Introduction
Weathering and erosion of sandstone produces a great variety
of natural landforms (Young et al., 2009). Some of the finest
ancient monuments are carved in sandstone (e.g. Petra,
Jordan, Heinrichs, 2008), and sandstone is also one of the most
frequently used dimension stones (Siegesmund and Snethlage,
2011). Thus, understanding the weathering and erosion of
sandstone is crucial in the preservation of unique natural landforms
as well as in protecting ancient monuments. However,
those factors affecting sandstone weathering and erosion are
still not well understood, especially on the field scale as the
following authors demonstrate on numerous examples (Viles,
2001, 2013; Turkington and Paradise, 2005; Warke et al.,
2006; Ruedrich and Siegesmund, 2007; Mol, 2014).
Salt and frost weathering are probably the most common
processes responsible for sandstone degradation on bare sandstone
surfaces (Steiger et al., 2011). This is nicely demonstrated
by Huinink et al. (2004) showing that the location where salt
weathering will occur is controlled by the moisture flux, and
hence by the hydraulic field and its boundary conditions. Mol
and Viles (2012) documented a dynamic relationship between
surface hardness (i.e. case hardening), degree of weathering,
and internal moisture regimes. Laboratory experiments on
sandstone specimens several centimeters in size, make it
possible to compare the durability of various sandstone types
to a specific salt, to frost weathering, or to test the degree of
damage created by various salts to a single sandstone lithology
(Robinson and Williams, 2000; Williams and Robinson, 2001).
However, these experiments are on too small a scale and too
simplified to account for effects of the hydraulic field and other
factors occurring on the field scale. Also, moisture content and
other conditions used in laboratory tests are often far from the
conditions present in the field (Dorn et al., 2013).
As has recently been shown by laboratory experiments
performed on many different sandstone types and under
controlled conditions, increased gravity-induced stress (hereafter
‘stress’) significantly decreases the salt and frost weathering
rate (Bruthans et al., 2014; Rihosek et al., 2016). Via physical
modeling it was also demonstrated that stress plays a crucial
role in the origin of natural sandstone landforms such as pillars,
arches, alcoves, and pedestal rocks. Numerical modeling
revealed how erosion and weathering are coordinated by the
stress field to reshape the original rock exposure geometry into
a new one, where portions with low loading have been
removed (Bruthans et al., 2014). Nevertheless, it is still difficult
to carry out a relevant quantitative study of those factors affecting
the weathering and erosion rates in the field for the following
specific reasons: Erosion is too slow to be measured over
the same time period as those factors affecting it. Contrast in
weathering intensity at a single locality is too small to see clear
differences in erosion rates. Too many weathering factors occur
at a single location and/or during the same time period. Finally,
lithology changes on the same scale as the weathering factors,
such that the effects of weathering and lithology on the erosion
rate can not be separated.
That is why the present study is focused upon a unique
sandstone overhang with extremely high erosion rates
(Figures 1A–1D, 2A, and 2B). During the last ~35 years, the
fallen material has gradually accumulated below the overhang,
forming a 1.8 m high pile (mean debris accumulation rate of
approximately 50 mm/yr; time period reported by a local
resident). Such rapid sandstone erosion only occurs in this
overhang, the only wet place in the studied sandstone cliff.
The source of the moisture at the overhang is suspected to be
a leak from a sewerage reservoir (Figure 1B). No other segments
of the sandstone cliff (including other overhangs) show traces of
such rapid erosion. A well-preserved engraving ‘1915’, which
was found just 5 m to the west of the overhang in the same
sandstone layer (Figure 2C), demonstrates that the erosion rate
does not exceed 1 mm/100 yr on dry surfaces.
The erosion rate at the seasonally wet overhang is at least
2000 times faster.
We consider this overhang as an appropriate field-scale and
decade-long ‘model’, which may facilitate an understanding of
rapid sandstone weathering and erosion under real conditions,
in general. As the rapid decay of sandstone in this locality is
confined to a limited place with neighboring stable sandstone,
the properties of both environments can be compared.
The purpose of this study was to characterize those factors
responsible for the rapid weathering and erosion of the sandstone
in the overhang in contrast to the stable sandstone in its
immediate surroundings and use this knowledge to interpret
relative rate of erosion of overhangs in the past in Central
Europe and elsewhere. To attain this, the following particular
tasks were carried out:
• sandstone and salt efflorescence were characterized from
the mineralogical point of view and the salt content was
determined;
• erosion rates were measured – at various locations in the
overhang, and the conditions under which the erosion rates
were maximal and minimal were analyzed;
• selected parameters including, tensile strength, moisture,
suction and water chemistry were monitored in the overhang
and its vicinity;
• the field data were complemented with a laboratory simulation
of frost weathering of the sandstone, and numerical
modeling of the stress in the studied overhang.
• erosion rates measured in the artificially wetted overhang
were compared with those derived for overhangs with ambient
moisture content in Central Europe and Colorado
Figure 1. Overall situation of the studied overhang: (A) location in Europe; (B) study area; (C) plan of the studied overhang, with positions of particular
sites; for details about the sections A–B and C–D see Figures 8A and 8B; (D) vertical section by the overhang.
712
Copyright © 2016 John Wiley & Sons, Ltd.
Plateau (USA) and used to develop ideas on the timing of
evolution of these overhangs.
Characteristics of the Study Area
The studied overhang developed in a sandstone cliff named
‘Čertova Kazatelna’ (‘Devil's Pulpit’ in English), and is situated
85 km west-southwest (WSW) of Prague in the western
part of the town of Plzeň, at an altitude of 315 m above sea
level (a.s.l.) (Figures 1A and 1B). The mean annual precipitation
in the area is 527 mm, with an average temperature of
7.5 °C (Station Plzeň-Bolevec, 1961–1990). The mean regional
evapotranspiration is ~90% (according to the Czech
Hydrometerological Institute).
The cliff is approximately 700 m long and up to 22 m high,
and it lines the alluvium of the Mže River (Vítek, 1987). The cliff
is formed by arkosic sandstone and conglomerate that belong
to the upper part of Nýřany member of the Kladno formation,
which is of Carboniferous age. The material of the cliff is fluvial
in origin, with common cross-bedding (Opluštil et al., 2005),
belonging to the Plzeň Basin, one of the post-Variscan intermountain
depressions, with a several hundred meter thick infill.
The sediments are faulted but not folded, the beds are
subhorizontal. The cliff hosts several overhangs, numerous
tafoni-like forms, and honeycombs (Vítek, 1987).
The easternmost part of the cliff, in close proximity to the
overhang, is studied; it is located at 49°45′28.1″N, 13°20′
24.3″E (Figure 1B). Here, the cliff is from 6 to 10 m high. The
studied overhang, approximately 3 m wide and up to 1.5 m
deep, is situated at the base of the cliff (Figures 1C, 2A, and
2B). The overhang faces southward. Its sides and ceiling are
made of coarse sandstone and conglomerate, with sporadic
pebbles up to 5 cm in size. Two thin (~1 cm) clay horizons
Figure 2. Field situation of the overhang and nearby surroundings: (A) front view of the overhang with the heap of fallen materials (photograph taken
prior to first freezing – December 2011) – note the vegetation cover at the heap top; (B) side view; (C) well-preserved engraving ‘1915’ situated 5 m to
the west of the overhang; (D) flakes parallel to overhang surface; (E) detail from part A showing case hardening covered by salt efflorescence outside of
the studied overhang area (left from the dashed line); (F) fiberglass pegs for measurement of erosion rate; (G) stable cliff wall with case hardening (left
in the photograph), and front side view of the overhang with heap covered by freshly fallen material – March 2013); (H) accumulation of flakes and
sand on the top of the heap. [Colour figure can be viewed at wileyonlinelibrary.com]
713QUANTITATIVE STUDY OF A RAPIDLY WEATHERING SANDSTONE OVERHANG
Copyright © 2016 John Wiley & Sons, Ltd.
cause abrupt changes in the ceiling geometry (Figure 1D). The
overhang's sides and ceiling are decaying into flakes from
several millimeters up to several tens of millimeters thick,
similar to those described by Dragovich (1967). Flakes are
parallel to the overhang surface (Figure 2D). The pile below
the overhang has formed by the accumulation of fallen flakes.
Outside the area of the studied overhang, the sandstone exposures
are mostly covered by case hardening 5–16 mm thick
(Figure 2E). Relics of the case hardening in some parts of the
overhang indicate that the case hardening originally occurred
there, as well. Salt efflorescences are common on sandstone
exposures shielded from the rain; by contrast, they occur only
temporarily on the studied overhang. In the colder periods of
the year, water drips from the western part of the overhang.
Materials and Methods
Solid sampling and sample treatment
Solid samples were taken from the overhang, and also from
other parts of the cliff; i.e. from a sandstone outcrop unaffected
by seeping moisture. For the locations of the sampling sites see
Figures 1C and 1D. Samples 5 mm thick (approximately 5–20 g
of material) were scraped off from the surface zone of the overhang.
This material was used for solid phase identification, as
well as periodic determination of moisture and salt contents.
Samples were taken from the wider surroundings of fiberglass
pegs (Figure 2F) in order to not affect the erosion rates. Additionally,
materials that fell onto plastic foil were sampled for a
mineralogical study and measurements of their tensile strength.
Sandstone blocks sampled outside the overhang area from the
same lithological horizon were used as background for both
the tensile strength measurements and frost weathering tests.
Phase identification and salt content
To determine the mineralogical composition of the matrix in
the sandstone cliff, the fine clay- and silt-fraction was separated
by sedimentation of the disintegrated flakes in distilled water.
Phase identification was determined by conventional X-ray
diffraction (XRD) analyses using a PANalytical X'Pert Pro diffractometer
under the following conditions: CuKα radiation 40 kV,
30 mA, step scanning at 0.013°/300 s in the range 2°–89° 2θ.
Qualitative analysis was performed using the HighScorePlus
software package (PANalytical, the Netherlands; HighScorePlus,
2011), Diffrac-Plus software package (Bruker AXS,
Germany), and JCPDS PDF-2 database (2004).
To determine the salt content, the solid material was dried at
105 °C until a stable weight was obtained, crushed by a rubber
pestle, and quartered. A small amount of material was mixed
with deionized water (1:100), shaken at 150 RPM for 24 hours,
filtered through a membrane filter (Millipore®), and stored in
HDPE bottles in the cold. The conductivity of leachates was
measured, and TDS (total dissolved solids) was calculated from
the electric conductivity of the leachate based on a calibration
using the chemical composition of the analyzed leachates.
Measurement of erosion rate
Two different approaches were applied to measure the erosion
rate of the sandstone overhang. The first approach consisted of
placing the plastic foil on the floor below the overhang ceiling,
and collecting the fallen material during several time periods
(Figures 1C and 1D). The fallen material was dried, weighed,
and its weight was divided by the surface area of the plastic foil,
and by the time period for which material was collected (in
kg/m2
/day). Finally, the erosion rate was calculated from the
accumulation rate using the following equation:
RR ¼ AR=dens*ACarea=EXarea;
where RR is the erosion rate (in m/day); AR is the accumulation
rate (in kg/m2
/day); dens is sandstone density (in kg/m3
); ACarea
is the extent of the accumulation area (in m2
); EXarea is the
extent of the exposure area (in m2
).
Between 2008 and 2011, the rate was occasionally measured
in 0.2–1 day long periods during different seasons of
the year. Between December 26, 2012 and March 14, 2015
the erosion rate was measured continuously in 1–3 month
long periods.
The second approach was based on placing fiberglass pegs
(0.8 cm in diameter, ~15 cm long, Figure 2F) into drill holes in
the sandstone, fixing them with epoxy resin, and measuring
the distance between the peg top and the rock surface in the
peg's surroundings with a caliper. Five fiberglass pegs (P1–P5)
were emplaced into the overhang on March 3, 2013, forming
a line from the innermost part of the overhang outwards
(Figures 1C and 1D). Two additional pegs (P6 and P7) were
placed on December 2, 2013 on the vertical wall west of the
overhang. Surface erosion around the fiberglass pegs was
measured in 1–3 months long periods. Due to the undulating
surface, the distance between the peg top and rock surface is
measured with a precision of ±2 mm.
Sandstone bulk density was measured on small samples by
the wax immersion method (ASTM, 2004). The bulk density
of debris which had accumulated in a pile below the overhang
was measured as the ratio of the weight of debris taken from a
hole dug into the debris pile, divided by the volume of the hole.
The hole bottom and sides were covered by thin plastic foil,
and the volume of the excavated material was measured by
the amount of liquid needed to fill the hole.
Erosion rates were compared with air temperatures measured
hourly at the Plzeň Mikulka station (Czech Hydrometeorological
Institute, situated 3 km east of the overhang, altitude of
370 m a.s.l.). Total duration of freezing was calculated for each
period. As a second parameter, those continuous freezing
periods exceeding 24 hours were selected, and their duration
was multiplied by the mean temperature during the period.
Resulting numbers were summed over each monitoring period.
This parameter, called FDI (o
C × day) later in the text, accounts
for both the duration and intensity of freezing.
Tensile strength measurement
Cementation cohesion of sandstone can be roughly approximated
by the tensile strength (Bruthans et al., 2012b, 2014).
The tensile strength is an important property showing resistance
of sandstone to disruptive impact of majority of weathering and
erosion processes (Steiger and Asmussen, 2008; Goudie,
2013). The tensile strength of dry and water saturated samples
was measured in the laboratory. Strips of T profile (aluminum
with a 2 by 2 cm surface area) were glued with epoxy to the
sandstone surfaces. After hardening, a tensile force was gradually
increased perpendicular to the sandstone surface until the
sandstone beneath the epoxy failed. The maximum (pull off)
force and effective area was measured (tensiometer) and calculated
in kPa (Bruthans et al., 2014). The dry tensile strength was
measured after drying the sample under laboratory conditions
(25 °C, 50% relative humidity). The saturated tensile strength
was measured by placing the sandstone blocks in a water tank
714 BRUTHANS J. ET AL.
Copyright © 2016 John Wiley & Sons, Ltd.
(with the surface measured facing upward). The block was
slowly saturated adding water, and the tensile strength was
measured after one week of saturation.
Determination of moisture content and drip
water/leachate chemistry
Moisture content is a key factor in many processes of
weathering and deterioration affecting sandstones (Viles and
Goudie, 2007; Mol and Viles, 2010). Moisture content in the
overhang subsurface was measured by the gravimetric method.
The material was weighed in the field, dried at 105 °C, and
weighed again until a stable weight was reached. As the surface
of sandstone outside of the region of the overhang is firm,
and collecting samples would thus result in unacceptable damage;
the moisture content outside of the area of the overhang
was monitored indirectly by suction, and thus its relationship
with moisture content. To monitor the suction, the miniature
tensiometer T5x (UMS Germany) was used. The T5x device is
designed to measure the capillary suction in soils and other porous
materials. The more negative the suction, the lower the
water content (Tindall and Kunkel, 1999). Suction was measured
in drilled holes 5 mm in diameter at depths 20–40 mm
below the sandstone surface. The position of the measuring
points are labeled as a–g in Figures 1C and 1D. The holes were
sealed by plastic foil between individual measurements to prevent
evaporation.
The air temperature and humidity was monitored by sensor
with a datalogger placed inside the overhang between March
3, 2013 and October 5, 2014. The potential evaporation rate
was measured by two plastic cups, 57 mm in diameter, filled
by distilled water, and weighed periodically (E1, E2 in
Figures 1C and 1D). The evaporation surface was protected
with a mesh, 2 mm in diameter, to prevent damage by animals.
The chemical composition of the liquid samples (dripping
water, local spring water, and solid-water leachates) were
analyzed for concentrations of nitrates, chlorides, and sulfates
by high performance liquid chromatography (HPLC) (Dionex
ICS-2000); cations by inductively coupled plasma optical
emission spectrometry (ICP OES) (Thermo Scientific) in the
laboratories of the Institute of Geology of The Czech
Academy of Sciences and the Faculty of Science, Charles
University.
Laboratory frost weathering tests
Frost weathering tests were performed to evaluate the effect of
gravity-induced stress on frost weathering rate. Sandstone
cubes with edge lengths of 40 ± 1 mm were cut from a block
of the collected sandstone using a diamond saw cooled by
water. Susceptibility to frost weathering was tested by repeated
cycles consisting of submerging the cubes in water (+20 °C) for
≥8 hours, followed by placing them in a freezer (À20 °C) for the
same period of time. To remove the loose grains, the cubes
were turned upside-down each 10th cycle for 10 seconds
and submerged again into the water. The cubes were weighed
in each cycle.
Frost weathering tests were done with three unconfined and
three uniaxially loaded cubes. Unconfined cubes were left
bare, touching the base on its lower side. Uniaxially loaded
cubes were compressed by placing square stiff steel plates of
the same size as the cube sides onto the opposite side of the
cube. The plates were compressed against the cube sides by
tightening the nuts on a steel casing constructed around the
cube. The nuts were tightened by a torque screwdriver to
0.75 Nm, which corresponds to a confinement of ~680 kPa,
based on calibration with a tensiometer.
Numerical modeling
Numerical modeling was used to predict stress distribution
within the rock mass of the overhang and its effect on the erosion
rate. PLAXIS 3D geotechnical software was used. Model
dimensions were exactly the same as the dimensions of the
overhang, and the model geometry closely resembled the
geometry of the physical models. The three-dimensional (3D)
shape of the overhang was created by means of photogrammetry.
The finite element mesh of the model consisted of 618 000
tetrahedral finite elements with second-order interpolation of
displacements. Boundary displacements were assigned for the
bottom boundary, zero horizontal displacement in the normal
direction to boundary planes (free displacement in the other
two axes) for left and right vertical model boundaries with the
rock mass, fixed horizontal displacement (free vertical displacement)
for the back vertical boundary with the rock mass,
and free displacement in all three axes for the free superficial
boundary (frontal and upper boundary).
The sandstone was described using a Mohr–Coulomb constitutive
model. Model parameters for weak sandstones were
used: Young modulus of E = 4 GPa, Poisson's ratio nu =0.3, internal
friction angle ϕ = 30°, angle of dilatancy was estimated
as ψ = 15°, cohesion =600 kPa, density as ρ = 2390 (wet)
kg/m3
. The stress state within the physical model was generated
using the gravity loading procedure available in the PLAXIS 3D
software. Stress was generated during finite element simulation.
The initial stress was prescribed as zero alongside of zero
gravity acceleration. During the stress generation stage, the
gravity acceleration was gradually increased to the value of
g = 9.81 m/s2
under the condition of constant density.
Results and Discussion
Solid phase characterization and salt content
The fine fraction (silt and clay) of the sandstone, separated from
the overhang material, is formed of quartz, kaolinite, illite, and
K-feldspar. Dark surficial case hardening zone contains a significant
portion of gypsum and most probably also organic matter
and trapped dust particles (Gorbushina, 2007, and references
cited therein). Gypsum is also the major constituent of all the
efflorescence samples analyzed. Occurrence of efflorescence
is strongly connected to those parts of the case hardening situated
outside of the overhang area in those places shielded from
rain. By contrast, in the overhang itself, the efflorescence is only
present temporarily, due to the fast erosion of the material.
The measured salt content in the overhang samples varied
noticeably during the year between 0.1 and 2 wt% in depths
from 0 to 5 mm (Figure 3A). The salt content abruptly decreased
during November, as the overhang became wet, and salts were
dissolved and leached by seeping water (Figure 3A). Salts
started to accumulate in spring, and reached its maximum
content either in summer or in autumn, depending on the time
when pore water ceased to flow towards the sandstone surface
at a specific part of the overhang. Flakes which fell from the
overhang during the winter showed only low content of salts
(0.06–0.1 wt%).
A tight correlation was found between the concentration of
sulfate and calcium (R2
= 0.88), sulfate and magnesium
(R2
= 0.87), sodium with both chlorides (R2
= 0.67), and nitrates
(R2
= 0.63) in leachates from the overhang. Gypsum was
715QUANTITATIVE STUDY OF A RAPIDLY WEATHERING SANDSTONE OVERHANG
Copyright © 2016 John Wiley & Sons, Ltd.
identified as the main dissoluble mineral based on the ionic
composition of the leachates. Mean gypsum content in the
leachates is ~40% of dissoluble matter. Other potentially occurring
minerals are nitratine (NaNO3) and halite, with the
mean content in the leachates of about 20 and 10%, respectively.
Unlike gypsum, halite and nitratine were not confirmed
by XRD. Both of these phases, especially the unstable nitratine,
may only occur intermittently during the warmest periods of the
day, since the increasing air moisture will dissolve such phases.
Erosion rate of sandstone in the overhang
Seven short-term measurements (0.2–1 day long) of the accumulation
rate of the fallen materials from the overhang were
made between 2008 and 2011. The maximum measured accumulation
rate (60–200 g/m2
/day) occurred in the winter during
an intense thaw (Table I, Figures 2G and 2H). Sandstone flakes
were falling with a frequency of one every several seconds
when the overhang was exposed to direct solar radiation. The
peak intensity of the accumulation rate was estimated to be
5000 g/m2
/day (December 26, 2008). This event lasted just a
few hours, but removed a considerable amount of weak material
from the overhang's ceiling and walls. A much lower accumulation
rate was measured when the air temperature was
slightly below the freezing point (10–20 g/m2
/day). A negligible
accumulation rate was measured prior to the first freezing
(< 1 g/m2
/day). In spring and summer, the accumulation rate
was low (2–10 g/m2
/day).
Between December 26, 2012 and March 14, 2015 the erosion
rate was continuously measured in periods 1–3 months
long. Within this period, three different winters occurred. Winter
2012/2013 was unusually long and cold, with freezing periods
lasting 68 days in total. The longest continuous freezing
period lasted 15 days. In contrast, winters 2013/2014 and
2014/2015 were mild, with freezing periods lasting a total of
34 and 27 days, respectively. The mean accumulation rate
was extremely high during a cold winter (1047 g/m2
/day), while
during a mild winters the mean accumulation rate was only
55–89 g/m2
/day (Table II). In spring, summer, and autumn the
mean accumulation rate did not exceed 8 g/m2
/day. Erosion
rate timings, measured on the fiberglass pegs, correspond to
the accumulation rates of materials on the plastic foil. The
highest erosion was observed between March 3, 2013 and
April 11, 2013 (Table III). The erosion rate in 2013–2015, normalized
to one year, was lowest in the inner part of the overhang
(P1 and P2, with a rate of 3 to 4 mm/yr), but increasing
outwards to 31 mm/yr at P5.
All the measurements clearly demonstrated that 87–99% of
the material is accumulated during the winter period, and that
its' falling is connected to the thaw after freezing periods. A
Figure 3. Temporal changes of moisture and salt content in overhang:
(A) salt content; (B) moisture content; (C) suction.
Table I. Accumulation and erosion rates based on materials collected on plastic foils over short time periods
Start time of accumulation
rate measurement
Period of
measurement (days)
Accumulation
area (m
2
)
Accumulation
rate (g/m
2
/day) Climatic conditions
December 23, 2009 16:10 1.0 2.6 214 Cloudy, +3 °C; 24 hours after onset of thaw
December 25, 2010 14:00 0.9 4.5 62 Freezing in shade, but melting on insolated places
December 26, 2008 18:00 0.6 4.8 17 Cloudy; À1 °C
August 15, 2009 10:00 1.0 5.1 11 Clear sky, 25 °C
December 27, 2008 09:00 0.2 4.8 10 Cloudy; À1 °C
May 1, 2009 12:00 1.0 5.0 2.1 —
December 24, 2011 12:00 1.0 5.0 0.1 5 °C, no freezing yet
716 BRUTHANS J. ET AL.
Copyright © 2016 John Wiley & Sons, Ltd.
cold winter with long freezing periods produces ~10 times
more accumulated material than a mild winter with limited
freezing (Table II). Considering the sandstone dry density of
2.3 g/cm3
, and average accumulation rate of 169 g/m2
/day,
the mean erosion rate of the protruding part of the sandstone
exposure is 14 mm/yr (Table II). However, at P6 and P7, where
no protruding sandstone mass occurs, the erosion rate did not
exceed 2 mm/yr (Table III).
The density of the debris pile below the overhang is 1.6 g/
cm3
. As the original sandstone density is 2.3 g/cm3
, the 1.8 m
high pile of debris was derived from a 1.25 m thick zone of
sandstone. Thus the sandstone in the overhang has been
eroded at a mean rate of 36 mm/yr over a period of 35 years.
The long-term erosion rate is 2.6 times higher than the rate
measured on foil 1 during 2013–2015, which is in agreement
with the statement of a local resident that in the past the erosion
rate was even greater.
The measured erosion rate is much higher than the reported
erosion rates on sandstone exposures from various locations:
0.6–9 mm/yr in different coastal areas (Gill, 1981; Mottershead,
1994; Sunamura, 1996), 0–0.35 mm/yr in an arid climate
measured on the sandstone tombs at the Petra Monument
(Paradise, 2005; Heinrichs, 2008), or approximately
0.3 mm mm/yr in a mountainous precipitation-rich area of
Japan (Imaizumi et al., 2015).
Tensile strength
Tensile strength was measured on larger sandstone flakes
having fallen from the overhang during the winter (Table IV).
Most of the measurements were done on the flat sides of the
flakes (A flakes); other measurements were also done in a
direction perpendicular to the flat sides of the flakes (B flakes).
A large sandstone block sampled outside of overhang area was
measured to determine whether the materials of the flakes are
weakened compared to the sandstone in the vicinity of the
overhang. Results showed that the mean dry tensile strengths
of both flakes and block are relatively low (120–270 kPa).
Table III. Erosion in given time period (in millimeters), and in the last row the average erosion rate (in mm/yr) measured at the plastic pegs
Period From To
Freeze
duration (days)
FDI
(°C × day) P1 P2 P3 P4 P5 P6b
P7b
Cold winter March 3, 2013 April 11, 2013 17 (68a
) –37 (À167a
) 3 2 10 26 47 — —
Warmer period April 11, 2013 October 28, 2013 0 0 ≤2 ≤2 ≤2 ≤2 ≤2 — —
Mild winter October 28, 2013 March 8, 2014 34 –34 3 ≤2 ≤2 ≤2 ≤2 ≤2 ≤2
Warmer period March 8, 2014 October 5, 2014 2 0 ≤2 ≤2 ≤2 ≤2 ≤2 ≤2 ≤2
Mild winter October 5, 2014 March 14, 2015 27 –35 ≤2 ≤2 19 ≤2 16 ≤2 ≤2
Total March 3, 2013 March 14, 2015 81 –106 8 7 34 26 63 ≤2 ≤2
Normalized to one year 40 –53 4 3 17 13 31 ≤2 ≤2
Note: FDI, factor of frost duration and intensity.
a
Entire winter.
b
Placed December 2, 2013.
Table II. Accumulation and erosion rates based on materials collected on plastic foils over long time periods
Period From To
Foil 1
(g/m
2
/day)
Foil 1, mean
(g/m
2
/day)
Erosion rate
foil 1 (mm/yr)
Foil 2
(g/m
2
/day)
Cold winter December 26, 2012 March 3, 2013 1070 93 —
March 3, 2013 April 11, 2013 1020 1047 88 —
Warmer period April 11, 2013 May 21, 2013 <26 <2.3 —
May 21, 2013 July 8, 2013 <4 <0.3 —
July 8, 2013 August 22, 2013 0 0 —
August 22, 2013 October 28, 2013 1.2 <8 0.1 —
Mild winter October 28, 2013 December 2, 2013 39 3.4 —
December 2, 2013 December 25, 2013 24 2.1 8
December 25, 2013 February 1, 2014 103 8.9 90
February 1, 2014 March 8, 2014 54 55 4.7 54
Warmer period March 8, 2014 May 11, 2014 0 0 0
May 11, 2014 July 2, 2014 0 0 0
July 2, 2014 August 16, 2014 26 2.3 1.6
August 16, 2014 October 5, 2014 0 7 0 0
Mild winter October 5, 2014 December 25, 2014 11 0.9 2.3
December 25, 2014 March 14, 2015 168 89 14 146
Mean December 26, 2012–December 25, 2013 315 27
Mean December 26, 2013 À December 25, 2014 22 1.8
Mean December 26, 2012–December 25, 2014 169 14
Table IV. Tensile strength of sandstone in overhang (flakes), and
outside of the overhang area (block)
Surface type Tensile strength
Mean
value (kPa) SD (kPa) n
Sandstone block Dry 266 28 9
Flakes A Dry 194 105 17
Flakes B Dry 122 38 14
Flakes A Saturated 27 20 9
Note: n, number of measurements; SD, standard deviation.
717QUANTITATIVE STUDY OF A RAPIDLY WEATHERING SANDSTONE OVERHANG
Copyright © 2016 John Wiley & Sons, Ltd.
Flakes A showed nearly the same dry tensile strengths as the
sandstone block, while flakes B had dry tensile strengths only
50% of flakes A. A drastic decrease of tensile strength was observed
when the material was saturated. The mean saturated
tensile strength of flakes A dropped to 27 kPa, which is nearly
one order of magnitude less than the same material under dry
conditions (Table IV).
Moisture content and drip water chemistry
Rock moisture content in the overhang subsurface (0–5 mm
depth) varies over time. The highest values (12–18 wt%) were
measured in winter, while values of 1–5 wt% were measured
during late spring and summer. High moisture values in winter
correspond well with water dripping from the overhang ceiling
and the presence of icicles when the temperature falls below
zero. The largest variations in moisture content were determined
in the outer parts of the overhang (spots P6, P7); and
lowest in the innermost part (P1). This probably can be explained
by differential evaporation rates in the two parts
(Figure 3B).
High variations of the moisture content at a depth of 20 to
40 mm below the sandstone surface of the overhang are visible
in the oscillation of the suction data (Figure 3C). The suction in
the overhang was close to zero during winter, proving that the
pore space was nearly fully saturated with water. As evaporation
dominated during the summer, moisture was lost from
the pore spaces, and suction reached up to 90 kPa in the outer
parts of the overhang (Figure 3C, point d). In the inner part of
overhang, the suction reaches between 40 and 60 kPa in summer
(Figure 3C, points a and b). Suction measured outside the
overhang area clearly showed low moisture content in the pore
space in winter (35–45 kPa). Variation of the suction was much
lower outside the overhang area. Based on the relationship between
moisture content and suction, a moisture content of
40 kPa occurring in the winter outside of the overhang area is
equal to a moisture content of ~5wt%. (Figure 4).
The mean potential evaporation rate in the innermost part of
the overhang is up to 13 mm/month in late spring and early
summer, and about ~5 mm/month in late autumn and winter
(mean annual rate ~ 6–7 mm/month), based on plastic cups
with water (Table V). In the outer parts of the overhang, the
evaporation rate will be considerably higher. The mean annual
air temperature and relative humidity at the overhang is 11.6 °C
and 78%, respectively.
Drip water in the western side of the overhang was sampled
in 2009, 2011, and 2015. TDS were markedly higher in the drip
water than in the local groundwater (1007–1152 mg/l, and
276 mg/l respectively), pH was neutral (7.1–7.5). Furthermore,
Ca2+
, Na+
, and K+
are the major cations determined, and
NO3
À
, SO4
2À
, and ClÀ
the major anions (Table VI). Concentrations
of Na+
, K+
, ClÀ
, NO3
À
, and boron are considerably higher
in the drip water than in the local groundwater, which indicates
that drip water is strongly affected by leakage from a sewage
reservoir; therefore, the wetting is artificial in origin. Based on
the solubility of gypsum (2.2–2.6 g/l in the temperature range
0–25 °C), and Ca2+
and SO4
2À
concentrations in drip water,
the evaporation of 80 to 90% of the pore water in the overhang
is necessary to initiate precipitation of gypsum.
Factors responsible for sandstone decay
The erosion rate of the overhang was compared with the: (a)
freezing duration, (b) freezing duration and intensity (FDI); (c)
moisture content; and (d) salt content. There is a close relationship
between the erosion rate and both the freezing duration
Figure 4. Relationship between moisture content and suction obtained
from the overhang. The measured data fall within the region between
the dashed lines.
Table V. Potential evaporation rate measured at locations E1 and E2
From To Number of days E1 (mm/month) E2 (mm/month)
May 11, 2014 May 24, 2014 13 10 8
May 24, 2014 July 2, 2014 39 13 12
July 2, 2014 August 16, 2014 45 6 10
August 16, 2014 October 5, 2014 50 6 7
October 5, 2014 December 25, 2014 81 5 5
December 25, 2014 March 14, 2015 79 5 5
Table VI. Concentration in drip water from overhang and local
groundwater (in mg/l)
Ion
Overhang drip
December 24,
2009
Overhang drip
December 25,
2011
Overhang drip
March 14,
2015
Čumperka
Spring
(Radčice)
Cl
À
179 198 128 20.0
SO4
2À
240 200 162 38.0
NO3
À
318 309 417 139
HCO3
À
À 66 À À
Ca
2+
220 132 160 42
Mg
2+
22 18 21 14
Na
+
98 89 81 6,0
K
+
62 53 47 11
Si 12 8 10.0 6.0
Fe 0.04 0.02 0.033 <0.005
Al <0.05 >0.07 À À
B À À 0.13 <0.005
718 BRUTHANS J. ET AL.
Copyright © 2016 John Wiley & Sons, Ltd.
and FDI. However, no relationship was found between the erosion
rate and the moisture or salt content (Figure 5).
Based on our measurements and observations, the major factor
responsible for the rapid erosion of the sandstone overhang
is clearly the strong wetting of the sandstone pore space by
seeping water in the wintertime (up to 18 wt% in the overhang,
in contrast to ~5 wt% outside the overhang area). High moisture
content in the pores can facilitate decay of the sandstone
in several ways. First, the compressive and tensile strength decreases
greatly when sandstone is wetted (cf. also Lin et al.,
2005; Bruthans et al., 2012b, 2014). Weaker sandstone fails
under lower disruptive pressures (caused by weathering) than
does dry sandstone. Second, repeated wetting and drying is
capable of deteriorating sandstone (Ruedrich et al., 2011).
Third, frost weathering strongly increases with the degree of
pore saturation (Ruedrich and Siegesmund, 2006). In winter,
the overhang pore space contains much more water than sandstone
exposures in the nearby surroundings (Figure 3C).
The perfect match between the freeze–thaw periods and the
erosion rate (Tables II and III; Figures 5A and 5B) strongly indicates
that frost weathering is the major (if not the only) important
weathering process operating at the studied site. The ice
acts as a temporary cement, falling down of flakes is delayed
until the next thaw. The highest falling intensity observed immediately
after long freezing periods supports this idea. In accordance,
Stark (1989) observed that longer freeze periods
result in much greater destruction than the more frequent but
shorter periods.
The susceptibility to frost weathering was tested with repeated
cycles consisting of submerging the sandstone cubes
in distilled water, followed by placing them in a freezer.
Unconfined cubes lost 70% of their original weight within
62–153 cycles [mean and standard deviation (SD) 102 ± 38 cycles].
However, uniaxially loaded cubes lost 70% of their
original weight after 316–384 cycles (mean and SD 347
± 28 cycles; Figure 6). Thus, the frost weathering rate was 2–6
times higher than in the case of unconfined cubes. These results
clearly document that gravity loading stress significantly affects
the intensity of the frost weathering. All uniaxially loaded cubes
weathered into pillars, which is also an indication of the stress
effect on weathering (Figure 7A, Bruthans et al., 2014).
The number of weathering cycles needed to fully disintegrate
sandstone even without any load (78–187) seems to be
relatively high. Nevertheless, the number of frost cycles may,
in reality, be lower as the material in the overhang does not
disintegrate completely, but decomposes into flakes several
centimeters thick. Flakes of such a size cannot be replicated
on the small cubes used in the laboratory; however, small-scale
flaking was commonly observed on the cubes (Figure 7B).
Based on our monitoring (see Figure 3A) the gypsum and
other salts were leached from the sandstone overhang prior to
the winter period. Therefore, low salt content most probably
did not affect the frost weathering. In addition, the engraving
‘1915’, which is covered by gypsum efflorescence, shows no
retreat after 100 years. In light of these observations, it is unlikely
that gypsum/salt crystallization under low moisture conditions
(~5 wt%) plays an important role in the rapid decay of
the studied sandstone. However, gypsum and potentially occurring
halite and nitratine (or other salts of Na, K-Cl, NO3)
may possibly play some role in weakening of the sandstone
during the spring and summer, when evaporation dominates
and salts crystallize in the sandstone pore spaces. Both halite
and nitratine are deliquescent, so these salts dissolve in atmospheric
moisture if the relative humidity exceeds 68–76%, dependent
on their proportions (for 22 °C, Gupta et al., 2015).
The measurements show that the relative humidity in the overhang
during the period October 1, 2013 to September 30, 2014
Figure 5. Relationship between erosion rate and various factors: (A)
total duration of freezing; (B) absolute value of FDI – frost duration
and intensity (sum of multiples of length of freezing periods and their
mean temperature); (C) mean moisture content; (D) mean salt content.
Erosion rates are calculated from accumulation on foil 1.
719QUANTITATIVE STUDY OF A RAPIDLY WEATHERING SANDSTONE OVERHANG
Copyright © 2016 John Wiley & Sons, Ltd.
decreased below the 68 and 76% thresholds 175 and 201
times per year, respectively. During each such event, the halite
and/or nitratine may crystallize, and dissolve once again after
the event, which would considerably increase potential sandstone
damage. Nitratine is reported to be 10 to 100 times more
effective in degradation than gypsum (Zehnder, 1996). Halite
was found to be two times more destructive than gypsum (Robinson
and Williams, 2000). Additionally, the combination of
gypsum with halite, and more generally with any deliquescent
salt, is appreciably more destructive than either of the salts
alone (Robinson and Williams, 2000; Charola et al., 2007).
Nevertheless, salt weathering cannot be a major weathering
factor, otherwise the accumulation of the materials on the foil
would occur during the first wetting of the sandstone after the
summer period, and not after the thaw. The same consideration
may be applied for weathering by repeated wetting and drying
as well as by biodegradation.
Effects of stress on erosion rate
The erosion rate measured at the fiberglass pegs was compared
with the calculated stress in the sandstone subsurface (Figure 8).
The highest erosion rate (13–31 mm/yr, P3–P5) corresponds to
the overhanging surface with a very low compressive stress,
and even tension (stress À10 to +50 kPa). However, a low erosion
rate (< 2 mm/yr) was measured at a vertical wall, with
compressive stress exceeding 150 kPa (high stress, P6, P7). An
erosion rate of 3 to 4 mm/yr was measured at the overhanging
Figure 6. Frost weathering of loaded (L) and unconfined (U) cubes. Loaded cubes survived many more weathering cycles than unconfined cubes.
Figure 7. Laboratory frost weathering of 4 cm cubes (A) weathered
into pillar, (B) small scale flaking of pillar surface. [Colour figure can
be viewed at wileyonlinelibrary.com]
Figure 8. Numerical models showing principal stress magnitude on sections from Figure 1C. (A) Section A–B is directed across the overhang; (B)
section C–D is situated across the cliff outside of the overhang area. Note that the erosion rate is decreasing with increasing stress. [Colour figure
can be viewed at wileyonlinelibrary.com]
720 BRUTHANS J. ET AL.
Copyright © 2016 John Wiley & Sons, Ltd.
surface with a stress of 70 to 200 kPa (P1, P2). The erosion rate
is thus affected by stress in the rock subsurface. Zones with low
or even negative stress (tension) erode at the fastest rate; while
zones with the stress exceeding 70 kPa erode at a slower rate.
This is in agreement with the laboratory frost weathering experiment
with the sandstone cubes, where the stress also decreased
the weathering rate (Figure 6).
The magnitude of stress in the overhang subsurface is dictated
by its geometry, including the overburden. The central
part of overhang ceiling acts as a large protrusion (Figure 1C).
For protruding parts of the sandstone massive, lower stresses
are typical (Bruthans et al., 2014). The protruding shape of this
particular overhang is therefore strongly increasing the effect of
the frost weathering. Sandstone landforms are commonly
stress-controlled; this includes the overhangs lack of protruding
parts, as these would have been quickly eroded (Bruthans et al.,
2014; Rihosek et al., 2016). It is possible that the studied overhang's
geometry was changed in the past during the building of
a road, and the protrusion is a recent result of cutting of the
western side of the overhang, which originally had a fully buttressed
ceiling.
Broader consequences
Dramatic differences in the erosion rate, dependent on sandstone
moisture content, can help to elucidate the development
of overhangs and caves during the Holocene in the studied
sandstone area (Central Europe), but also in sandstone terrains
in arid temperate climate (e.g. Colorado Plateau, USA).
Radiocarbon dating of sedimentary infillings at archaeological
sites of sandstone overhangs and shallow caves in the northwest
part of the Bohemian Cretaceous Basin has indicated that
walls of these overhangs and caves recessed less than
2–5 mm/1000 years after 8–11 kyr BP (Cílek, 2007; Cílek and
Žák, 2007). Radiocarbon and U-series dating of secondary
carbonates demonstrated that the walls and ceilings in the
north part of the Bohemian Cretaceous Basin have not recessed
more than 1 mm/1000 years during the last 8 kyr BP (Bruthans
et al., 2012a). Similarly, the radiocarbon dating of ‘fossil dung’
in sandstone overhangs and caves from the Colorado Plateau
have also shown that these sites were created prior to 13 kyr
BP (Jennings, 1980; Davis et al., 1984), and that they were not
enlarged in the Holocene, except for occasional block collapses
of the ceilings. It is likely that neither salt weathering
nor biota can effectively disintegrate sandstone inside relatively
deep and dark cavities, and that frost weathering is the most
significant factor.
Our study demonstrated that sandstone, the surface of which
does not measurably retreat under the present climate
(rate < 0.01 mm/yr), can retreat at a rate of ~18–36 mm/yr if
the pore space is nearly saturated. Similarly, increased groundwater
seepage in sandstone areas can be expected in the Last
Glacial period, when low temperatures greatly decreased the
evapotranspiration rate. More water was left for surface flow
and infiltration, as demonstrated by the high level of lakes in
closed basins (Licciardi, 2001). Overhangs in Central Europe
and the Colorado Plateau might thus be a result of rapid frost
weathering of nearly saturated sandstone during the Last
Glacial period. Seeping groundwater and possibly wind
could excavate the fallen material from overhangs, and explain
why there is nearly no material dated prior to 12 kyr BP in
sandstone caves in both regions (Bruthans et al., 2012b).
The erosion of overhangs and caves followed the critical
partings (Oberlander, 1977), such as clay or shale layers.
The erosion of critical partings is not affected by stress, and
as a consequence, the stress was not effective in decreasing
the weathering rate (Bruthans et al., 2014). Enlargement was
strongly reduced after a change of climate in the Holocene,
when relatively dry caves and overhangs became traps for sediments
(cf. Cílek and Žák, 2007, p. 138).
Similar drastic changes in the erosion rate can be expected in
many other situations where the climatic changes resulted in
changes of the sandstone moisture content. At rare places with
high moisture content the frost weathering can be relatively fast
at the present time. In Adršpach (northeast Bohemian Cretaceous
Basin) the frost weathering up to 240 g/m2
/yr was measured
below strongly wetted sandstone blocks (unpublished).
This observation does not rule out the possibility of locally relatively
fast weathering/erosion by increased content of aggressive
salts at the present time.
Conclusion
Characterization of those processes controlling weathering
and erosion of the specific sandstone overhang near Plzeň
joins together valuable data of general significance. Long-term
monitoring of the parameters important for weathering was
performed, and data were compared against the relevant published
information. The overhang ceiling is being eroded at a
mean rate of 14 mm/yr, which is unusually high compared to
other measured sites elsewhere. It was revealed that 87–99%
of the erosion occurs during melting after periods of freezing.
Frost weathering combined with wetting weakening play
major roles in the degradation of the studied sandstone
overhang. Salt weathering might possibly contribute to some
extent. In addition, the erosion rate is strongly affected by
the geometry of the overhang and overburden, which controls
the stress field in the sandstone massive. With increasing stress
the erosion rate decreases as clearly demonstrated both on
laboratory experiments and in the field. Exposure geometry
and stress should thus be considered as important factors in
weathering and erosion studies concerning sandstones (cf.
Bruthans et al., 2014).
The salt content in the overhang varies over the year,
with the minimum in the winter, and maximum in summer or
autumn. Variation in salt content inversely correlates with
the oscillation of moisture content. This oscillation is adequately
measurable by suction, which can therefore be used
for monitoring the moisture content, with minimum damage
of the rock. The results clearly show that a strong increase of
the moisture content in the pores (12–18 wt% in winter) increases
sandstone weathering and erosion by nearly four
orders of magnitude, compared to the same material under
natural moisture content (5 wt%). This finding suggests that
sandstone landforms may potentially develop very rapidly if
the pore space is nearly saturated by water; yet remain stable
in later phases when the moisture content decreases below
the critical threshold. We believe that overhangs in Central
Europe, but also in other regions like in the Colorado Plateau
(USA), might be the result of rapid frost weathering of nearly
saturated sandstone during the Last Glacial period; and that
the ‘present’ state is long-standing, until the next significant
change of climate.
721QUANTITATIVE STUDY OF A RAPIDLY WEATHERING SANDSTONE OVERHANG
Copyright © 2016 John Wiley & Sons, Ltd.
reviews of two anonymous referees and also help of editors in improving
the manuscript. Hourly air temperatures were provided by the
Czech Hydrometeorological Institute.
References
ASTM. 2004. C914–95: Standard Test Method for Bulk Density and Volume
of Solid Refractories by Wax Immersion. West Conshohocken,
PA: ASTM International; 188–190.
Bruthans J, Schweigstillová J, Jenč P, Churáčková Z, Bezdička P. 2012a.
14
C and U–series dating of speleothems in the Bohemian Paradise
(Czech Republic): retreat rates of sandstone cave walls and implications
for cave origin. Acta Geodynamica et Geomaterialia 9:
93–108.
Bruthans J, Světlík D, Soukup J, Schweigstillová J, Válek J,
Sedlackova M, Mayo AL. 2012b. Fast evolving conduits in claybonded
sandstone: Characterization, erosion processes and significance
for origin of sandstone landforms. Geomorphology
177–178: 178–193.
Bruthans J, Soukup J, Vaculiková J, Filippi M, Schweigstillova J, Mayo
AL, Masín D, Kletetschka G, Rihosek J. 2014. Sandstone landforms
shaped by negative feedback between stress and erosion. Nature
Geoscience 7: 597–601.
Charola AE, Pühringer J, Steiger M. 2007. Gypsum: a review of its role
in the deterioration of building materials. Environmental Geology 52:
207–220.
Cílek V, Žák K. 2007. Late Glacial and Holocene sedimentation under
sandstone rock shelters of Northern Bohemia (Czech Republic). In
Sandstone Landscapes, Härtel H, Cílek V, Herben T, Jackson A, Williams
R (eds). Academia: Praha; 133–138.
Cílek V. 2007. Climate, microclimate and paleoclimate of sandstone
areas of Central and Northern Bohemia (Czech Republic). In Sandstone
Landscapes, Härtel H, Cílek V, Herben T, Jackson A, Williams
R (eds). Academia: Praha; 97–103.
Davis OK, Agenbroad L, Martin PS, Mead JI. 1984. The pleistocene
dung blanket of Becham Cave, Utah. In Contribution in Vertebrate
Paleontology: Volume in Memorial to John E. Guilday, Genoways
HH, Dawson MR (eds). Carnegie Museum Natural History Special
Publication 8. Carnegie Museum: Pittsburgh, PA; 267–282.
Dorn RI, Gordon SJ, Allen CD, Cerveny N, Dixon JC, Groom KM, Hall
K, Harrison E, Mol L, Paradise TR, Sumner P, Thomson T, Turkington
AV. 2013. The role of fieldwork in rock decay research: case studies
from the fringe. Geomorphology 200: 59–74.
Dragovich D. 1967. Flaking, a weathering process operating on cavernous
rock surfaces. Geological Society of America Bulletin 78:
801–804.
Gill ED. 1981. Rapid honeycomb weathering (tafoni formation) in
Graywacke, S.E. Australia. Earth Processes and Landforms 6: 81–83.
Gorbushina AA. 2007. Life on the rocks. Environmental Microbiology 9
(7): 1613–1631.
Goudie AS. 2013. Arid and Semi-arid Geomorphology. Cambridge University
Press: Cambridge; 454 pp.
Gupta D, Kim H, Park G, Li X, Eom JH, Ro CU. 2015. Hygroscopic
properties on NaCl and NaNO3 mixture particles as reacted inorganic
sea-salt aerosol surrogates. Atmospheric Chemistry and Physics
15: 3379–3393.
Heinrichs K. 2008. Diagnosis of weathering damage on rock-cut monuments
in Petra, Jordan. Environmental Geology 56: 653–675.
HighScorePlus. 2011. version 3.0d, PANalytical b.v., Almelo, NL.
Huinink HP, Pel L, Kopinga K. 2004. Simulating the growth of tafoni.
Earth Surface Processes and Landforms 29: 1225–1233.
Imaizumi F, Nishii R, Murakami W, Daimaru H. 2015. Parallel retreat of
rock slopes underlain by alternation of strata. Geomorphology 238:
27–36.
JCPDS PDF2. 2004. Sets 1-54. International Centre for Diffraction Data:
Newtown, PA.
Jennings JD. 1980. Cowboy Cave, University of Utah Anthro;pological
Papers 104. University of Utah Press: Salt Lake City, UT.
Licciardi JM. 2001. Chronology of latest Pleistocene lake-level fluctuations
in the pluvial Lake Chewaucan basin, Oregon, USA. Journal of
Quaternary Science 16(6): 545–553.
Lin ML, Jeng FS, Tsai LS, Huang TH. 2005. Wetting weakening of tertiary
sandstones-microscopic mechanism. Environmental Geology
48: 265–275.
Mol L. 2014. Investigations into the relationship between changes in internal
moisture regimes and rock surface deterioration in cavernous
sandstone features. Earth Surface Processes and Landforms 39:
914–927.
Mol L, Viles HA. 2010. Geoelectric investigations into sandstone moisture
regimes: Implications for rock weathering and the deterioration
of San Rock Art in the Golden Gate Reserve, South Africa. Geomorphology
118: 280–287.
Mol L, Viles HA. 2012. The role of rock surface hardness and internal
moisture in tafoni development in sandstone. Earth Surface Processes
and Landforms 37: 301–314.
Mottershead DN. 1994. Spatial variations in intensity of alveolar
weathering of dated sandstone structure in a coastal environment,
Weston Super mare, UK. In Rock Weathering and Landform Evolution,
Robinson DA, Williams RBG (eds). John Wiley & Sons: Chichester;
151–174.
Oberlander TM. 1977. Origin of segmented cliffs in massive sandstones
of southeastern Utah. In Geomorphology of Arid Regions, Doehring
DO (ed). Allen and Unwin: Boston, MA; 79–114.
Opluštil S, Martínek K, Tasáriová Z. 2005. Facies and architectural analysis
of fluvial deposits of the Nýřany Member and the Týnec Formation
(Westphalian D-Barruelian) in the Kladno-Rakovník and Pilsen
basins. Bulletin of Geosciences 80(1): 45–66.
Paradise TR. 2005. Petra revised: An examination of sandstone
weathering research in Petra, Jordan. Geological Society of
America, Special Paper 390: 39–49.
Rihosek J, Bruthans J, Masin D, Filippi M, Carling GT, Schweigstillova J.
2016. Gravity-induced stress as a factor reducing decay of sandstone
monuments in Petra, Jordan. Journal of Cultural Heritage 19:
415–425.
Robinson DA, Williams RBG. 2000. Experimental weathering of sandstone
by combinations of salts. Earth Surface Processes and Landforms
25: 1309–1315.
Ruedrich J, Siegesmund S. 2006. Fabric dependence of length change
behaviour induced by ice crystallization in the pore space of natural
building stones. In Heritage, Weathering and Conservation, Fort R,
Alvarez de Vuegro M, Gomez-Heras M, Vazquez-Calco C (eds). Taylor
& Francis Group: London; 497–505.
Ruedrich J, Siegesmund S. 2007. Salt and ice crystallisation in porous
sandstones. Environmental Geology 52: 225–249.
Ruedrich J, Bartelsen T, Dohrmann R, Siegesmund S. 2011. Moisture
expansion as a deterioration factor for sandstone used in buildings.
Environment and Earth Science 63: 1545–1564.
Siegesmund S, Snethlage R. 2011. Stone in Architecture: Properties, Durability.
Springer Verlag: Berlin.
Stark D. 1989. Effect of length of freezing period on durability of concrete.
PCA Research and Development Bulletin RD096.01T. Portland
Cement Association: Skokie, IL: 9.
Steiger M, Asmussen S. 2008. Crystallization of sodium sulphate phases
in porous materials: the phase diagram Na2SO4·H2O and generation
of stress. Geochimica et Cosmochimica Acta 72: 4291–4306.
Steiger M, Charola AE, Sterflinger K. 2011. Weathering and deterioration.
In Stone in Architecture: Properties, Durability, Siegesmund S,
Snethlage R (eds). Springer Verlag: Berlin; 227–316.
Sunamura T. 1996. A physical model for the rate of coastal tafoni development.
Journal of Geology 104: 741–748.
Tindall JA, Kunkel JR. 1999. Unsaturated Zone Hydrology for Scientists
and Engineers. Prentice Hall: Englewood Cliffs, NJ.
Turkington AV, Paradise TR. 2005. Sandstone weathering: a century of
research and innovation. Geomorphology 67: 229–253.
Viles HA. 2001. Scale issues in weathering studies. Geomorphology 41:
63–72.
Viles HA. 2013. Linking weathering and rock slope instability: nonlinear
perspectives. Earth Surface Processes and Landforms 38:
62–70.
Viles and Goudie. 2007. Rapid salt-weathering in the coastal Namib
desert: implications for landscape development. Geomorphology
85(1–2): 49–62.
Vítek J. 1987. Pseudokrasové tvary v karbonských sedimentech sz. od
Plzně. Český kras 38: 125–127 .in Czech
722 BRUTHANS J. ET AL.
Copyright © 2016 John Wiley & Sons, Ltd.
Warke PA, McKinley J, Smith BJ. 2006. Variable weathering response in
sandstone: Factors controlling decay sequences. Earth Surface Processes
and Landforms 31: 715–735.
Williams RBG, Robinson DA. 2001. Experimental frost weathering of
sandstone by various combinations of salts. Earth Surface Processes
and Landforms 26: 811–818.
Young RW, Wray AL, Young ARM. 2009. Sandstone Landforms. Cambridge
University Press: Cambridge.
Zehnder K. 1996. Gypsum efflorescences in the zone of rising damp. Monitoring
of slow decay processes caused by crystallizing salts on wall paintings.
In Proceedings, 8th International Congress on Deterioration on
Conservation of Stone, Riederer J (ed). Möller Druck: Berlin; 1669–1678.
723