to be the weathered product of rock or
surfi cial deposits contain ing vary ing proportions of organic material. A civil engi­
neer could look at a IOO-m-th ick exposure of glacial till and call the entire ex po­
sure "soil,' whereas a pedologist may recognize a soil (or soil "profile") in only
the top I m of the til!. To most geologists, soil is simply the product of in situ
weathering processes. lt is regrettable that semantical difficulties such as these
exist. The problem is related to old scienl;llc translations from German and R us­
sian. In r cent years soil terminology problems have been further compou nded
by the introd uction of numerous cumbersome terms by American pedologists
(Soil Survey Staff 1975). [Olson ( 1981), Birkeland (1984) , and Schlemon (1985)
have practical explanations of thesc terms.]
What constitutes " fi rm bedrock" is also a moot point. To a geologist familiar
with granite terrain, the Pierre Shale (a Cretaceo us, semi-consolidated shale in
the mi dwest) may not appear to be bedrock. But midwestern geologists refer to it
as bedrock.
Hansmire (1 981 , p. 78) proposes the fo ll owing useful quantitative c1assification
for the definition of soil and rock:

For very soft soils an uncon fined compression str ngth of 100 psf(4800 N / m l ) would
be typical for a frcsh a ll u ¡al c\ay deposit t hat wo uld not support a person stand ing
o n it. Líghtly loaded building can be supported by hallow foundatíons on a
medi um to tiff c\ay with desígn loadings of 2000 psf (96,000 N/m2) and would be
di fficult 10 cut with a kn ife. An u ncontlned com pressive st rength ofabout 20,000 psf
(960,000 N/ m ) is norma lly the division betweeo soil and rac k. The transition from
soft t higher strength roc k is not as well cstablished and ma y range from 200,000
psf (9,600,000 N/ ml) to 500,000 psf (24.000,000 N / m ' ).

Other met hods of classification of soil and rocks inelude weatherability, hard­
ness, durability and jolot spacing (Goodman 1976).

3.3.1 Residual Soils

Soils and surticia! deposits (referred to as soils by civil engi neers) can be either
residual or transported in origin (Fig. 3. 7).
Residua l soils develop in sit u, and thei r characteristics depend on the kind of
bedrock from which they are derived. Examples of resid ua l soils are gi en below.

Granite (and other massive igneous rocks). T he predominant physical weath­

ering processes which act on granite rocks include unloading and frost action.
Chem ical weathering causes a decomposition of the rock mass. Wahrhaft ig (1965)
and K.rank and Watters (1983) believe that the maj r process of weathering of
44 Cha ptcf 3 Weathering and Soil-Forming Processes

Residual Soils
Massive Igneous Rocks
Arid Climate
Humid Climate
Spheroidal boulder

~--¡~!/
r, ­ ", ' . '. , " " ~ , Corestone
Claysoil ' . ~ .', . '~' , ~
r: . J ,. . ~ ~~'. " ~ .-. - .
o "·

Exfoliating~
shells ..

Limestone

Sharp_
Contact

Sandstone and Shale

Poor soil development
Weathered shale o

ltI.

Figure 3,7 Idealized sketch es showing (A) resid ual and (B) transported soils,

Sierra Nevada granodíori tes is the expansion of bioti te in contact with ground
water, Tbe expansion of bio tite produces microfractures which progressively
break down the original rock to a saprolite.
Figure 3. 7A contains a sketch showing tbe appearance of igneous rock out­
crops. Tbree sta tes of weathering may occur: sound rock, exfoliation shells, and
grus. Grus is disin tegrated granite; the size of tbe fragments is typi cally coarse
 3.3 Soil and Bedrock Con laCIS 45

Transportad Seils
Alluvium

Terrace

Glacial Dril!
(1) TiJl

¡. .. 1 .. il -
Paleasel • ...::. -':;-. ;-. :0 , ":.... _ - ,
4 . ' - ' '.......,--r;:­
"l ' :;--' - 'o-' -, .
Sharp . .. - " --,. .~
contac! ·....-" - ~ -===---~
Nar!heas! USA ---- -----------
-- USA
-Midwes!
(2) S!ra!ilied Dril!

sand about t he same size as the crystals in the granite (Rah n, 1966). Hydrat ion
a nd bydrolysis are largely r sponsible fo r the weathering and res ulting di lation of
the sound gra nite altho ugh the actual change in chemical eleme nts bctween the
thr sta tes is small (Larsen, 1948),
hemical weathering is slower in arid cli ma tes than in humid cli m ates; thus
the depth of "rotti ng" is less. Due to ¡ntense thunderstorms cha racteristic of
46 Chapter 3 Weathering and Soil-Form ing Processes

deserts, grus tends to get washed away. causing the spheroidal boulders to pro­
trude aboye th e land surface. In hu m id climates, the depth of chemical weatber­
ing is great, and grus can be 100 m tlú ck or more. Ha theway (1 980) reports granite
weathered to 20-m deptbs in tropical Singapore. Ruxton and Berry (1957)
describe engineering problems associated with excavations in the decomposed
granite of Ho ng Kong.
Weathering of crystall ine rocks in tbe Piedmon t area of the eastern U nited
States has resulted in a saprolite averaging about 30 m thick in upland areas. The
saprolite is decomposed rock cbaracterized by the retention of parent rock struc­
tures and textures, but primary minerals are mostl y replaced by clays and iron
oxides. These saproli tes can be excavated by a shovel. In the Piedmont, most
water wells are located in saprolite because wells in the underlying bedrock are
not very productive.
The main engineering problem in resid ual soils formed on igneous rock is the
identi fication of a sou nd horizon on which to seat a fou ndatio n. Due to the gra­
dational nature, there is no depth al which one can uneq uivocally say that soil
ends and bed rock begi ns. Tbe presence of indivi d ual boulders of sound rock (core
stones) in residual soils presents special proble ms in fou ndation engineering. Tbc
problem is acute where the residual soil possesses a collapsible grain structure.
but the core stones are too large to be excavated by earth-moving eq uipment or
cannot be penetrated by an a ugered pile hole (Brink, 1979). Another problem is
the mislead ing identifi ca tion of sound bedrock in an exca vatio n where tbe top of
an isolated core stone is exhumed.

Limestone. In contrast lO granite, the soil-bed rock conlact in carbonate (li me­
stone or dolomite) terrains is typically qu ite sharp. Figure 3.8 shows tbe sharp
contact exposed in the waU of a limesto ne quarry near Philadelph ia, P A. The
thickness of soil in a carbonate terrain varies considearbly (-0- 100 m) due to
the formation of caves and sinkholes. Aft er a sinkhole forms, slope wash will
gradually fill in the sin khole causing a t hick soil of small areal extent.
Theoretically, beca use limestone wea th ers by soJ ution, there are no weathered
products; the entire rack is carried away by ground water. In reality, however,
because of small beds of clay, cbert, and sand, wh ich are usually present in li me­
stone, a residuum is left aft er solution of the carbonate rock. T he red, clayey soil
typically derived from limestone is called terra rossa. Over geologic time the
thickness ofthe residual soil is maintained by the conti nued chemical weathering
of the bed rock below an d the stripping of tbe surface by erosiono
The irregular nature of a limesto ne/ soil conlact was drama tically evident in
the excavations for the foundations ofthe Ten nessee Valley Authori ty dams. Ver­
tical solution channels over 60 m deep were discovered in Mississippian lime­
stone. Although filled with terra rossa and cave-fill deposits, tbe solution zones
leaked enormous amou nts ofwater from the reservoirs. Following lbe appearance
of downstream springs and "sand boiJs," wh ich threatened the concrete dams,
grouti ng programs were undertakeo. Al Hales Bar Dam , fo r example, from 19 19
to 1921 an attempt was made to stop leakage by pum ping hol asphalt in lo the
cavities functioning as conduits. Sixty-eight boles were drilled and over 220 m 3
o[ asphalt was pumped into them (Burwell aod Moneymaker. 1950). T be Ha les
Bar Da m project, estimated at $3 mili ion and scheduled for com pletion in 2 yr,
cost $ 12 million and req uired 8 yr for eom pletion. In 1940 the entire foundat ion
bad to be grautcd agai n d ue to conti nued enorm ous leakage.
 3.3 Soil and Bedrock Contacts 47

Figure 3.8 Sharp soil-bedrock

conlact on limeslone al Con­
shohockon, PA. The terra rossa
soi l is O-10m lhick. T he water
table in lhis abandoncd quarry
is al about 50 m in deplh .

Differential settlement is com mon in thick residual soils overlying weathered

carbonate rocks. Residual soil over solution cavities may be quite thick; building
foundations or other ma n-m ade struClures located in the residual soil should be
rigorously analyzed (Lifrieri and Raghu, 1982). A similar problem occurs where
a "floating bould r (a core sto ne) is encountered and is assumed to be part of
the bedrock. Even greater hazards occur in areas containing solution cavities at
depth. Sowers ( 1975) att ributes the 1969 bridge collapse in Tarpon Springs, FL,
to failure of a roof of a cavern. The settlement of a junior college complex near
Miami, FL, was believed due to the crushing of highly weathered porous
li mestone.
Sudden siokhoJe coLlapse is a geologic bazard. Below the residual soil, solution
selectively dissolves layers of limestone along joints, bedd ing planes, fossiliferous
layers, ete., until a cave develops. The sudden coUapse of the roof of the cave
results in a si nkhole. This ge logic process happens naturally, or can be acceler­
ated by ma n's activities sucb as ground water withdrawal or the construction of
parking lots which all ow for abnormal discharg of rain water into the ground
(Parizek, 197 1a). (See Chapter 9 for further discussion of sinkholes.)
The sudden collapse of a sink.hole can cause more than lhe inconvenience of
disrupted farmJand, fo rest, or a broken highway. Si nkhole collapse resulted in the
rupture of a gas majn in Allentown , PA. whjch fi lled caves with gas, threatening
to ignite. In the same regíon, near Saucon Vallcy PA, sinkholes ha ve developed
due to the lowering of the water table due to a deep z.inc mine (Fig. 3. 9).
 48 Chapter 3 Wealhering and Soil-Forming Processes

Figure 3.9 Sinkhole at Saucon Valley, PA, developed due to water table decline
associated with underground zinc mining.

In an arid to scmiarid clima te, limestone typically form s a resistant unit, form­
ing an obstruction to construction of pipel ines, sewage systems, etc. One of the
more sobering reminders of the engineering significance of soil-bedrock contacts
to m ilitary com bat is offered by Brooks (1920, p. 87) in his discussion of the
tre nch warfare in the Western Front d uring World War 1:

During the great battlc ofVerd un a body of lroops was ordered lo inlrenc h itself on
the high plateau of the Cotes de Meuse. Even a casual exam ination of the geologic
map would have shown lhat the platea u was underlain by hard limestone with less
tha n a root ofsoi!. This material could nol be excavaled with the light tools furnished
or even with proper eq uiprnenl in tbe time a vailable. As a consequence there was a
large a nd needless loss of life.

Shale- expansive soils. Very little residual soil develops over sandstone or
shale because th e bulk of the minerals contained in the rocks (quartz, clay min­
erais) are relatively imm une from weathering processes (Table 3. 1). Thick trans­
ported soils can accum ulate, however, as transported soils due to mass wasting
processes, e.g. , colluvi um as described below. White (1962) describes the weath­
ering of the Cretaceous Pierre Shale as first forming loose platey fragments and
tben forming a clay-rich 5011. Shrink.ing and swelli ng of the clay as it dries and
becomes wet causes mass movements, wh ich gives a smoothly rounded lopog­
raphy to the western South Dakota prairie.
CIay is a very fi ne-gra ined deposi t with plastic propert ies when moist, and
composed largely of hydrous alumjnum and magnesium silicates (Gillott, 1968).
Clay exhi bits considerable strength when air-d ried. (The term clay has been used
widely to designate the percentage of soil fragments finer lhan 0.002 m m (0.005

.;¡¡¡¡¡¡¡¡
 3.3 Soil and Bed rock Contacts 49

mm in sorne cases), but it is recom mended that this usage be di sconti nued since
from an engineering pojnt of view, its strength properties are m ore im portant
than its size per se.) Clays wb ich have resistance to deformation (low plasticity
and compressibility, typical of th kaolinite group) are called lean days. Less
desi rable engineeri ng properties such as highly plastic and compr ssible da 's
(typical of tbe mon tmorillo nite group) are called fat clays.
Shale is similar to clay, but it has a fissile or laminated structure parallel to the
bedding. Slate is well indurated, possessing cleavage. Weathering processes tend
to make shale and slate revert ba k to clay, so from an engineering point ofview,
weathered slate or shalc possess properties of day.
Engineeri ng problems associated with shale or clayey soil s are not due to t heir
soJubl nature, but to their low shear strengtbs, which make t hem haza rdous for
fou ndatio ns or for any construction particularly in landslide-prone areas. Shales
used in high way 611s often settle due to slaking of the shale pieces. U nfort unate
past experiences deati ng wit h tbe nonhomogeneous natur of shale and clay and
their poor strength and weath ering characteristics have led most engi neers to treat
sbales as problem material.
At the ti me of deposi tio n of subaqueous clay, the water content is ery large.
After burial the water con t nl decreases. The process of vo id rati o reductio n is
terrned comp ti o n by geologists and consolídation by engineers. The importance
ofsettlement oftbe land surface (especially differential settlemen t) due to loadj ng
of soils has led to detailed theoretical and empirical st ud ies. Not only is water
content important in the abitity of soils to sustain loads but it has been found
that sorne deposit may become "q uick" or completel flui d under rising ground
water cond itio ns, as in a seepage area. Other fine-gra in d soíls are thixolropÍc in
that they lose t heir strength when disturbed as by a v ibrat ing load. Laboratory
tests show that clay e res may support a considerable load in an undi stu rbed con­
d ition, but wben remolded (a t tbe same moisture content) behave like a flu id a nd
may be literally poured from a beaker (Gi11ot1., 1968). (See Section 5.5.4.)
The physical properties of shale vary considerably. Sorne shales have strengths
comparable to concrete, and as they havc low permeabili t y, make ideal dam sites.
Other shales are not very consoljdated, and have attained a degree of consolida­
ti on only sügbtly greater than clay soils. Low-grade shales 1hat have settlem ent
problems are also subject to rebound. If lhe superincu m beDt weight of materi als
removed in an excavation appre iablyexceed the weight ofthe man-made struc­
tUTe placed in the excavation, rebound may occur. Accordi ng to Burwell and
Moneymaker (1950), "The rebo und characteristics o f shales tike the Bear Paw,
Pi rre, and Fort Union of North Dakota and Soutb Dakola are su h that deep
spiLlway cuts may cause buckling of spiI1way Iinings and suffi cient diffi ren tial
rebound movements in the foundations of spilIway structu res to require special
design prov isions. For example, the Fort Peck Dam spillway, located o n lhe Bear
Paw shale, is anchored deeply in to its foundation with a series of 60-inch d iam­
eter reinforced steel dowel ."
Clayey soils re aJso hazardous due to expansion a nd contraction accompa­
nying soil moisture changes. U Dlike kaolinite clay minerals which fo rm from t he
weathering of ti ldspars, shale and cla ey soiIs typical1y inelude large percentages
of illile or montmorilIonite clay. The kaolinites fo rm v ry stabl clay because
their tight, ínexpandabIe crystal structure resists the in troduction of water into
their lattices and the conseq uent expansion or heavi ng when saturated. On the
50 Chapter 3 Weathering an d Soil-Forming Proccsses

Figure 3.10 Slaking of Cretaceous shale block whicb had been excavated I year earlier.

other hand , montmorillonite crystal sheets are bound rather loosely, allowing for
space water molecules to insert themselves between tbe sheets, causing expansion
or swelling (Tourtelot, 1974). When a saturated montmori llon ite dries out, it is
subject to shrinking or crack ing. In pure form, mo ntmori llonite c1ays may swell
to over J 5 times their dry voJu me, although most soils contain o nly small
am ounts ofmontmorillonite so that not many swell to more than 1.5 times t heir
dry voJu me (Jones and Holtz, 1973). Expansive soils cause d amage when they
shrink upon drying or when they expand upon wetting, and the resul ting soil
movem en ts disrupt houses, m ul tistory buildings, sidewalks, streets, and utilities
(Bui ldi ng R esearch Advisory Board, 1968). Leaking sewers or swimm ing pools,
improper storm drains, or poorly buil! bouses with in adeq uate foundatio ns
develop damage ranging fro01 sticking doors and ha irline plaster crack s to co m­
plete destruction.
The process of shrinking and swelling can result in expansive soil problems, or
in slaking, the crumbling of sou nd outcrops or rock specimens in to fl akes or gran­
ular parti cles (Fig. 3. 10). For cxamplc, because ofs laking of th e shale in the Mj o­
cene Castaic Formation, it was necessary 10 place earth till witl1in 24 hr of expo­
su re of fresh bedrock surface dunng the construction of l he Castaic dam in
Californ ia (Hanegan, 1973). Poorly cemented shales may be quickly red uced 10 a
pile of ch ips upon several cycles of wetting an d d rying, while weU-ce men ted
sha les are qui te resistant to such changcs. (Note: Sorne soi] enginecrs define slak­
ing as lhe disintegration of dry soi) when it is subrnerged in water, whereas other
soil engineers define slaking as lhe disintegration of soil under alternating wetting
and drying. )
 3.3 Soil and Bedrock Contacts 51

A slaking test has been devised by highway engineers to eval uate slaki ng p rob­
lems w hich are typical of calcareous shales. T he usual proced ure is to im m erse
rack sam ples in water and observe a n y di sintegration that occurs. Slaking dura­
bility depends 011 1) the a bility of flu ids to penetrate the rock, 2) the a bility of the
flu ids to cause d isruptio n by solution of cernent or disruption of bo nds or pre­
pressure relationships, and 3) the inability o f the rock skeleton to resist the d is­
ru ptive forces. Clay-bearing rocks and sandstones cemented by soluble minerals
are the most suscep tible to slake deterioration. Among the several mechanism s
that may account for slak ing of shale, ion exchange appears to be domi nant. Ca p­
illary effects, d esiccation u pon drying, and unloading probably also play an
im portan t ro le in slake deterioration .
Another example of soi1 problems was described by Baker (1975). U rban plan­
ning in the Bo ulder, ca, ar a is high ly dependent upon bedrock and surficial
deposits, a nd their derived so ils. Three ty pes of parent materia l exist: loess, a
terrace gravel a nd the Cretaceous Pierre ShaJe. The terrace gravel forms the best
engineering su bstra te beca u e it has good grain-ro-grain strength pro perties and
con tains few swell ing clays. Plastic clays are present in the Pierre Shale, however,
and swelljng a nd expa nsion exert d ifferential pressure on overlying foundations
which can b reak them up. Soils developed o n loess a re subject to solution of the
cal ite cemen t, ca using collapse of structures; the chain of events culminating in
loess fa ilure may be ini tiated by something a s innocuous as watering the lawn.
Expans ive soils rep resent one of the cost liest engineering geologic hazards,
exceeding fl oods, landslides, and earthquakes in terms of average yearly destruc­
tion in tbe Un ited States. Damage to family bo rnes cost homeowners about $300
mili ion ann u.all . Still, each year over 250,000 new homes are built in expansi ve
soil.
Expansive soils are fo und tbroughout the western U nited Stat s, pa rticularly
those d rived fro m Cretaceous shalcs. Sorne o f these shales contain volcanic clay
(bentonite) beds which are thick enough 10 be mined commercially as low perme­
11 abil ity sealan ts for d a m s or bouse basemen ts. Regions underlain by expansive
11 soi ls can be recognized in some prairie rangel and areas by su btle but distinctive
r topography called gilgai. Gilgai are networks of small downslope ridges and
grooves ha ving wavelengths on the order of 2- 3 m and amplitudes of about 0. 1
m. Gustavson ( 1975) used aerial p hotograph ic interpretation of gilgai on undis­
turbed p rairie lands as a means of mapp ing potcntial hazards from clay
expansiono
Expansive soils are capa ble of developing tremendous pressures when mois­
ture is introduced. Laboratory pressure devi es have recordcd l A X 10 6 N/m l in
montmorillo nlte shale soils from Dalias, TX ( ryn ine and J udd 1957). These
r forces are capable of buckling concrete floor slabs. W hen expansive shale d ries it
n- shrinks and can cause wall cracking or other forms of d istress in struct ures. The
wetting of expansive soils typically occurs every season when the frost lea ves the
grou nd and spri ng rains add a surge o f moisture.
n Moderately swell ing soils are present througho u t m uch of the D enver area.
a Swelli ng soi ls (typica lly pos essing liquid Jim its o f a bout 55% and plasticity
ind í es of about 30%) well 3-10% under normal loads of 500 N/m 2 (Costa and
BiJodeau, 1982). Swelling pressures can be as great as 15,000 N/ml. Structural
r dam age can occu r when swelling i ' as little as 1%. Bo th shrinking and swelling
ng can occur with moisture variation, and eit her can cause da mage to streets or
structures.
52 Chapler 3 Wcathering and Soil-Fomling Processes

AvaiIability of water io expansive soils is controlled by climatic cooditions;

bowevcr, surface draioage, amouot of vegetation, yard rnainleoance performed
by l he homeowner, and depth oftbe water table act lo modify Ihe moisture avail­
abiJity. Any of these factors wbich help to maintain a constan! moisture content
act 10 reduce the sbrink-swell activity. One way to reduce structural damage io
xpansi ve soiIs is to dig Ihe structural walIs or piers below the zone of scasonal
moisture change. Factors such as improper drainage, leaky swirnming pool s, elc.
that make large fl uctuatioos in the soil moisture conten t result io ao incrcased
shri nk-swelI activity. Harnmer aod Thompsoo (1966) found that large trees
planted near houses could cause structural failure. Becausc tree rOOls may cause
local soil moist ure deficiencies and shrinkage, it was recommended lhal trees
shouId be planted nO closer to the foundation than one-half of thei r mature
height. A study of expansive soil damage to hornes in Texas, Mathewson et al.
(1980) showed that yard maintenance and the truckness ofan active soil zone (soil
experiencing seaso nal moislure changes) were the mOSI sigoificant paramelers
affecting ho use dam age. In Rapid City, SD, houses most sigoifican tIy damaged
a re ones which have roof drain spouts ernptying near th e house fo undation;
houses are less affected where drains extend 3 m or more away from the house.

S andSlone. Sandstone typically weathers slowly in any climate and forms

highlands such as mesas, buttes, and hogbacks. For example, the silica-cemented
T uscurora Sandstone form s the most prominent ridges in the central Appalachian
Mou ntains. T here is virtually no residual soil on tbis formatioo, only large boul­
ders of sandstone form in g block fields or talus. Calcite-cemented sandstone, on
the other hand, weathers relatively quickly to sand, and in many places forms
commercial deposits of sand.
In general, sandstone 1S a good foundationa l material, altho ugh unweathered
sandstone may be a n impediment for foundation or service line excavation.

3.3.2 Transported Soils

Transported 50ils are surfi cial deposits which accum ulate due lo the erosion,
transportation, and deposition ofweathered residual soi l or bed rock. Four com­
mon surficial processes and their resulting deposits are described below (Fig.
3. 7B).

Colluvium. Collu vium results [rom the process of creep, whereby soi! and
weathered bedrock slowly move downslope due lO gravity. The contact with Ihe
underlying bedrock can be sharp or gradational (Fig. 3. 11). Colluv ium resembles
glacial till in appearance. Typically colluvium consists of unstratified, seemingl y
randomly oriented angular blocks ofbedrock io a clayey matrix. In geologic ti me,
colluvium is actively movi og. CoUuvium is very widespread and can reach 100
m thick in h umid areas. In unglaciated humid areas, practicall y the enti re land
surface is covered by a mantle of colluvium, ooI y interrupted by bedrock out­
crops, or other surficial deposits such as glacial drift or alluvium. Thick accu­
mulations of colluvium may have formed during past cl imatic regimes, such as
in periglacial areas which may have had permafrost conditions.
Engineering geologists should view coUuvium with suspect because any disturb­
ance co uld cause accelerated movement. Although in the life of a structure
 3.3 Soil and Bedrock Conlacts 53

Figure 3. 11 Colluv ium exposed aboye timestone bedrock in a slrea m cut near
Hermosa, SD.
red

(-1 00 yr) th d irect structural damage d ue to natural rat s of movement may be

small, ma n's activities such as excavations may accelerate creep, or ca us a land­
slide, thereby leading 10 building distress or failure. Radbruch-Hall ( 1978) cites
on, numerous exam ples of creep failure in the United States and Europe.
m­ Creep differs rom lh process of lan dslidi ng in that creep opera tes slowly.
Fig. Creep has more displacement near the surface of 1h , land, typically causing slow
rotation of telepbone poles, tombstones, walls, etc. (The curved tru nks of trees,
attributed to creep by sorne geologists, may in fact be d ue to site- pecifi c growing
nd conditions unrelated to creep (Prupps, 1974).) Landslides also di ffer from creep
he in tha t landslides occur as massive displacements of soU or rock alo ng a specifi c
es failure surface. (See Chapter 8 for a more detailed d íscussion of mass wasti ng.)
:>Iy
lfl1e, Alluvium. Alluvium ineludes alJ sedimenl deposited by streams. The deposits
100 are stratified into layers of silt, sand, gravel, and cla r. In mountainous areas, alIu­
nd vium consists largely of boulders. In areas dominated by sluggish streams, allu­
ut­ vium t pically is clayey sil1. Figure 3. 12 shows aIl uvium xposed in a stream ter­
u­ race. Coarse-grained all uvium is an excellen t aquifer wh re saturated due lo the
as sorted, stratified nature of thc debris. Along tbe edges of /lood plains in humid
areas, colluvium may constitute a large part of the valley fill (Latt man, 1960).
rb­ Two processes act to produce alluvial deposits. During intense flo ods, coarse­
ure grained debris such as cobbles are transported as bed load. T he strea m may scour
S4 Chapter 3 Weathering and Soil- Form ing Processes

Figure 3.12 Alluvium exposed in a kame lcrracc, Willi man tic, CT. Brunto n com pas
indicates scale.

down as far as bedrock, and upon dirninution of discharge cobbJes are deposited
back to or iginal grade. At the same time, overbank flooding occurs, and silt is
deposited on the flood plain (Vanoni, 1975). Thus alluvium typically consists of
two components: the lower part is a bedload-derivcd coarse fraction mantIed by
overbank siJt aboye. It is fortuitous that nature provides this dual process. Allu­
vial valleys are the most arable places on earth, thanks to lhe faet that arable silt
masks the sand and gravel below. The sand and gravel is also an aquifer wh.ich
can supply ground water or can supply rnoisture to the flood-plain crops as
"subirrigation. "
The engineering properties of aUuvium vary over a wide range. Oxbow lakes
resulting from meander cutoffs eventually fill with fine-grai ned all uvium. The
subsequent meander pattern which develops in the lower Mississippi River Val­
ley is affected by the location of (bese clay plugs wh.ich resist erosion (Fisk, 195 1).
Uniform deposits of sand and gravel can support high footing loads and are not
very susceptible to shrinkage or swelling. One problem in all uvi al foundations is
the deli neati on of specific sand , gravel, or c1ayey lenses; if slab or corumn loads
are distri buted onto beterogeneous layers, di fferential settlement can occur.
Anoilier problem ofaIl uvial soils occurs in areas such as New Orleans. where very
thick deposits of saturated clayey silt occur, giving rise to lack of beari ng strength
for large loads. In these areas, pilings are extensively used (see Chapter 5). (Note:
by far the biggest environmental hazard associated with man's utilization of allu­
vial soils is fl oods; this subject is dicussed in Chapter 8.)
Terraces are abandoned flood plains no longer related to the preseat stream
regimc. Most terraces are Pleistocene or Holoccne in age, probably formed by
climatic and/or stream regi me changes associatcd witb Quaternary climate and
 3.3 Soil and Bedrock Contacts 55

8
Figure 3. 13 Sharp contact of gneissic bedrock a nd glacial drino Storrs, CT. A. Contact
exposed in an cxcavation. B. Close-up of glaclally polished bedrock showing truncatcd
pegmatite veí n in gneiss.

glaciation. Terraces are ideal construction sites as they are rela tively fl at, not sub­
ject to flood , and are typically underlain by sand and grave!.

Glacial drij!. G lacial drift ¡neludcs al! deposits formed by glaci rs. Much of
the land area of the northern hemisphere was covered by continental glaciers dur­
ing lh Pleistocene epoch . An engineering geologist working in these areas should
acquire a sound knowledge ofglacial geology. The contact between drift and the
underlyi ng bedrock i usuaIly very sharp (see Fig. 3.7B and 3.l3).
56 Chapter 3 Weath eri ng and Soil-Forming Processe

T here are two lypes of glacia! drift: till (no nstratified drift) and stratified drift.
Ti lJ is deposited d irectly by gIacier ice, either plastered down and compacted by
a n advancing glacier (lodgement til)) or left as loose piJes ofdebris as the ice melts
al the tem1Ínus of a stationary or retreating glacier (ablation ti U). Lodgement (or
"basal") ti ll is fo und under drumlins, and consists of still, hard till possessing
stones orien ted with their long axes generally parallel to the direction ofice move­
menl. Ablation till is typicaIly cohesionless and sloughs natural1y to the angle of
repose (Hatheway, 1983). Lodgement tiU generally has better geolechmcal prop­
erties in terms of strengtb, but may be severa! orders of magnitude less in perme­
ability than ablation ti ll, thus affecting the yields ofwater wells or the operation
of sewage drain fields. The contrast in engincering properties of these two types
of ti 11 s is exem plified by the construction of the Se Lawrence Seaway and Power
project. On the basis ofborings, ablation tiU was expected. Soon after excavation,
a den se (p = 2.7 g/cm 3) lodgement till was encountered, and costs increased by
about 400%. T he contractors involved either went bankrupt, defaulted. or entered
litigation fo r extra pa yment (Legget, 1979).
T ill is typicaliy nonstratified, unsorted, and contalns angulaJ' to sub-rounded
rock pa rticles of a ll sizes (Fig. 3. 14). Platey fragments within lodgement tiU may
be oriented roughly paraUel to the ice movement, gi ving the tiU a texture referred
10 as Hlbric (MiIls, 1977). The compositioD of till fragment s reflects the type of
bedrock over which the glacier passed. Thus in New England, due to the hard
bedrock (e.g., gneiss and granite), the till typica lly contai ns large boulders ofgran­
íte. These boulders pose a severe impediment lo fa rmers plowing the land (Fig.
3.1 4A). T hrough out much of the midweslern United Sta tes the bedrock Ís pre­
dom ínately shale, and henee the tiIJ is very c1ayey wíth only a few erratic boul dcrs
lO bother the farmers. TiIl generally has favorable engíneering characleristics.
Lodgement till is un iform in nature, and is very compact, almost impenetrable 10
a ba nd shovel.
Sediment deposited by streams ofwater derived from melting glaciers is caUed
stratified drift. Tbe debris is reworkcd till; the particles ha ve beeo sorted,
rou nded, an d stratified by runni ng water. Stratified dri ft is, in fact, a ll uvi urn and
as such has the properties of alluvium. In New England the con trast in tbe engi­
neering properties of ti ll and stratified drift is exemplified by the 10catÍon of ceme­
leries. Cemeteries which were used roughly 100-300 years ago are seattered rather
randomly across the landscape, whcrcver small vi lJages or a fam ily plot happened
to be, in both tiJl and stratified drift. Tbe occurrence of large q uantities of boul­
ders in the ti ll is asevere obstacle to digging in tiU; as a consequence, modern
cemeteries (as well as most landlills) are generally fo und only in slratified drift
because the sand and gravel is easier to excavate (Raho, 197 1b).

Lacustrine and marine deposits. The fine-grained sedi ments typically depo::.­
ited in lakes are limited in arcal extent but can be very troublesome for fou nda­
lions. A dramatic example of the engíneering ínfl uence ofc\ayey lacustrine depos­
its (andent glac ial Lakc Agassiz) occurred in 191 1 when a newly constructcd gra in
elevator near Winn ipeg, Canada was filled for the ftrst time. The cntire concrete
structure, placed on a fo undation only 6 m into the underlying clay, sank and
ti lled 27°. Little damage occurred lo the grai n elevator, and so it was righ ted and
placed on a good fo und ati on . Lacustrine depos its are associated wltb the Great
Lakes. Many of the fou ndation problems in the Chicago area originate from tbi n
 3.3 Soil and Bedrock Contacts 57

drut.
d by
melts
nt(or

alian,
ed by
otered

uoded
I may
-..'ferred
pe of
~ bard

called
orted,
m and

Figure 3.14 GlacialtiLl, near Storrs, CT (fro m Rahn , 1971 b). A. Surface appearance
of pasture land . 8. Exposure showing poorly sorted texture.

lacustrine sediments deposited in Lake Michigan when the Great Lakes were
more extensivc than today.
ncrete Many shallow marine coastal deposits that accumu lated du ri ng th P leistocene
.,k and are now above sea level because ofisostatic rebou nd or sea levellow ringo Exam­
ed and pies include the Boston blue c1ay, 25 m of mari ne cJ ay underlying the low parts
Great ofBoston. The sensitive Leda c1ay, fou nd in the Sto Lawrence Vall y, is of marine
58 Chaptcr 3 Weathering and Soil-Forming Proccsses

origi no Other Pleistocene marine deposits, such as along the southern California
coast, inelude sand and gravel, which generally ma ke acceplable foundation
material.
The engineering properties of beach sand were emphasized during World War
11 in the Pacific campaign. U.S. Marines were able to easi ly dig foxholes for pro­
tection upon landing on sorne beaches. But on sorne islands, such as Iwo Jima,
Marines found it impossible. to dig foxho les in volca nic ash deposits. The Japa­
nese took advantage of local geology by building caves in coral Iimestone to shel­
ter their men and cannons.

Eolian deposits. Eolian deposits include windblown sand (sand dunes) and
windblown silt (loess).
The geomorphology and d ynamics of sand dunes are described by Bagnold
(1941) and Sharp (1966). Sand dunes invariably have a nearby source of sand.
Extensive sand dune fields are usually related to a nearby source of unconsoli­
dated to semiconsolidated sandstone. For example, the so urce of sand du nes for
the Sand Hills of Nebraska is the Ogallala Formation of Pliocene age. The vast
Sand Hills area (35,000 km 2) is co vered by sand du nes averagin g 8 m in th ickness.
The sand dunes of the Sahara Desert are derived fro m the Cretaceous Nubian
Sandstone. Smaller areas of sand dunes are usually foun d within a short distance
(1 km) of a source of sand . Examples ¡nelude the southeastern shoreli ne along
Lake Michigan ; Cape Hatteras, Ne, areas south ofbraided stream beds along the
Platte River in Nebraska ; and areas downwi nd of Playa lakes al W hite Sands
National Monument, New Mexico.
Some dune fields have traveled considerable distance. T he Al gones d unes l1ear
Yuma, AZ, for example, are a nuisance for the operation of the All-America n
Canal, which carries water from the Colorado R iver to the Imperial Valley. T he
source of sand is believed to be reworked sand from shoreli nes of Pleistocene
pluvial lakes sorne 5-30 km distant (Smith, 1978).
Sand dunes rarely serve as sites of construction due lo the arid and / or inhos­
pitable environment. Active san d dunes are su bject to conti nual movement and
are totally unsuitable for const ruction. Even inactive sand dunes have a delicatc
vegetal cover which can be broken down by man's activities causing tbem to
become active (Péwé, 1982). In 1he Salton sea area of Califo rn ia, ho uses have
been overwhelmed by sand dunes which migrate easterly al about 12 m/yr (Shel­
ton, 1966). The Sahara Desert dunes move up lO 100 m/yr.
Fine wind-blown sand grades ¡nto loess, which is much more widespread than
sand dunes. Loess is ubiquitous throughout the midwest of the Un ited Sta tes,
capping bluffs along the maj ar rivers (Fig. 3.15) and forming ferti le soils over
broad areas (Ruhe, 1969). In the upper Mississippi R iver Basin, loess was depos­
ited mai nly between 14,000 and 22,000 years ago (Ruhe, 1973).
A peculiar engineering property of loes s is its ability to sta nd in vertical culs.
For example, vertical railroad cuts made over 100 years ago in M ississippi are
sti Ll stable. At depths greater than 3 m , calca reous cement may occur in loess
beca use infiltrating meteoric water has nOl dissolved tbe delicate calcareous Ínter­
nal skeleton. Clay minerals a lso act as ceme nting agen ts (Gibbs and Holland,
1960). This eementing material apparently develops along aneient root holes. The
calca reous skeleton combined with evaporatio n of m inerahzed water 0 0 the sur­
fa ce of a man-made excavation helps bond the loess so tha1 il is stable in vertical
 59

Figure 3. 15 Loess exposed al Sioux City, l A.

n
e
g Figure 3. 16 Geologic cross section showing loess al Vicksburg, MS (after Kolb and
e Steinriede, 1967).
d West Eost
12 0

110

100

90

60

nd 70

le .J 60
w
>
W
.J 50
<t
UJ
(f)
40
W

n o> 30
m
<t

Z 20
er S'
J­
<t 10
s­ >
W
.J
W
O
lS. Grav el

~re - 10
~
s -20
w
u
O
cr­ w

~~------=------'----~
-----::
.,d, - 30

he - 40 I
ur­ O 5 VERT. EX . :..@Q
I
I
cal km
60 Chapter 3 Wcathcring and Soil-Forming Processes

cuts. Continued evaporation of soil water precipitates minerals on the vertical

walls, thus strengthening the exposure with time. Loess is susceptible to fail ure
by landslide or gully erosion where gcntly sloping cuts are made.
Loess is generally acceptable media for foundations from the standpoint that
it is uniform in composition and has very low swelling propcrties. Individual
cases ofland settlement after water has been added to loessal foundat ion soils are
reported. Loess dry densities of less than 1.3 g/cm 1 are especially suspect for set­
tlement (hydrocompaction) when t hey become wet.
At Vicksburg, MS, loess has been extensively studied by engi neering geologists
(Fig. 3.16). Loess blankets the eastcrn side of the Mississippi Rivcr for most of
its length, deposited by northwestern prevailing wi nds d uring Pleistocene time.
At Vicksburg th e loess is 10-30 m thick at the bluffs to abo ut 1 m thick 100 km
to the east. The loess is composed of weU-sorted clayey silt (Fig. 3.17), which
beco mes more fine-grained with increasi ng distance from its poi nt of origi noT he

Figure 3. 17 Typical grain-size curves for loess easl of Vicksburg, MS (from Kri nitzsky
and Turnbull, 1967).
SIEVE H YO RO M ETER

50 70 100 140 200

o
I I

~~
I

10
'\ Ro
1\\
20
i -:
~
f-
30
[
'\. i

\1
I
t.:J
W

\.~\
1
~
I
>­ 40 ---"
m i
a: I 40 .9 MILES
w
<JI 1
a: 50
«
o I \
u
f­ 50
\ ~\
Z
W
U
a:
w 70
1
I

I \\\ 't\r-.. 2 6.3 MI L ES

~ \
-­-­
n.. 1. 0 MILE .......---l ")
"

"'­
80
11
I , i ''\ :, r-.....
r-
I ,

I ! ~ 1/
:--- 7.4 MILES
I
90
-- I
100
0. 1
11
0 . 05 0. 01 0.005 0.001
GRAIN S I ZE I~J MM

FINE SANO SIL T C LA Y

 3.4 Soil-Forming Processes 61

""lical loess deposits at Vicksburg formed such an ideal merua for construction of under­
ground shelters that the Civil War Genera l U lyssess S. Gran t was unable to dis­
lodge ¡he Confederate troops by bombardment and won the Battle of Vicksbu rg
only by starving ou t the Confederate troops.

3.4 Soil-Forming Processes

The wealhering and surficial geologic processes described aboye prod uce weath­
ered rocks and surficial deposits that are commonly referred to by civil engineers
and others as' soil." Nevertheless, soil scientists and most geologists think of soíl
as onl y the thin profi le developed on weathered rocks or surficial deposits. The
study of soil-form ing proeesses and soil profiles is a scienee unto itself, whose
seope is beyond the limits of this book. Nevertheless, there are lessons for the
engin eeri ng geologist to learn from a cursory examina tíon of soil-forming
proc sses.
Soil scientists have developed a cli matic c1assification of soils. It has long been
re ognized that topsoils tend to be th ick and c1ayey, rich in red iron minerals in
tropical a reas, but tend to be thin and contain a lot of calcium carbonate and
cal ium sul fate in deserts. For this reaso n tropical soils used to be called "pedal­
fers" because of high alum inum and iron content, and desert soils were called
"pedocals" because of high calcium cont n1. This cli mate ("zonal") c1assification
of soils has beeo expa nded into "Great Soil Groups" (T able 3.2). Essentially,
pedalfers are now called laterites, aod pedocals are called caliches. More rece nt
revision of soil classification by pedologists involves omplex recognition of clÍ ­
matic as well as vegetative factors (Soil Survey Slaff, 1975). The recognition of an
ancient soil (paleoso!) in Tertiary rocks (Retallack, 1983) or Q uaternary deposits
(Birkeland, 1984) can be used as a cli matic indicator.

Table 3.2 Soil Classification Accord ing to the Great Soil Groups ("Zonal" oils)
Amount 01 moisture
Zone Arid Semiarid ! Wet and dry Subhumid Humid
I Wet

Cold Tundra

Cool and Podzol

al ~E
temperate !!l E 'OE
tQJ..c: "'QJ '0 S QJ
N QJ QJ QJ
Brown podzolic
QJN f/) e o .¡:: 'ON
00.2 (jj e .~ <11 0
'00 e
e~
~
QJ - >-~ QJ
a~ «J~
.... f/)
..c: (¡¡
.c ct a.c
QJQJ
O>~
Gray-brown podzolic
Temperate (,!)~
Ü
ID Ü u
Gray wooded soils

Warmand I ()
'O
QJ
=o
... ES lamperale a:
~ .~
N
Reddish Reddish 'O
'O .­ '00 Red-yellow
'0<11 QJa.
brown chestnut QJ .... podzolic
Red a:c. ~~
Hot 0>=
desert QJQJ
0>­
O. 01

Tropical Laterite
 '"
N

Organic debris lodged on the soil;

olly a b.ent an .ail. developed by gra ..e •.
U!u­
{ Orgonic debri•.

A dark-calored harizon conlaining a relalively high content 01

Al argonic marrer bu' mixed with mineral matter. Thick in chernozem
ond very Ihin in padzaL

A light-colored horizon, representing the region of mcximum

lone' of eluv¡otion.
leaching (or reducllan) where podzolized' or .olodized. " The
Al bleicherde al Ihe podzoi. Ab.enl in chernozem,t brown,t .iero­
zem, t and some other soils.
The ,olum .
(Thi. portion Tron,ilional lo B bul more like A Ihan B. Somelimes absenl .
inelude, Ihe
A3
Irue soil de ­ Tron.iliona l lo 8 bul more lik" 8 Ihan A. Sametl me s ab.e"'.
velo ped by
'011- build in g
pra, e"e,.) A deeper-color ed lu,ually) horizon repre,enting Ihe regian 01
mo ~i mum ill u"iotion where pod zo li zed ar .olodi z e d . Th e orstei n
lone of illuvio lion. (Exclu.ive of carbonole.
01 ¡he podzo l a nd Ih e cl a ypa n 0 1 Ine solodized . 0 10 ne1L In cher ­
or sulphotes os in c.hernozem, brown, and
noze m, brown, a nd ~ierozem sotls, this re gio n has a definHe
~ietolem soi b . In luch so'" this horizon is to
structural cho ro cter , fre quenl1y p rismotic, b ut does no' hO'le much
be considered a. essentiolly Iran.iliona l be­
Iwefln A o nd C.) if an y ¡lI u....ial malerial,, ; H represents a tro n!ition between A ond
C_ freq ue ntly ob.enl in In" i ~lra zon Cl I soils o f Ihe humld re g iol'l5.

83 Tr an.iliona I lo C.

The pare ni maleri ol. e Paren t material.

1
{
Any ,'rolum un d erneath Ihe p are nl maleriol, such os hard rack
ar a loyer 01 d oy or . ond, Ihol i. nol parenl materia' bul ma y
ho ve sig nificonce la Ih e ove rlying ,oil.
D Underl ying ,Iratum.

.Proce u of woler leoching dow nw a rd throvgh A ond B ho rizons.

"Procen of occumuloling svrf oct! min er al , through leoching upword, pro d vced by It voporotion in or e o' of lo ..... roinfall couling moi1olure movem enh lo be Pow ord 'h. lur-foce .
tM embefl o f gre o' ioil grovps (se e Table 3.2).

Figure 3_1 8 A hypothetical soil profile havi ng all the soil horizons.
 3.4 Soil-Forming Processes 63

T he solum includes thc soil honzons modified by soil-fo rming processes. Ide­
ally these processes form a distinct A and B horizon. As shown in Fig. 3.18, D is
th source material, wnich could be anything from sound bedrock to glacial till
aod C is weathered parent material, which may include decaycd gra nite, terra
rossa over li mestone, etc. CLimatic contro'ls on the development ofthe solum lead
to the folI owing threefold classificatio n of soils (soil groups).

3.4.1 Podzol
Podzol soils develop in moist, cool climates such as the forested areas of north­
eastern Uni ted States. The soil homons ilIustrated in Fig. 3.1 8 are those of a
podzol soil. The process of formi ng a podzol soil is calIed podzolization. In tbis
process the most solubl ions (calcium, sodium, magnesium as welI as iron an d
aluminu m orga nic compounds called micelles) are peptized (become mobile) in
the A horizon. They are transported downward by percolating water, and floc­
culate (precipitate) in the B horizon. As a result, the A horizon is friabie and is
bleached to a grey color, and lhe B horizon is still , blocky, and clayey, often with
a columnar structure. The A horizon is acidic due to organic acids fro m the forest
litter; the B horizon begins where the pH increases to about 6.5.
Podzol soils have no special engineering significance. On many residual surfi­
cial deposits or surficial deposits oftill the podzol soil horizons may be unnoticed
Vi in excavations. In well-developed podzol soils, the B horizon may be very hard
¡::
O
.~ and troublesome to excava te .
O
..c: Table 3.2 show the range in climate soil names from podzol soils to the trop­
O
ical laterite soils to the desert caliche soils. It is these la11er two soil groups that
'"
<!)
deserve the most attention by engineering geologists because of their unusual
-5 properties.
<Q
00
e
;; 3.4.2 Laterite
0=
oC
o Residual red soils (Ioosely called laterites) form in tropical (wet and hot) climates.
..1
cao In North America, laterite soils fo rm in Florida, Cuba, and Mexico. Laterites
....o.
extend over large areas of Brazil lndia southeast Asia, Australia, and centra l
'0 Africa. T hey ar best developed as resid ual soils developed on bedrock rich in
'" iron and alu minu m (such as basalt) and in tropical ch mates with alternating wet
]
QJ and dry periods (Hun t, 1983).
- oC
Laterites are products of intense chemical weathering, where the residual soil
oo.
;>, is higher in iron and aluminum and lower in silica than merely kaolinized rock.
..c
Typica ll organic accumulation is min imal, and silica and mobile ions (such as
-< alkali nes) are leached through lhe entire soil horizon. Laterization is the pro ess
00
whereby the precipitation of iron and aluminum compounds in the B horizon
:.., ....¡
•
-!. ~00
becomes excessive, leading to the formation of a brick-red, clayey B horizon
which is called a laterite or more properly a cuirasse. The laterite horizon is typ­
i.i: ically composed of kaoLin ite, goethite, gibbsite, and quartz, but may contain up
to 80% Fc 20 J in the form of goethite or hematite (Goudie, 1973). In sorne places
the presence of a water table promotes a very tough cuirasse (iron and alumi num
hydrates called "hardpan" or "duricrust").
Laterite soils can be an impedimen t to construction. In Vietnam , excavation
of airport ru nways during the Vietnam War was hampered by these tough soils.
64 Chapter 3 Weathering and Soil-Forming Processes

Figure 3.1 9 Laterite cxposed in a roadcut In northeastern N icaragua. The soil , devel­
oped in alluvium , has a bright red color.

rn Nicaragua (Fig. 3.19), 1 ha ve seen attempts at han d-d igging of water wells fru ;:,­
trated by laterite soils developed in c1ayey alluv ium. In the N icaragua n wet sea­
son, well-digging is impossible due 10 swampy conditions; in the dry season, dig­
ging is extreme ly difficult because the latente dries to a brick-li ke consistency. In
central Afnca, saprolites approximately 30 m thick, con tain ing laterite soils in the
upper approxim ately 10 m, mask tbe bedrock lithologies below and make pros­
pecti ng difficult.
Laterites are resistant to erosion and may cap hUI tops or slopes where vigorous
erosion has produced a truncated soil horizon (Cho rley et al., 1984). Laterites
have a long history of use as bujlding O1aterials 5uch as the bricks used in the
constructi on of walls in Thailand or the Ankor Wat temple complex ofCa01bod ia
(Banerj i, 1982).

3.4.3 Caliche
In hot, arid places, weathering is slow, and a soil profile is virtuaUy nonexistent.
Nevertheless, calcium carbonate and calcium sulfate compounds accum ulatc in
the top few meters of desert surfaces, particularly in all uvium. The format ion of
these desert soils is called calci fi cation, and the resulting soil is called caliche.
Calci fication is restrictcd to subarid and arid c1imates where precipitatíon ís
typically insufficient to allow for the movement of meteoric water downward to
tbe wa ter table. Caliche at or near tbe surface of alluviaJ fans decreases water
 3.5 Soil Maps 65

infiltratíon rates (Cooley el a l., 19 73). In Nevada, Lattman ( 1973) descri bes tbe
caliche near Las Vegas and its resulting impediment to the recharge of gro und
water. Calic he m ay act as a cap rock to preserve alluvial fan su rfaces (Gard ner,
1972) a nd affect arroyo geometry (Van Arsdale, 1982).
Gile et al. ( 1981) shows that caliche thickness is proportional to the age of the
terrace alo ng the Río G rande Ríver in New Mexico. Higher terraces a re ol der and
have thicker cali ches. Due to excavation problems, caliche caused the aba ndon­
ment of construction o f a NASA facility airport on a high terrace of the Rio
Grande where th e caliche is about 5 m thick.
In part s of sou tbern Adzona, caliche resembles concrete; 1 have found places
where prospectors with bulldozers were unable to break through a caliche­
e ncrusted coll uv ial slope where malachite float gave promise of copper veins in
the bedrock.
In the Atacama Desert in Chile, nitrate-rich caliche has been mined as a source
of nitrate used in explosives and fertilizers (Ericksen, 1983).

3.5 Soil Maps

So il surveys are routinely pu bJished by the U.S. Department of Agricul ture; these
m aps provide informatíon on climate, soil-forming processes a nd parent ma te­
rial. These repo rts can be used by engine ri ng geologists to determine depth to
bedrock, water table conditions, perm eability, shrink-sweU potenti aJ, a nd other
soil propert ies. The so il units mapp d generally are confined to the to p 2 m. Soil
survey m aps are printed into a n aerial p hotograph mosaie at 1:3 1,000, 1:24,000,
1:20,000, or 1: 15,840 seale.
The soil map uses a soil "series" in the same manner that a geologist calls a
r ck un it a " group. " All soil series are developed by the same genetic combina­
5­ tion. Subdivision of the soil series is do ne into m appable soil un its called soil type
3­ and soil phase, which ineludes characteristics such as erodibility, slope, and
19­ stoniness.
In
he
- s­

. u
'es
he
dla

nc
'. in
·of

i
~ to
h;r