Impact of a large flood on mountain river habitats, channel morphology, and valley infrastructure

    Impact of a large flood on mountain river habitats, channel morphology, and valley infrastructure Hanna Hajdukiewicz, Bartłomiej Wy˙zga, Paweł Miku´s, Joanna Zawiejska, Artur Radecki-Pawlik PII: DOI: Reference:

S0169-555X(15)30138-0 doi: 10.1016/j.geomorph.2015.09.003 GEOMOR 5369

To appear in:

Geomorphology

Received date: Revised date: Accepted date:

13 April 2015 3 September 2015 7 September 2015

Please cite this article as: Hajdukiewicz, Hanna, Wy˙zga, Bartlomiej, Miku´s, Pawel, Zawiejska, Joanna, Radecki-Pawlik, Artur, Impact of a large flood on mountain river habitats, channel morphology, and valley infrastructure, Geomorphology (2015), doi: 10.1016/j.geomorph.2015.09.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Impact of a large flood on mountain river habitats, channel morphology, and valley

T

infrastructure

IP

Hanna Hajdukiewicza,*, Bartłomiej Wyżgaa, Paweł Mikuśa, Joanna Zawiejskab, Artur

a

SC R

Radecki-Pawlikc

Institute of Nature Conservation, Polish Academy of Sciences, al. Mickiewicza 33, 31-120

b

Institute of Geography, Pedagogical University, ul. Podchorążych 2, 30-084 Kraków, Poland

Department of Hydraulic Engineering and Geotechnique, University of Agriculture, al.

MA

c

NU

Kraków, Poland

AC

CE P

TE

D

Mickiewicza 24/28, 30-059 Kraków, Poland

*

Corresponding

author.

Tel.:

+48-123703524;

Hajdukiewicz).

1

E-mail:

[email protected]

(H.

ACCEPTED MANUSCRIPT Abstract The Biała River, Polish Carpathians, was considerably modified by channelization and

T

channel incision in the twentieth century. To restore the Biała, establishing an erodible

IP

corridor was proposed in two river sections located in its mountain and foothill course. In

SC R

these sections, longer, unmanaged channel reaches alternate with short, channelized reaches; and channel narrowing and incision increases in the downstream direction. In June 2010 an 80-year flood occurred on the river; and this study aims at determining its effects on physical

NU

habitat conditions for river biota, channel morphology, and valley-floor infrastructure.

MA

Surveys of 10 pairs of closely located, unmanaged and channelized cross sections, performed in 2009 and in the late summer 2010, allowed us to assess the flood-induced changes to physical habitat conditions. A comparison of channel planforms determined before (2009) and

TE

D

after (2012) the flood provided information on the degree of channel widening as well as changes in the width of particular elements of the river’s active zone in eight stretches of the

CE P

Biała. The impact of the flood on valley-floor infrastructure was confronted with the degree of river widening in unmanaged and channelized river reaches. Before the flood, unmanaged

AC

cross sections were typified by finer bed material and greater lateral variability in depthaveraged and near-bed flow velocity than channelized cross sections. The flood tended to equalize habitat conditions in both types of river cross sections, obliterating differences (in particular physical habitat parameters) between channelized and unmanaged channel reaches. River widening mostly reflected an increase in the area of channel bars, whereas the widening of low-flow channels was less pronounced. A comparison of channel planform from 2009 and 2012 indicated that intense channel incision typical of downstream sections limited river widening by the flood. Active channel width increased by half in the unmanaged cross sections and by one-third in the channelized cross sections. However, damage to the valleyfloor infrastructure was practically limited to the channelized river reaches with reinforced

2

ACCEPTED MANUSCRIPT channel banks. This indicates incompetent management of riparian areas rather than the

T

degree of river widening as a principal reason for the economic losses during the flood.

SC R

IP

Keywords: mountain river; physical habitat conditions; channel morphology; flood damages

1. Introduction

Literature on floods in mountain watercourses abounds in information concerning

NU

geomorphic effects of the floods and hazards associated with these events (Wohl, 2010, 2011,

MA

and papers cited therein). During the past century mountain streams and rivers in densely populated areas were considerably modified by human activities, such as extensive channel regulation, in-channel sediment mining, and the construction of dams and weirs (Bravard and

TE

D

Petts, 1996; Wohl, 2006; Comiti, 2012). Recognition of the adverse effects of these impacts on channel stability of the rivers and the condition of their ecosystems has stimulated a large

CE P

amount of river restoration activities worldwide (e.g., Hillman and Brierley, 2005; Habersack and Piégay, 2008; Rinaldi et al., 2013). The identification of the geomorphic, ecological, and

AC

economic effects of floods on rivers undergoing restoration and their comparison with changes in the nearby reaches where human constraints are maintained are still lacking because of the recent nature of restoration activities. As a result of large-scale anomalies in the oceanic–atmospheric interaction, such as the North Atlantic Oscillation, increased flooding may occur at the scale of catchments (Foulds et al., 2014) or large regions (Pociask-Karteczka, 2006). Such was the situation in 2010 when large floods occurred at various times during the warm half-year over vast areas of central Europe (Lóczy, 2013), including catchments in the Western Carpathians where they resulted in considerable flood damage and channel adjustments (Frandorfer and Lehotský, 2011; Gorczyca et al., 2013). In the Polish Carpathians, the Biała was among the rivers with

3

ACCEPTED MANUSCRIPT the most extreme flood incidence in 2010. Interestingly, in the late 2000s an erodible corridor was established in mountain and foothill sections of this river. This restoration measure

T

consists in allowing free channel development in a delimited portion of the valley floor to

IP

reestablish a natural course of fluvial processes (e.g., Piégay et al., 1996; Nieznański et al.,

SC R

2008). This circumstance thus provided an opportunity to study the effects of the extreme flood of 2010 in the unmanaged reaches of the Biała and to compare them with those recorded in the adjacent channelized reaches.

NU

The importance of flooding in maintaining the function and integrity of aquatic

MA

ecosystems has long been recognised (Wydoski and Wick, 2011). According to the Flood Pulse Concept (Junk et al., 1989), seasonal or regularly recurring flooding is considered a driving force in a river–floodplain system, allowing lateral exchange of water, nutrients, and

TE

D

organisms between river channel and the connected floodplain. Extreme flood events are generally considered to periodically pose a threat on aquatic ecosystems (Wydoski and Wick,

CE P

2011). However, the unpredictable nature of such events is a cause for very few studies comparing post-flood condition of riverine biocoenoses with that recorded shortly before the

AC

flood. A study on benthic invertebrates in the Alpine river Isar indicated no significant changes resulting from an extreme flood (Hering et al., 2004), but the reliability of this finding was reduced by the fact that the initial survey had been performed a few years before the flood. Studies on various river reaches (from mountain to lowland) indicated relatively minor effects of extreme floods on fish fauna, with the greatest decline in abundance recorded among young-of-the-year individuals (Bischoff and Wolter, 2001; Lojkasek et al., 2005; Jurajda et al., 2006) and the individuals of eurytopic (Bischoff and Wolter, 2001) or pelagic and benthic species (Jurajda et al., 2006). Washing down of fish individuals, especially juveniles, by floodwaters (Bischoff and Wolter, 2001; Lojkasek et al., 2005) and fish mortality caused by retreat from the floodplain of water with very low concentration of

4

ACCEPTED MANUSCRIPT dissolved oxygen (Lusk et al., 2004) were indicated as the reasons for adverse changes among fish communities. While these effects are transient and reflect direct action of floodwaters on

T

fishes, only the study by Bischoff and Wolter (2001) considered flood-related changes in

IP

physical habitat structure with more prolonged effects on fish fauna and found that some

SC R

species benefited from increased habitat heterogeneity after the flood.

Flood damage and flood losses are intrinsic to the occurrence of major floods (Merritts, 2011). Economic losses resulting from major flood events have increased during

NU

recent decades at various spatial scales, from local to global scale (Kundzewicz et al., 2013).

MA

In Poland material damage caused by the flood of 2010 equalled 1% of gross national product of the country (Kundzewicz et al., 2012), and a considerable proportion of the damage occurred in the valleys of Carpathian tributaries to the Vistula River (Biedroń et al., 2011)

TE

D

where floods typically have high energy and are thus highly erosive (Kundzewicz et al., 2014). Accelerated water runoff from catchments caused by urbanization and an increase in

CE P

impervious surfaces (Hollis, 1975) and the loss of floodwater storage in floodplain areas resulting from river channelization and incision (Wyżga, 1997, 2001b) or flood embankment

AC

construction are frequently invoked to explain the high economic losses caused by major flood events. However, these factors do not apply to the sections of the Biała River considered in this study. During the last few decades the mountain part of the Biała catchment exhibited a marked increase in forest cover (Kozak et al., 2007) and no significant urbanization occurred there. The studied sections of the river are not embanked and channelized reaches constitute not more than 15% of their total length; channel incision apparent on a proportion of the sections length reduces floodwater storage in the floodplain areas (Czech et al., 2015, this issue), but the resultant increase in the magnitude of flood flows must affect downstream river sections rather than the incised sections themselves. It is thus interesting to consider factors that determined the amount and spatial pattern of the economic

5

ACCEPTED MANUSCRIPT losses caused by the extreme flood of 2010 in the studied sections of the Biała. Using the example of the Biała, this study analyses hydromorphological, geomorphic,

T

and economic effects of the extreme flood on a mountain river subjected to restoration. In

How did the flood affect physical habitat conditions for riverine biota in the

SC R



IP

particular, it aims to answer the following questions:

unmanaged, freely developing river reaches and the neighbouring channelized reaches?

What was the effect of the alternation of the unmanaged and the channelized reaches

NU



MA

and the differing degree of channel incision along the Biała on the spatial pattern of the changes in planform river geometry caused by the flood? What was the principal reason for the economic losses caused by the flood in the river

D



TE

valley?

CE P

2. Study area and the flood of June 2010 The gravel-bed Biała River drains a catchment with an area of 983 km2 in the Polish

AC

Carpathians (Fig. 1). In its upper course, this 102-km-long river flows through the low mountains of Beskid Niski where it is fed with coarse and medium-sized sandstone material. As only shallow, slope aquifers occur in this part of the catchment, here the river is typified by very high flow variability; at the Grybów gauging station with a 50-year-long record period (Fig. 1B), the coefficient of runoff irregularity (ratio of the highest and the lowest flow on record) equals 7500. The high flow variability and the delivery of coarse bed material explain that in unmanaged channel reaches the river forms a wide, multithread channel. In the middle course within the Ciężkowice foothills, the Biała runs across alternating sandstone and shale complexes that supply the river with cobble to pebble material and large volumes of fines. Fed with such material, the Biała maintains a gravel bed, but tends to form a sinuous

6

ACCEPTED MANUSCRIPT channel in its unmanaged reaches. In the lower part of its mountain course and along the foothill course, the Biała incised

T

by up to 2 m and its channel was considerably narrowed (up to a sixth of the original value) in

IP

the second half of the twentieth century. River training works (Wyżga, 2001a, 2008) and

SC R

uncontrolled, widespread in-channel gravel mining (Rinaldi et al., 2005; Wyżga et al., 2010) were the main causes of these channel changes, although a reduction in catchment sediment supply (Lach and Wyżga, 2002) following a considerable increase in forest cover in the

NU

mountains in the second half of the century (Kozak et al., 2007) might have intensified the

MA

scale of the channel adjustments.

Several studies have demonstrated that channelization and incision of Polish Carpathian rivers bring about a range of adverse effects, such as a reduction in floodplain

TE

D

inundation at given flood discharges (Wyżga, 2001b; Czech et al., 2015, this issue) and the resultant increase in flood hazard to downstream river reaches (Wyżga, 1997, 2008) or the

CE P

degradation of physical habitat integrity (Wyżga et al., 2009) and the ecological state of the rivers (Wyżga et al., 2013). An erodible river corridor was thus delimited in the mountain and

AC

foothills sections of the Biała (14.5 and 5.9 km in length, respectively) (Fig. 1) in order to mitigate the adverse effects of the channel changes and improve hydromorphological quality of the river. Because the delimited streamway has to encompass the existing infrastructure, both sections of the corridor consist of alternating longer (1-3 km) unmanaged reaches and shorter (0.1-0.3 km) channelized reaches in the vicinity of bridges. Annual precipitation in the catchment ranges from about 950 mm in its highest parts to 650-700 mm in the lowest parts (Niedźwiedź and Obrębska-Starklowa, 1991), and the respective values of the coefficient of runoff vary from 50% to <30% (Dynowska, 1991). At the Grybów gauging station located just downstream of the mountain river course, (catchment area of 210 km2; Fig. 1B), mean annual discharge amounts to 2.9 m3 s-1 and the average for

7

ACCEPTED MANUSCRIPT the maximum annual discharges to 108 m3 s-1. Frequent, moderate floods caused by snowmelt typically occur in early spring, and rare large floods can occur between May and August as a

T

result of frontal rains.

IP

A flood of the latter type occurred on the Biała in early June 2010. Late spring 2010

SC R

was remarkably wetter than usual over the entire central Europe (Bartholy and Pongrácz, 2013). Between 1 May and 4 June, 450-550 mm of rain fell in the mountain part of the Biała catchment, whereas in the foothill part 400-450 mm of precipitation was recorded (Cebulak et

NU

al., 2011). The first flood wave formed in mid-May: at the Grybów gauging station it had a peak discharge of 211 m3 s-1 with an 8-year recurrence interval. Between 30 May and 4 June,

MA

precipitation totals in the Biała catchment amounted to about 200 mm and half of that fell on a single day. With soil retention in the area exhausted by the preceding precipitation, the

TE

D

heavy rain caused the second flood wave that at Grybów had a peak discharge of 600 m 3 s-1

3. Methods

CE P

with an 80-year return period. The impacts are analysed in this study.

AC

3.1. River study units

The study was carried out in three types of longitudinal river units. First, the river course was divided into unmanaged reaches and channelized reaches (in the vicinity of bridges) where surveys of physical habitat parameters and observations of flood damages were conducted. In turn, analyses of changes in river planform were conducted in parts of the river course between successive bridges, here called river stretches. Particular stretches thus comprise an unmanaged reach and parts of adjacent channelized reaches. River stretches were further subdivided into 100-m-long segments in which GIS measurements were carried out.

3.2. Changes in physical habitat parameters

8

ACCEPTED MANUSCRIPT Physical habitat parameters in the river were investigated during base-flow conditions in late July and August 2009 and again after the flood at similar flow conditions in July 2010.

T

Ten sites were selected for the study: 7 sites were located in the longer section of the erodible

IP

corridor in the mountain course of the Biała and 3 other in the shorter section of the corridor

SC R

in the foothill river course (Fig. 1). Each site consisted of a pair of cross sections in the neighbouring unmanaged and channelized river reaches that were located between significant tributaries (Fig. 1).

NU

In the cross sections, elevation profiles were surveyed and active channel width and

MA

aggregated width of low-flow channels were determined. Subsequently, water depth, near-bed and depth-averaged flow velocity, and mean grain size of surface bed material were measured at equal intervals across the low-flow channels. Because of a substantial increase in river size

TE

D

between the first and last investigated sites, different spacing of measurement points in lowflow channels had to be applied to obtain a similar number of measurements from all sites; the

CE P

measurements were spaced 0.5 m apart at sites 1–7, 0.75 m apart at sites 8–9, and 1 m apart at site 10. Flow velocity was measured using an Ott Nautilus C 2000 electromagnetic current

AC

meter; the measurements at 0.6 of the depth indicated depth-averaged velocity, and those performed 1 cm above the bed surface indicated near-bed velocity. We used transect sampling (that yields results equivalent to bulk sieve analysis; Diplas and Sutherland, 1988) to determine the grain size of gravelly bed material, with 15 particles measured at each point; this sample size allowed us to limit sampling to the area characterized by hydraulic measurements. The distribution of the b axis diameters of measured particles in each sample was subsequently determined and mean grain size of the sample was calculated as an average of the 3rd, 8th, and 13th particles in the sequence. These diameters were equivalent to the 20th, 50th, and 80th percentiles of the grain-size distribution; and their average was the closest available approximation of the formula of Folk and Ward (1957). Samples of sands

9

ACCEPTED MANUSCRIPT and muds were taken to a laboratory, where sieving and hydrometer analyses, respectively, were used to determine their grain-size distribution. Mean grain size of the sediments was

T

then calculated from the same percentiles of the distribution. Finally, means and coefficients

IP

of variation of water depth, near-bed and depth-averaged flow velocity, and bed material grain

SC R

size were calculated for each cross section.

Values of most of the analysed physical habitat parameters changed with changing river discharge. It was thus necessary to verify comparability of flow conditions for the

NU

surveyed cross sections in particular years as well as between years. First, a power regression model between river discharge measured in particular cross sections and catchment area was

MA

estimated (Wyżga et al., 2013) separately for 2009 and 2010. The obtained relationships explained a very high proportion of the variation in the measured discharges (R2 = 0.99 for

TE

D

2009 and for 2010), thus confirming comparability of flow conditions among the cross sections surveyed in a given year. Second, values of base-flow discharges during the

CE P

measurement campaigns in either year were considered independent statistical samples, and thus the Mann-Whitney test was used to determine significance of the differences between the

AC

discharges at which physical habitat parameters were measured in the cross sections in 2009 and 2010. A lack of statistically significant differences in the discharges between these years was found for channelized cross sections (p = 0.76) and unmanaged cross sections (p = 0.68). A Wilcoxon signed rank test was used to examine the differences in physical habitat parameters typifying the river in 2009 before the large flood and in 2010 after the flood. This test for dependent statistical samples was used because morphological and sedimentary characteristics of the river recorded in 2010 were dependent on their initial state existing prior to the flood. The test was used to separately analyse the flood-related changes in unmanaged river cross sections and in channelized cross sections. A two-way analysis of variance was subsequently employed to verify whether the mutual relation in particular physical habitat

10

ACCEPTED MANUSCRIPT characteristics between unmanaged and channelized cross sections changed as a result of the flood. We used this parametric test despite non-normal distribution of some of the analysed

T

statistical samples as it has no nonparametric equivalent. We should note that its use to test

IP

the samples lacking a normal distribution does not question the occurrence of recognized

SC R

relationships but may inhibit recognition of other relations (because non-normal distribution of the analysed sample makes it difficult to reject the null hypothesis of the test, although it is

NU

actually incorrect) (Lindman, 1974).

3.3. Changes in planform geometry of the river

MA

Changes in the planform geometry of the Biała River caused by the flood of June 2010 were analysed along 20.4 km of the river course with delimited erodible corridor. On that

TE

D

length, eight river stretches were distinguished between successive bridges, with six stretches located in the upper part of the erodible corridor and two in its lower part. Each stretch was

CE P

subsequently divided into river segments 100 m in length. On the orthophoto from 2009, morphological elements of the river’s active zone: low-flow channels, side bars, mid-channel

AC

bars, and forested islands were digitized in an ArcGIS environment. With pixel size of the orthophoto of 0.25 m and 0.3 m estimated error of orthorectification of the aerial photos, the linear accuracy of the digitization of channel forms on the orthophoto can be estimated at 0.5 m. The area of the polygons of particular morphological elements of the river was determined for each river segment, and the average width of each element together with their aggregated width (indicating the width of the active river zone) in the segment was calculated by dividing the area by the segment length. Data for successive river segments were subsequently averaged for the eight distinguished stretches of the erodible corridor. In summer 2012 particular morphological elements of the river were delineated in the field using a ‘Trimble GeoXT’ GPS receiver. It allowed us to obtain precision of about 0.4 m

11

ACCEPTED MANUSCRIPT after correction with base station data, which was comparable with the accuracy of measurements on the orthophoto. The GPS measurements were next elaborated in ArcGIS,

T

yielding data about the width of particular morphological elements of the river in successive

IP

river segments and their averages for the distinguished river stretches. Differences in the

SC R

width of low-flow channel, channel bars, and active river zone between 2009 and 2012 were subsequently examined with a Wilcoxon test; and the contribution of changes in the average width of particular morphological elements of the river to the total change in active zone

MA

NU

width was determined.

3.4. Infrastructure damage caused by the June 2010 flood After the flood we identified damages to valley-floor infrastructure and allocated them

TE

D

to either unmanaged or channelized river reaches. The damages were classified into two categories: (i) resulting from the erosional action of channel flow, and (ii) resulting from

CE P

bridge clogging with wood and inundation of the adjacent valley floor. Data about approximate costs of infrastructure repairs were subsequently obtained from the local

AC

authorities. Comparisons of the damages/costs of repairs with data about the average river widening in unmanaged and channelized river reaches allowed us to infer about the primary cause of the damages and recommend changes in the future management of the valley floor.

4. Results

4.1. Changes in physical habitat parameters in channelized and unmanaged cross sections The flood did not change the single-thread character of channelized cross sections, while in the unmanaged cross sections the average number of low-flow channels decreased from 2 prior to the flood to 1.6 after the flood (Table 1). The width of the active river zone

12

ACCEPTED MANUSCRIPT increased in both types of cross sections. The increase was greater in the unmanaged cross sections (Fig. 2): the average width of the active zone increased by 51% from 52.4 m in 2009

T

to 79.3 m after the flood of June 2010 and the increase was statistically significant (Wilcoxon

IP

test, p = 0.005). A significant increase of the active zone width was also recorded in the

SC R

channelized cross sections (p = 0.01): the width increased on average by 32% from 17.4 m in 2009 to 23 m after the flood (Table 1). The width of low-flow channels in unmanaged cross sections amounted to 14.3 m in 2009 and to 16.7 m in 2010, whereas the respective values in

NU

channelized cross sections equalled 10.3 and 12.1 m (Table 1). However, the changes in this

MA

parameter were not significant statistically (p = 0.24 for the unmanaged cross sections and p = 0.17 for the channelized ones).

The flood did not cause a significant change in mean flow depth or the cross-sectional

TE

D

variability of flow depth in either of the cross section types (Table 1). A similar lack of a significant change was also recorded with respect to mean values of depth-averaged and near-

CE P

bed flow velocity in either type of the river cross sections (Table 1). However, opposite tendencies of change in the variability of depth-averaged and of near-bed velocity were

AC

recorded in the unmanaged and in the channelized cross sections (Table 1). In the unmanaged cross sections, the lateral variability of flow velocity was reduced: the coefficient of variation of near-bed velocity by 29% (p = 0.03) and that of depth-averaged velocity by 21% (with the change close to the boundary of significance, p = 0.06). In the channelized cross sections, the cross-sectional variability of flow velocity increased (Fig. 3): the coefficient of variation of depth-averaged velocity by 49% (p = 0.03) and that of near-bed velocity by 21% (p = 0.07). The flood resulted in a decrease in mean grain size of bed material (Table 1). In unmanaged cross sections the decrease by 16% (from 39.1 mm in 2009 to 33 mm in 2010) was not statistically significant (p = 0.39). In channelized cross sections, mean grain size of bed material decreased from 60.5 mm in 2009 to 38.4 mm in 2010, that is by 37%, and the

13

ACCEPTED MANUSCRIPT change was statistically significant (p = 0.005). The variability in mean grain size of bed material exhibited small, yet contradictory, changes in the two types of cross sections,

T

decreasing in the unmanaged and increasing in the channelized cross sections.

IP

Results from two-way analysis of variance indicated changes in the physical habitat

SC R

parameters resulting from the flood of June 2010 that markedly altered the previous relations of the parameters between the two types of river cross sections (Table 1; Fig. 4). The flood caused the active channel width to increase in both types of cross sections, but the increase

NU

was larger in the unmanaged ones. Before the flood, unmanaged cross sections were typified

MA

by markedly coarser grain size of bed material and lower variability in depth-averaged and near-bed velocity than channelized cross sections. As a result of the opposite changes in the variability of depth-averaged and near-bed flow velocity in the two types of cross sections and

TE

D

the larger decrease in mean grain size of bed material in channelized cross sections, the former differences in the values of these parameters between both cross section types became

CE P

obliterated.

AC

4.2. Changes in planform geometry of the river along the erodible corridor The increase in the width of the river’s active zone (Figs. 2 and 4) caused by the floods of May and June 2010 (with 8- and 80-year recurrence interval, respectively) resulted in a reactivation of numerous valley-side and riverbank undercuts, particularly in the mountain river course (Fig. 5). Delivery of coarse material from the now active undercuts and the energy of the flood of July 2011, with a peak discharge of 110 m3 s-1 and a 3-year recurrence interval, enabled the channel change initiated by the flood of May 2010 to continue. Comparison of planform geometry of the river recorded in the summer of 2012 with that presented on the orthophoto from 2009 provides information on the change resulting from the three flood events; however, as field observations indicated that a vast majority of the

14

ACCEPTED MANUSCRIPT recorded changes had occurred during the flood of June 2010, below they are considered to have been caused by that flood event.

T

Comparison of the planform geometry of the Biała River before and after the large

IP

flood indicated a statistically significant increase in the width of the active river zone in all

SC R

eight studied stretches (Wilcoxon test, p value between 0.002 and 0.000001; Fig. 6). The greatest increase (by 95%-105%) occurred in the first three stretches of the erodible river corridor (ERC) in the mountain course of the Biała. Farther downstream, the average width of

NU

the active river zone increased by 48%-76% within the mountain part of the ERC (stretches 4-

MA

6) and by 40%-55% in its foothill part (stretches 7-8). The increase in the average width of the active river zone was accompanied by an increase in the difference between minimum and maximum river width in a given stretch. The largest such change (an increase by 67% to

TE

D

132%) was recorded in stretches 3 and 4, where the river flows within a wide valley floor and its channel is not or only slightly incised into bedrock. In stretch 5 where the river flows away

CE P

from the valley sides and its channel is deeply (up to 2 m) incised, the difference between the minimum and maximum river width increased less – by 43%. In the remaining stretches with

AC

the river flowing along valley side and the channel incised into bedrock, the increase in the variation of the active zone width ranged from 5% to 35%. Before the flood the river in stretches 4-6 (i.e., in the lower part of the mountain section of the ERC; Fig. 1) exhibited a progressive downstream decrease in the average and the variation of active zone width (Fig. 6), reflecting the increasing amount of channel incision coupled with lateral river confinement in stretch 6. Although the flood increased the average width and the width variation in the river stretches, the pattern of these parameters remained unchanged; and after the flood, downstream of stretch 3, the river was still typified by a progressive decrease in average width and width variability (Fig. 6). The flood increased the width of low-flow channel(s) in nearly all stretches of the

15

ACCEPTED MANUSCRIPT ERC (Fig. 7). The average low-flow channel width increased the most (by 72%-79%) in stretches 1-2 (p value between 0.003 and 0.0002), whereas in stretches 3-5 and 7-8 the width

T

increase ranged from 22% to 39% (p value between 0.04 and 0.0001). In stretch 6 no

IP

significant change in low-flow channel width was recorded (Fig. 7). In all stretches of the

SC R

ERC the relative increase in low-flow channel width caused by the flood was smaller than that of the active river zone.

The flood increased the aggregated width of lateral bars in all the studied stretches,

NU

while the width of mid-channel bars increased in some stretches and decreased in others,

MA

except stretches 7-8 where such bars ceased to occur (Fig. 7). As a result of these diverse tendencies, the aggregated width of channel bars increased by 59% to 144% (p value between 0.04 and 0.000002). The greatest increase was recorded in stretch 3 where the river flows

TE

D

within a wide valley floor (Fig. 8) and is only sporadically constrained by the valley side. Only in the two uppermost stretches was the relative increase in channel bar width slightly

CE P

lower than that in the active zone width, and it was greater in the remaining stretches. This indicates that river widening caused by bank erosion mainly led to the increase in the width of

AC

channel bars (Fig. 8).

Prior to the flood, 26 forested islands occurred in stretches 2-4 where they covered between 1.3% and 13.7% of the river area. The flood eroded almost all the islands. However, the avulsion of side channels and fragmentation of forested floodplain during the flood caused the formation of new islands that were incorporated into the active river zone, and after the flood a few pioneer islands developed from resprouting living wood deposited on gravel bars. As a result, in 2012 forested islands were present in five river stretches where they represented from 0.1% to 6.1% of the area of the active river zone (Fig. 8).

4.3. Damage caused by the flood of June 2010

16

ACCEPTED MANUSCRIPT The flood of 2010 originated in the catchments of Carpathian tributaries to the Vistula and subsequently the flood wave propagated toward lowland reaches of the river, causing

T

inundation even at distances of up to 400 km from the Carpathians (Kozak et al., 2013).

IP

About 35% of the total flood damage in the country occurred in the Małopolska Province

SC R

covering about half of the area of the catchments of Carpathian tributaries to the Vistula, with damage to the infrastructure as well as communal and private buildings representing 88% of the economic losses in the province (Biedroń et al., 2011). Costs of the flood damage to the

NU

infrastructure and buildings in the 20.4-km-long studied sections of the Biała River were

MA

estimated at 10.2 million PLN, about 0.36% of the economic losses in the province, and they deserve attention not because of the value but of the spatial distribution of the damage. In the unmanaged reaches of the erodible corridor, active channel width increased by

TE

D

half on average during the flood of July 2010 (Table 1; Figs. 2, 9). However, effects of river widening were practically limited to the erosion of riparian land overgrown with young to

CE P

submature alluvial forest (Figs. 2, 9). The only infrastructural object destroyed by the flood was a steel foot-bridge in stretch 5, with the reconstruction cost estimated at 0.1 million PLN,

AC

i.e. about 1% of the total economic losses in the studied river sections. In the channelized reaches, the river widened by one-third on average (Table 1). However, in three channelized reaches widening of the river resulted in a destruction of bridges and bank reinforcements in the vicinity of the bridge cross sections (Fig. 10); subsequent costs of rebuilding of this infrastructure (with new bridges constructed with considerably wider span than previously) amounted to 10 million PLN, i.e., 98% of the total losses in the studied sections. One of the bridges in the lower section of the ERC trapped considerable amounts of wood debris on its pillars. Wood jams that formed on the upstream side of the pillars caused a blockage of about 40% of the bridge cross section (Fig. 11A). The increased water level upstream of the bridge led to markedly higher elevation than would occur at undisturbed flow

17

ACCEPTED MANUSCRIPT conditions and directed a considerable proportion of the flood flow onto the inhabited valley floor (Fig. 11B). As a result, basements of several houses in the vicinity of the bridge were

T

inundated and a local road damaged where a hydraulic jump that formed on the downstream

IP

side of the road embankment caused intensive turbulence of floodwater and scour of the

SC R

ground (Fig. 11B). The economic losses at that site amounted to 0.1 million PLN – 1% of the total losses in the studied reaches.

NU

5. Discussion

This study has demonstrated that the flood of June 2010 modified physical habitat

MA

conditions in the mountainous Biała River. This confirms an earlier observation by Bischoff and Wolter (2001) that extreme floods may cause a significant change in habitat heterogeneity

TE

D

in rivers. Before the flood of 2010, unmanaged cross sections of the Biała supported finer bed material and displayed greater variability in depth-averaged and near-bed flow velocity in the

CE P

low-flow channels than channelized cross sections. The flood markedly reduced the former difference in mean grain size of surface bed material and obliterated differences in the cross-

AC

sectional variability of flow velocities between the unmanaged and the channelized river cross sections. The changes were beneficial for the channelized river reaches where habitat heterogeneity increased, but they negatively affected the unmanaged reaches where this heterogeneity diminished. While the above mentioned changes in habitat conditions within low-flow channels affect the functioning of riverine biota during low to medium flows, the increase in active channel width caused by the flood radically changed the conditions for the organisms during future flood events. The resultant changes encompass slowing down of the increase in flow velocity with increasing discharge (Negishi et al., 2002), a reduction in flow velocity at given flood discharges (cf. Czech et al., 2015, this issue) and increased availability of shallow-

18

ACCEPTED MANUSCRIPT depth, slow-velocity refugia over channel bars and in marginal channel areas (Rempel et al., 1999) in the wider channel. These changes will thus help fish and benthic invertebrates to find

T

flow refugia and escape the shear forces, hence reducing their washing down during floods.

IP

As the widening of the Biała caused by the flood of 2010 was greater in unmanaged river

SC R

reaches, its beneficial influence on the resistance of riverine communities to flood disturbances will also be greater in the unmanaged than in the channelized reaches. The substantial widening of the Biała channel during the flood of June 2010 was a

NU

typical morphological change caused by extreme floods in mountain rivers (e.g., Naef and

MA

Bezzola, 1990; Fuller, 2008; Krapesch et al., 2011; Nardi and Rinaldi, 2015). However, local conditions exerted a considerable influence on the spatial pattern of the river widening. The alternation of short, channelized reaches where the channel width increased less and of longer,

TE

D

unmanaged reaches with the larger width increase markedly amplified downstream differences in active channel width in comparison with the pre-flood situation. Moreover, the

CE P

largest increase in average channel width and in the width variation took place in the stretches in which the river is unconfined and its channel is entirely formed in alluvium. River

AC

confinement by a valley side and channel incision into bedrock diminished the increase in channel width and in its downstream variability. The survey of river cross sections carried out shortly after the flood of June 2010 indicated some increase in the width of low-flow channels that was not, however, significant statistically. The GPS measurements performed two years later documented a greater increase in the width of low-flow channels between 2009 and 2012. The continuation of the trend of width increase of these channels after the major flood of June 2010 probably reflected the adjustment of the river to increased sediment delivery from reactivated valley-side and riverbank undercuts and the resultant increase in bedload transport during the small flood of July 2011 and snowmelt freshets in the spring of 2011 and 2012. However, the essential

19

ACCEPTED MANUSCRIPT morphological change of the river caused by the flood of June 2010 consisted in the widening of channel bars that resulted from bank retreat and the widening of the active channel of the

T

river. The increase in active channel width and the associated bar widening reflected river

IP

adjustment to convey the flow of the extreme flood.

SC R

Floods result in the turnover of vegetation patches in the multithread channels, eroding a proportion of islands and resetting their formation (Mikuś et al., 2013; Surian et al., 2015). In the Biała, the flood of June 2010 eroded most hitherto existing islands in stretches 2-4. At

NU

the same time, the flood reactivated functioning of some former channels that had previously been incorporated into the floodplain, including patches of riparian forest in the widened

MA

active river zone. As a result, stretches 1 and 5 — previously lacking islands — now support such forms; and in stretch 3 the share of islands has markedly increased (Fig. 7).

TE

D

Geomorphological significance of extreme floods is considered to increase upstream in a river network (e.g., Froehlich and Starkel, 1987; Miller, 1990), and the inventory of flood

CE P

damage on the Biała confirmed a general observation that economic losses caused by floods on mountain rivers are mostly associated with bank erosion and rapid channel widening rather

AC

than with valley-floor inundation. Such flood losses in mountain valleys reflect high stream power of flood flows conditioned by steep channel slope and the rapid runoff from mountain catchments. The occurrence of intense bank erosion is more likely in confined sections of the channel (Shroba et al., 1979), provided that higher unit stream power typical of such sections (Magilligan, 1992) is associated with erodible river banks. Several studies demonstrated that erosion in confined channel sections during large floods led to major changes in channel configuration (e.g., Fuller, 2008; Thompson and Croke, 2013). In the Biała, unit stream power of the flood flow must have been especially high in channelized river reaches where the channel was ~ 3 times narrower than in the unmanaged reaches (Table 1) and where the floodplain in the vicinity of bridges is constricted (Czech et al., 2015, this issue). Despite the

20

ACCEPTED MANUSCRIPT higher energy of the flood flow in the channelized reaches, greater resistance of the reinforced banks to erosion resulted in a lower degree of channel widening than in the unmanaged

T

reaches as the bank retreat could have only been initiated after the reinforcements had been

IP

destroyed by the flood. However, almost the whole material damage caused by the erosional

SC R

action of the flood occurred in the channelized reaches of the Biała. This indicates that the principal reason for the economic losses during the flood was inappropriate management of riparian areas and the construction of infrastructures unadjusted to the flood hazard on the

NU

river (cf. Arnaud-Fassetta et al., 2005) rather than intensive river widening. The artificially

MA

reduced channel width at bridges resulted in insufficient channel conveyance; this, together with the high flow energy in the channelized reaches, facilitated bank scour and the destruction of the bridges.

TE

D

The alternation of reaches with a different degree of channel confinement and thus with markedly different unit stream power of flood flows is common in European rivers as a

CE P

result of varying management of riparian areas along their valleys. Understanding of the coexistence and functioning of such dissimilar but adjacent river morphologies is important in

AC

order to design and apply appropriate river management, especially in mountain rivers where flood events of high magnitude may cause significant damage and induce long-term channel change. As the damage is likely to concentrate in intensively managed valley reaches with a regulated river channel, it is vital to adjust the infrastructure in these reaches so that it can sustain flood flows and transport of sediment from the upstream, mountain part of the catchment, thus limiting the risk of material loss. In turn, delimiting an erodible river corridor with the riparian areas left unmanaged is useful where sufficient room exists on the valley floor as the presented case study demonstrated the effectiveness of such solution in reducing flood losses on mountain rivers. The expansion of riparian forest that took place in the valleys of Polish Carpathian

21

ACCEPTED MANUSCRIPT rivers in the twentieth century (Wyżga, 2007; Wyżga et al., 2012) considerably increased the delivery of large wood to their channels. A significant proportion of the wood is rapidly

T

recruited from retreating channel banks during flood events and may become a significant

IP

factor of flood hazard if subsequently accumulated at critical sections such as bridges (Ruiz-

SC R

Villanueva et al., 2013; Kundzewicz et al., 2014). This threat was realized during the flood of 2010 when one of the bridges on the Biała became partially blocked with accumulated wood and the flood flow was directed onto the adjacent valley floor that otherwise would not have

NU

been inundated. Also here the flood damage resulted from improper management decisions as

MA

the wood was trapped by the bridge with the distance between pillars shorter than the height of the riparian trees and thus the length of transported wood pieces. This indicates that river management must take into account the delivery, transport, and deposition of large wood

6. Conclusions

CE P

2012; Rinaldi et al., 2015).

TE

D

during floods to eliminate or reduce the related material losses (Wyżga, 2007; Comiti et al.,

AC

The analysis of the effects of the extreme flood of June 2010 on the mountainous Biała River highlighted changes to the river morphology, physical habitat conditions for riverine biota and the valley-floor infrastructure, and allowed for their comparison between unmanaged and channelized river reaches. The analysis of changes to the river morphology confirmed the ability of high-magnitude floods to considerably widen the channels of mountain rivers, indicated by previous studies. However, the pattern of the river widening reflected the legacy of human impacts on the Biała and the recent introduction of an erodible corridor in mountain and foothill sections of the river. A larger scale of the river widening in unmanaged reaches within the erodible corridor reflected rapid retreat of the banks lacking reinforcements. The high energy of the flood flow confined in the narrow, regulated channel

22

ACCEPTED MANUSCRIPT caused channelized reaches also to be widened, but bank reinforcements reduced the scale of the river widening. Another factor limiting river widening during the flood was channel

T

incision into bedrock, which generally increases downstream in the Biała.

IP

Changes to physical habitat conditions for river biota caused by the extreme flood

SC R

seem to have different effects at low to medium flows and at flood flows. The flood increased the variability of physical habitat conditions during low to medium flows in channelized river reaches and decreased it in unmanaged reaches, hence obliterating the previous differences in

NU

the hydromorphological state of the river in its reaches with different style of channel

MA

management. The simplification of physical habitat conditions in the unmanaged reaches will thus require a recovery in the following years with lower flood flows. At the same time, river widening will reduce flow velocity and shear forces during future flood events (cf. Czech et

TE

D

al., 2015, this issue), decreasing washing down of river biota during floods. This hydraulic effect of the geomorphological river change, beneficial for the resistance of river communities

CE P

to flood disturbance, will be greater in the unmanaged reaches of the Biała where the relative increase in river width was larger.

AC

With current population density and increasing development in mountain valleys historically concentrated along rivers, restoration of entire watercourses is not feasible for many reasons. Wherever possible, schemes such as the erodible river corridor can be effectively introduced; but the remaining reaches, supporting infrastructure such as bridges, will continue to be maintained and function as artificial reaches. As flood damage is likely to concentrate in the nonrestored reaches, adjusting the infrastructure in these (often narrow) reaches to the flood hazard is crucial to limit the risk of material loss.

Acknowledgments This study was completed within the scope of the Research Project DEC-

23

ACCEPTED MANUSCRIPT 2013/09/B/ST10/00056 financed by the National Science Centre of Poland. The paper makes use of hydrometric data provided by the Institute of Meteorology and Water Management –

T

State Research Institute (IMGW-PIB). We also gratefully acknowledge helpful comments of

SC R

IP

Francesco Comiti and an anonymous reviewer.

References

Arnaud-Fassetta, G., Cossart, E., Fort, M., 2005. Hydro-geomorphic hazards and impact of

NU

man-made structures during the catastrophic flood of June 2000 in the Upper Guil catchment (Queyras, Southern French Alps). Geomorphology 66, 41–67.

MA

Bartholy, J., Pongrácz, R., 2013. Spring and summer weather in 2010: regular or exceptional? In: Lóczy, D. (Ed.), Geomorphological Impacts of Extreme Weather. Case Studies from Central and Eastern Europe. Springer, Dordrecht, pp. 3–19.

D

Biedroń, I., Bogdańska-Warmuz, R., Kwiecień, M., 2011. Szkody i straty powodziowe w

TE

dorzeczu Wisły (Flood damages and losses in the Vistula River drainage basin). In:

CE P

Maciejewski, M., Ostojski, M.S., Walczukiewicz, T. (Eds.), Dorzecze Wisły – monografia powodzi maj-czerwiec 2010. Instytut Meteorologii i Gospodarki Wodnej, Warszawa, pp. 171–182.

AC

Bischoff, A., Wolter, C., 2001. The flood of the century on the River Oder: effects on the 0+ fish community and implications for floodplain restoration. Regulated Rivers: Research and Management 17, 171–190. Bravard, J.P., Petts, G.E., 1996. Human impacts on fluvial hydrosystems. In: Petts, G.E., Amoros, C. (Eds.), Fluvial Hydrosystems. Chapman and Hall, London, pp. 242–262. Cebulak, E., Kilar, P., Limanówka, D., Mizera, M., Pyrc, R., 2011. Wysokość, natężenie i przestrzenny rozkład opadów atmosferycznych (Amount, intensity and spatial distribution of precipitation). In: Maciejewski, M., Ostojski, M.S., Walczukiewicz, T. (Eds.), Dorzecze Wisły – monografia powodzi maj-czerwiec 2010. Instytut Meteorologii i Gospodarki Wodnej, Warszawa, pp. 27–41. Comiti, F., 2012. How natural are Alpine mountain rivers? Evidence from the Italian Alps. Earth Surface Processes and Landforms 37, 693–707.

24

ACCEPTED MANUSCRIPT Comiti, F., D’Agostino, V., Moser, M., Lenzi, M.A., Bettella, F., Dell’Agnese, A., Rigon, E., Gius, S., Mazzorana, B., 2012. Preventing wood-related hazards in mountain basins: From wood load estimation to designing retention structures. In: 12th Congress Interpraevent 2012

T

– Grenoble/France, pp. 651–662.

IP

Czech, W., Radecki-Pawlik, A., Wyżga, B., Hajdukiewicz, H., 2015, this issue. Modelling the

SC R

flooding capacity of a Polish Carpathian river: A comparison of constrained and free channel conditions. Geomorphology.

Diplas, P., Sutherland, A.J., 1988. Sampling techniques for gravel sized sediments. Journal of

NU

Hydraulic Engineering 114, 484–501.

Dynowska, I., 1991. Bilans wodny. In: Dynowska, I., Maciejewski, M. (Eds.), Dorzecze

MA

górnej Wisły. PWN, Warszawa-Kraków, pp. 223–227, in Polish. Folk, R.L., Ward, W.C., 1957. Brazos River bar: a study in the significance of grain

D

sizeparameters. Journal of Sedimentary Petrology 27, 3–26.

TE

Foulds, S.A., Griffiths, H.M., Macklin, M.G., Brewer, P.A., 2014. Geomorphological records of extreme floods and their relationship to decadal-scale climate change. Geomorphology

CE P

216, 193–207.

Frandorfer, M., Lehotský, M., 2011. Channel adjustment of a mixed bedrock-alluvial river in response to recent extreme flood events (the upper Topl’a River). Geomorphologia Slovaca et

AC

Bohemica 2011, 59–71.

Froehlich, W., Starkel, L., 1987. Normal and extreme monsoon rains – their role in the shaping of the Darjeeling Himalaya. Studia Geomorphologica Carpatho-Balcanica 21, 129– 158. Fuller, I.C., 2008. Geomorphic impacts of a 100-year flood: Kiwitea Stream, Manawatu catchment, New Zealand. Geomorphology 98, 84–95. Gorczyca, E., Krzemień, K., Wrońska-Wałach, D., Sobucki, M., 2013. Channel changes due to extreme rainfalls in the Polish Carpathians. In: Lóczy, D. (Ed.), Geomorphological Impacts of Extreme Weather. Case Studies from Central and Eastern Europe. Springer, Dordrecht, pp. 23–35. Habersack, H., Piégay, H., 2008. River restoration in the Alps and their surrounding: past experience and future challenges. In: Habersack, H., Piégay, H., Rinaldi, M. (Eds.), Gravel-

25

ACCEPTED MANUSCRIPT Bed Rivers VI: From Process Understanding to River Restoration. Developments in Earth Surface Processes 11, Elsevier, Amsterdam, pp. 703–737. Hering, D., Gerhard, M., Manderbach, R., Reich, M., 2004. Impact of a 100-year flood on

IP

T

vegetation, benthic invertebrates, riparian fauna and large woody debris standing stock in an alpine floodplain. River Research and Applications 20, 445–457.

SC R

Hillman, H., Brierley, G., 2005. A critical review of catchment-scale stream rehabilitation programmes. Progress in Physical Geography 29, 50–76.

Hollis, G.E., 1975. The effect of urbanization on floods of different recurrence interval. Water

NU

Resources Research 11, 431–435.

Junk, J.W., Bayley, P.B., Sparks, R.E., 1989. The flood pulse concept in river flood plain

MA

systems. Canadian Special Publications of Fisheries and Aquatic Sciences 106, 110–121. Jurajda, P., Reichard, M., Smith, C., 2006. Immediate impact of an extensive summer flood

D

on the adult fish assemblage of a channelized lowland river. Journal of Freshwater Ecology

TE

21, 493–501.

Kozak, J., Estreguil, C., Troll, M., 2007. Forest cover changes in the northern Carpathians in

CE P

the 20th century: a slow transition. Journal of Land Use Science 2, 127–146. Kozak, J., Ostapowicz, K., Bytnerowicz, A., Wyżga, B., 2013. The Carpathian Mountains: Challanges for the Central and Eastern European landmark. In: Kozak, J., Ostapowicz, K.,

AC

Bytnerowcz, A., Wyżga, B. (Eds.), The Carpathians: Integrating Nature and Society Towards Sustainability. Springer, Berlin, pp. 1–11. Krapesch, G., Hauer, C., Habersack, H., 2011. Scale oriented analysis of river width changes due to extreme flood hazards. Natural Hazards and Earth System Sciences 11, 2137–2147. Kundzewicz, Z.W., Dobrowolski, A., Lorenc, H., Niedźwiedź, T., Pińskwar, I., Kowalczak, P., 2012. Floods in Poland. In: Kundzewicz, Z.W. (Ed.), Changes in Flood Risk in Europe. International Association of Hydrological Sciences Special Publication 10, pp. 319–334. Kundzewicz, Z.W., Pińskwar, I., Brakenridge, G.R., 2013. Large floods in Europe, 19852009. Hydrological Sciences Journal 58, 1–7. Kundzewicz, Z.W., Stoffel, M., Kaczka, R.J., Wyżga, B., Niedźwiedź. T., Pińskwar, I., RuizVillanueva, V., Łupikasza, E., Czajka, B., Ballesteros-Canovas, J.A., Małarzewski. Ł., Choryński, A., Janecka, K., Mikuś, P., 2014. Floods at the northern foothills of the Tatra

26

ACCEPTED MANUSCRIPT Mountains – A Polish-Swiss research project. Acta Geophysica 62, 620–641. Lach, J., Wyżga, B., 2002. Channel incision and flow increase of the upper Wisłoka River, southern Poland, subsequent to the reafforestation of its catchment. Earth Surface Processes

IP

T

and Landforms 27, 445–462.

Lindman, H.R., 1974. Analysis of Variance in Complex Experimental Designs. Freeman, San

SC R

Francisco (297 pp.).

Lóczy, D. (Ed.), 2013. Geomorphological Impacts of Extreme Weather. Case Studies from Central and Eastern Europe. Springer, Dordrecht (373 pp.).

NU

Lojkasek, B., Lusk, S., Halacka, K., Luskova, V., Drozd, P., 2005. The impact of the extreme floods in July 1997 on the ichthyocenosis of the Oder catchment area (Czech Republic).

MA

Hydrobiologia 548, 11–22.

Lusk, S., Halacka, K., Lusková, V., Holub, M., 2004. Impact of extreme flood on fishes in

D

rivers and their floodplains. Ecohydrology and Hydrobiology 4, 173–181.

TE

Magilligan, F.J., 1992. Thresholds and the spatial variability of flood power during extreme floods. Geomorphology 5, 373–390.

CE P

Merritts, D., 2011. The effects of variable river flow on human communities. In: Wohl, E.E. (Ed.), Inland Flood Hazards: Human, Riparian, and Aquatic Communities. Cambridge Univ. Press, New York, pp. 271–290.

AC

Mikuś, P., Wyżga, B., Kaczka, R.J., Walusiak, E., Zawiejska, J., 2013. Islands in a European mountain river: Linkages with large wood deposition, flood flow and plant diversity. Geomorphology 202, 115–127. Miller, A.J., 1990. Flood hydrology and geomorphic effectiveness in the central Appalachians. Earth Surface Processes and Landforms 15, 119–134. Naef, F., Bezzola, G.R., 1990. Hydrology and morphological consequences of the 1987 flood event in the upper Reuss valley. In: Sinniger, R.O., Monbaron, M. (Eds.), Hydrology in Mountainous Regions II. Artificial Reservoirs, Water and Slopes. International Association of Hydrological Sciences Publication 194, pp. 339–354. Nardi, L., Rinaldi, M., 2015. Spatio-temporal patterns of channel changes in response to a major flood event: the case of the Magra River (central-northern Italy). Earth Surface Processes and Landforms 40, 326–339.

27

ACCEPTED MANUSCRIPT Negishi, J.N., Inoue, M., Nunokawa, M., 2002. Effects of channelisation on stream habitat in relation to a spate and flow refugia for macroinvertebrates in northern Japan. Freshwater Biology 47, 1515–1529.

T

Niedźwiedź, T., Obrębska-Starklowa, B., 1991. Klimat. In: Dynowska, I., Maciejewski, M.

IP

(Eds.), Dorzecze górnej Wisły. PWN, Warszawa-Kraków, 68–84, in Polish.

SC R

Nieznański, P., Wyżga, B., Obrdlik, P., 2008. Oder border meanders: A concept of the erodible river corridor and its implementation. In: Gumiero, B., Rinaldi, M., Fokkens, B.

River Restoration, Venice, pp. 479–486.

NU

(Eds.), IVth ECRR International Conference on River Restoration 2008. European Centre for

Piégay, H., Barge, O., Landon, N., 1996. Streamway concept applied to river mobility/human

MA

use conflict management. In: Rivertech 96: 1st International Conference on New/Emerging Concepts for Rivers. International Water Resources Association, Chicago, pp. 681–688.

TE

review. Ambio 35, 312-314.

D

Pociask-Karteczka, J., 2006. River hydrology and the North Atlantic Oscillation – a general

Rempel, L.L., Richardson, J.S., Healey, M.C., 1999. Flow refugia for benthic

CE P

macroinvertebrates during flooding of a large river. Journal of the North American Benthological Society 18, 34–48. Rinaldi, M., Wyżga, B., Surian, N., 2005. Sediment mining in alluvial channels: physical

AC

effects and management perspectives. River Research and Applications 21, 805–828. Rinaldi, M., Wyżga, B., Dufour, S., Bertoldi, W., Gurnell, A., 2013. River processes and implications for fluvial ecogeomorphology: a European perspective. In: Shroder, J., Butler, D., Hupp, C.R. (Eds.), Treatise on Geomorphology. Ecogeomorphology, vol. 12. Academic Press, San Diego, pp. 37–52. Rinaldi, M., Surian, N., Comiti, F., Bussettini, M., 2015. A methodological framework for hydromorphological assessment, analysis and monitoring (IDRAIM) aimed at promoting integrated

river

management.

Geomorphology,

http://dx.doi.org/10.1016/j.geomorph.2015.05.010 Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Eguibar, M.A., Pardo-Igúzquiza, E., 2013. Reconstruction of a flash flood with large wood transport and its influence on hazard patterns in an ungauged mountain basin. Hydrological Processes 27, 3424–3437.

28

ACCEPTED MANUSCRIPT Shroba, R.R., Schmidt, P.W., Crosby, E.J., Hansen, W.R., Soule, J.M., 1979. Storm and flood of July 31-August 1, 1976 in the Big Thompson River and Cache la Poudre River Basins, Larimer and Weld Counties, Colorado. Geologic and geomorphic effects in the Big

IP

T

Thompson Canyon area, Larimer County. US Geological Survey Professional Paper 1115B, 87–152.

SC R

Surian, N., Barban, M., Ziliani, L., Monegato, G., Bertoldi, W., Comiti, F., 2015. Vegetation turnover in a braided river: frequency and effectiveness of floods of different magnitude. Earth Surface Processes and Landforms, 40, 542–558.

NU

Thompson, C., Croke, J., 2013. Geomorphic effects, flood power, and channel competence of a catastrophic flood in confined and unconfined reaches of the upper Lockyer valley,

MA

southeast Queensland, Australia. Geomorphology 197, 156–169. Wohl, E., 2006. Human impacts to mountain streams. Geomorphology 79, 217–248.

D

Wohl, E., 2010. Mountain Rivers Revisited. American Geophysical Union, Washington (573

TE

pp.).

Wohl, E. (Ed.), 2011. Inland Flood Hazards: Human, Riparian, and Aquatic Communities.

CE P

Cambridge Univ. Press, New York (518 pp.). Wydoski, R.S., Wick, E.J., 2011. Flooding and aquatic ecosystems. In: Wohl, E.E. (Ed.), Inland Flood Hazards: Human, Riparian, and Aquatic Communities. Cambridge Univ. Press,

AC

New York, pp. 238–268.

Wyżga, B., 1997. Methods for studying the response of flood flows to channel change. Journal of Hydrology 198, 271–288. Wyżga, B., 2001a. A geomorphologist’s criticism of the engineering approach to channelization of gravel-bed rivers: case study of the Raba River, Polish Carpathians. Environmental Management 28, 341–358. Wyżga, B., 2001b. Impact of the channelization-induced incision of the Skawa and Wisłoka Rivers, southern Poland, on the conditions of overbank deposition. Regulated Rivers: Research and Management 17, 85–100. Wyżga, B., 2007. Gruby rumosz drzewny: depozycja w rzece górskiej, postrzeganie i wykorzystanie do rewitalizacji cieków górskich (Large woody debris: storage in a mountain river, perception, and use in the restoration of mountain watercourses). Instytut Ochrony

29

ACCEPTED MANUSCRIPT Przyrody PAN, Kraków ( 176 pp.) in Polish, with English summary. Wyżga, B., 2008. A review on channel incision in the Polish Carpathian rivers during the 20th century. In: Habersack, H., Piégay, H., Rinaldi, M. (Eds.), Gravel-Bed Rivers VI: From

IP

T

Process Understanding to River Restoration. Developments in Earth Surface Processes 11, Elsevier, Amsterdam, pp. 525–556.

SC R

Wyżga, B., Amirowicz, A., Radecki-Pawlik, A., Zawiejska, J., 2009. Hydromorphological conditions, potential fish habitats and the fish community in a mountain river subjected to variable human impacts, the Czarny Dunajec, Polish Carpathians. River Research and

NU

Applications 25, 517–536.

Wyżga, B., Hajdukiewicz, H., Radecki-Pawlik, A., Zawiejska, J., 2010. Eksploatacja osadów

MA

z koryt rzek górskich – skutki środowiskowe i procedury oceny (Exploitation of sediments from mountain river beds – environmental impact and evaluation procedures). Gospodarka

D

Wodna 6, 243–249, in Polish, with English summary.

TE

Wyżga, B., Zawiejska, J., Radecki-Pawlik, A., Hajdukiewicz, H., 2012. Environmental change, hydromorphological reference conditions and the restoration of Polish Carpathian rivers. Earth

CE P

Surface Processes and Landforms 37, 1213-1226. Wyżga, B., Oglęcki, P., Hajdukiewicz, H., Zawiejska, J., Radecki-Pawlik, A., Skalski, T., Mikuś, P., 2013. Interpretation of the invertebrate-based BMWP-PL index in a gravel-bed

AC

river: insight from the Polish Carpathians. Hydrobiologia 712, 71–88.

30

ACCEPTED MANUSCRIPT Figure captions: Fig. 1. (A) Location of the Biała River in relation to physiogeographic regions of southern Poland. (B) Drainage network of the upper and middle parts of the Biała catchment and

T

detailed setting of the studied sites and river stretches. 1 – Mountains of intermediate and low

IP

height; 2 – foothills; 3 – intramontane and submontane depressions; 4 – boundary of the Biała

SC R

catchment; 5 – flow-gauging stations; 6 – study sites (each consisting of a pair of unmanaged and channelized cross sections) and the river sections with delimited erodible river corridor; 7 – river stretches with analysed changes in planar geometry of the river.

NU

Fig. 2. View and morphology of the Biała River in the surveyed unmanaged cross section from site 6 before (2009) and after the major flood (2010). For low-flow channels, mean flow

MA

velocity is indicated at 0.5-m intervals.

Fig. 3. View and morphology of the Biała River in the surveyed channelized cross section

D

from site 6 before (2009) and after the major flood (2010). For low-flow channels, mean flow

TE

velocity is indicated at 0.5-m intervals.

Fig. 4. Changes in active channel width, mean grain size of surface bed material, and the

CE P

coefficients of variation of depth-averaged and near-bed velocity in the unmanaged and channelized cross sections of the Biała River caused by the major flood of June 2010. Whiskers indicate minimum and maximum values of the parameters among the cross sections

AC

and points show the means. Statistical significance of temporal changes of the parameters is shown by a Wilcoxon signed-rank test, while that of the combined changes along temporal and channel management-related gradients is tested by two-way analysis of variance. Fig. 5. Considerable widening of the Biała River during the flood of June 2010 increased channel–slope coupling and activated many sources of coarse material delivery to the channel along the mountain river course. Fig. 6. Mean and the range of active zone width in eight studied stretches of the Biała River before (2009; light grey diagrams) and after (2012; dark grey diagrams) the major flood of June 2010. Statistical significance of the differences between the mean active zone width recorded in 2009 and 2012, determined by the Wilcoxon test, is indicated for each river stretch. p values < 0.05 are indicated in bold. Fig. 7. Comparison of the average width of particular elements of the active river zone in

31

ACCEPTED MANUSCRIPT eight studied stretches of the Biała River before (2009; light grey columns) and after (2012; dark grey columns) the major flood of June 2010. Fig. 8. Comparison of the elements of the active zone of the Biała River in stretch 3 before

T

(2009) and after (2012) the major flood of June 2010. 1 – Low-flow channel; 2 – side bar; 3 –

IP

mid-channel bar; 4 – vegetated island.

SC R

Fig. 9. View of the Biała River in an unmanaged reach before (2009) and after the major flood (2010).

Fig. 10. View of the Biała River in an artificially constricted reach before (2009) and after the

NU

major flood (2010).

Fig. 11. (A) Large woody debris trapped during the flood of June 2010 on pillars of the bridge

MA

at Jankowa in the foothill course of the Biała River, and (B) damage caused by floodwaters

AC

CE P

TE

D

directed onto the valley floor as a result of the bridge blockage.

32

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

Fig. 1. (A) Location of the Biała River in relation to physiogeographic regions of southern Poland. (B) Drainage network of the upper and middle parts of the Biała catchment and detailed setting of the studied sites and river stretches. 1 – Mountains of intermediate and low height; 2 – foothills; 3 – intramontane and submontane depressions; 4 – boundary of the Biała catchment; 5 – flow-gauging stations; 6 – study sites (each consisting of a pair of unmanaged and channelized cross sections) and the river sections with delimited erodible river corridor; 7 – river stretches with analysed changes in planar geometry of the river.

33

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

Fig. 2. View and morphology of the Biała River in the surveyed unmanaged cross section from site 6 before (2009) and after the major flood (2010). For low-flow channels, mean flow velocity is indicated at 0.5-m intervals.

34

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

Fig. 3. View and morphology of the Biała River in the surveyed channelized cross section from site 6 before (2009) and after the major flood (2010). For low-flow channels, mean flow velocity is indicated at 0.5-m intervals.

35

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

Fig. 4. Changes in active channel width, mean grain size of surface bed material, and the coefficients of variation of depth-averaged and near-bed velocity in the unmanaged and channelized cross sections of the Biała River caused by the major flood of June 2010. Whiskers indicate minimum and maximum values of the parameters among the cross sections and points show the means. Statistical significance of temporal changes of the parameters is shown by a Wilcoxon signed-rank test, while that of the combined changes along temporal and channel management-related gradients is tested by two-way analysis of variance.

36

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

Fig. 5. Considerable widening of the Biała River during the flood of June 2010 increased channel–slope coupling and activated many sources of coarse material delivery to the channel along the mountain river course.

37

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

Fig. 6. Mean and the range of active zone width in eight studied stretches of the Biała River before (2009; light grey diagrams) and after (2012; dark grey diagrams) the major flood of June 2010. Statistical significance of the differences between the mean active zone width recorded in 2009 and 2012, determined by the Wilcoxon test, is indicated for each river stretch. p values < 0.05 are indicated in bold.

38

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

Fig. 7. Comparison of the average width of particular elements of the active river zone in eight studied stretches of the Biała River before (2009; light grey columns) and after (2012; dark grey columns) the major flood of June 2010.

39

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

Fig. 8. Comparison of the elements of the active zone of the Biała River in stretch 3 before (2009) and after (2012) the major flood of June 2010. 1 – Low-flow channel; 2 – side bar; 3 – mid-channel bar; 4 – vegetated island.

40

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

Fig. 9. View of the Biała River in an unmanaged reach before (2009) and after the major flood (2010).

41

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

Fig. 10. View of the Biała River in an artificially constricted reach before (2009) and after the major flood (2010).

42

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

Fig. 11. (A) Large woody debris trapped during the flood of June 2010 on pillars of the bridge at Jankowa in the foothill course of the Biała River, and (B) damage caused by floodwaters directed onto the valley floor as a result of the bridge blockage.

43

ACCEPTED MANUSCRIPT Highlights We analyse effects of a large flood on physical habitat conditions, channel morphology and

T

infrastructure.

IP

The flood equalized habitat conditions in channelized and unmanaged river reaches.

SC R

The flood widened the river mostly by expanding the area of channel bars. River widening in unmanaged reaches was greater than in channelized reaches.

AC

CE P

TE

D

MA

NU

Flood damage to valley-floor infrastructure was limited to channelized reaches.

44

ACCEPTED MANUSCRIPT Table 1 Results of a Wilcoxon signed-rank test for the significance of difference between particular physical habitat parameters recorded in a given type of the studied cross sections of the Biała River before (2009) and after (2010) the major flood of June 2010, and the significance of changes along the temporal and the channel management-related gradients shown by two-way analysis of variance

Number of low-flow channels

2.0

Channelized

1.2

1.2

p = 1.00

52.4

79.3

p = 0.005

17.4

23.0

p = 0.01

14.3

16.7

p = 0.24

10.3

12.1

p = 0.17

Unmanaged

0.26

0.21

p = 0.58

Channelized

0.24

0.22

p = 0.58

Unmanaged

0.559

0.465

p = 0.14

Channelized

0.437

0.504

p = 0.28

Unmanaged

0.28

0.30

p = 0.72

Channelized

0.34

0.30

p = 0.80

Unmanaged

0.899

0.711

p = 0.06

Channelized

0.509

0.757

p = 0.03

Unmanaged

0.16

0.20

p = 0.14

Channelized

0.20

0.21

p = 0.88

Unmanaged

0.897

0.654

p = 0.03

Channelized

0.630

0.765

p = 0.07

Unmanaged

39.1

33.0

p = 0.39

Channelized

60.5

38.4

p = 0.005

Unmanaged

0.548

0.475

p = 0.58

Channelized

0.187

0.256

p = 0.28

NU

Channelized

MA

Flow depth: coefficient of variation

CE P

Near-bed velocity: mean (m s-1)

TE

Depth-averaged velocity: coefficient of variation

D

Depth-averaged velocity: mean (m s-1)

Near-bed velocity: coefficient of variation

AC

Grain size: mean (mm)

Grain size: coefficient of variation

T

p = 0.14

Unmanaged

Flow depth: mean (m)

2010

Significance of the Wilcoxon test

1.6

Channelized Low-flow channel width (m)

2009

Unmanaged

Unmanaged

Active channel width (m)

Year

IP

Type of cross section

SC R

Physical habitat parameter

45

Significance of two-way Anova p = 0.27

p = 0.21

p = 0.88

p = 0.64

p = 0.09

p = 0.47

p = 0.005

p = 0.45

p = 0.04

p = 0.02

p = 0.30