Bank erosion history of a mountain stream determined by means of
anatomical changes in exposed tree roots over the last 100 years
(Bílá Opava River -- Czech Republic)
Ireneusz Malik , Marcin Matyja
Faculty of Earth Sciences, University of Silesia, Będzińska St. 60, Sosnowiec, 41-200, Poland
Received 19 January 2006; received in revised form 4 January 2007; accepted 21 February 2007
Available online 13 May 2007
Abstract
The date of exposure of spruce roots as a result of bank erosion was investigated on the Bílá Opava River in the northeastern
Czech Republic. Following the exposure of roots, wood cells in the tree rings divide into early wood and late wood. Root cells
within the tree rings also become smaller and more numerous. These processes permit dating of the erosion episodes in which roots
were exposed. Sixty root samples were taken from seven sampling sites selected on two riverbed reaches. The results of root
exposure dating were compared to historical data on hydrological flooding. Using the root exposure dating method, several erosion
episodes were recorded for the last 100 years. The greatest bank erosion was recorded as consequence of an extraordinary flood in
July 1997. In the upper, rocky part of the valley studied, bank erosion often took place during large floods that occurred in the early
20th century. In the lower, alluvial part of the valley, erosion in the exposed roots was recorded only in 1973 and has been intensive
ever since. It is suggested that banks in the lower part are more frequently undercut, which leads to the falling of trees within whose
roots older erosion episodes were recorded. Locally, bank erosion is often intensified by the position of 1- to 2-m boulders in the
riverbed, which direct water into the parts of the banks where erosion occurs. Selective bank erosion could be intensified by debris
dams and hillslope material supply to the riverbed.
 2007 Elsevier B.V. All rights reserved.
Keywords: Bank erosion; Mountain stream; Dendrogeomorphology; Exposed roots
1. Introduction
The rate of bank erosion in mountain streams mainly
depends on the bank material erodibility and the history
of fluvial processes, particularly the character and
frequency of high water episodes (Gregory and Walling,
1973; Krzemień, 1976; Thorne, 1982; Starkel, 2002).
What is also of considerable importance for the erosion
of banks composed of loose rocks is the activity of
needle ice and ice floats, as well as the nature of the
riparian vegetation (Klimek, 1989). Banks strengthened
by root systems are more resistant to washing out and
less undercut as compared to those without vegetation
cover (Sttot, 1997; Rowntree and Dollar, 1999,
Abernethy and Rutherfurd, 2000). Banks not strengthened
by tree roots are by half more erodible than banks
with root systems. Roots of riparian trees protect banks
against erosion more effectively than species inhabiting
non-riparian zones (Pollen and Simon, 2005).
Available online at www.sciencedirect.com
Geomorphology 98 (2008) 126­142
www.elsevier.com/locate/geomorph
 Corresponding author. Tel.: +48 32 2918381; fax: +48 32 1915865.
E-mail address: irekgeo@wp.pl (I. Malik).
0169-555X/$ - see front matter  2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2007.02.030
In order to define the history of bank erosion in
riverbeds of mountain streams, a number of field methods
are used (Thorne, 1981). Erosion pins and geodetic
methods, e.g., repeated riverbed cross-section (Harvey,
1974; Lawler, 1978; Hooke, 1980; Lane et al., 1994) are
used in short, several-year-long, study periods. Another
way of estimating bank erosion intensity is the analysis
of riverbed morphologies on maps and air photos of
various ages (Trafas, 1975). However, the method is
applied most frequently for medium-sized and large
rivers with a riverbed at least several meters wide. Also,
photo-electronic erosion pins (PEEP) can be applied,
these based on photographs of the bank taken automatically
during erosion episodes (Lawler, 1991, 2005).
The analysis of anatomical changes within the wood
of trees growing on the undercut levels allows flood
frequency, and indirectly bank erosion rate, to be reconstructed
(Sigafoos, 1964; Hupp and Osterkamp, 1996).
A lot of information may be provided by the analysis
of coarse woody debris (CWD) lying in situ under
erosional undercuts (Malik, 2005, 2006). However, high
discharges occur in mountain streams and CWD is most
often redeposited, which prevents the drawing of
conclusions about bank erosion intensity at particular
sites (Wyżga and Zawiejska, 2005). It seems that the
exposed roots of trees overgrowing rock-and-regolith
banks, floodplains, and terraces, are a much more valuable
material for the analysis of bank erosion rate in
mountain streams.
Dendrogeomorphological studies focused on root
exposure dating as a tool for calculating erosion rate are
few. LaMarche (1966) found a correlation between the
distance from the exposed root to actual soil level and
by age of the root and used this information to calculate
the slope erosion rate. Alestalo (1971) noted the morphological
changes in the wood anatomy of roots occurring
in reaction to changes from aerial to subaerial
environments or vice versa. During successive tree root
exposure, ring reduction (suppression) often occurs,
sometimes masked by release resulting from the elimination
of competition from nearby trees (Alestalo, 1971).
Shroder (1980) and Strunk (1997) have also described
the effect of sediment burial on the production of adventitious
roots. The age of the adventitious roots enables
the year in which the deposition occurred to be identified.
Different event types occur within root cross-sections,
such as width, density, reaction wood, and corrasion scars
(Shroder, 1978, 1980; Shroder and Butler, 1987). Those
types of events can provide information about erosion
intensity.
Carrara and Carroll (1979) also calculated erosion
rates by examining exposed roots. They used tree root
anatomical features to identify the years in which
erosion occurred. The study was based on the time of
initial cambium dieback, interpretation of annual ring
growth patterns, and the earliest occurrence of reaction
wood. Cross-dating of root samples was used to estimate
gully erosion rates by Vandekerckhove et al. (2001). The
authors suggest that a large number of samples are
required to understand gully evolution. In order to estimate
precisely the gully erosion volume, samples should
be collected from various places on the exposed root,
depending on its position (Vandekerckhove et al., 2001).
Cross-dating of tree ring series from trees and roots
allows identification of the year when an erosion episode
occurred. Scars on roots produced by transported
material may also inform about the occurrence of geomorphic
events.
Roots document erosion episodes that have led to
their exposure. Exposed parts of the root undergo anatomical
changes. Tree rings occurring in roots after
exposure are wider, cells are clearly smaller and there
are many more of them. In addition the division into
early and late wood is often clearly marked (Gärtner
et al., 2001). Clear visible anatomical changes are
recorded after exposure of coniferous tree roots (Hitz
et al., 2006). Erosion analysis is best performed on live
exposed roots documenting the exact year where root
exposure occurred. Dead roots only document the minimum
time passed from exposure and a simultaneous
erosion episode (Malik, 2006).
In this study, the authors assumed that it is possible to
reconstruct bank erosion history by examining the
anatomical changes in exposed tree roots. The purpose
of this study is (i) to determine how many and which
exposed roots ­ depending on their position related to
the river bank and soil ­ are particularly useful in
riverbank erosion dating; (ii) to explain anatomical
changes in exposed roots in relation to bank erosion
processes; (iii) to identify sources of error in the root
exposure dating process; and (iv) to determine the bank
erosion history in mountain streams using exposed roots
dating based on the example of the Bílá Opava River.
2. Study area
The Bílá Opava River valley is located in the
northeastern part of the Czech Republic, in the Hrubý
Jeseník massif (Eastern Sudetes). The elevation of the
massif is 1000­1500 m a.s.l. (Fig. 1a­c). Along the
upper course of the Bílá Opava River drainage basin, the
Hrubý Jeseník massif consists primarily of Devonian
orthogneiss, local migmatites and ­ in the upper parts fine-
and medium-grained paragneiss. The northern
127I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
foothills of the Sudetes were probably twice covered by
the Scandinavian ice sheet. At that time the study area
was subject to a periglacial climate and intensive mechanical
weathering occurred. As a result of this process
regolith covers were formed. Regolith is mainly transported
by debris flow to the riverbeds of streams flowing
in the Hrubý Jeseník Massif. Debris flow delivery
zones can be several meters wide and up to 800 m long
(Gába, 1992; Migoń et al., 2002).
The Bílá Opava River valley, in its upper course, cuts
about 200­300 m into the eastern face of the Praded
Massif. Seven sites were selected here in two river
Fig. 1. (a) Location of study area and sampling sites in Central Europe, (b) in the Jeseniki Mountains, (c) in the Bílá Opava River Valley.
128 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
reaches (Fig. 1c). Sites I­V were located in the upper
reach studied A; (Fig. 1c) where the average valley
gradient amounts to 13°, valley hillslopes even reach a
gradient of 50°, and the valley width reaches up to 20 m.
Numerous waterfalls on quartz vein outcrops occur in
the riverbed. On steep hillslopes distinct debris flow
tracks occur which document sediment yield to the
riverbed. Banks are usually formed of rock; hence
factors enhancing root exposure may here include biomechanical
weathering. Growing roots also expand in
width, which leads to loosening rocky material. As a
result of bank erosion the material is carried away.
Sites V­VII are situated in the lower part of the reach
studied B, where the average valley gradient amounts to
8°, hillslopes reach a gradient of 35°, and the valley
width reaches up to 80 m. The majority of banks are
formed of alluvial material. Numerous quartz, quartzite
and gneiss boulders up to 2 m across are deposited in the
riverbed forming a typical braided pattern. The riverbed
is characterized by a large quantity of CWD.
The average annual precipitation in the Hrubý
Jeseník Massif is about 1500 mm. A large part of the
intensive precipitation is linked to specific meteorological
situations, in which cyclones arriving from the
direction of Central or Northeastern Europe create
favourable conditions for continuous heavy precipitation,
especially between June and August. Maximal
intensity of rainfall (300 mm/2 h) occurred in Hrubý
Jeseník on the 1st of June 1921 (Štekl et al., 2001).
During heavy precipitation events large floods occur,
dozens of which have taken place in the Hrubý Jeseník
Massif in the last 500 years (Polách and Gába, 1998).
Particularly frequent floods occurred at the turn of the
20th century. One of the largest was the flood of 1903,
when the valley floor morphology of Bílá Opava was
transformed. Numerous sub-fossil braided river patterns
with residual boulders were dated dendrochronologically
to the first decade of the 20th century (Klimek et al.,
2003). The analysis of post-flood damage performed by
Polách and Gába (1998) in historical sources showed
that floods and high precipitation events occurred within
a radius of several kilometers from the reach studied in
the years 1910, 1913, 1914, 1921, 1930, 1931, 1940,
1948, 1958, 1966, 1972 and 1977. Water flow from the
drainage basin of the Bílá Opava, Stedni Opava, and
erná Opava streams is recorded at the hydrological
gauge in Karlowice, located 15 km east of the reach
of the Bílá Opava studied (Fig. 1b). This station recorded
high maximum monthly discharge rates in 1977
(76.1 m3
/s), 1980 (77.1 m3
/s), 1985 (46.1 m3
/s), and
1996 (50.1 m3
/s). However, the highest discharge
(320 m3
/s) was registered in July 1997 during an extraordinary
flood which ranged over the entire Central
Europe.
Forest clearance has been practised in the mountain
reaches of the Bílá Opava River drainage basin since the
Middle Ages when copper smelters existed there. Hence
the area studied was periodically deforested in the past
(Klimek, 2005; Latocha, 2005). At the present time, the
tree line (1400­1430 m a.s.l.) is composed of natural
communities of spruce that gradually grade into mountain
pine (Pinus mughus). The reaches of the valley
studied are vegetated with spruce plantation (Picea
abies), the oldest of which are about 110 years old.
According to a map of 1780, there was a water
reservoir in this reach of the Bílá Opava fed by springs
above the study sites. Timber from trees felled above
was rafted downstream when water was released from
the reservoir. The reservoir no longer existed in the mid-
19th century, as testified by the map of 1841.
3. Methods
The roots of spruce growing on the valley bottom
of the Bílá Opava River are exposed due to erosion
(Fig. 2). Exposure dating made it possible to determine
the time when intensive erosion occurred at the sites
under examination. In March 2004, before collecting
samples in the study area, the authors exposed dozens of
spruce roots in order to monitor changes in anatomical
features occurring after exposure. The experimental
trees grew in the Silesian Highlands in southern Poland
in limestone soil. One year after exposure, every tree
produced clearly marked tree rings divided into early
and late wood with small, numerous cells. This means
that spruce are capable of producing annual rings with
anatomical changes after exposure.
At 7 sites located on the valley floor of the Bílá Opava,
60 samples were taken from 23 roots among the exposed
roots of spruce trees growing in the vicinity of undercut
banks. Roots were mostly exposed in the bank area (sites
I, III, IV Vand VI), in the case of a spruce tree at site II, on
the hillslope over a dry riverbed, and at site VII, in an area
of terrace levels. All samples were collected in November
2004, that is after the trees had developed full
annual tree rings. Using a pole and measuring tape, forms
of valley floor were mapped 15­20 m above and below
each place where exposed root samples were collected.
Using a clinometer, hillslope was defined at the sites
where exposed root samples were collected.
Samples of 5- to 10-cm fragments of living exposed
spruce roots were taken using a handsaw. When more
roots at the site were exposed in a similar position
(distance from the bank and altitude above riverbed),
129I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
samples were taken from the thickest roots possible.
Where conditions permitted, samples were taken from
exposed parts of the roots 10 to 30 cm apart. From two
to five samples were taken from each root depending on
its length. Moreover, root orientation was compared
to the bank line in three directions: vertical, horizontal,
oblique. Also, root height above riverbed level was
defined as was its distance from the undercut surface.
All samples were cut by knives to examine the
structure of the wood in cross-section. The location
where the root became exposed was determined using a
binocular microscope, based on the assumption that the
moment of exposure was marked by a sharp decline in
cell size and an increase in the number of cells. The time
of exposure was calculated by counting the number of
tree rings younger than the first at which anatomical
changes were found. Information from previous works
shows the possibility of multiple counting directions
across samples (Shroder, 1980; Shroder and Butler,
1987). In this study calculation of tree rings was performed
along the longest possible radius of the root
cross-section, so as to include the wedging tree rings
(Schweingruber, 1996).
In the case of 18 samples taken from roots, the use of
a binocular microscope with a rather small magnification
proved insufficient to precisely define the year of
exposure. The year of exposure of these samples was
identified using a traditional microscope. Slices of wood
15­20 m thick were cut by means of a microtome. The
length of slices of wood was calculated from the
distance between the root centre and the last of its tree
rings on the longest radius possible. The width of each
wood slice amounted to approximately 0.5 cm. Therefore,
in the largest samples, several slices were prepared.
The number of tree rings with anatomical changes
visible under the microscope defined the root exposure
time.
4. Results and discussion
Sites I­V are located in reach A, 500­2050 m below
the source of the Bílá Opava. The width of the valley
floors varies from 2 to 20 m (Figs. 5 and 6; Table 1). The
gradient of the valley hillslopes is 30­45°. The riverbed
cuts into bedrock and is mainly composed of gravelly
deposits. Sometimes boulders up to 2 m in diameter
accumulate in the riverbed. Two terraces of 0.8 and
Table 1
Morphological features of the studied sites
Reach and
site number
Location of sites
studied below
river spring (m)
Gradient of
adjacent
hillslopes
Valley
width (m)
AI 500 30­40° 2­8
AII 750 30° 2­8
AIII 1150 30­40° 4­10
AIV 1500 30­45° 8­20
AV 2050 35­40° 8­15
BVI 2300 25­35° 20­45
BVII 2620 30­35° 50­80
Fig. 2. Photograph showing exposed spruce roots in the Bílá Opava riverbed.
130 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
2­2.2 m occur at sites IV and V. Numerous steps on
bedrock outcrops, boulders, and CWD are located in the
riverbed (Figs. 3 and 4; Table 1).
In total, 27 samples from 12 roots were taken at sites
I­V. Most were horizontally oriented. The sampled
roots grew from 0.2 to 0.8 m above the riverbed floor
and the distance between roots and cutting banks was
0.05­0.6 m (Table 2). As many as 12 root samples were
exposed in 1997, whereas the oldest root exposure took
place in 1912 (Table 2).
Sites VI and VII are located in reach B, 2300­
2620 m below the source of the Bílá Opava River.
The width of the valley floor varies from 20 to 80 m.
The gradient of the valley hillslopes is 25­35°. Terraces
Fig. 3. Geomorphological sketches of the Bílá Opava River Valley floor at sampling sites I­IV. 1 -- boulders above 1 m in diameter, 2 -- gravelly
deposits, 3 -- coarse woody debris, 4 -- spurs and rock ribs, 5 -- pools, 6 -- rocky steps, 7 -- debris steps, 8 -- terraces to 1.5 m, 9 -- terraces of
1.5­2.5 m, 10 -- erosional edges, 11 -- bank rocky outcrops, 12 -- hillslopes, 13 -- sampled spruces, 14 -- flow direction.
131I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
Fig. 4. Geomorphological sketches of the Bílá Opava River Valley floor at sampling sites V­VII. 1 -- boulders above 1 m in diameter, 2 -- gravelly
deposits, 3 -- coarse woody debris, 4 -- spurs and rock ribs, 5 -- pools, 6 -- rocky steps, 7 -- debris steps, 8 -- terraces to 2 m, 9 -- terraces of
2­4 m, 10 -- erosional edges, 11 -- bank rocky outcrops, 12 -- hillslopes, 13 -- sampled spruces, 14 -- flow direction.
132 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
at 2­2.8 m and 3­3.8 m occur in this reach. The
riverbed contains numerous boulders of over 2 m in
diameter, which form steps and in which debris dams
often occur (Fig. 4; Table 1). At sites VI and VII, 33
samples from 11 roots were collected. Most were
oblique and horizontally oriented. The roots grew 1.8­
2.8 above the riverbed floor and were longer (0.25­
1.4 m) than roots in reach A (Table 3). The roots
recorded erosion episodes between 1973 and 1997; most
often root exposure took place in 1977, 1983 and, as in
reach A, in 1997.
4.1. Validity and reliability of bank erosion dating using
anatomical changes in tree roots after exposure
The sampling of sites/trees should be conducted at
regular distances from each other in the longitudinal
profile of the studied valley. This sampling strategy
helps to identify real differences in erosion intensity in
the studied valley. The minimum tree density to ensure
a sufficient number of sample trees depends on the
exposure intensity in the valley floor. It appears that
sampling about 25% of the exposed roots is sufficient to
measure bank erosion intensity in the studied valley
because the remaining roots are in a similar or the same
position as the sampled roots and most likely recording
the same erosion events. If trees are lacking in certain
riverbank sections, observations of bank width and
structure are needed. If a riverbed section without trees
growing on its banks ­ and thus without exposed roots is
relatively wider than sections up- and downstream, it
is likely that the trees growing on the banks were
undercut and transported. Additionally, if erodible rocks
or especially older alluvium occur in the banks, the
probability that the trees that had grown on the banks
were transported downstream increases. By assuming
that part of the logs give a testimony as to bank erosion
in valleys where CWD is not cleared, in the future root
exposure dating can be connected with log dating to
improve the identification of the bank erosion process.
The Bílá Opava valley floor contains places where
the trees are sufficiently distant from the banks being
studied that bank erosion does not cause root exposure.
When analyzing the degree of afforestation of the valley
floor, it appears that root systems cover at least half of
the valley floor surface in reach A. In reach B, sometimes
trees do not grow on the valley floor or are scarce
and do not cover more than 20% of the surface. In the
Table 2
Position, height, distance and the year of exposure of the roots in reach A
Site
number
Root number and
piece of root number
Exposed root
length (m)
Root
orientation
Height above
the riverbed (m)
Distance from
cutting surface (m)
Most probable
year of exposure
AI 1 a 0.6 Vertical 0.2­0.5 0.3 1997
b 0.5 1997
2 a 0.4 0.7­0.9 0.4 1997
b 0.2 1967
AII 3 a 0.5 Horizontal 2.4 0.5 1912
b 0.1 1958
AIII 4 a 0.7 Vertical 0.9­1.5 0.05 1997
b 0.05 1997
c 0.15 1993
5 a 0.6 Horizontal 1 0.5 1949
b 0.3 1997
6 a 0.5 Oblique 0.6­0.8 0.5 1931
b 0.2 1997
7 a 0.7 Horizontal 0.8 0.4 1958
b 0.2 1997
8 a 0.4 Oblique 0.6­0.8 0.3 1964
b 0.1 1993­95
AIV 9 a 0.6 Horizontal 1 0.6 1988
b 0.4 1990
c 0.2 1992­94
10 a 0.3 Vertical 0.3­0.5 0.4 1957­58
b 0.3 1960­62
AV 11 a 0.7 Oblique 0.4­0.8 0.3 1988­1989
b 0.5 1997­1998
12 a 1.1 Horizontal 0.8 0.3 1997
b 0.4 1997
c 0.2 1997
133I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
treeless sections located in reach A, numerous rock
outcrops occur within the banks and the banks without
trees are not wider than the banks covered with trees.
This indicates that intensive erosion did not take place in
these sections. Most often, the treeless sections in the
Bílá Opava valley in reach B are significantly wider than
the forested sections. The banks in reach B are composed
mostly of erodible gravels. The probability is substantial
that the trees that grew on the banks were undercut and
transported downstream.
Living roots with ends still growing in the soil are
particularly useful for bank erosion dating because after
exposure these roots usually form wide rings with
clearly visible anatomical changes. More than half the
roots sampled in the Bílá Opava valley floor formed
clear rings after exposure. Sometimes roots are damaged
during exposure occurring during erosion. While these
roots produce suppressed rings after exposure, most
often the anatomical changes are clear. This form of
growth was observed in 14 root samples in the Bílá
Opava valley floor. Alestalo (1971) and Shroder (1980)
described suppression occurring after root exposure.
If roots are exposed at one site in different horizons
or strata of bank sediments, it is important to sample all
of them to identify the possibility of occurrence of more
erosive events in one riverbed/valley cross-section.
Samples of the thickest roots should be collected from
the same horizon, as these roots are potentially the oldest
and have recorded a relatively large number of erosion
episodes.
Samples should be taken at a distance of 10 cm from
one another. This precise sampling enables the capture
of gradual exposure as occurs when the root is exposed
in sections during floods. Thus, one exposed root may
document several erosion episodes. As a result of
gradual exposure, root fragments sampled from the soil
neighbouring the direction of exposure are younger and
younger. Gradual exposure was identified in 20 of 23
Table 3
Position, height, distance and the year of exposure of the roots in reach B
Site
number
Root number and
piece of root number
Exposed root
length (m)
Root
orientation
Height above
the riverbed (m)
Distance from
cutting surface (m)
Most probable
year of exposure
BVI 13 a 0.7 Oblique 1.7­2.2 0.4 1982
b 0.3 1983
c 0.4 1983
14 a 0.8 Oblique 1.5­1.9 0.2 1973­75
b 0.6 1974
c 0.3 1977
15 a 0.4 Horizontal 2 0.2 1977
b 0.3 1997­79
16 a 0.8 Horizontal 2 0.3 1983­85
b 0.6 1978
c 0.4 1977
BVII 17 a 1.4 Horizontal 2.1 0.25 1982
b 0.1 1997
c 0.3 1997
d 0.15 1997
18 a 1 Horizontal 2.1 0.4 1974
b 0.6 1973­75
c 0.15 1977
19 a 0.6 Oblique 1.8­2.1 0.4 1983
b 0.3 1991
c 0.45 1997
20 a 0.5 Oblique 2.6­2.8 0.2 1980­84
b 0.35 1986
c 0.2 1993­94
21 a 0.4 Vertical 2.3­2.7 0.1 1980­82
b 0.3 1977­79
c 0.4 1983
22 a 0.25 Oblique 2.6­2.7 0.05 1997
b 0.05 1997
23 a 0.9 Horizontal 2.7 0.45 1990­92
b 0.3 1977­79
c 0.3 1983
d 0.4 1973­74
134 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
sampled roots in the Bílá Opava valley (Tables 2 and 3).
Only three roots (1, 12 and 22) were exposed in the July
1997 flood alone.
Sometimes, gradually exposed roots feature an
illegible record of the years in which exposure occurred
within one sample (Fig. 5b). The 2­5 rings that follow
root exposure record gradual anatomical changes. At
first, the rings recording exposure are wider than previously,
though often the division between early and late
wood is very weakly visible. The next rings record more
and more visible anatomical changes. Hence, in the case
of such root samples, it is only possible to estimate their
exposure time with a precision of up to several years.
The first of the tree rings recording anatomical changes
is to be considered as the moment when root fragment
exposure began. While it is possible that the same anatomical
changes from gradual exposure could occur in
soil, more experimental study is needed on tree ring
formation in relation to gradual exposure.
Gradual exposure in the root samples collected in
Bílá Opava valley occurred especially in case of roots
exposed in reach B where the erodible alluvium banks
occur (root samples 14a, 15b, 16a, 18b, 20a,c, 21a,b,
23a,b,d). This form of root growth commonly testifies to
less intensive bank erosion events. For example, root
sample (11b) was exposed gradually, thus the change in
wood anatomy also occurred gradually (Fig. 5b). On the
other hand, root (1b) from site I was dramatically exposed
in 1997, which is reflected in the sudden change
in its wood anatomy (Fig. 5a).
Apart from anatomical changes in roots that occur
immediately after their exposure, changes also frequently
occur after longer periods of exposure, consisting of a
dramatic decrease in tree-ring size, often accompanied
by an accumulation of resin ducts as in case of root 20b
(Fig. 5c). At least two suppressed tree rings (more
than 50% thinner than previous rings) occurring after a
longer period of exposure were observed in 17 roots
fragments sampled in the Bílá Opava valley. Such
changes are often the result of damage to the exposed
roots by the transported material. They may also be the
result of exposure of the end of the root, which causes
root necrosis. In order to date such roots, their structure
must be analyzed under a microscope.
The oldest spruces growing in the Bílá Opava River
valley are up to 110 years old. The exposed roots are
most commonly much younger, which limits conclusions
on bank erosion to about the previous 100 years.
Fig. 5. Microphotograph showing part of root sample cross-sections with anatomical changes, (a) tree rings formed before exposure, (b) first year with
anatomical changes (tree ring formed simultaneously with exposure), (c) tree rings formed after exposure.
135I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
Roots testifying to bank erosion can be broken by the
transported material. Trees with exposed root systems
are often undercut and fall into the riverbed. Therefore,
the quantity of exposed roots testifying to floods in the
earlier years of this period is rather small, while later
erosion episodes are well recorded in the wood of exposed
roots. Hence, the large majority of roots exposed
in the Bílá Opava valley floor document bank erosion
episodes from the last 40 years.
False rings, missing rings or wedging rings occur in
stems in a similar manner as in roots (Schweingruber,
1988). They may distort the number of years that have
elapsed since the erosion episode. However, errors
relating to that limitation are not large, since, in contrast
to cores collected from trees, the whole cross-section is
examined in the case of roots. This makes identification
of false, missing or discontinuous rings easier. Tree
rings formed in roots after exposure are similar to the
rings produced in stems, meaning that root form growth
after exposure is very clear. Before exposure lots of
wedging, as well as false and probably missing rings in
roots occur (Fig. 5b). Cross-dating can potentially
eliminate false and missing rings within the roots. The
authors attempted to cross-date roots samples, but the
relatively short tree ring series occurring after exposure
and the numerous growth disturbances such as the
dramatic decrease occurring after longer periods of
exposure and a long series of compression wood made it
impossible to reach conclusions. Perhaps cross-dating
root ring series with stem ring series would be more
accurate. Before cross-dating, site/local chronology
should be reconstructed to eliminate as much error as
possible because of missing and false rings that may
occur within the stems.
Compression wood occurs within numerous root
cross-sections; in 21 root samples compression wood
occurs in the first ring after exposure. In about 38 cases,
compression wood occurs after a longer period of exposure.
This suggests that compression wood is not a
precise marker for establishing when erosion events
took place.
The time of root exposure as identified from the
analysis of anatomical changes within its tree rings is
not always the same as the actual year of exposure of
that root. Tree rings are formed in a temperate climate in
the period from May to November (Zielski and Kr1piec,
2004). When a root is exposed at this time, anatomical
changes are visible in the area of a tree ring. If the root
exposure occurred from January to May, the tree ring
may be changed in the same year as the exposure
occurred. If the exposure occurred between November
and December, the tree ring with anatomical changes
will appear the following year. A marker of exposure
may be thus recorded in roots 1 year after an erosion
episode.
The authors decided not to calculate bank erosion
rates, which depends on a rather precise time that the
erosion rate is measured. The recording of erosion is
more precise when more recent events are analyzed.
This means that is not possible to select a representative
period for measuring erosion rates in the Bílá Opava
valley and the results of such calculations could be
distorted.
4.2. Bank erosion of the Bílá Opava River in reach A
The root exposure analysis shows that bank erosion
took place in reach A of the Bílá Opava valley during
the last 100 years (Fig. 6). The earliest recorded bank
erosion episodes occurred in 1912, 1931 and 1949.
Later erosion periods were concentrated in the years
1956­1967, 1988­1995 and 1997. The best recorded
erosion episode is the one in 1997, when as many as 12
of the 28 root fragments were exposed.
The earliest registered bank erosion episodes were
recorded in site II (1912) and site III (1931 and 1949).
At site II, exposed spruce roots are growing as much as
2.4 m above the level of the riverbed. One can suspect
that only extreme precipitation episodes would cause
flows permitting the hillslope to be undercut at such a
level. The date of root exposure of this spruce in 1912
corresponds to the floods of 1913 and 1914. During the
latter, in Mikulowice 20 km to the north, 128 mm of
precipitation was recorded during 1 day in 1913 year
(Polách and Gába, 1998). Partial exposure of root 6 at
site III may be related to the intense precipitation that
occurred in the area of the Hrubý Jeseník in 1930 or
1931. The root at site III was exposed in 1949. The root
documents bank erosion which occurred during intense
precipitation in 1948, when 126 mm of rain was recorded
at Praded massif.
A large group of root fragments exposed at sites I­IV
in reach A in the period between 1956 and 1967 are a
result of bank erosion during two floods which took
place in the Jeseniki mountains in 1958 and 1966. The
earlier resulted from rainstorms registered in Jeseniki, as
it is known that the Desná River flooded then (Polách
and Gába, 1998). The analysis of root exposure shows
that precipitation also covered the Bílá Opava valley,
where bank erosion took place. At that time the bank at
site II was eroded, which would demonstrate a considerable
flow during that flood. Also, fragments of
roots 6 and 7 at site III were exposed, as well as root 10
at site IV. In 1966, a large flood covered the area around
136 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
the Jesenik massif (Polách and Gába, 1998). The flood
in the Bílá Opava riverbed caused erosion at site II,
where a fragment of root 2 was exposed.
The group of root fragments exposed in the years
1988­1995 covers the period in which both historical
and hydrological sources give no information about
significant floods/rainstorms in the Jeseniki area. At this
time fragments of root 8 were exposed at site III, root 9
at site IV and root 11 at site V. Perhaps erosion in this
period occurred as a result of rapid precipitation that
was so short that it was not registered at the Karlowice
gauge. However, in such a case roots would inform us
about a rapid erosion episode, while 4 out of 7 roots
were exposed gradually.
Root exposure in the years 1988­1995 may be a
result of local changes in the riverbed morphology. Bank
erosion may be started by redeposition of boulders, the
transport of which is testified to by imbrication in the
Bílá Opava riverbed. The boulders are usually transported
short distances (Baker and Ritter, 1975, Pizzuto
et al., 1999), producing erosion by local redirection of
the current towards the banks, which have not suffered
erosive activities before. Such a situation occurs in the
case of roots exposed at site IV, where boulders which
accumulated in the upper zone direct the current towards
the left bank where erosion occurs. Systematic washing
out of the left bank resulting from the change in the
riverbed configuration may help bank erosion that takes
place regardless of high water episodes. This is confirmed
by gradual exposure of as many as three root
pieces at this site. A similar situation was observed at
site V, where boulders deposited in the centre of the
riverbed divide the current into two parts. The left part
intensively undercuts the bank as a result of bank
erosion occurring in the years 1988­89.
Changes in the riverbed morphology may be caused
by hillslope material supply. The supply zone, to which
the river is predisposed due to the steeply sloping valley
sides, occurs in the valley studied above sites III and IV.
There is evidence for this in visible debris tracks on the
Fig. 6. Years of tree roots exposure in the Bílá Opava riverbed.
137I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
hillslope. In the Jeseniki, debris flows occur periodically.
Recently they occurred in the Keprník Massif, 10 km
to the north west, where in the summer of 1991 two
large debris flows, and at least several smaller ones,
were produced (Gába, 1992). It is most likely that the
material had recently been loosened above the sites
under study, which is testified to by the large amount of
angular material deposited in the Bílá Opava riverbed.
The change of riverbed morphology related to material
supply to the riverbed could produce changes in the flow
direction. As a result, bank erosion started to take place
in new locations and erosion occurred even with relatively
small discharge.
Erosion of the banks may be enhanced by debris
dams formed of CWD positioned perpendicular to the
river axis (Gregory et al., 1985; Gurnell et al., 2000).
Debris dams may change the flow direction during
floods (Marston, 1982; Robinson and Beschta, 1990), in
particular in the riverbeds of mountain streams. In such
cases, overflow channels even form in the valley floor
(Kaczka, 1999). It is possible that bank erosion at sites
III­V of the Bílá Opava riverbed in the years 1988­
1995 was produced by debris dams that were destroyed
during a later flood in 1997. This is all the more likely
because in the reach where the sites are located many
dams are caught by trees and boulders and deposited
transverse to the river direction. It seems that a width of
valley floor similar to CWD length (about 20 m) is a
decisive factor in promoting intensive CWD deposition
on this reach. One such debris dam is situated 5 m below
the roots exposed at site V.
Significant transformations of riverbed morphology
were observed in the areas affected by precipitation in
July 1997 (Zieliński, 2003). Large undercuts of banks,
slopes and terrace levels were formed at that time
(Owczarek, 2004a,b). Water in the upper reach of the
Bílá Opava valley floor flowed in a narrow riverbed
during the flood of July 1997, which probably enhanced
bank erosion. This occurred here at sites I, III and V. At
site I, 3 out of 4 root fragments studied were exposed,
while at site V as many as 4 out of 5 fragments were so
affected. All three roots studied at site III were partially
exposed during the flood of 1997.
Due to a relatively large number of roots affected,
bank erosion at site III can be interpreted separately. It
occurred in the years 1931, 1949, 1958, 1993 and 1997.
Such a big difference of time between the exposure of
particular parts of the roots testifies to lateral stability of
the riverbed at this site. Bank erosion may occur here
only during big floods. This is demonstrated by the
possible correlation of all erosion episodes which took
place at site III, except for 1993, with large floods.
4.3. Bank erosion of the Bílá Opava River in reach B
In reach B, signals of bank erosion have only appeared
since 1973 (Fig. 6). At this reach, due to greater
intensification of water flow and alluvial banks, the
probability of bank erosion is greater than in the upper,
rocky reach A of the river. Therefore, trees growing in
the lower reach, which feature older erosion episodes in
their exposed roots, are felled to the river, which is
testified by a high number of CWD deposited in the
riverbed in reach B. Lack of older signals of bank
erosion in the lower reach should be interpreted as a
symptom of the high rate of bank erosion in reach B.
The majority of root exposure episodes in reach B are
found within two periods of time, the years 1973­1983
and 1986­1995. Most root exposure episodes, including
several episodes a year, were recorded in 1977, 1983
and 1997. It is hard to define the floods responsible
for bank erosion in the years 1973­1983 precisely. As
many as 9 out of 23 root fragments recorded gradual
exposure. Moreover, signals of root exposure are scattered
across the period analyzed. Bank erosion of the
Bílá Opava could be enhanced by rainstorms in late
August 1972, when Praded gauge recorded 83.7 mm of
rain (Polách and Gába, 1998). Bank erosion in reach B
can also be the result of heavy precipitation which
occurred in 1977 in a significant part of Europe, including
the Šumperk and Jeseník districts. At that time
intensive bank erosion took place at site VI, where three
out of four roots analyzed were exposed.
As many as 7 out of 23 root fragments from the
period analyzed were exposed in 1983, although no
precipitation data document such an erosion episode.
The reason for root exposure during several years following
floods may be the high erodibility of previously
exposed alluvial sediments. Undercuts formed during
high floods are susceptible to erosion, which makes
relatively small floods occurring after the much bigger
ones cause erosion. Gradually the erosional undercut is
stabilized by the flora, which limits erosion. The high
erodibility of sediments exposed in the lower part of the
valley floor is testified by frequent root exposure periods
at particular sites. Almost all roots at sites VI and VII
were exposed with a frequency of several years, while
roots in the upper, rocky reach are exposed with a
frequency of several dozen years (Fig. 7).
What is characteristic is the fact that in the period
between 1973 and 1983 roots were only exposed in
reach B. This may mean that precipitation episodes responsible
for bank erosion did not cover the upper part of
the valley. It is more probable, however, that precipitation
was not heavy enough to erode the rocky banks
138 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
in reach A, yet was sufficient to undercut terrace levels in
reach B.
The period of bank erosion, registered in the lower
reach in the years 1986­1994, also occurred in reach A.
Bank erosion was also recorded by three roots at site
VII. Signals about root exposure are gradual here, which
suggests several episodes of erosion. Similarly as in
reach A, in the lower part of the reach, erosion in the
years 1987­1985 could have been caused by debris
dams that appear periodically. Dams may also cause
water flow in the terrace levels zone where even slight
but systematic floods result in erosion.
Dating of root exposure shows that during the flood
of 1997 erosion in the lower reach was insignificant.
Exposure only referred to roots 15, 17 and 19 of the
spruce growing at the edge of the terrace level of 2.8 m
at site VII. It should be noted, however, that trees
growing on lower banks were felled onto the riverbed
during the extraordinary flood in 1997.
The results of dating of root exposure of spruce
growing on the edge of terrace levels at site VII allow
for analysis of bank erosion in the cross-section of the
valley floor. It started in the years 1973­74 within the
upper terrace at 2.8 m. In the years 1977­1979, 1982­
1988, and 1991­1994, bank erosion occurred several
times in both the lower and the upper terraces. It also
took place in the upper terrace in 1997. The significant
bank erosion in the terrace levels that are situated high
above the riverbed is a result of the heavy flood of 1977.
Since then, erosion has very often occurred in the area of
Fig. 7. Time and frequency of tree root exposure in the Bílá Opava riverbed.
139I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
both terrace levels. During the flood already mentioned,
a debris dam was probably formed, which during the
next small floods made water flow into the area of the
terrace levels, undercutting them. Presently, there is a
small debris dam above exposed roots, yet it is not
significantly obstructing water flow and simultaneously
undercutting the upper terrace levels. It seems that the
debris dam producing erosion at the site studied was
destroyed during the flood of 1997. Since then undercuts
around the terrace levels have become inactive.
5. Conclusions
This study has shown that a number of factors need to
be taken into account when carrying out a dendrogeomorphological
assessment of riverbank erosion history.
In particular, the sampling of roots representing different
heights above the riverbed is needed because this will
capture the recording of different erosion episodes.
Samples should be taken from the thickest root along the
same strata; these roots are potentially the oldest and
have recorded a relatively large number of erosion
events. Sub-samples should be taken at 10-cm intervals,
such a precise sampling can reveal gradual exposure.
One problem with the methodological approach is that
false rings, missing rings or wedging may distort the
number of years that have elapsed since the erosion
episode. Cross-dating can potentially eliminate false and
missing rings within the roots, however, in this study
relatively short tree-ring series occurring after exposure
and numerous growth disturbances such as a long series
of compression wood precluded its application. Future
research should focus on cross-dating root ring series
with stem ring series to improve accuracy.
The main conclusions regarding the bank erosion
history of the Bílá Opava River are the following:
* Root exposure analysis showed that one can distinguish
four bank erosion periods: the oldest root
exposure signals are significantly dispersed over the
period 1912­1967, and later signals are less dispersed
within two periods, 1973­1984 and 1986­1995, with
the most numerous group of signals in 1997.
* Bank erosion related to root exposure episodes in the
first half of the 20th century was only recorded in
reach A of the Bílá Opava River. Only three bank
erosion episodes were observed here in the years
1912, 1931 and 1949, corresponding to heavy floods
documented in historical sources. These episodes
must have been significant enough for bank erosion
to become active in the rocky riverbed. Simple
erosion signals in the upper part of the valley studied
took place in the years 1956­1967. Perhaps they are
related to two heavy floods that occurred in Jeseniki
in 1958 and 1966.
* In the period between 1973 and 1984, roots were
only exposed in reach B. Bank erosion occurred
gradually here, which is demonstrated by the overlapping
of several erosion episodes, the first of which
triggered sediment exposure. Perhaps this was encouraged
by the floods of 1972 and/or 1977. In the
period 1973­1984 bank erosion was affected by the
riverbed morphology at particular sites. Situations
have been defined where the location of boulders
enforces local bank erosion. Therefore, it is not necessarily
heavy precipitation that causes erosion, it
may occur after a change of bed configuration. Then
water systematically undercuts the bank even during
relatively small floods.
* In the period 1986­1994, banks eroded in reaches A
and B. Historical sources or meteorological data do
not confirm heavy precipitation/floods in the area of
the Jeseniki. Roots were most frequently exposed
gradually, which would suggest systematic bank undercutting.
Erosion in reach A could be caused by
changes of the riverbed configuration of the Bílá
Opava, influenced by hillslope material supply from
the slope, which was particularly intensive in the
Jeseniki in 1991. Bank erosion in reach B could also
have been caused by the formation of debris dams
and the undercutting the terrace levels.
* Bank erosion of the Bílá Opava was most intensive
during the flood of 1997. At the time rocky banks
were undercut at four out of five sites in reach A.
Root exposure analysis shows that during the flood
of 1997 erosion in the lower part of the valley in
reach B was not as intensive as in the upper part.
However, here undercutting referred to the erodible
sediments of terrace levels, which must have intensified
the erosion. Perhaps some trees were felled
and transported downstream, which is testified to by
the many CWD in the lower part of the river.
Acknowledgments
This article presents the results of research conducted
under a grant from the Committee of Scientific Research,
KBN, 3 PO4E 023 25. The authors wish to
thank Prof. F.H. Schweingruber and Dr. Holger Gärtner
from the Swiss Federal Institute for Forest, Snow and
Landscape Research and Dr. Ingo Heinrich from the
University of Fribourg in Switzerland for teaching us to
understand the roots studied. Thanks go to two anonymous
reviewers for critical reading and suggestions.
140 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
References
Abernethy, B., Rutherfurd, I.D., 2000. The effect of riparian tree roots
on the mass-stability of riverbanks. Earth Surface Processes and
Landforms 25, 921­937.
Alestalo, J., 1971. Dendrochronological interpretation of geomorphic
processes. Fennia 105, 1­140.
Baker, V.R., Ritter, D.F., 1975. Competence of rivers to transport coarse
bedloadmaterial.GeologicalSociety ofAmericaBulletin89,975­978.
Carrara, P.E., Carroll, T.R., 1979. The determination of erosion rates
from exposed tree roots in the Piceance basin, Colorado. Earth
Surface Processes and Landforms 4, 307­317.
Gába, Z., 1992. Mury pod Keprnikiem v ervenci 1991. Severní Morava
64, 43­50.
Gärtner, H., Schweingruber, F.H., Dikau, R., 2001. Determination of
erosion rates by analysing structural changes in the growth pattern
of exposed roots. Dendrochronologia 19, 81­91.
Gregory, K.J., Walling, D.E., 1973. Drainage Basin Form and Process.
Arnold, London, pp. 234­287.
Gregory, K.J., Gurnell, A.M., Hill, C.T., 1985. The performance of
debris dams related to river channel processes. Hydrological
Sciences Journal 30, 3­9.
Gurnell, A.M., Petts, G.E., Hannah, D.M., Smith, B.P.G., Edwards, P.J.,
Kollmann, J., Ward, J.V., Tockner, K., 2000. Wood storage within
the active zone of a large European gravel-bed river. Geomorphology
32, 601­619.
Harvey, A.M., 1974. Gully erosion and sediment yield in the Howgill
Fells, Westmorland. In: Gregory, K.J., Walling, D.E. (Eds.), Fluvial
Processes in Instrumented Watersheds. Institute of British Geographers
Special Publications, pp. 45­58.
Hitz, O., Gärtner, H., Monbaron, M., 2006. Reconstruction of erosion
rates in Swiss Mountain Torrents. TRACE -- Tree Rings in
Archaeology. Climatology and Ecology 4, 196­202.
Hooke, J.M., 1980. Magnitude and distribution of rates of river bank
erosion. Earth Surface Processes and Landforms 5, 143­157.
Hupp, C.R., Osterkamp, W.R., 1996. Riparian vegetation and fluvial
geomorphic processes. Geomorphology 14, 277­295.
Kaczka, R.J., 1999. The role of coarse woody debris in fluvial
processes during the flood of July 1997, Kamienica Łącka valley.
Beskidy mountains. Poland. Studia Geomorphologica CarpathoBalcanica
32, 119­129.
Klimek, K., 1989. Flood plains activity during floods in small
mountain valleys, The Bieszczady Mts, The Carpathians, Poland.
Quaestiones Geographicae 2, 93­100.
Klimek, K., 2005. Eastern part of Hruby Jesnik as an example of
human­environment interactions in mid-mountain ecosystem.
Human Impact on Mid-Mountain Ecosystem 1, 3­7.
Klimek, K., Malik, I., Owczarek, P., Zygmunt, E., 2003. Climatic and
human impact on episodic alluviation in small mountain valleys.
The Sudetes. Geographia Polonica 76, 55­64.
Krzemień, K., 1976. Recent riverbed transformation processes in
Konina stream in Gorce mountains. Folia Geographica. Series
Geographica-Physica 10, 87­122.
LaMarche Jr., V.C., 1966. An 800-Year History of Stream Erosion as
Indicated by Botanical Evidence. U.S. Geological Survey Professional
Paper, vol. 550-D, pp. 83­86.
Lane, S.N., Chandler, J.H., Richards, K.S., 1994. Developments in
monitoring and modelling small scale river bed topography. Earth
Surface Processes and Landforms 19, 349­368.
Latocha, A., 2005. Geomorphic evolution of mid-mountain drainage
basins under changing human impacts, East Sudetes, SW Poland.
Studia Geomorphologica Carpatho-Balcanica 39, 71­93.
Lawler, D.M., 1978. The use of erosion pins in river banks. Swansea
Geographer 16, 9­18.
Lawler, D.M., 1991. A new technique for the automatic monitoring
of erosion and depositions rate. Water Resources Research 27,
2125­2128.
Lawler, D.M., 2005. The importance of high-resolution monitoring in
erosion and deposition dynamics studies: examples from estuarine
and fluvial systems. Geomorphology 64, 1­23.
Malik, I., 2005. Rates of lateral channel migration along the Ma3
a
Panew River (southern Poland) based on dating riparian trees and
Coarse Woody Debris. Dendrochronologia 23, 29­38.
Malik, I., 2006. Contribution to understanding the historical evolution
of meandering rivers using dendrochronological methods: example
of the Ma3
a Panew River in southern Poland. Earth Surface
Processes and Landforms 31, 1227­1245.
Marston, R.A., 1982. The geomorphic significance of log steps in
forested streams. Annals of the Association of American Geographers
72, 99­108.
Migoń, P., Hrádek, M., Parzóch, K., 2002. Extreme events in the
Sudetes Mountains, their long-term geomorphic impact and
possible controlling factors. Studia Geomorphica CarpathoBalcanica
36, 29­49.
Owczarek, P., 2004a. Evolution of alluvial channel forms under the
influence of hillslope sediment delivery (S Poland). Interpraevent
2004 -- Changes within Natural and Cultural Habitat and
Consequences, Tagungspublikation Band, Garmisch-Partenkirchen,
pp. 225­232.
Owczarek, P., 2004b. The influence of hillslope sediment delivery on
the morphology of gravel bed rivers, Polish Flysh Carpathians.
Geomorphologia Slovaca 4, 40­45.
Pizzuto, J.E., Weeb, R.H., Griffiths, P.G., Elliot, J.G., Melis, T.S.,
1999. Entrainment and transport of cobbles and boulders from
debris fans. Geophysical Monographs 110, 53­70.
Polách, D., Gába, Z., 1998. Historie povodní na šumperském a
jesenickém okrese. Severní Morava 75, 3­29.
Pollen, N., Simon, A., 2005. Estimating the mechanical effects of
riparian vegetation on stream bank stability using a Fiber-Bundle
Model. Water Resources Research 41, 572.
Robinson, E., Beschta, R.L., 1990. Coarse woody debris and channel
morphology interactions for undisturbed streams in Southern Alaska,
USA. Earth Surface Processes and Landforms 15, 141­156.
Rowntree, K.M., Dollar, E.S.J., 1999. Vegetation controls on channel
stability in the Bell River, Eastern Cape, South Africa. Earth Surface
Processes and Landforms 24, 127­134.
Schweingruber, F.H., 1996. Tree rings and Environment. Dendroecology,
Swiss Federal Institute for Forest, Snow and Landscape
Research, pp. 222­226. Berne, Stuttgart, Vienna, Haupt.
Shroder Jr., J.F., 1978. Dendrogeomorphologic analysis of mass movement
on Table Cliffs Plateau, Utah. Quaternary Research 9, 168­185.
Shroder Jr., J.F., 1980. Dendrogeomorphology: review and new
techniques of tree-ring dating. Progress in Physical Geography 4,
161­188.
Shroder Jr., J.F., Butler, D.R., 1987. Tree-ring analysis in the earth
sciences. In: Jacoby, G.C., Hornbeck, J.W. (Eds.), Proceedings of
the International Symposium on Ecological Aspects of Tree-ring
Analysis. Lamont-Dougherty Geological Observatory & U.S.
Department of Agriculture, pp. 186­212.
Sigafoos, R.S., 1964. Botanical evidence of floods and flood-plain
deposits. US Geological Professional Paper, vol. 485-A, pp. 1­26.
Starkel, L., 2002. Change in the frequency of extreme events as the
indicator of climatic change in the Holocene (in fluvial systems).
Quaternary International 91, 25­32.
141I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142
Strunk, H., 1997. Dating of geomorphological processes using
dendrogeomorphological methods. Catena 31, 137­151.
Sttot, T., 1997. A comparison of stream bank erosion processes on
forested and moorland streams in the Balquhidder catchments,Central
Scotland. Earth Surface Processes and Landforms 22, 383­399.
Štekl, J., Brázdil, R., Kakos, V., Je, J., Tolasz, R., Sokol, Z., 2001.
Extrémní denní srákové úhrny na území R v období 1879­2000 a
jejich synoptické píiny. eský hydrometeorologický ústav,
Praha.
Trafas, K., 1975. Changes of the Vistula river bed east of Cracow in
light of archival maps and photointerpretation. Prace Geograficzne
40, 1­85.
Thorne, C.R., 1981. Field measurements of bank erosion and bank
material strength. International Association of Hydrological Sciences,
Florence, Italy, pp. 503­512.
Thorne, C.R., 1982.Process and mechanism of river bank erosion. In: Hey,
R.D., Bathurst, J.C. (Eds.), Gravel-Bed Rivers. Wiley, Chichester,
pp. 227­271.
Vandekerckhove, L., Muys, B., Poesen, J., De Weerdt, B., Coppé, N.,
2001. A method for dendrochronological assessment of mediumterm
gully erosion rates. Catena 45, 123­161.
Wyżga, B., Zawiejska, J., 2005. Wood storage in a wide mountain
river: case study of the Czarny Dunajec, Polish Carpathians. Earth
Surface Processes and Landforms 30, 1475­1494.
Zieliński, T., 2003. Catastrophic flood effects in alpine/foothill fluvial
system (a case study from the Sudetes Mts, SW Poland). Geomorphology
54, 293­306.
Zielski, A., Krąpiec, M., 2004. Dendrochronologia. Polskie Wydawnictwo
Naukowe, Warszawa.
142 I. Malik, M. Matyja / Geomorphology 98 (2008) 126­142