Rates of lateral channel migration along the Mala Panew River (southern Poland) based on dating riparian trees and Coarse Woody Debris

ARTICLE IN PRESS Dendrochronologia 23 (2005) 29–38 www.elsevier.de/dendro ORIGINAL ARTICLE Rates of lateral channel migration along the Mala Panew River (southern Poland) based on dating riparian trees and Coarse Woody Debris Ireneusz Malik Department of Quaternary Paleogeography and Paleoecology, University of Silesia, Sosnowiec 41–200, ul., Bedzinska 60, Poland Received 31 December 2003; accepted 12 July 2005 Abstract Two methods of estimating the lateral migration of a river channel have been proposed. The first method is based on Coarse Woody Debris (CWD) dating. The terraces of the Mala Panew River are mainly covered with plantations of Pinus sylvestris where individual trees grow at equal distances to each other. During times of high discharges trees fall onto the riverbed providing information on the extent of flood plain erosion. The ages of CWD and the surface of the eroded flood plain provide an estimation of the rate of lateral migration. The erosion rates measured at two sites in the Mala Panew River were between 0.24 and 0.36 m/year. Another way of reconstructing the rate of lateral migration is by dating trees growing on different-aged sandy meander bars. These levels are primarily covered with Alnus glutinosa and Alnus incana. The oldest trees growing on each level give information about the minimum age of that level, which allows us to reconstruct the rate of lateral migration. The lateral migration of the channel has also been estimated dating the oldest trees growing on mid-channel islands separated from the lateral banks. The values obtained for 10 sites of the Mala Panew channel oscillate between 0.07 and 1.83 m/year. Tree ring analyses also allow us to determine the impact of individual high discharges on the lateral migration rate of the Ma"a Panew channel. The lateral migration of the channel was most rapid in the years 1953–57 and 1966–68 as well as during the extraordinary flood in 1997. r 2005 Elsevier GmbH. All rights reserved. Keywords: Dendrogeomorphology; Alluvial development; Lateral migration; Riparian forest; Coarse woody debris; Poland Introduction One of the major factors controlling the patterns of meandering river channels is the presence of riparian forests covering the valley floor (Hickin and Nanson, 1984; Thorne, 1990; Abernethy and Rutherfurd, 2000; Brooks and Brierley, 2002). Tree root systems protect the Corresponding author. Tel.:+ 48 032 3809104. E-mail address: irekgeo@wp.pl. 1125-7865/$ - see front matter r 2005 Elsevier GmbH. All rights reserved. doi:10.1016/j.dendro.2005.07.004 soil and older alluvial sediments against erosion and reduce the supply of material to the riverbed. Hence, rivers flowing through forested areas are not sufficiently charged with material coming from erosion of the valley sides. Large accumulation forms dividing the main channel into a number of minor streams are uncommon in these riverbeds in contrast to rivers flowing through deforested areas. Lowland and upland rivers crossing the forested areas flow, very often, in a single channel. Frequent dams are formed from trees and Coarse Woody Debris (CWD). ARTICLE IN PRESS 30 I. Malik / Dendrochronologia 23 (2005) 29–38 CWD is defined as fragments of dead trees longer than 1 m and not less than 10 cm in diameter at the fragment middle point (Van Sickle and Gregory, 1990). These fragments fill the riverbed and make the water flow around them, contributing to increase river bendiness as well as the generation of numerous irregularly shaped meanders (Rachocki, 1978; Gregory, 1992; Gurnell et al., 1995; Gurnell and Sweet, 1998; Kaczka,1999). The meandering channel pattern of a river depends on the progress of lateral migration (Hickin and Nanson, 1984). During high discharge, the concave banks are systematically undercut, whereas at times of retreating flood waters, mineral material is accumulated on the convex sides of the riverbed. By comparing the meander routes on historical maps, the average lateral migration rate can be estimated (Trafas, 1975; Brooks, 2001). This method has some limitations imposed by the frequency and accuracy of cartographic materials reflecting possible changes in the meandering channel pattern. Only an average rate of channel lateral migration can be calculated. No information is obtained from the course and intensity of individual flood episodes contributing to the lateral migration of a river channel. Dendrochronological research makes it possible to estimate the lateral migration rate of the river channel more accurately. Dendrochronological methods have been applied in fluvial geomorphology (Alestalo, 1971; Shroder, 1980; Butler, 1987; Schweingruber, 1988; Hupp, 1988; Krapiec, 1995). The lateral migration of a channel becomes more rapid as discharges become larger and more frequent. During high discharges, concave banks are undercut and the trees growing upon them fall into the riverbed where they stay as CWD. As the discharge wave decreases mineral and organic material accumulate along the convex banks of the channel, which are later colonized by trees. The dendrochronological dating of trees and CWD makes it possible to determine the periods of increased lateral migration with an accuracy of 1 year. In this study, two dendrochronological methods were used to estimate the rate of lateral migration in the Mala Panew River, Poland. The first method was based on CWD dating. Individual discs from the CWD were cross-dated against the local chronology. The second method was based on flood plain levels and mid-channel islands dating. The oldest trees growing on every level and individual mid-channel island provide a minimal age of these deposits. Material and methods Mala Panew River The Mala Panew River is a sandy-bed meandering river which flows through the southern part of Poland (Fig. 1). The river drains an area of 2037 km2 and flows for 20 km through a closed forest. The forest is dominated by alder (Alnus glutinosa, Alnus incana) (42%), pine (Pinus sylvestris) (35%), willow (Salix purpurea, Salix fragilis) (21.5%) and spruce (Picea abies) (2%). In addition, ash (Fraxinus excelsior), larch (Larix decidua) and birch (Betula pendula) occur sporadically. Pine monocultures grow on poor sandy habitats at the bottom of the Ma"a Panew valley. Fragments of a marshy riverside meadow overgrown with ash and alder have been preserved in the area. The study area was located in the upper Mala Panew River basin, where 12 sites were selected in reaches A–B and C–D (Fig. 1). The slope gradient, alluvia, and channel width are similar in the study reaches, but the highest banks are in the A–B reaches and the lowest in the C–D one (Fig. 10). The valley bottom is filled with glacial and fluvioglacial sediments (410 m thick) from the Middle Polish Glaciation (Gilewska, 1972). The alluvia are composed of sands of varying grain size. A 2–3 km wide Pleistocene and three Holocene terraces (3–4, 2–3, 0.5–2 m) are observed in the Ma"a Panew River valley. The mean annual precipitation in the area ranges from 650 to 750 mm. On average, the riverbed is 10 m wide and up to 2 m deep. The most frequent water stages in the Mala Panew riverbed are 40–70 cm deep.The highest discharges in the Mala Panew channel occurred in the years 1953, 1966, 1968, 1970, 1982, 1985 and 1997 (Fig. 2). Coarse Woody Debris During a channel migration in a meandering river, the CWD are buried and remain in the sediment for a long time. Consequently it is a valuable material for reconstructing the stages of the valley formation (Wronski, 1974; Becker and Schirmer, 1977; Goslar, 1987; Kalicki and Krapiec, 1991, 1994, 1995). In this case, however, dendrochronological dating is valid when CWD are found in situ. A number of methods have been proposed for identifying the CWD embedded in situ in a channel deposit or a riverbed. One can assume that most CWD whose root systems and trunks rest in a riverbed and tree-tops on the bank are in situ. Redeposited CWD can be recognized by the lack of bark and sapwood, which has been separated from the rest of the log during the river transport. To distinguish re-deposited CWD, the analysis of the log orientation against the riverbed axis may also be helpful. Trees in situ are usually placed across on the riverbed, whereas re-deposited and reoriented stems occur in accordance with the river axis direction. ARTICLE IN PRESS I. Malik / Dendrochronologia 23 (2005) 29–38 31 Fig. 1. Location of study area. Methods Two dendrochronological methods for measuring lateral migration rate of the channel have been used in this paper. The first one is based on dating the pine CWD laying in situ in the concave bank. Standard dendrochronological methods were used for developing a local chronology and cross-dating the CWD (Fritts, 1971; Schweingruber, 1988). A local tree ring chronol- ogy for Pinus sylvestris was constructed for the past 70 years. The chronology was based on tree ring analysis from 10 cores collected from pines currently living in the Mala Panew River valley. Another stage of the study was the collection of discs from CWD pines laying in the riverbed. The discs were collected from about 0.5 m above the root system. The tree rings from CWD discs were cross-dated against the local chronology. The percentage consistent coefficient ARTICLE IN PRESS 32 I. Malik / Dendrochronologia 23 (2005) 29–38 Fig. 2. Water stages in Krupski Mlyn guage in the Mala Panew River. ARTICLE IN PRESS I. Malik / Dendrochronologia 23 (2005) 29–38 (Huber, 1943) was used to evaluate the similarity between the dendrochronological series obtained from individual discs and the chronology. Another method consists of dating trees growing on meander bar levels and mid-channel islands. The age of the oldest trees growing on a genetically homogenous level provides a minimum age for these deposits. The location of mid-channel islands, which relates to the river banks, as well as the dating of trees growing on the islands can also be helpful to determine the direction and intensity of the lateral migration of the meandering channel. Trees growing on meander bar levels and midchannel islands were sampled using a 40 cm Pressler’s borer. Three black alders with the largest diameters at breast height were selected from each of the meander bar levels. 33 45.7 m2. The erosion of the bank connected with the logs falling into the river channel at this site was estimated at 2.5 m, that means half of the distance needed to include 12 trees in a 18.3 m long delivery belt (i.e. 5 m, Fig. 4). The oldest CWD has been lying in the channel since Results Lateral migration rate based on Coarse Woody Debris dating Most of the CWD in the Mala Panew channel are from Pinus sylvestris. At site 1, six overthrown pines lay under the eroded edge of the concave bank (Fig. 3). The pines have been overthrown as a result of the erosion of the undercut surface of the terrace level. The position of CWD in relation to the riverbed and the presence of bark indicate that the trees were in situ. This is also supported by a very strong undercutting of the bank at the site examined. Semicircular erosion hollows are visible within the undercut – these are traces of the root systems of trees that currently lie in situ in the channel. The pine delivery belt (understood as the eroded bank below which the CWD lays) is 18.3 m long. Twelve trees grow in an area 18.3 m long
5 m wide, (i.e. 91.5 m2), which allocates an area of 7.62 m2 per tree. There are six CWD in the channel, representing a total eroded area of Fig. 3. CWD lying in situ in the Mala Panew riverbed. Fig. 4. Pines growing on eroded terrace and CWD lying in the Mala Panew riverbed in site 1. Fig. 5. Pines growing on eroded terrace and CWD lying in the Mala Panew riverbed in site 2. ARTICLE IN PRESS 34 I. Malik / Dendrochronologia 23 (2005) 29–38 1993, i.e. for 7 years. Consequently, the mean lateral migration rate of the Ma"a Panew channel for the interval 1993–2000 was 0.36 m/year. At site 2, as many as 12 Pinus sylvestris CWD were found in the riverbed. The CWD are stuck in situ in the riverbed. The tree supply belt is 47.5 m long. Twenty seven trees grow in an area of 239.5 m2 (47.5 m long
5 m wide), representing an area of 8.8 m2 per tree. The 12 CWD in the channel represent a total eroded area of 105.6 m2. Consequently, the erosion of the bank at this site was estimated in 2.02 m (Fig. 5). The oldest CWD lay in the channel since 1991 (9 years at the sampling time), which implies a mean lateral migration rate of 0.24 m/year. Fig. 6. Meander bar levels in the Mala Panew valley floor. Lateral migration rate based on dating riparian trees Meander bar levels were studied at site 3 (Fig. 6). Individual levels were formed as a result of successive withdrawal of the river’s concave bank and the simultaneous accumulation of alluvium within the convex bank. The minimum age of the youngest level of meander bar (A) can be estimated to be 26 years, being the age of the oldest alder (A. incana) growing on this level. The author has observed that new meander bars, formed after a recent flood, were colonized 3 years after formation. The oldest alders growing on the levels B and C of meander bars are 38 and 45 years old, respectively. Therefore, the highest level of the meander bar (C) is estimated to be at least 45 years old. It can be deducted from the 1:25 000 maps of 1912 that levels A, B and C of the meander bars did not exist before that year (Fig. 7). Based on the previous information, the period of individual level formation of the meander bars can be Table 1. Fig. 7. Varying in age levels of meander bar in site 3. Rate of minimal lateral migration of the Mala Panew channel in site 3 Levels Age of the oldest tree overgrowing the level (years) Level formation period Maximum level width (m) Minimum lateral migration rate A 26 6 6 m : 12 years ¼ 0.5 m/year B 38 7.5 7.5 m : 7 year ¼ 1.07 m/year C 45 Year 1999 – 26 years (age of the tree) – 3 years (min. age of the level not overgrown by alders) ¼ 1970. The level was formed between 1958 and 1970, i.e. it has a minimum of 12 years. Year 1999 – 38 years (age of the tree) – 3 years (min. age of the level not overgrown by alders) ¼ 1958. The level was formed between 1951 and 1958, i.e. it has a minimum of 7 years. Year 1999 – 45 years (age of the tree) –3 years (min. age of the level not overgrown by alders) ¼ 1951. The level was formed between 1912 and 1951. 6.5 ARTICLE IN PRESS I. Malik / Dendrochronologia 23 (2005) 29–38 roughly determined. The analysis indicate that level C was formed in the period 1912–1951, level B between 1951 and 1958, and level A during the interval 1958–1970. Based on the maximum width of individual levels, the minimum rate of lateral migration of the channel can be determined between 0.5 and 1.07 m/year (Table 1). At site number 4, the two lowest levels A and B of the meander bar are covered by trees of different ages. The age of the oldest tree within level A is 24 years. The oldest tree growing on level B is 46 years old (Fig. 8). The minimum rate of lateral migration at this site amounts to 0.48 m/year (Table 2). The lateral migration rate was also calculated by determining the age of the oldest trees growing on midchannel islands. The mid-channel islands are often located in channel axes (Fig. 9). Podzolic soils with an elluvial horizon in the islands were recorded at sites 35 6, 7, 9 and 11. Its thickness was from 20 to 35 cm. This horizon was formed over a thousand years (Prusinkiewicz et al., 1980). Therefore the islands were cut off from the concave banks and not formed in the riverbed. Alders growing on mid-channel islands are seldom older than 50 years. The youngest alders growing on spurs were 14 years old. Therefore, it was assumed that islands are colonized by trees of a minimum 14 years age. Hence the age of the oldest trees growing on the islands together with the distance from the receding bank may testify the channel’s lateral migration rate. Based on the island distances from the banks and the age of the oldest trees growing on the midchannel islands, the lateral migration rate at 8 sites has proceeded at a rate between 0.07 and 1.83 m/year (Table 3). Discussion The rate of lateral migration of the Mala Panew channel calculated with the two dendrochronological methods is different. The results obtained by means of CWD indicates a lateral migration rate between 0.24 Fig. 8. Varying in age levels of meander bar in site 4. Table 2. Fig. 9. Mid-channel island in the Mala Panew riverbed. Rate of minimal lateral migration of the Ma"a Panew channel in site 4 Levels Age of the oldest tree overgrowing the level (years) Level formation period Maximum level width (m) Minimum lateral migration rate A 24 10.5 10.5 : 22 years ¼ 0.47 m/ year B 46 Year 1999 – 24 years (age of the tree) – 3 years (min. age of the level not overgrown by alders) ¼ 1972. The level was formed between 1950 – 1972, i.e. it has a minimum of 22 years. Year 1999 – 46 years (age of the tree) – 3 years (min. age of the level not overgrown by alders) ¼ 1950. The level was formed between 1912 – 1950. 8.5 ARTICLE IN PRESS 36 Table 3. I. Malik / Dendrochronologia 23 (2005) 29–38 The channel lateral migration rate as based on the ages of alders growing on mid-channel islands Minimum distance from the eroded concave bank from which the island has been cut off (m) Age of the oldest tree growing on every islands (years) Lateral migration rate Site 5 3.5 35 Site 6 1.9 41 Site 7 6.2 22 Site 8 5.5 17 Site 9 6.3 24 Site 10 4.1 31 Site 11 6.8 20 Site12 3.2 19 35 – 14 ¼ 21 years 3.5 m : 21 years ¼ 0.17 m/year 41 – 14 ¼ 27 years 1.9 m : 27 years ¼ 0.07 m/year 22 – 14 ¼ 8 years 6.2 m : 8 years ¼ 0.77 m/year 17 – 14 ¼ 3 years 5.5 m : 3 years ¼ 1.83 m/year 24 – 14 ¼ 10 years 6.3 m : 10 years ¼ 0.63 m/year 31 – 14 ¼ 17 years 4.1 m : 17 years ¼ 0.24 m/year 20 – 14 ¼ 6 years 6.8 m : 6 years ¼ 1.13 m/year 19 – 14 ¼ 5 years 3.2 m : 5 years ¼ 0.64 m/year Table 4. Years Site 1 Site 2 The intensity of lateral erosion at the sampling sites based on the CWD fall into the riverbed 1991 8.8 m 1992 2 8.8 m 1993 2 6 m2 8.8 m2 1994 1995 2 8.8 m and 0.36 m/year. The lack of CWD in the Ma"a Panew riverbed before 1990 indicates that earlier discharges probably did not cause a migration of the channel at sites 1 and 2. However, older CWD could have been removed by the local population for heating purpose. Dating of the CWD shows that periods of most intense lateral erosion occurred between the years 1997 and 1993 (Table 4). The largest flood of the century in the Ma"a Panew valley occurred in 1997, which explains the large number of CWD dated 1997 in the riverbed. On the other hand, 1993 belongs to the period of lowest water levels in the Ma"a Panew channel in the 20th century. The drying of the banks and the fall in the ground water level could probably lead to the loosening of the sediments in the banks and the subsequent fall of CWD into the riverbed. Different lateral erosion rates have been obtained from the dates of trees growing on levels of meander bars and mid-channel islands (0.07–1.83 m/year). The results are presented in Fig. 10. The slower rate of lateral migration of the Mala Panew channel in sites 1, 2, 5 and 6 (sections A–B) probably reflects the higher banks at those sites (Fig. 10). In these sites the mean bank height was 2.4 m, almost double the height at the remaining sites (section C–D). 1996 1997 1998 12 m2 61.6 m2 8.8 m2 1999 The time formation of subsequent meander bar levels corresponds roughly with the dates of flood episodes in the Mala Panew riverbed as measured in the gauge record. During high water stages (1953–1957) levels B and C were probably formed at site 3 as well as level B at site 4. Similarly high water levels were recorded in the years 1966–1968 which might be related to the formation of meander level A at both sites. During the 1997 flood, the banks in sites 3 and 4 were not eroded. Probably in these sites water flowed outside of the riverbed, onto the floodplain and the higher terraces. Only sections with high banks, where water flowed in the riverbed as sites 1 and 2, were eroded. CWD and mid-channel island were restricted to some sections of the river. Results from the study sites should not be considered as an indicator of the average migration rate of the whole Mala Panew River. They provide information about lateral migration rates at the individual selected sites. The methods used in this study have same limitations. CWD remains in the absence of large floods which remove dead trees downstream. In small meandering rivers such as the Mala Panew River, large CWD deposits are rarely removed. However, local populations ARTICLE IN PRESS I. Malik / Dendrochronologia 23 (2005) 29–38 37 Fig. 10. Different results of lateral migration rate of the Mala Panew channel. pull out CWD from the riverbed for heating purpose. CWD lying in riverbeds above the water level also decay very quickly, reducing the amount of material for dating. The estimation of lateral migration rates based on dating trees growing on flood plain levels is more precise, but still has limitations. In this study, a 3 years minimal age of tree colonization on new surfaces was assumed. However, this value is highly variable depending on sites, species and climate conditions (Schweingruber, 1996). Similarly, lateral migration rates based on dates of mid-channel islands have large uncertainties related to the age of the oldest tree growing on individual form. Trees may die during the cutting of the island from the bank. As a consequence, the methods provide a minimal age for mid-channel islands, which could affect the determination of the lateral migration rates in individual sites. Conclusions The lateral migration rates of meandering rivers can be estimated by applying two different tree-ring methods:   Dating the CWD laying in situ at bank undercuts. Dating trees growing on meander bars and midchannel islands. In spite of the limitations mentioned above, the rates of channel lateral migration are more precisely determined using dendrochronological methods than historical maps. Individual flood episodes which have contributed to the channel’s lateral migration can be detected by dendrochronological methods. Acknowledgements The author wishes to thank Prof. Dr. hab. Kazimierz Klimek for helping to organize the dendrochronological laboratory. Special thanks to Prof. Dr. hab. Marek Krapiec who taught me dendrochronology. References Abernethy B, Rutherfurd ID. The effect of riparian tree roots on the mass-stability of riverbanks. Earth Surface Processes and Landforms 2000;25:921–37. Alestalo J. Dendrochronological interpretation of geomorphic processes. Fennia 1971;105:1–140. Becker B, Schirmer W. Paleoecological study of the Holocene valley development of the River Main, Southern Germany. Boreas 1977;6:303–21. Brooks AP, Brierley GJ. Mediated equilibrium: the influence of riparian vegetation and wood on the long-term evolution and behaviour a near-pristine river. Earth Surface Processes and Landforms 2002;27:343–67. ARTICLE IN PRESS 38 I. Malik / Dendrochronologia 23 (2005) 29–38 Brooks GR. Channel changes along the lower reaches of major Mackenzie River tributaries. Geological Survey of Canada 2001;547:159–66. Butler DR. Teaching general principles and applications of dendrogeomorphology. Journal of Geological Education 1987;35:64–70. Fritts HC. Dendroclimatology and dendroecology. Quaternary Research 1971;1:419–49. Gilewska S. Wyzyny Sla˛ sko-Ma"opolskie. In: Klimaszewski M, editor. Geomorfologia Polski. PWN: Warszawa; 1972. p. 233–50. Goslar T. Dendrochronological studies in the Gliwice Radiocarbon Laboratory, equipment, first results. Annalar Academiae Scientiarum Fennicae 1987;A(145):97–104. Gurnell AM, Gregory KJ, Petts GE. The role of coarse woody debris in forest aquatic habitats: implications for management. Aquatic Conservation: Marine and Freshwater Ecosystems 1995;5:143–66. Gurnell AM, Sweet R. The distribution of large woody debris accumulations and pools in relation to woodland stream management in a small, low-gradient stream. Earth Surface Processes and Landforms 1998;23(12):1101–21. Gregory KJ. Vegetation and river channel process interactions. In: Boon PJ, Calow PJ, Peth GE, editors. River Conservation and Management. New York: Wiley; 1992. p. 225–69. Hickin EJ, Nanson GC. Lateral migration rates of river bends. Journal of Hydraulic Engineering, American Society of Civil Engineers 1984;110(11):1557–67. Huber B. Über die Sicherheit jahrringchronologischer Datierung. Holz als Roh – und Werkstoff 1943;6:263–8. Hupp CR. Ecological aspects of flood geomorphology and paleoflood history. In: Baker VR, Kochel RC, Patton PC, editors. Flood Geomorphology. New York: Wiley; 1988. p. 335–6. Kalicki T, Krapiec M. Subboreal black oaks identified from the Vistula alluvia at Grabie near Cracow (South Poland). Kwartalnik AGH- Geologia 1991;17:155–71. Kalicki T, Krapiec M. Problemy datowan form i aluwiow metoda dendrochronologiczna na przyk"adzie doliny Wis"y ko"o Krakowa. Zeszyty Naukowe Politechniki Slaskiej 1994;1229:173–89. Kalicki T, Krapiec M. Problems of dating alluvium using buried subfossil trunks: lesson from the black oaks of the Vistula Valley, Central Europe. The Holocene 1995;5(2):243–50. Kaczka RJ. The role of coarse woody debris in fluvival processes during the flood of the July 1997, Kamienica Ła˛ cka valley, Beskidy Mountains, Poland. Studia Geomorphologica Carpatho-Balcanica 1999;33:119–29. Krapiec M. Metodyka badan dendrochronologicznych. In: Mycielska-Dowgiallo E, Rutkowski J, editors. Badania osadow czwartorzedowych. Warszawa: Wydzia" Geografii i Studiow Regionalnych; 1995. p. 318–28. Rachocki A. Wp"yw roslinnosci na ukszta"towanie koryt i brzegow rzek. Przeglad Geograficzny 1978;3:469–79. Prusinkiewicz Z, Bednarek R, Pokojska U. The soil of the podzolozems-class in Poland. Polish Geographical Review 1980;52:103–13. Schweingruber FH. Tree rings. Basics and Applications of Dendrochronology. Reidel: Dodrecht; 1988 276 pp. Schweingruber FH. Tree rings and environment. Dendroecology. Berne: Paul Haupt Verlag; 1996. p. 133–4. Shroder Jr JF. Dendrogeomorphology: review and new techniques of tree-ring dating. Progress in Physical Geography 1980;4:161–88. Thorne CR. Effects of vegetation on riverbank erosion and stability. In: Thornes JB, editor. Vegetation and Erosion. Chichester: Wiley; 1990. p. 125–44. Trafas K. Changes of the Vistula river bed east of Cracow in light of archival maps and photointerpretation. Prace Geograficzne 1975;40:1–85. Wronski J. Wiek bezwzgledny aluwiow niektorych rzek Dolnego Slaska. Przeglad Geologiczny 1974;12:602–6. Van Sickle J, Gregory SV. Modeling inputs of large woody debris to streams from falling trees. Canadian Journal of Forest Research 1990;20:1593–601.