Catena 106 (2013) 5467

Contents lists available at SciVerse ScienceDirect

Catena
journal ho me p a g e: ww w. elsevi er. c om / l ocate/ cate n
a

Clay mineralogy as a crucial factor in badland hillslope processes

Milica Kasanin-Grubin
Soil Erosion Laboratory, Military Trail 1265, University of Toronto at Scarborough, ON, Canada M1C 1A4

a r t i c l e
Keywords:
Badlands
Weathering
Lithological properties
Clay mineralogy
Smectite
Hillslope processes

i n f o

a b s t r a c t
Comprehensive and important work has been done in understanding badland processes on both the big
scale, and more recently on smaller scale in order to understand the subtle differences between sites.
However, most studies lack important information regarding clay mineralogy and their importance for
weathering pro- cesses. Type and rate of weathering depend on material physico-chemical properties,
climatic regimes and slope characteristics. As weathering progresses, the surface conditions change,
inuencing erosion processes. Usually badland lithologies are reported as sodium-rich, highly dispersive
materials but that is not invariably true. The primary objective of this study is to show that clay minerals
have an important role in dening sur- face and subsurface processes on badland hillslopes.
Field studies conducted in Dinosaur Park badlands, Alberta, Canada, Chinguacousy badlands, Ontario, Canada
and Sestino badlands, Tuscany, Italy, have shown that there are two dominant clay-rich lithologies:
smectite-rich and smectite-poor mudrocks. Different surface crust and weathering proles were observed
on smectite-rich mudrocks during two eld years with distinct precipitation inputs. After a dry fall-winter
season, the popcorn surface was very well developed (~ 23 cm). The subsurface layer, 1015 cm thick,
overlaid the layer of smaller shards, followed by the unweathered material. After a very wet fall-winter season the surface layer was denser and thinner (~ 0.5 cm), with wider cracks and a thin subsurface layer attached to the crust. Weathering proles on smectite-poor lithologies did not show any differences with
variable climatic conditions. A thin surface layer was composed of soft, thin and small (0.5 0.3 0.1 cm)
shards. A layer of larger shards, overlying unweathered material was beneath it.
In order to better understand the described occurrences badland materials were monitored in controlled laboratory conditions during an incremental weathering experiment. Six samples of each material were
subjected to 10 cycles of simulated rainfall which duration ranged 1060 min. Very short rainfalls
(1020 min) can cause swelling of clay minerals in smectite-rich materials and formation of popcorn
sur- face during drying periods. During subsequent wetting-drying periods, the crust becomes atter and
denser, and desiccation cracks become wider and deeper. Samples subjected to 50 and 60 min of rainfall
showed maximum swelling after only one cycle of rainfall. After this swift swelling, samples became unstable
and dis- persive. During subsequent wetting, dispersion became dominant, resulting in atter surfaces,
thinner crust and narrower desiccation cracks.
Smectite-poor mudrocks appeared to be resistant to variable wetting and drying. Large shards broke down
into smaller due to differential swelling of illite and chlorite. Once they were reduced to tiny, aky shards
the surface became compacted. The surface cracks served as the initial pathways for sediment and water
movement.
Knowing if the material is smectite-rich or smectite-poor can help in prediction of processes and the erosion
rates. The amount, intensity and duration of precipitation will determine not only the rate of weathering but
also the type of the weathered surface on smectite-rich surfaces. Smectite-poor materials appear to be more
resistant to change in precipitation.
2012 Elsevier B.V. All rights reserved.

Faculty of Environmental Governance and Corporate Responsibility, EDUCONS

University, Vojvode Putnika 87, Sremska Kamenica, Serbia. Tel.: + 381 214893706;
fax: + 381 214893660, + 1 416 971 3077.
E-mail address: m.kasanin.grubin@utoronto.ca.
0341-8162/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.catena.2012.08.008

1. Introduction
A complex combination of physico-chemical properties of the material and climatic conditions determine the type and amount of
weathering and subsequent crust formation. Different materials respond differently in the same environments, but also the same materials can respond in different ways to the same processes in different
environments. Usually more than one process is active at one site and

M. M.
Kasanin-Grubin
Kasanin-Grubin
/ Catena
/ Catena
106106
(2013)
(2013)
5467
5467

it is often hard to determine the exact weathering process that

resulted in a certain weathering product (Robinson and Williams,
2000).
Badlands, as ideal
eld
laboratory are
intriguing
geomorphologists for decades. Comprehensive and important work
has been done in un- derstanding processes on the big scale (e.g.
Bryan and Yair, 1982; Campbell, 1997; Howard, 1997). Recently
more detail work has been done on smaller scale in order to
understand the subtle differences be- tween sites (e.g. Faulkner et
al., 2000, 2003, 2004; Kuhn et al., 2004). However, most studies
oversimplify important information regarding mineralogy, especially
content and type of clay minerals in badland ma- terials. Usually, clay
minerals are reported as a grain sizes factor or gen- erally badland
materials are characterized as smectite shales.
Badland slopes have complex slope forms mostly because of
contrasting lithology (Howard, 1997). Regolith properties, together
with the dominant process will determine the slope shape and slope
angle. Savat and De Ploey (1982) suggest that the critical slope
angles for rill initiation depend on the
material textural
characteristics and also that on clay-rich materials, rills occur on
gentler slopes. Slope angles thresh- olds for rill initiation are mostly
around 2-3 for loamy soils, and around
9 for loamy sands (Savat and De Ploey, 1982).
Schumm (1956) found that creep is the main process that
forms round slopes on the clay-rich Chadron formation, while wash
processes produce straight slopes on the Brule formation in the South
Dakota bad- lands, USA. The difference in the dominant processes is
obviously a con- sequence of different lithologies, because both units
are exposed to the same conditions, and the Chadron slopes
lowered, while the Brule slopes retreated (Campbell, 1997). The data
presented on the composi- tion of the two units are limited.
Schumm (1956) described the Brule formation as having a hard,
cracked, rilled and impermeable surface, while the Chadron surface is
more permeable, covered with loose pop- corn aggregates that
become sticky mud when wetted. However, in arid climates slope
convexity can be caused by rainsplash as well (Carson and Kirkby,
1972) which can also obliterate rilling (Dunne and Aubry, 1986).
Arid and semi-arid climates are most often associated with
badlands. However, since not all landscapes in this climate develop into
badlands, and badlands can form in different climates as well, the
fundamental at- tention should be given to badland materials
(Campbell, 1997). In bad- land areas where two or more different
lithologies are present, the erosion rates vary, as well as slope
properties and processes. Besides the differences in processes on
the Brule and Chadron formations in South Dakota badlands, USA,
(Schumm, 1956) due to lithological proper- ties, another example is
the development of calanchi and biancana in badlands in Italy.
Calanchi have steep rilled and gullied slopes, with mass movement
being a main process on the side-slopes, while biancana have gentler
slopes, with equally active surface and subsurface networks (pipes and
subsurface cracks), having besides rill erosion, mass move- ment and
sheet wash as dominant processes (Alexander, 1982). Calanchi,
usually much larger landforms, generally form in coarser sedi- ments
like clayey silts and sandy clayey silts, while smaller dome type
biancana have very high clay content (Battaglia et al., 2002; Torri
and Bryan, 1997). Biancana sediments have higher Na contents
compared to calanchi (Alexander, 1982; Battaglia et al., 2002).
Certain minerals play a crucial role in the weathering of rocks at
sur- face conditions. Generally these minerals can be divided into two
main groups: soluble minerals (e.g. gypsum, calcite) and clays
(Gallart et al.,
2002). Dissolution of gypsum or calcite adds calcium to the
solution, which can replace Na on clay exchange sites and in that way
can sup- press clay swelling (Bridge and Tunny, 1973). Yair et al.
(1980) con- rmed the role of gypsum in limiting the swelling of
mudrock in Zin badlands, Israel. In Tabernas badland, Spain,
dissolution of moderately soluble gypsum which lls the cracks is
responsible for mudrock break- down (Canton et al., 2001).

Iron bearing minerals are often present in sediments as secondary

minerals and they can have
important inuence on the
weathering

M. M.
Kasanin-Grubin
Kasanin-Grubin
/ Catena
/ Catena
106106
(2013)
(2013)
5467
5467

process. Chigira and Oyama (1999) described the different inuence of

pyrite in the mudrock and sandstone weathering processes. Pyrite
dis- solves and produces sulphuric acid, which in turn dissolves
rock- forming minerals. The authors further conclude that during this
process mudrock is weakened, but the sandstone is strengthened
with iron oxide cement. Pyrite was found to play an important role
in smectite rich claystone (Sadisun et al., 2005). Free sulphuric acid
that is released after pyrite oxidation reacts with calcite and transfers
it into gypsum pro- moting slow increase in rock volume followed by
rock cracking (Sadisun et al., 2005).
Most badland lithologies are clay-rich materials and their behaviour is controlled by the type and amount of clay minerals present.
The effect of clay minerals on the behaviour of materials is variable.
The permeability of smectites depends greatly on the exchangeable
cations and it generally increases with the increasing Na content
(Quirk and Schoeld, 1955), kaolinite has a moderate effect on permeability, while illite clogs pores and reduces permeability signicantly (Appelo and Postma, 1993). However, the response of the
clay-rich material to contact with water does not solely depend on
the dominant clay mineral but also on the whole mineral assemblage.
For example, small amounts of smectite in kaolinitic soils cause it to
be more dispersive and unstable (Levy et al., 1993). On the other
hand, Yair et al. (1980) found that the swelling of smectite is
suppressed in the presence of substantial quantities of kaolinite in
the Zin badlands, Israel, materials, while calcite and gypsum are also
present. The limited swelling together with the highly stable clay
rich aggregates can then result in a high inltration rate and reduce
runoff (Yair et al., 1980).
In studies of badland processes, mineral composition and clay
min- eralogy of lithologies was usually reported, but not studied in
detail. So far, existing linkages of clay mineral type to specic
processes were not tested and dened. Fine-grained sediments are
frequent constitu- ents of badlands, and
depending on the
classication used, reecting mainly the clay mineral content and
lamination they are most often called mudrocks or shales. There are
many uncertainties about the proper classication of clay-rich

sediments (Carozzi, 1993; Schieber et al., 1998). According to Lewan

(1978) shales contain more than 65 % of material b 5 m in size and
mudstones have from 45 to 65 % of ne particles, while Carozzi
(1993) denes shales as the terrigenous rocks composed of clay and
silt-sized particles, but indicates that it is impor- tant to distinguish
clay shales from
silt shales. Probably the classica- tion of
sedimentary rocks which is still most widely used has been done
by Folk (1974) who identies mudrocks as a general term for rocks
that contain more than 50 % of silt and/or clay. Mudrock with a
distinct structure like lamination or graded bedding is called shale
(Folk, 1974). In badland literature, mudrocks and shales are the
terms most often mentioned which are probably inherited from the
geological literature for that area. Even if the same classications
were not always
used, especially when they are so broad, a complete textural analysis
to- gether with the clay mineral identication should provide enough
in- formation about the lithology.
Sorting of the material is an important factor in material
susceptibil- ity to erosion. Well-sorted ne sediments are prone to
disintegration and piping, while sediments with a large range of sizes
and higher de- gree of packing are more resistant (Terzaghi and Peck,
1967). For exam- ple, clays in sandstone in Dinosaur Park badlands,
Alberta, turn into gels when wetted and ll the pores in the sandstone,
making it impermeable (Bryan et al., 1978). Silt and sand grains
together with pores and voids can disrupt swelling and shrinking of
clay-rich sediments (Chertkov,
2003).
In previous studies conducted on badland lithologies, electrical
con- ductivity (EC), sodium absorption ratio (SAR), and cation
exchange ca- pacity (CEC) were determined (De Boer and Campbell,
1990; Faulkner et al., 2003; Hodges and Bryan, 1982; Imeson et al.,
1982; Kuhn et al.,
2004; Sole et al., 1992). The chemical properties of the surface and
sub- surface material, the changes that occur with progressive
weathering

and their importance has been considered with more attention by

for
(e.g. Canton et al., 2001; Faulkner et al., 2000,
2003).
Faulkner et al. (2000) showed that changes that occur in site geochemical characteristics over time in Almeria badlands, Spain, were
involved in badland stabilisation. The authors proposed the
functional relationship between EC and SAR as a useful tool
characterising pip- ing prone sites. In the weathering prole in this
area clay content is found to increase with depth (LopezBermudez and Romero-Diaz,
1989) which coincides with the nding that the subsurface material
is more dispersive than the surface (Gerits et al., 1987). Progressive
change in material dispersion with depth is supported by ndings
that the materials are depleted in Na relative to Ca from the surface
downwards (Faulkner et al., 2000). The leached crust leads to slope
stabilisation (Faulkner et al., 2000, 2003) which can later be colonized
by algae (Finlayson et al., 1987; Yair et al., 1980).
Important work has been done by Faulkner et al. (2004) regarding
the changes in geochemical properties in surface and subsurface layers
in ma- terials prone to piping in Almeria badlands, Spain. Generally
dispersive materials are considered to be mostly affected by piping
(e.g. Gutierrez et al., 1997), however, the potentially dispersive soils
(SAR b 3) can be piped as well (Faulkner et al., 2004). Faulkner et al.
(2004) suggests a top-down model for pipe development in
subsurface layers. According to this model, calcium replaces sodium in
the crust during leaching, leav- ing a calcic crust and sodic subcrust which is prone to piping. Throughow originating from high
inltration rates through the silty crust surface is redirected by
subsurface swelling into dispersive less clay-rich subsurface layers.
Limited swelling of this layer makes it more vulnerable to erosion
(Faulkner et al., 2004).
The crust characteristics, such as mineralogical and geochemical
composition, cracks and thickness of the surface and subsurface
layers inuence the hillslope processes. The type of crust that develops on the exposed material depends on its physico-chemical
characteristics, the percentage and mineralogy of the clay present,
and on the magnitude and frequency of precipitation. In sandy materials, with low clay content, even if clay minerals are reactive, inltration rates will not be substantially reduced. In clay rich materials, the
inltration rates will depend on type of clay minerals and the amount
of precipitation. Inltration capacities of the clay-rich materials are
low (Imeson, 1986) but with the development of cracks and popcorn
surface the inltration capacities increase (Yair et al., 1980). Torri and
Bryan (1997) have shown that the initial inltration rates of the
unweathered biancana corestone were very low because of low
po- rosity, high bulk density and small pore size. However, the
inltration rates increased rapidly with repeated rainfalls during the
weathering experiment because of increase in pore diameter due to
slaking, dis- persion and solution. However, with prolonged wetting
and change from plastic to liquid state, inltration capacities of the
popcorn surface decrease and considerable changes in strength
occur and run- off can occur (Hodges and Bryan, 1982).
Intense shrink-swell activity in smectite-rich sediments can produce desiccation cracks and loose popcorn regolith that has high
macroporosity (Imeson, 1986; Schumm, 1956). The popcorn
sur- face was described in Chadron formation, South Dakota,
USA (Schumm, 1956), Dinosaur Park badlands, Alberta (Bryan et
al.,
1978), Vallcebre badlands, Spain (Regues et al., 1995), Val D'Orcia
badlands, Italy (Torri and Bryan, 1997). In the Zin badlands, Israel,
the swelling of smectite clays was suppressed in the presence of kaolinite and calcite and instead of a typical popcorn crust with
loose aggregates, a dense crust, rich in desiccation cracks with a
subsurface coarse shard layer developed (Yair et al., 1980). A similar
crust devel- oped on marls in badlands in SE Spain (Canton et al.,
2001). The weathering rind
on the
Dinosaur Park badland
sandstone is only few mm thick (Bryan et al., 1978).
Even though the response of the material to precipitation has
been proven to be signicant, so far there have been only few

studies that monitored weathering processes of badland materials

in controlled

laboratory conditions. It is essential to achieve conditions in the laboratory which reproduce eld conditions as closely as possible. So far
only a few weathering experiments have been conducted using badland materials. However, they have not fully represented natural conditions, and did not observe changes in detail. In addition, clay
minerals have not been identied in detail, and link to processes
has not been established.
Furthermore, material response to seasonal changes is reported to
be an important factor in erosion processes over time. Therefore it is
essential to study in detail how materials behave in different climatic
conditions. Natural systems have many variables acting at the same
time, mainly due to climate (amount and type of precipitation, prevailing winds, etc.) and slope characteristics. In order to minimize
the number of variables it is essential to study progressive
weathering of fresh material in controlled laboratory conditions.
Therefore, the primary objective of this paper is to show the
important role of clay minerals, both their content and mineralogical
composition, in den- ing surface and subsurface processes on
badland hillslopes.
2. Materials and methods
2.1. Field sites
Three badland sites were selected in order to assure variety of
mudrock with different content and type of clay minerals: Dinosaur
Park badlands, Canada, Sestino badlands, Italy and Chinguacousy badlands, Canada. The average annual precipitation in the Dinosaur Provincial Park is 350 mm (75% of rain and 25% of snow). Typically two
main types of storms occur in the badland area: the low intensity
rainfalls that can occur throughout the year and the high intensity
rainstorms mostly during July and August (Bryan and Campbell,
1986).
Badland formation started after deglaciation about 15,000 years BP
when Red Deer River started incision into clastic sediments
(Campbell,

1970). Bryan et al. (1987) differentiated four erosion surfaces in the

Dino- saur Park badlands. The rst two were caused by glacial
meltwaters and the third by uvial discharges. The authors suggest
that surface 1 devel- oped almost 14,500 years BP during the early
postglacial period, surface
2 incised within a few hundred years, surface 3 was developed
very soon after the glaciation (between 13,000 and 9000 years BP),
and it was generated by runoff, and nally surface 4 that coincides
with the present oodplain. All surfaces were blanketed with aeolian
sands and silts around 5500 years BP, which subsequent stripping by
streams, ac- companied by piping and tunnelling, has re-exposed
vulnerable Creta- ceous sediments and initiated high erosion rates
that still occur in this area (Bryan et al., 1987). Erosion rates in
1
Dinosaur Park badlands are on average about 3 mm y
and with
1
extreme values up to 13 mm y
(Campbell, 1997). The vegetation is
sparse and slopes of high range of in- clinations, often with pipes and
tunnels, are dissected by numerous rills, gullies and valleys (Bryan et
al., 1984).
Sestino badlands cover an area of approximately 4 km 2 and are located on the border of the provinces of Tuscany and Marche, Italy
(Fig. 1b). The climate is humid-temperate with an annual precipitation of 12201330 mm. The geology of the area where badlands
have formed is very complex and includes several very different
units. The oldest unit is Tortonian marly siltstone (MA), represented
by clastic sediments (sandstones and siltstones) changing upward
to the marly shale. This is also the structural base level. The overlying
unit is the Cretaceous deep marine varicoloured shales. They were
folded over the MA, during which their chemical composition has
changed and became highly fractured. The major types of shales distinguished by colour are dark grey, light grey and red. Olistostrome,
consisting of poorly-sorted clastic sediments, which were transported
by deep marine landslides overlie both the MA and varicoloured
shale. Bulging of the limestone massif (Sasso de Simone) into
olistostrome and landslides that followed triggered the rapid erosion

Fig. 1. a) Alberta badlands, Canada, b) Sestino badlands, Italy, c) Chinguacousy badlands, Canada.

on both olistostrome and varicoloured shale. The slopes on

olistostrome are steep and straight, while on varicoloured shale are
convex.
The Chinguacousy badlands, Ontario, 140 km northeast of Toronto,
2
cover an area of 0.4 km (Fig. 1c). Mean annual precipitation
is
~ 890 mm. The badlands were triggered by land clearance at the
begin- ning of the twentieth century. Chinguacousy badlands are
developed en- tirely on the Ordovician Queenston shale. Two facies of
Queenston shale outcrop in Chinguacousy badlands. The lower facies is
the dark red shale, which is the transitional unit between lower,
more marine and supratidal facies and represents input of
terrestrial muddy deposits (Brogly et al., 1998). The upper facies,
light red shale, was deposited in the coastal-plain and it is highly
bioturbated, contains bioclasts and reworked shards from the
underlying dark red facies.
2.2. Materials
Six major lithological units can be distinguished in these three
bad- land areas: sandstone and mudrock from Dinosaur Park
badlands, Alberta, varicoloured shale and marly shale from Sestino
badlands, Tuscany, and dark and light red shales from the
Chinguacousy badlands, Ontario (Table 1). According to the clay
content and the dominant clay mineral, tested badland lithologies can
be divided into three main cate- gories: sandstone, smectite-rich
lithologies (SR) and smectite-poor lithologies (SP). Although all
Alberta badland mudrock varieties have smectite as the dominant
clay mineral, the smectite:illite ratio de- creases from yellow, over
brown to grey mudrock. All three types of varicolour shales have
similar mineralogical composition. Besides the dominant smectite,
they also contain illite and chlorite.
The lithological units described as Alberta badland sandstone
actual- ly include a range of texturally different varieties from clayey
siltstone to silty sandstone. The average chemical composition of
sandstone types is presented in Table 2. The content of clay fraction
is on average
20%, and smectite and kaolinite are the dominant clay
minerals.
When dry, the sandstone is white, hard and compact. It is
essentially impermeable, and during precipitation water inltrates
only few millimetres, mostly along joints (Bryan et al., 1978, 1984;
Sutherland and Bryan, 1988). Inltrated water causes clay dispersion
and gel for- mation that precludes further moisture penetration
(Bryan et al.,
1984). Released individual sand
grains are easily entrained.
However, when a piece of sandstone was immersed in water for two
days, it dis- persed completely (Fig. 2).
Alberta badland mudrocks and Sestino badland varicolour shales
contain > 60% of clay minerals, with dominant smectite. The main
va- rieties can most easily be distinguished by colour and in
previous studies (e.g. Bryan et al., 1978) the colour of the Alberta
badland mudrocks was distinguished by the colour of the wet
mudrock sur- face; for the purpose of this study, mudrocks were
distinguished by colour of the unweathered material. To facilitate
distinction of lithologies, hereafter they will be referred to by their
colour. In addi- tion, to stay consistent with the literature, the
Alberta
badland smectite-rich
lithologies will
be
called
mudrocks, while Sestino

badland lithologies, shales. As indicated above, the most typical varieties of mudrocks are yellow, brown and grey, while varieties of
shales are red, dark grey and light grey. All mudrocks and both the
dark grey and red shales have similar chemical composition which
is in accordance with their mineralogical composition (Table 2). The
light grey shale has a high Ca content indicating abundant calcite.
Three lithological units were distinguished as smectite-poor:
Sestino badland marly shale, and Chinguacousy badland dark red and
light red shales (Table 1). Besides illite and chlorite, marly shale
contains calcite, while dark and light red shales contain quartz.
Chemical composition of dark and light red is also identical (Table 2).
Marly shale has signi- cantly higher Ca content indicating abundant
presence of calcite.
2.3. Weathering proles
2.3.1. Sandstone
A thin weathering rind that develops on the sandstone surface
after drying was observed and described in previous studies (e.g.
Bowyer- Bower and Bryan, 1986; Bryan et al., 1984; Hodges and
Bryan, 1982). It can form on the sandstone surface in one storm
(Bowyer-Bower and Bryan, 1986) and beneath it, the sandstone is
unweathered. On the sandstone surface rainfall also
leaves
splash pits, which are destroyed during the next rainfall (Hodges,
1982).
The thickness of the weathering rind depends mainly on the
textural characteristics of the material. On the laminated clayey
siltstone the rind is 24 mm thick, and only on this material could the
35 mm thick sub- surface layer be differentiated from the underlying
unweathered sand- stone. On clayey sandstone and coarse sandstone
the weathering rind is thinner (b 2 mm) and no subsurface layer can
be differentiated. The weathering rind gets deeper (57 mm) along
joints because when water inltrates cracks to the subsurface area it
does not evaporate as quickly as from the surface and the weathering
process penetrates the material.
2.3.2. Smectite-rich lithologies
Typical weathering proles, as developed on mudrock in the Dinosaur Badlands Park, Alberta (Hodges and Bryan, 1982), marls from the
Guadix Basin badlands, Spain (Gerits et al., 1987), mudrocks from the
Chadron Formation, Utah (Howard, 1994a) and mudrocks from the
Zin Valley badlands, Israel (Yair et al., 1980) have the following
layers: a) 12 cm thick porous crust with desiccation cracks that is
leached of highly soluble components; b) ~ 10 cm subsurface
compact layer rich in micropores; c) 1040 cm thick transitional
layer, with partly weathered shards and d) unweathered material.
A typical weathering prole on mudrock surfaces consists of 4
basic layers: crust, subsurface layer, layer of thin small shards and
unweathered layer (Gerits et al., 1987). Swelling and shrinking of
smectite-rich material usually develops a specic surface colloquially
called popcorn. This type of surface causes great variations in surface microrelief and antecedent moisture conditions (De Boer and
Campbell, 1990). Depending on the rainfall intensities popcorn
sur- face crust can develop on the unweathered material after two
rain- falls (Bowyer-Bower and Bryan, 1986). It is loose and porous
and the aggregates are resistant to splash detachment at the
beginning of the rainfall (Bryan et al., 1984).

Table 1
Main lithological units in the badland areas and clay mineral content.
Lithology

Area

Clay
content (%)

Sandstone

Dinosaur Park
badlands

24

Dominant clay
minerals

Smectite,
kaolinite

According to eld observations, Kasanin-Grubin and Bryan (2007)

in- dicated that the weathering prole and surface crust of
smectite-rich
materials depends on the climatic conditions. Authors described
that after a dry winter and dry spring, the popcorn surface was
very well

Smectite-rich
lithologies

Smectite-poor
lithologies

Mudrock (yellow,
brown, grey)
Varicolour shales
(red, dark-grey,
light-grey)

72
Sestino
badlands

Marly shale
Dark red shale
Light red shale

Chinguacousy
badlands

63

Smectite, illite,
kaolinite
Smectite, illite,
kaolinite

65

Illite, chlorite

68
58

Illite, chlorite
Illite, chlorite

developed on all mudrock surfaces. The popcorn surface was

around
23 cm thick, and the surface popcorn aggregates contained
abundant voids (Fig. 3). The subsurface layer was between 10 and 15
cm thick. Be- neath was a layer of smaller shards, followed by the
unweathered material. A similar weathering prole was described by Bryan et al. (1984)
and Bowyer-Bower and Bryan (1986). The popcorn surface can also
form with repeated freezing-thawing cycles and expansion that occurs

Table 2
Chemical composition of unweathered materials (%).
Lithological type

Fe

Ca

Si

Al

Mg

Na

Sandstone
Smectite-rich lithologies

1.20
3.14
3.45
4.27
5.59
5.80
3.61
5.68
5.25
2.10

9.16
0.36
0.19
0.34
0.49
0.12
20.73
5.14
5.47
24.25

1.42
2.04
2.45
2.34
2.99
3.07
0.93
4.20
4.16
1.05

24.40
30.00
30.94
28.59
29.78
29.44
17.84
24.92
24.57
13.67

5.33
8.34
8.88
8.90
10.77
10.43
5.62
8.06
8.13
3.38

0.74
1.16
1.08
1.00
1.70
1.66
1.20
2.50
2.44
1.10

0.94
1.28
1.02
1.22
0.77
0.68
0.79
0.08
0.10
0.08

Smectite-poor lithologies

Yellow mudrock
Grey mudrock
Brown mudrock
Red shale
Dark grey shale
Light grey shale
Dark red shale
Light red shale
Marly clay

due to ice crystal growth and clay swelling can induce stronger
alteration of mudrock than caused with wetting and drying (Pardini
et al., 1995). Regues et al. (1995) indicate that the sequence of
seasonal processes is strongly
effecting
characteristics
and
appearance of regolith and Pardini et al. (1996) emphasis that the
frost-defrost action is most de- structive process for materials
containing swelling clays.
When the same sites were examined after a very wet winter
and spring the surface layer was denser and thinner (~ 0.5 cm), more
com- pact and with wider cracks. The weathering proles also
differed with the type of mudrock (Fig. 3). The surface crust on
mudrocks was thin, averaging 0.5 cm in depth. A thin subsurface
layer attached to the crust which was 4 cm thick. Underlying was
the 4 cm thick ne shard layer, overlying the unweathered material.
This uniform, dense, com- pact crust covered most of the surface
homogeneously (Fig. 3).
2.3.3. Smectite-poor lithologies
The weathering proles on three materials differ in thickness of
layers. Marly shale has the thinnest surface layer composed of soft,
thin and small (0.5 0.3 0.1 cm) shards. Beneath is the layer of larger shards, followed by unweathered material. The weathering
proles of the dark red and light red shale are similar in appearance
but differ in the thickness of the layers. The surface layer on dark
red shale is about 2 cm, while on the light red only 1 cm. The
subsurface on the dark red shale is ~ 10 cm thick, and on the light
red it is only 2 cm.

increase in the inltration rates with the crust development it

would be interesting to observe the rates at which these changes
occur during a number of wetting and drying experiments with
constant wetting and drying intervals.
Experiments described by Kasanin-Grubin and Bryan (2007) have
clearly shown that the moisture input has substantial inuence on
behaviour of the yellow Alberta mudrock. As it was stated above,
the aim of this study is to test other clay rich sediments. For that purpose ve other smectite-rich materials and three smectite-poor materials were chosen.
Despite the fact that weathering proles on smectite-poor lithologies are not as well developed as on smectite-rich materials, these
materials were subjected to the same experiments in order to test
the inuence of dominant clay mineral on behaviour of materials.
Alberta badlands sandstone could not be used in this type of experiments since the unweathered material was compact and breaking it
into shard would not represent the realistic eld conditions.
Incremental weathering experiment was designed to simulate realistic conditions for the eld site. The unweathered shards were
placed in circular aluminium sample trays (radius = 12.5 cm, depth
4 cm) above funnels. Samples were subjected to 10 cycles of simulat1
ed rainfall at intensity 45 mm h
with duration ranging from 10 to
60 min. Drying periods were 23 h of which 10 h were under lamps
which produced an average temperature of 35 C. The temperature
and the duration were chosen to accelerate drying, but to minimise
the impact of transformation of minerals. In the experiment Lechler

2.4. Methodology
Even though the response of the material to precipitation has
been proven to be signicant, so far there have been only few
studies that monitored weathering processes of badland materials in
controlled lab- oratory conditions. It is essential to achieve conditions
in the laboratory which reproduce eld conditions as closely as
possible. Unweathered material should be as intact as possible
during sampling. Pardini et al. (1995) developed a good sampling
technique by mechanically inserting aluminium boxes into the
regolith and in this way the original structure of the material was
preserved. However, this is not always possible be- cause badland
sediments are often fractured and break easily into shards.
Both Pardini et al. (1995) and Canton et al. (2001) wetted
samples with capillary movement by placing them in sand baths
saturated with water. With clay-rich materials it is hard to determine
the rate of water intake. In the experiment done by Pardini et al.
(1995) the wetting cycles lasted 3 days followed with air-drying for
12 days, while Canton et al. (2001) left samples in the sand bath until
the whole surface appeared wet and oven dried them at 40 C.
With the wetting from the bottom, leaching of elements and crust
development, which are essential for ero- sional processes, will not
evolve exactly as they do in the natural conditions. Torri and Bryan
(1997) used rainfall simulators in their ex- periment which provided
more realistic conditions. The storm intensity over 27 storms was
constant (50.4 mm h 1), however the duration in- creased from
initial 15 min to 120 min, while drying varied from 1.25 to 73 h (~
21.13 h). Although this experiment revealed the stages in

Fig. 2. Sandstone immersed in tap water for 2 days and oven-dried overnight.

Fig. 3. (a) Difference in the rill network and surface characteristics on the same site in 2001 (left) and 2003 (right). (b) Photos and SEM images of popcorn surface in 2001 (left)
and surface crust in 2003 (right) (Kasanin-Grubin and Bryan, 2007).
1

nozzle 460 778 30 BC at 45 mm h

intensity, with fall-height 3 m
was used, giving a kinetic energy of 28.3 J m 2 mm 1. Rain intensity
and drop fall size was determined prior to the experiment. Sample
trays were positioned in the area under the rainfall were rainfall distribution was uniform. Volume, sediment concentration and specic
conductance were measured in the leachate.
Mineralogical composition was determined using powders by
X-ray diffraction (XRD) using a Philips model PW 1050/37. Phases
were identied with on automated peak search and mineral identication software using the ICDD database. Clay mineralogy was determined by XRD using the standard procedure (Brindley and
Brown, 1980).
Content of major elements was determined using the X-ray orescence (XRF) Philips model PW 2404. Samples were analysed as
pressed powder pellets with calibration using synthetic standards,
in-house rock standards and standard reference materials.

Grain size analysis was conducted by sieves and the pipette

method. One hundred grams of sample was left overnight in 300 ml
distilled water with 20 ml 0.5% sodium-hexametaphosphate as a
dispersant to prevent aggregation of particles. Wet sieving
procedure was used with the range of sieves from 1 mm, 0.5 mm,
0.25 mm, 0.125 mm,
0.063 mm. Sieved material was oven-dried and weighed. Ten
grams of b 0.063 mm fraction was subjected to the pipette analyses.
The proce- dure described by Loveland and Whalley (2001) was
followed.
3. Results
3.1. Surface characteristics
Incremental laboratory experiments testing the surface response
to simulated rainfall were carried out to explore progressive change
in physical and chemical properties when unweathered materials

from different lithological units were exposed to same conditions.

The aim of these experiments was to understand how the processes
active respond to different material characteristics. Surface crust evolution and leachate characteristics were closely monitored.
3.1.1. Smectite-rich lithologies
The Dinosaur Park badlands yellow, brown and grey mudrocks
and Sestino badlands dark grey, light grey and red shale are
smectite-rich materials (SR). In these smectite-rich lithologies a
marked difference in surface crust and desiccation cracks emerged
in samples subjected to simulated rainfall. The shard structure with
dened margins was maintained throughout the experiment with
rainfall durations of 10 and 20 min up to 30 min, especially in the
grey mudrock and in the dark grey shale (Fig. 4). The shard
swelling potential was limited by water availability and the short
duration of the wetting period. During drying cycles minute cracks,
b 1 mm in width, often appeared on the shard surfaces. The type of
smectite in the grey mudrock requires sub- stantial input of water
to achieve full swelling potential, and even where water supply
is unrestricted, swelling takes place over a prolonged period. In
the yellow and brown mudrocks, and red and light grey shales
short rainfall (1020 min) durations were enough to
produce substantial clay swelling and
popcorn
surface.
The effect of water availability is shown by all samples subjected
to 50 and 60 min of rainfall, which showed maximum swelling after
one cycle of rainfall (Fig. 5). After this swift swelling, samples became
unstable and dispersive, and after the third rainfall cycle, dispersion
became dominant, resulting in atter surfaces, thinner crust development and narrower desiccation cracks. In the 60 min samples the
fastest transition from swelling to dispersion occurred in the grey
mudrocks and the dark grey shale. In all other samples the transition
was gradual. In samples subjected to 30 and 40 min of rainfall, dispersion was much less effective and samples showed maximal swelling
during all subsequent rainfall cycles. This is most evident in the yellow and brown mudrock and the light grey shale.
Surface
crusting and desiccation cracking are caused by
physico- chemical changes of the mudrock material due to wetting and
subsequent sunshine. Addition of
water initiates swelling,
progressive closure of pores, and eventually, breakdown of shards
by dispersion and slaking

(Imeson and Verstraten, 1988), releasing abundant ne clays which

can clog macropores. Swelling reduces hydraulic conductivity by
closing des- iccation cracks, while dispersion releases ne clay particles
which clog mi- cropores (enhanced in this case by the very small size
of clay particles: b 0.5 m).
3.1.2. Smectite-poor lithologies
The Sestino badlands marly shale and Chinguacousy badlands
dark red and light red shales are smectite-poor lithologies (SP).
These lithologies contain illite and chlorite as dominant clay minerals.
Generally, in the smectite-poor materials there was no apparent
dif- ference in samples caused by rainfall duration (Figs. 6, 7). When
exposed to rainfall, marly shale and dark red shale shards broke apart
after each drying cycle (Figs. 6, 7). The mixed-layer non swelling
clay minerals can cause very high pressures under repeated wetting
and drying condi- tions that can result in brittle failure of material
(Wust and McLane,
2000). Cracks that form during slaking promote further permeability
and expose more rock surface to water (Sadisun et al., 2005).
The rst rainfall left the surface of the light red shale rough in
all samples (Figs. 6, 7). During subsequent wetting and drying cycles,
the surface became compact and smooth. Surface cracks developed in
all samples early in the experiment and became deeper and
increased in density over subsequent cycles.
3.2. Leachate characteristics
The leachate was collected from all samples, and volume, specic
conductance and sediment concentration were measured. Hereafter
only results for 10 min and 60 min samples will be presented as the
end members of tested durations of rainfall (Fig. 8a,b).
Smectite-rich samples subjected to 10 min rainfall showed an increase in leachate volume after the rst two or three cycles, which
then remained almost constant until the end of the tenth cycle (Fig.
8a).
Most smectite-rich samples subjected to 60 min rainfall showed
maximal increase in leachate volume after the rst cycle, which then
in- creased slowly or remained constant until the end of ten cycles
(Fig. 8b). Certain variations in the leachate volume between the
samples that re- ceived different amounts of precipitation depend on
the mineralogical

Fig. 4. Incremental weathering experiment for the smectite-rich lithologies. Samples subjected to cycles 1, 5 and 10 of 10 min long rainfall (intensity 45 mm h

).

Fig. 5. Incremental weathering experiment for the smectite-rich lithologies. Samples subjected to cycles 1, 5 and 10 of 60 min long rainfall (intensity 45 mm h

and chemical compositions of the sediment. The grey mudrocks had

the highest leachate volume rate of smectite-rich lithologies with two
sud- den peaks in volume rates once in the 30 min sample and twice
in the
60 min sample (cycles 3 and 8) (Fig. 8b). By contrast, the leachate
vol- ume increased rapidly from the rst to the third cycle and
then remained almost constant in other smectite-rich mudrocks
(Fig. 8b). The deep surface cracks that developed in the early
weathering stages in grey mudrock facilitate faster throughow and
consequently a lot less water was available for retention. In other
smectite-rich sediments

).

the hydraulic conductivity was limited by swelling which

occurred mostly during the rst three cycles.
Release of sediments showed more variations between different
smectite-rich types. Generally, the grey mudrock released more sediment than other smectite-rich materials. The clay structure in the
grey mudrock was easily destroyed when it received longer rainfalls
(50 and 60 min) and particles were dispersed and transported. This
is evident as the 60 min sample released substantially more sediment
than the 10 min samples.

Fig. 6. Incremental weathering experiment for the smectite-poor lithologies. Samples subjected to cycles 1, 5 and 10 of 10 min long rainfall (intensity 45 mm h

).

Fig. 7. Incremental weathering experiment for the smectite-poor lithologies. Samples subjected to cycles 1, 5 and 10 of 60 min long rainfall (intensity 45 mm h

In smectite-rich materials subjected to 60 min rainfall the sediment concentration was highest at the beginning of cycles in all
samples and slowly decreased towards the tenth cycle (Fig. 9b).
There was no signicant difference between the sample intervals
because of the swift change from swelling to dispersion and the
clogging of pores. Higher smectite content prevented rapid sediment release. Grey mudrock showed sudden decrease in the 30
and the 60 min samples in sediment concentration in same cycles
during which a sudden increase in the volume occurred (cycles 3
and 8) (Figs. 8b, 9b). This conrms that higher volumes of water
passed through subsurface cracks, but that they were lled with sediment in the following cycle and cracks closed again. However, it
must be expected for this to happen again, once the sediment is
washed out.
Leachate specic conductance was higher in samples subjected to
10 min samples than for 60 min samples. Specic conductance also
showed distinct trends between 10 and 60 min samples (Fig. 10a
and b). In all 10 min smectite-rich lithologies, the specic conductance increased during rst three to four rainfalls and then remained
at an almost constant (Fig. 10a). This conrms that most solutes are
released when the fresh material is exposed to rainfall. With time,
surface and subsurface cracks form and act as important water and
sediment conduits. In 60 min samples initially high conductance decreased and remained almost constant in all smectite-rich samples.
In smectite-poor lithologies, leachate characteristics were substantially different from the smectite-rich lithologies. In the marly
shale and the dark red shale, volume rates increased from the 10 to
the 60 min samples (Fig. 8a and b). The light red shale was distinct
from other smectite-poor samples because the 10 and 20 min samples had higher volume rates than the other samples (Fig. 8a).
Sediment concentration and
specic conductance were
extremely low in all smectite-poor samples (Figs. 9a and b, 10a and
b). However, in marly shale volume and sediment concentration
suddenly decreased in certain cycles (cycles 5 and 8), then increased
again during the fol- lowing cycles (cycles 7 and 9), while the
specic conductance had the opposite pattern. This happened
during shard breakdown and when water used new pathows.
Sediment concentrations were also highest

).

in these two cycles. These changes are due to compaction which is

least evident in the 10 and 20 min samples. The surface was
compacted after the rst cycle in samples that were under rainfall
for longer increments. This was marked with highest sediment concentration. The surface cracks served as the steady pathways for
sediment and water movement.

4. Discussion
In each lithology, depending on its properties, a certain threshold
level of weathering is necessary to make the material vulnerable to
erosion processes. The type and rate of weathering depend on material physico-chemical properties, climatic regimes and slope characteristics. As weathering progresses, the surface conditions change,
inuencing erosional processes. The amount and type of clay minerals are critical for weathering processes in badland materials and
will determine whether chemical or physical weathering processes
are dominant.

4.1. Alberta badlands sandstone

The Alberta sandstone is extremely susceptible to rill erosion
for two main reasons, both related to clay mineralogy. The rst is
the presence of both smectites and kaolinites which have very
different properties. Materials containing both smectite and
kaolinite have about the same shear strength at low water contents
(b 20%). As the water content increases smectites lose their strength
more rapidly than the kaolinite because of the high cation
exchange capacity (Taylor and Eggleton, 2001).
The second reason making Alberta sandstone particularly susceptible to erosion is the critical amount of clay fraction present in these
materials (~ 20%). Dalrymple and Jim (1984) indicate that when the
clay content in sandstone is 1020%, clay particles coat silt and sand
particles and are linked by the bridging clay particles that can easily
be broken.

a) Leachate volume - 10 min rainfall

30
25
SR-yellow
SR-brown
SR-grey
SR-dark grey
SR-red
SR-light grey
SP-marly shale
SP-dark red
SP-light red

ml min-1

20

15
10
5
0
C1

C2

C3

C4

C5

C6

C7

C8

C9

C10

cycle number

b) Leachate volume - 60 min rainfall

30

25
SR-yellow
SR-brown
SR-grey
SR-dark grey
SR-red
SR-light grey
SP-marly shale
SP-dark red
SP-light red

15

ml min-1

20

10

0
1

10

cycle muber
Fig. 8. Leachate volume during incremental weathering experiment. a) 10 min samples, b) 60 min samples (SR smectite-rich; SP smectite-poor).

4.2. Smectite-rich materials

Six materials are distinguished as smectite-rich: yellow, brown
and grey mudrock from Dinosaur Park badlands, Alberta and dark
grey, light grey and red shales from Sestino badlands, Tuscany. All
tested litholo- gies have smectite as the dominant clay mineral.
There are a number of varieties of mudrocks, mainly reecting the
amount and type of clay minerals. Furthermore, mudrocks are often
interbedded with ma- terials of different physico-chemical
properties.
Weathering has a major inuence on the processes on smectiterich materials. The rate of weathering depends partially on climatic
and slope characteristics, but mainly on material physico-chemical
proper- ties. Weathering processes make smectite-rich materials
susceptible to erosion through swelling, shrinking, dispersion and
slaking of clays, surface sealing and cracking, transport of ne
materials and migration of elements.
From the presented data it is obvious that even rainfall of very
short duration (1020 min) can cause swelling of clay minerals during wetting and formation of popcorn surface during drying
pe- riods. During subsequent wetting-drying periods, the crust
becomes atter and denser, and desiccation cracks became wider and
deeper.

Depending on size, continuity and reappearance after wetting cracks

can become owpaths and may evolve into rills. The rate at which
cracks re-appear after sealing is very important for runoff generation
particularly in typical infrequent rainstorms of low intensity and duration (Bryan et al., 1978).

4.3. Smectite-poor materials

Three materials were distinguished as smectite-poor lithologies:
marly shale from Sestino badlands, Tuscany, and dark red and light
red shale from Chinguacousy badlands, Ontario. Sediments in which
smectite is not the dominant clay mineral behave differently from
smectite-rich materials. During repeated wetting and drying large
shards break down into smaller shards due to differential swelling
of illite and chlorite. Once they are reduced to tiny, aky shards
(0.5 cm 0.2 cm) the surface becomes compacted. The Mancos
shale in Utah has similar characteristics (Fam and Dusseault, 1998).
The Mancos shale decomposes after few tens of hours into aky
shards with only slight swelling (Howard, 1997). These materials
are salt-rich and each time shards disintegrate they yield a yellowish

a) Sediment concentration - 10 min rainfall

40
35
30
SR-yellow
SR-brown
SR-grey
SR-dark grey
SR-red
SR-light grey
SP-marly shale
SP-dark red
SP-light red

g ml-1

25
20

15
10
5
0

10

cycle muber

b) Sediment concentration - 60 min rainfall

40
35
30
SR-yellow
SR-brown
SR-grey
SR-dark grey
SR-red
SR-light grey

20

g ml-1

25

15

SP-marly shale

SP-dark red
SP-light red

10
5
0

10

cycle muber
Fig. 9. Leachate sediment concentration during incremental weathering experiment. a) 10 min samples, b) 60 min samples (SR smectite-rich; SP smectite-poor).

liquid rich in Na and Ca sulphates (Laronne, 1981). If salt leaching

does not occur, shard disintegration will not occur (Howard, 1997).
Smectite-rich sediments are most common in badland areas,
how- ever, lithologies of similar composition to Chinguacousy shales
can be found in other badlands. Illite, kaolinite and illite-smectite
mixed layers comprise 26% of the clay fraction in Mancos shale, Utah
(Forsans and Schmitt, 1994). Fam and Dusseault (1998) showed
that Queenston shale and Mancos shale are similar materials with
high density, low cat- ion exchange capacities and low volumetric
water content. Due to the absence of swelling clays these materials
are not as reactive as the ma- terials in the previous group, however,
the presence of two clay min- erals with different properties makes
these materials susceptible to erosion. Due to different clay minerals
present, water is nonuniformly distributed on the surface of the
shards inuencing the ow of water through shales (Fam and
Dusseault, 1998). Differential swelling, higher dispersion rates of
illite, and
explosion of air
trapped in
micropores cause
disintegration of shards during wetting. Shards brake down from
~ 5 cm to ~ 0.5 cm in diameter after only a few rainfalls in the controlled laboratory conditions. The continuous disintegration of shards
produces small amounts of sediment for transport during rainfall.
The

ne material that is released usually becomes the cementing agent

for remaining shards.
Low sediment yield from these materials is in accordance with the
detachment-limited model that Howard (1994b) developed for the
Mancos shale. Convex slopes in Chinguacousy can be attributed to
the dominance of creep and shallow slumping, as Howard (1994a)
also described for the Morrison Formation in Utah and the Chadron
Formation in South Dakota.
5. Conclusions
Weathering does not have the same effect on all badland materials. On the Alberta sandstone only a thin weathering rind develops.
It can develop after one rainfall on the unweathered sandstone and
does not change its properties with rainfall intensity and duration.
Weathering is an important process on smectite-rich mudrocks.
The weathering rate
depends on intensity and
duration of
precipitation and length of drying periods. Shard
swelling and
surface sealing occurs after only 1020 min of rainfall. Drying often
produces a characteristic popcorn surface. During
subsequent
rainfalls and drying periods

a) Specific conductance - 10 min rainfall

2500

uS cm-1

2000
SR-yellow
SR-brown
SR-grey
SR-dark grey
SR-red
SR-light grey
SP-marly shale
SP-dark red
SP-light red

1500

1000

500

10

cycle muber

b) Specific conductance - 60 min rainfall

2500

2000
SR-yellow
SR-brown
SR-grey
SR-dark grey
SR-red
SR-light grey
SP-marly shale
SP-dark red
SP-light red

uS cm-1

1500

1000

500

10

cycle muber
Fig. 10. Leachate specic conductance during incremental weathering experiment. a) 10 min samples, b) 60 min samples (SR smectite-rich; SP smectite-poor).

desiccation cracks form on the surface. Over time cracks become

wider and deeper and crack re-appearance can produce owpaths
that can evolve into rills.
When unweathered smectite-rich material is exposed to
surface conditions sediments and solutes are mostly released during
the rst one or two rainfalls and usually reach a constant value
after three more rainfalls. The peak of sediment yield usually
shows a delay
of one to two cycles compared to the solute
concentration. This is in accor- dance with the crack appearance.
Most solutes are washed through cracks that formed initially, causing
high specic conductance and sed- iment concentration in the
leachate. After that, the surface, and espe- cially the subsurface cracks
serve as the preferential ow conduits, so that specic conductance
and sediment concentration decreased and the leachate volume
increases during subsequent weathering cycles.
Clay-rich materials that contain illite and chlorite as dominant
minerals have different weathering processes. During repeated wetting and drying shards break down into smaller shards due to differential swelling of illite and chlorite. Once they are reduced to tiny,
aky shards (0.5 cm 0.2 cm) the surface becomes compacted.

During wetting these shales release a constant small amount of sediments, with regular decrease in specic conductance and leachate
volume over rainfalls.
This study has
shown that clay mineralogy is extremely
important for the behaviour of different materials. According to their
response to precipitation, type of weathering, erosional process and
rill network characteristics, badland lithologies can be grouped by
clay mineralogy: sandstones and siltstone that typically have ~ 20%
of clay minerals, smectite-rich mudrocks and shales, and shales that
do not have smec- tite as the dominant clay mineral. In previous
badland studies clay min- eralogy was very rarely tested in detail and
it is crucial information for understanding the processes active and
prediction of slope
evolution. Obtained results indicate that
determining the detail mineralogical composition of materials is
essential for understanding geomorphic processes and should not be
neglected in future studies.
So far physically-based models of soil erosion and
rill
development have not incorporated lithological characteristics in
an appropriate way. This study has proved that lithological
properties, particularly clay mineralogy, are of great importance for
hillslope processes. In this

study it has been shown that smectite-rich and smectite-poor

litholo- gies behave distinctively and this classication should be
used as a basis for the model improvement.
Acknowledgments
I would like to express deep gratitude to Dr. Rorke Bryan for his
support and guidance throughout the research. I would also like to
gratefully acknowledge support for research on which this paper is
based from Natural Science and Engineering Research Council,
Canada.
References
Alexander, D., 1982. Difference between calanchi and biancane badlands in Italy. In:
Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. GeoBooks, Norwich,
pp. 112.
Appelo, C.A.J., Postma, D., 1993. Geochemistry, Groundwater and Pollution. A.A.
Balkema, Brookeld, USA, p. 536.
Battaglia, S., Leoni, L., Sartori, F., 2002. Mineralogical and grain size composition of
clays developing calanchi and biancane erosional landforms. Geomorphology 49, 153
170. Bowyer-Bower, T.A.S., Bryan, R.B., 1986. Rill initiation: concept and experimental
evaluation on badland slopes. Zeitschrift fur Geomorphologie, N.F. 60, 161175.
Bridge, B.J., Tunny, J., 1973. The effect of gypsum treatment on the swelling of natural
clods of a clay soil. Soil Science 115 (6), 414419.
Brindley, G.W., Brown, G., 1980. Crystal Structure of Clay Minerals and their X-ray Identication. Min. Soc., London, pp. 2115.
Brogly, P.J., Martini, I.P., Middleton, G.V., 1998. The Queenston Formation:
shale- dominated, mixed terrigenous-carbonate deposits of Upper Ordovician,
semiarid, muddy shores in Ontario, Canada. Canadian Journal of Earth Sciences 35 (6),
702719. Bryan, R.B., Campbell, I.A., 1986. Runoff and sediment discharge in a
semiarid ephemeral
drainage basin. Zeitschrift fur Geomorphologie, N.F. 58, 121143.
Bryan, R.B., Yair, A., 1982. Badland Geomorphology and Piping. Geo Books, Norwich, p.
408. Bryan, R.B., Yair, A., Hodges, W.K., 1978. Factors controlling the initiation of runoff
and piping in Dinosaur Provincial Park badlands, Alberta, Canada. Zeitschrift fur
Geomorphologie, N.F. Supplement Band 29, 151168.
Bryan, R.B., Imeson, A.C., Campbell, I.A., 1984. Solute release and sediment
entrainment on microcatchments in the Dinosaur park badlands, Alberta,
Canada. Journal of Hydrology 71, 79106.
Bryan, R.B., Campbell, I.A., Yair, A., 1987. Postglacial geomorphic development of the
Dinosaur Provincial Park badlands, Alberta. Canadian Journal of Earth Sciences
24, 135146.
Campbell, I.A., 1970. Erosion rates in the Steveville badlands, Alberta. The Canadian
Ge- ographer 14 (3), 202216.
Campbell, I.A., 1997. Badlands and badland gullies, In: David, S.G. Thomas (Ed.), Arid
Zone Geomorphology: Process, Form and Changes in Drylands, 2nd edition. John
Wiley & Sons Ltd., pp. 261291.
Canton, Y., Sole-Benet, A., Queralt, I., Pini, R., 2001. Weathering of a gypsum-calcareous
mudstone under semi-arid environment at Tabernas, SE Spain: laboratory and
eld-based experimental approaches. Catena 44, 111132.
Carozzi, A.V., 1993. Sedimentary Petrology. Prentice Hall, Englewood Cliffs, New Jersey.
Carson, M.A., Kirkby, M.J., 1972. Hillslope Form and Process. Cambridge University
Press, p. 461.
Chertkov, V.V., 2003. Modelling the shrinkage curve of soil clay paste. Geoderma 112,
7195.
Chigira, M., Oyama, T., 1999. Mechanism and effect of chemical weathering of
sedimen- tary rocks. Engineering Geology 55, 314.
Dalrymple, J.B., Jim, C.M., 1984. Experimental study of soil microfabrics induced by isotropic stresses of wetting and drying. Geoderma 34, 4368.
De Boer, D., Campbell, I.A., 1990. Runoff chemistry as an indicator of runoff sources and
routing in semi-arid, badland drainage basins. Journal of Hydrology 121, 379394.
Dunne, T., Aubry, B.F., 1986. Evaluation of Horton's theory of sheetwash and rill erosion
on the basis of eld experiments. In: Abrahams, A.D. (Ed.), Hillslope Processes. The
Binghamton Symposia in Geomorphology, International Series, no. 16. Buffalo, NY,
USA, pp. 3153.
Fam, M.A., Dusseault, M.B., 1998. High-frequency complex permittivity of shales (0.02
1.30 GHz). Canadian Geotechnical Journal 35, 524531.
Faulkner, H., Spivey, D., Alexander, R., 2000. The role of some site geochemical processes in the development and stabilisation of three badland sites in Almeria, Southern
Spain. Geomorphology 35, 8799.
Faulkner, H., Alexander, R., Wilson, B.R., 2003. Changes to the dispersive characteristics
of soil along an evolutionary slope sequence in the Vera badlands, southeast Spain:
implications for site stabilisation. Catena 50, 243254.
Faulkner, H., Alexander, R., Teeuw, R., Zukowskij, P., 2004. Variations in soil dispersivity
across a gully head displaying shallow sub-surface pipes, and the role of shallow
pipes in rill initiation. Earth Surface Processes and Landforms 29, 11431160.
Finlayson, B.L., Gerits, J., Van Wesemael, B., 1987. Crusted microtopography on badland
slopes in southeast Spain. Catena 14, 131144.
Folk, R.L., 1974. Petrology of Sedimentary Rocks. Hemphill Publishing Co., University of
Texas. Forsans, T.M., Schmitt, L., 1994. Capillary forces: the neglected factor in shale
studies?
Eurock '94. A. A. Balkema, Rotterdam, pp. 7184.

Gallart, F., Sole, A., Puigdefabregas, J., Lazaro, R., 2002. Badland systems in the Mediterranean. Dryland Rivers. Hydrology and Geomorphology of Semi-arid Channels.
John Wiley & Sons, p. 388.
Gerits, J., Imeson, J.M., Verstraten, J.M., Bryan, R.B., 1987. Rill development and badland
regolith properties. Catena Supplement 8, 141160.
Gutierrez, M., Sancho, C., Benito, G., Sirvent, J., Desir, 1997. Quantitative study of
piping processes in badland areas of the Ebro basin, NE Spain. Geomorphology 20,
237253. Hodges, W.K., 1982. Hydraulic characteristics of a badland pseudo-pediment
slope sys- tem during simulated rainstorm experiments. In: Bryan, Rorke, Yair, Aaron
(Eds.),
Badland Geomorphology and Piping. GeoBooks, Norwich, pp. 127152.
Hodges, W.K., Bryan, R.B., 1982. The inuence of material behaviour on runoff initiation
in the Dinosaur badlands, Canada. In: Bryan, Rorke, Yair, Aaron (Eds.), Badland
Geomorphology and Piping. Geo Books, Norwich, pp. 1346.
Howard, A.D., 1994a. Badlands. In: Abrahams, Athol D., Parsons, Anthony J. (Eds.), Geomorphology of Desert Environments. Chapman & Hall, London, pp. 213242.
Howard, A.D., 1994b. A detachment-limited model of drainage basin evolution. Water
Resources Research 30, 22612285.
Howard, A.D., 1997. Badland morphology and evolution: interpretation using a simulation model. Earth Surface Processes and Landforms 22, 211227.
Imeson, A.C., 1986. Investigating volumetric changes in clayey soils related to
subsurface water movement and piping. Zeitschrift fur Geomorphologie, N.F. 60,
115130.
Imeson, A.C., Verstraten, J.M., 1988. Rills on badland slopes: a physico-chemically controlled phenomenon. Catena Supplement 12, 139150.
Imeson, A., Kwaad, F.J.P.M., Verstraten, J.M., 1982. The relationship of soil physical and
chem- ical properties to the development of badlands in Morocco. In: Bryan, Rorke,
Yair, Aaron (Eds.), Badland Geomorphology and Piping. Geo Books, Norwich, pp. 47
69.
Kasanin-Grubin, M., Bryan, R., 2007. Lithological properties and weathering response
on badland hillslopes. Catena 70, 6878.
Kuhn, N., Yair, A., Kasanin-Grubin, M., 2004. Spatial distribution of surface properties,
runoff generation and landscape development in the Zin Valley badlands, Northern
Negev, Israel. Earth Surface Processes and Landforms 29, 14171430.
Laronne, J.B., 1981. Dissolution kinetics of surcial Mancos shaleassociated alluvium.
Earth Surface Processes and Landforms 6, 541552.
Levy, G.J., Eisenberg, H., Shainberg, I., 1993. Clay dispersion as related to soil properties
and water permeability. Soil Science 155 (1), 1522.
Lewan, M.D., 1978. Laboratory classication of very ne-grained sedimentary rocks.
Geology 6, 745748.
Lopez-Bermudez, F., Romero-Diaz, M.A., 1989. Piping erosion and badland development in south-east Spain. Catena Supplement 14, 5973.
Loveland, P.J., Whalley, W.R., 2001. Particle size analysis. In: Smith, Keith A. (Ed.), Soil
and Environmental Analysis: Physical Methods, p. 637.
Pardini, G., Pini, R., Barbini, R., Regues, D., Plana, F., Gallart, F., 1995. Laser elevation
measurements of a smectite-rich mudrock following freeze-thawing and wetdrying cycles. Soil Technology 8, 161175.
Pardini, G., Vigna Guidi, G., Pini, R., Regues, D., Gallart, F., 1996. Structure and porosity
of smectitic mudrocks as affected by experimental wetting-drying cycles and
freezing-thawing cycles. Catena 27, 149165.
Quirk, J.P., Schoeld, R.K., 1955. The effect of electrolyte concentration on soil permeability. Journal of Soil Science 6, 163178.
Regues, D., Pardini, G., Gallart, F., 1995. Regolith behaviour and physical weathering of
clayey mudrock as dependent on seasonal weather conditions in a badland area at
Vallcebre, Eastern Pyrenees. Catena 25, 199212.
Robinson, D.A., Williams, R.B.G., 2000. Experimental weathering of sandstone by combinations of salts. Earth Surface Processes and Landforms 25, 13091315.
Sadisun, I.A., Shimada, H., Ichinose, M., Matsui, K., 2005. Study on the physical disintegration characteristics of Subang claystone subjected to a modied slaking index
test. Geotechnical and Geological Engineering 23, 199218.
Savat, J., De Ploey, J., 1982. Sheetwash and rill development by surface ow. In: Bryan,
R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich,
pp. 113126.
Schieber, J., Zimmerle, W., Sethi, P.S., 1998. Shales and Mudstones, I. Basin Studies, Sedimentology, and Paleontology. E. Schweizerbat'she Verlagsbuchhandlung (Nagel u.
Overmiller). Stuttgart.
Schumm, S.A., 1956. The role of creep and rainwash on the retreat of badland slopes.
American Journal of Science 254, 693705.
Sole, A., Josa, R., Pardini, G., Aringhieri, R., Plana, F., Gallart, F., 1992. How mudrock and
soil physical properties inuence badland formation at Vallcebre (Pre-Pyrenees,
NE Spain). Catena 19, 287300.
Sutherland, R.A., Bryan, R.B., 1988. Solute source and transport in an ephemeral badland catchment, southeastern Alberta, Canada. Canadian Journal of Earth Sciences
25, 167181.
Taylor, G., Eggleton, R.A., 2001. Regolith Geology and Geomorphology. John Wiley &
Sons, p. 375.
Terzaghi, K., Peck, R.B., 1967. Soil Mechanics in Engineering Practice, 2nd edn. John
Wiley & Sons, New York.
Torri, D., Bryan, R.B., 1997. Micropiping processes and biancana evolution in southest
Tuscany, Italy. Geomorphology 20, 219235.
Wust, R., McLane, J., 2000. Rock deterioration in the royal tomb of Seti I in the Valley of
the Kings, Luxor, Egypt. Engineering Geology 58 (2), 163190.
Yair, A., Lavee, R.B., Bryan, R.B., Adar, E., 1980. Runoff and erosion processes and rates in
the Zin valley badlands, Northern Negev, Israel. Earth Surface Processes and Landforms 5, 205225.