Introduction

In large river systems ecological demands and requirements for inland navigation are often conflicting. However, the needs for riverine waterways, such as sufficient water depth, are not necessarily contradicting ecological interests (Habersack et al., 2007). Modern river restoration projects often aim to improve waterway conditions or flood protection together with an improvement of the ecological situation. For the latter measures such as the removal of riverbank protection or sidearm reconnection are envisaged. Although these measures are mostly intended to be harmonized interdisciplinary, river engineers must be sensitive to avoid conflicts and heavy influences on riverine biota. Therefore, the implementation process should go along with an intensive integrative monitoring on various scales (Habersack, 2000). For example, the removal of bank protection initiates side erosion processes leading to a shift of bank lines, increased river widths and eventually to a changed bank gradient with an increased percentage of shallow areas. In many regulated rivers, a shift toward lateral activity and banks of gentle slopes means a shift toward the natural conditions corresponding to the “Leitbild” theory. The “visionary Leitbild” defines the natural, type-specific reference condition primarily to provide an objective benchmark for restoration efforts (Jungwirth et al., 2002). It is assumed that these changes normally go along with increased availability of suitable spawning and nursery habitats for the fish fauna and aquatic invertebrates, and has positive effects on natural production processes. But the existence of potentially suitable habitats due to a changed morphology is not the only criterion for the success of a restoration project. Within a navigable reach, the immediate effects of ship traffic—pollution, increase of shear stress, wave splash, and sediment turbulences—may lead to negative consequences particularly on riverine biota (Smart et al., 1985; Rütten, 1994; Zauner & Schiemer, 1994; Kucera-Hirzinger et al., 2009). Strong wash waves can be generated by a small ship at high speeds or by a large ship at moderate speeds, operating on a near-shore fairway or on an inland waterway (Jiang et al., 2002). Bank erosion is one of the foremost concerns related to boat waves, but disruption of habitat, resuspension of bottom sediments, and damage to aquatic plants (macrophytes) are other areas of concern (Maynord, 2005). Thus, positive effects of river restoration may be partly reduced when potentially suitable habitats are not useable because of disturbances by wave splash or abrupt drawdown, leaving riverine biota onto dry banks (Holland, 1986; Adams et al., 1999) or exceeding their metabolic abilities (Barett et al., 1992; Wolter & Arlinghaus, 2003; Wolter et al., 2004). Habitat loss caused by ship induced drawdown emerged as the most pervasive threat to biodiversity, contributing to the endangerment of 85% of the species analyzed by Wilcove et al. (1998). Shallow areas of large rivers—often presenting optimal instream nursery habitats regarding feeding and growth—are more affected by a high temporary arial loss caused by vessel-induced drawdown (Keckeis & Schiemer, 2001; Arlinghaus et al., 2002; Kucera-Hirzinger et al., 2009).

Therefore both objectives are graded as important: the augmentation of potentially suitable habitats and their evaluation in terms of actual usability.

Some studies aimed to describe vessel-induced waves, indicating that large vessels generate large drawdown and small wave heights, whereas small vessels such as pleasure crafts generate small drawdown and large wave heights (Parchure et al., 2001). Other studies show that the highest ship-generated waves are generally induced by the smaller vessels that operate at higher speeds rather than the large tankers and cargo ships (Sorensen, 1967). Many recent studies are addressed to specific vessels (Dam et al., 2008; Gharbi et al., 2010) or the waves that are generated by one single vessel at different speeds (McConchie & Toleman, 2003) for detailed determination of the impact of certain parameters. A lot of studies give complex insight on sediment suspension and subsequent river-bank erosion through wash waves (Nanson et al., 1994; Parchure et al. 2001; Bauer et al., 2002; McConchie & Toleman, 2003; Gharbi et al., 2010) or wave influence on aquatic plants (Magdi et al., 1999).

The movement of a vessel through water and the subsequent dynamic displacement of water, results in a velocity field around the hull. This velocity field is responsible for the pressure distribution along the surface and hence, the generation of waves. Depending on the surrounding boundaries, especially the shallow water effect and the blockage effect, the Kelvin wave pattern is transformed during propagation. (Oebius, 2000; PIANC, 2003). Hence the overall displacement of water by the hull of relatively large ships sailing in relatively small canals—as we are confronted with in the presented study—leads to large displacement waves, also called Bernoulli wakes, characterized by a significant drawdown and surge of water (PIANC, 2003). When operating close to critical speed, conventional vessels and high-speed vessels can produce solitary type waves, which are of very long period and can travel several vessel lengths ahead (Dand et al., 1999; Whittaker et al., 2001; PIANC, 2003). These displacement waves seem to have the greatest impact on bank-near drawdown effects (area loss, bank erosion) and therefore have been systematically recorded and described in many previous studies (e.g. Schoellhamer, 1996; Rapaglia et al., 2011). A 2-D mathematical model was also developed to describe displacement-wave’s propagation (Di Silvio et al., 2011). Bhowmink et al. (1992) presented a detailed description of controlled vessel passages at the Upper Mississippi River System; however, a characterization of vessel-induced waves actually occurring in rivers with a description of the spectrum of the waves and their effect on different bank types during varying discharge stages is lacking. Hence, the present paper aims to (i) typify the wave characteristics of the dominant ship types at the Danube River, (ii) describe the interaction of ship induced waves with different bank types including gentle and steep slopes, (iii) analyze the effect of the highest generated drawdown on the different shore types, (iv) develop river engineering strategies suitable for state-of-the-art restoration projects and (v) provide feasible and applicable basic information and data for biotic studies. A companion paper (Schludermann et al., 2013) is discussing the effects of ship-induced wave wash on the early stages of the offspring (YOY—young of the year) of riverine fish species.

Methods

Study site

The survey was conducted at a 3-km long river reach near the municipality of Hainburg at the Danube River, east of Vienna, situated between river kilometers 1,884.50 and 1,887.50. According to the classification of Nanson & Knighton (1996) the investigation area is classified as “gravel-dominated, laterally active anabranching river” (Hohensinner et al., 2005) and characterized by vigorous lateral activity (Nanson & Knighton, 1996).

The stretch contains a river bend, curved to the orographic right, with an apex near km 1,886.25, followed by a bend curved to the left, with an apex at km 1,885.50. The Danube between Vienna and Bratislava was heavily regulated and channelised at the end of the nineteenth century and thus both banks are rip-rap protected and a total of 15 groyne structures influence the riverine morphology. Two gravel bars have developed downstream of the two bends apexes. Within the stretch the Johler side arm branches off on the orographic right side and joins the Danube again downstream at the end of the stretch. Mean water depths within the waterway area lie at 3.0 m at low flow conditions, 4.5 m at medium flow and 8.0 m at the highest navigable flow. Extreme values range from 2.0 up to 6 m (at scour regions) within the waterway at low flow conditions. Average bed slope is about 0.0004, the average arithmetic mean subsurface sediment diameter is d m = 27.5 mm, the median d 50 = 21.2 mm and the d 90 = 59.9 mm near Hainburg.

The Danube near Hainburg can be described as moderate nival (snowmelt-dominated) water regime, where high flows occur during the snowmelt and storm periods (often raised by summer and autumn rains) and low flow periods are prevailing during cold seasons. Figure 1 shows the Danube hydrogram pointing out that the highest mean discharge can be found during spawning periods (spring to early summer). The absolute discharge minimum and maximum as well as the average discharge monthly maximum and minimum are depicted to show the high variability over the years.

Fig. 1
figure 1

Danube hydrogram showing the average monthly discharge (thick line), the average monthly maximum (dashed line) and minimum (dotted line) and the absolute discharge maximum and minimum (gray lines) for each month for the hydrograph from 1977 to 2010

Full size image

The regulated low discharge (RNQ, 94% exceedance duration) is determined at 915 m3 s−1, the annual mean discharge (MQ) at 1,930 m3 s−1, and the highest navigable and bank-full discharge (HSQ, 1% exceedance duration) at 5,060 m3 s−1.

The wave measurements were conducted at three different bank types within the 3-km long reach, including gravel bars and different types of groyne fields (Fig. 2).

Fig. 2
figure 2

Study site within Austria, catchment area, detailed view of the study site (Danube River-km: 1,884–1,888) and sampling locations (black dots) with brief description of the different bank types; Inset with schematized morphology of different bank types. Visualization of ship traces (yellow lines) of vessel type 1 (passenger ships) over a period of 70 days; the fairway is delimited by white lines

Full size image

Bank type 1 is a 200-m long groyne field with a groyne-length of about 130 m. It is characterized by gravel accumulation and a two vortex flow field. The groyne field at bank type 2 is much smaller (120 m long and 100 m wide) with higher water depth and a single vortex flow field. Bank type 3 is located at the beginning of a 500-m long gravel bar characterized by a gentle slope toward the shipping fairway.

Wave height measurements

To provide times series of the wave heights (strictly speaking water depths), two DOBIE wave gauges (NIWA Instrument Systems, 2001) were deployed at each sampling location. The DOBIE wave gauge consists of a pressure transducer (0–5 bar, with 0.04% accuracy) connected to a micro-controller, a battery power supply and a flash card, all housed in a pressure case, allowing an autonomous deployment. Usually they are used for marine purposes and were adapted in the scope of this study for the needs of continuous sampling in a river. As the DOBIE samples in bursts (periods of sampling activity separated by periods of inactivity), two devices had to be coupled in order to insure continuous sampling over a period of 3 weeks. Therefore, the two wave gauges were mounted onto a massive steel plate to insure a stable device position entailing constant measurement conditions (Fig. 3) and programmed to sample at an alternating time interval. Each burst consists of at most 2,048 sampling points, leading to a maximum sampling duration of 409.6 s, at a sampling interval of 0.2 s.

Fig. 3
figure 3

DOBIE Wave gauges and the massive steel plate used to provide stable device position

Full size image

In order to achieve a sufficient overlapping time the burst interval of each DOBIE was set to 12 min, resulting in an overlap period of 49.2 s. The overlap is required as the wave gauges do not support being started at an exact, user-provided time. The final time difference between the two wave gauges was calculated from a correlation of both reported wave heights during the common overlap period, considering a constant elevation difference between the time series. Software was developed to automatically detect the offset in time and height and merge the sampled series to one continuous series.

Ship monitoring

In order to conduct a ship/source-specific wave analysis the River Information Services (RIS) of “via donau” (Simon & Troegl, 2011), the Austrian waterways authority, was used. The regular function of the so-called DoRIS system is to record and display vessels on an electronic navigational chart (ENC) containing the most important nautical information concerning the waterway and traffic regulations. While RIS makes cargo transport and passenger services on waterways more efficient and increases traffic safety (Simon & Troegl, 2011), automatic identification system (AIS) transponders are part of the mandatory equipment for vessels passing the Austrian Danube and are used to identify the current position of every vessel using GPS technology. Every ship continuously transmits its ID, position, course over ground, speed over ground, and other data to a central register on a common VHF radio channel, forming the basis for plotting a digital navigational chart useable for all shipping companies and external users (Simon & Troegl, 2011). The data provided by this system can be used to plot the ship traces as depicted in Fig. 2.

The DoRIS data were used to determine ship type, speed, distance from shore, direction (upstream or downstream), time before, and time after last passage for all events. Software was developed to match the available data with the events detected with the wave gauges. From this information, a database was created which is capable of comparing all information of the passing vessels to other basic parameters, such as the Danube discharge or the bank type morphology, with the observed wave data.

All events were classified by the ship types involved, whereas only waves of the three prominent ship types (pictures in Figs. 4, 5, 6) were used for the further analyzes: (i) ship type 1—large passenger ship; (ii) ship type 2—high speed passenger ship; (iii) ship type 3—bulk carrier.

Fig. 4
figure 4

Wave characteristics of ship type 1; the maximum event as well as the 75% quartile, the median and the 25% quartile event are depicted; box plot showing the wave heights of the first primary waves

Full size image
Fig. 5
figure 5

Wave characteristics of ship type 2; the maximum event as well as the 75% quartile, the median and the 25% quartile event are depicted

Full size image
Fig. 6
figure 6

Wave characteristics of ship type 3; the maximum event as well as the 75% quartile, the median and the 25% quartile event are depicted

Full size image

Large passenger ships are summarized as ship type 1 (exemplarily depicted in Fig. 4) with lengths between 60 and 110 m, widths between 9 and 13 m, typical draft of 1.5 to 2.2 m and ship speeds ranging from 4.8 to 13.1 km h−1 (mean value 9.3 km h−1) going upstream and 17.0 to 36.6 km h−1 (mean value 25.2 km h−1) going downstream. All types of high speed passenger ships are merged as ship type 2 (exemplarily depicted in Fig. 5), where lengths range typically between 30 and 40 m, widths are found between 8 and 10 m, draft is about 0.9 m and ship speeds are ranging from 12.2 to 58.1 km h−1 (mean value 47.1 km h−1) going upstream and 35.3 to 73.9 km h−1 (mean value 63.2 km h−1) going downstream. Bulk carriers of different types are summarized as ship type 3 (exemplarily depicted in Fig. 6) with typical lengths of 95 to 170 m, typical widths of 9 to 23 m, drafts up to 2.7 m and ship speeds ranging from 2.0 to 10.2 km h−1 (mean value 5.0 km h−1) going upstream and 16.3 to 28.9 km h−1 (mean value 22.7 km h−1) going downstream. Wave data was collected at different bank types to allow a comprehensive analysis of the various waves. Table 1 gives an overview on the sampling setup, showing the discharge range, the sampling period and the number of vessel passages for each bank type.

Table 1 Sampling setup: Sampling period, number of passages and discharge range at each bank type
Full size table

Around 2,000 ship passages were sampled to establish statistically relevant relationships between the parameters investigated. As waves are primarily characterized by their length, period, and height, these parameters were evaluated within the study. To avoid superposition between events, a substantial number of events were omitted because of a further passage within 10 min before or after the event, leading to a total of 909 unaffected ship passages.

Habitat assessment tool

An analysis was performed to determine drawdown effects for all potential YOY nursery areas within the study reach. Therefore, over the entire study reach the terrain slope was calculated along cross-sections aligned perpendicular to the rivers centreline, with a spacing of 5 m, leading to a complete map of terrain gradients. Based on this map, the potential shift of the water’s edge for a discharge of 2.230 m3 s−1 (estival mean discharge, typical for the nursery season) was calculated for a mean drawdown of 0.20 m (the calculated median of the 10% highest drawdown events for all bank types). This analysis forms the basis for the map presented in Schludermann et al. (2013) and can be used for habitat assessment studies before or after river restoration projects on navigable rivers.

Results

Parameters influencing the characteristics of the ship waves are vessel type (size and hull shape) and speed, water depth, distance to shore, load of the vessel, and morphology (Houser, 2010). Wake generation includes bow and stern waves, drawdown of water levels, drawdown currents, and propeller wash (Maynord, 1996).

Influence of vessel type on wave characteristics and effects

For each ship type, the recorded wave events were analyzed. Figure 4 shows the scope of the records of waves generated by ship type 1 (passenger ships). The median event as well as the lower and the upper quartile events are plotted for the three main ship types. Additionally the maximum event is visualized, being of high importance in terms of disturbance of the riverine biota (Schludermann et al., 2013).

Typical passenger ships (Type 1) at the Danube River generated waves with a pronounced first primary wave (Bernoulli wake) with a large drawdown (trough) followed by a large crest and a series of oscillatory waves. The subsequent primary waves were rapidly decreasing in wave height; often a secondary wave system was superimposed due to wave reflections at morphological structures. The maximum event detected for passenger ships was associated with a wave height of 0.65 m, this being the maximum event of all ship types. The 75% quartile of wave height could be observed at 0.24 m, the median at 0.11 m and the 25% quartile at 0.06 m. Wave length of the primary wave system ranged between 70 m for the large waves and up to 580 m for small events; wave periods ranged between 20 s and 2 min.

The same analysis was conducted for type 2 ships, which are high-speed passenger ships of different construction types. Figure 5 depicts the characteristic events for this ship type.

Vessels of type 2 exhibited entirely different wave characteristics compared to type 1 vessels. The oscillatory waves observed were associated with a much higher frequency and a wave period ranging between 3 and 6 s. The primary wave system showed a wavetrain with almost uniform and subsequently slowly decreasing wave heights. The maximum wave height was explicitly smaller (0.32 m) compared to type 1 vessels, but the 75% quartile (0.19 m), the median event (0.14 m) and the 25% quartile event (0.07 m) were ranging at similar wave heights or had slightly higher waves. Also the wavelengths were significantly smaller than the ones of type 1 vessels and were ranging between 10 and 30 m.

Moreover, the wave characteristics of bulk carriers (ship type 3) were determined. Waves of type 3 vessels were found to be associated with a similar pattern as the displacement waves of type 1 vessels. They were characterized by a pronounced drawdown (trough) of the first primary wave (Bernoulli wake), followed by a crest, which appeared less pronounced in comparison to the other vessel types (Fig. 6).

Accordingly, the maximum event (height 0.55 m) as well as the 75% quartile event (0.1 m), the median event (0.06 m) and the 25% quartile event (0.03 m) generated explicitly smaller wave height than type 1 vessels. However, wavelengths (90–500 m) and periods (30–130 s) were ranging at similar values compared to type 1 waves.

Influence of absolute vessel speed on wave characteristics

As vessel speed is one of the fundamental parameters influencing wave height (Maynord, 2005), an analysis comparing wave heights to vessel velocities was performed for only one ship type to omit other influencing parameters. As especially ships of type 2 but also those of type 3 operate at a narrower velocity range, all events of type 1 (passenger ships) going downstream were used for this analysis. At all bank types, wave heights were found to be rising with absolute ship velocity (Fig. 7).

Fig. 7
figure 7

Correlation of wave height and absolute ship velocity for type 1 ships navigating downstream at all bank types. Different symbols were used for each bank type, a regression line was calculated for all data (R 2 = 0.37)

Full size image

The interrelation between ship speed and wave height was tested by a general linear model (GLM). The prerequisites of normal distribution (K–S test P = 0.994) and variance homogeneity (Levene Test P = 0.15) were fulfilled. The analysis of variance showed a variance value of P = 0.037 and 41% of the data variability can be explained by the parameter vessel speed. Hence vessel speed significantly influences wave height.

Influence of shore distance on wave characteristics

The ships’ distance to shore was determined as another parameter influencing wave height. All type 2 vessels navigating downstream were used for analysis, since they operate within the smallest velocity spectrum. Independent of the shore type, a decreasing wave height with rising distance to the bank was found (Fig. 8).

Fig. 8
figure 8

Relationship between distance of the vessel to investigated shore and the corresponding wave height for type 2 vessels navigating downstream; Regression line and 95% confidence range are depicted (R 2 = 0.37)

Full size image

Influence of bank type on wave characteristics

The varying morphologies and discharge conditions have a huge effect on wave characteristics at the different bank types of the Danube. Especially the bank water depth seems to be a fundamental parameter, influencing both wave damping and wave breaking. As even constructed groyne fields have diverse morphologies (e.g., the bank type 1 groyne field is strongly aggraded and as such totally different from the bank type 2 groyne field), the influence of varying discharges is also distinct at different shore types.

In Fig. 9, the correlation of discharge and wave height is depicted for two different groyne fields (bank type 1 and 2). This analysis was performed on data from all passages of a certain ship (high speed ferry boat between Vienna and Bratislava; ship type 2) navigating downstream, in order to exclude other potentially influencing parameters.

Fig. 9
figure 9

Relationship between discharge (x-axis) and wave height (y-axis) for type 2 vessels navigating downstream for bank type 1 (groyne field at river km 1,886.8) and bank type 2 (groyne field at river km 1,886.0)

Full size image

Whereas in the single-vortex groyne field (bank type 2), a decreasing wave height with rising discharges was observed, the two-vortex groyne field (bank type 1) shows a contrary trend. The discharge range during observations at bank type 3 was too small for obtaining an adequate dataset to perform this analysis.

Influence of depth Froude number on wave characteristics

The depth Froude number (Fr h ) shows the relation between the ships’ speed (v ship) and the celerity of a progressive wave in still shallow waters, given in Eq. 1

$${\text{Fr}}_{h} = \frac{{v_{\text{ship}} }}{{\sqrt {gh_{\text{s}} } }}$$
(1)

where g is the acceleration due to gravity and h s is the still water depth in the sailing line.

When water is not still, the significant speed of the ship is its relative velocity with respect to the moving stream. For the following analysis the relative depth Froude numbers Fr hr

$${\text{Fr}}_{hr} = \frac{{v_{\text{r - ship}} }}{{\sqrt {gh_{s} } }}$$
(2)

were calculated, considering the flow velocity u of the Danube River at the particularly corresponding discharge. Note that displacement waves as defined in previous sections can only be produced by subcritical vessels (i.e., ships sailing at a relative speed lower than the celerity of a progressive wave). The relative ship speed v r-ship is given by

$$v_{\text{r - ship}} = v_{\text{ship}} \, \pm \,u$$
(3)

where the flow velocity of the river is positive for a ship navigating upstream and negative for one navigating downstream. All vessels operating at speeds resulting in Fr hr  < 1 are defined as subcritical and Fr hr  > 1 as supercritical. The range between 0.84 < Fr hr  < 1.15 is defined as transcritical range, where no clear distinction between sub- and supercritical range can be found (Hüsig et al., 2000).

The relative depth Froude numbers were plotted for all passages distinguishing between the three ship types going up- or downstream (Fig. 10).

Fig. 10
figure 10

Dimensionless depth Froude number for all passing vessels and their primary wave height; Different ship types (Type 1 green, Type 2 gray and Type 3 red) and direction are distinguished

Full size image

Two clearly delimited groups can be shown with all type 2 vessels being in the supercritical group and all type 1 and 3 vessels in subcritical range. Within the supercritical group, no difference can be seen between down- and upstream bound vessels whereas within the subcritical group there is a clear (more than linear) trend between maximum wave height and relative Froude number. Furthermore, the ships going downstream are responsible for the higher primary waves.

Shift in water’s edge

As described in further detail in the companion paper (Schludermann et al., 2013), drawdown is one of the most relevant parameters (larval drift densities of single families and developmental stages showed a significant increase at vessel-induced wave events with heights up to 0.1 m regardless of the mesohabitat) influencing the survival, especially of the early stages of YOY fish fauna. Therefore, special attention has been dedicated to this aspect. As shown in Figs. 4 and 6, the highest drawdown was observed for vessels of types 1 and 3. The highest waves were generated by passenger ships; but when the proportion of the drawdown to the wave height differed (73% of the wave height for the bulk carriers vs. 58% for the passenger ships), the absolute values for the drawdown were generally within the same range for ships of both types. The largest drawdown value for all passages was found to be 0.45 m; the median value for the 10% highest drawdown events was found at 0.20 m (this value was used as an assessment reference for the drawdown analysis in Schludermann et al., 2013). As the consequences of drawdown are directly linked to shore morphology, a morphological analysis was accomplished, based on the slope ratio determined for the study site. Depending on the slope, a drawdown of 0.4 m, which is equal to the reference water depth that delimits suitable instream nursery habitats (according to Winkler et al., 1997), corresponds to a horizontal shift of the water’s edge varying with river discharge, as depicted in Fig. 11.

Fig. 11
figure 11

Shift in water’s edge at the three bank types for varying discharges and a drawdown of 0.4 m

Full size image

Mainly depending on the surface gradient and discharge, the shift in water’s edge ranged between 0.5 and 35 m for the observed bank types. The plane gradient (due to aggradation processes) within the groyne field at bank type 1, and the gentle slopes of the gravel bar (bank type 3) led to a large shift of the water’s edge during ship-induced drawdown. For the highest drawdown detected (0.45 m), the water’s edge would shift over more than 30 m during discharges around 1.100 m3 s−1, thereby resulting in a significant habitat loss (as described in Schludermann et al., 2013).

Discussion

Maynord (2005) found wave height to be the most significant indicator regarding the characteristics of ship waves because it is influenced by most of the relevant parameters and, at the same time, strongly affects the riparian habitat. Hence for the characterization of the typical wave events, the wave heights of all passing ships of each type were analyzed. Clear and diverse wave characteristics could have been distinguished. Type 1 (passenger ships) and Type 3 (bulk carrier) vessels were found to be associated with a similar pattern characterized by a displacement wave with a pronounced drawdown (trough) of the first primary wave (Bernoulli wake), followed by a wave crest, which appeared more pronounced for ship type 1 and less pronounced for vessel type 3. Whereas ship type 2 typically produces a wave system with oscillatory waves with a much higher frequency having almost uniform and subsequently slowly decreasing wave heights.

While the wave height of small boats (length of 8.5 m or less) mainly depends on whether the boat is in semi-planing or planing mode, with a decreasing wave height at higher speeds (Maynord, 2005), wave heights of huge vessels at the Danube were found to be rising at higher speeds. Absolute ship speed was correlated against wave height showing that that vessel speed significantly influences wave height and that, wave heights were found to be rising with absolute ship velocity at all bank types. The regression line for all data is plotted in Fig. 7, but was also calculated separately for each single bank type showing that the regression gradients of the relationship between wave height and ship velocity are similar for all bank types. This indicates that the relation has a more common character and can be applied to different bank situations at the Danube.

When testing the distance to shore as a parameter influencing wave height, independently of the shore type, a decreasing wave height with rising distance to the bank was found (Fig. 8). This is in accordance with the findings of previous studies (e.g. Nanson et al., 1994; Brunke et al., 2002).

Morphology and discharge conditions are found to influence the wave characteristics substantially. Waves in shallow waters are strongly influenced by water depth (depending on shore type and discharge conditions): when reaching shallower areas their velocity decreases, leading to an increasing wave height to maintain a constant energy flux. The wavelength also decreases while the frequency remains constant. When the wave amplitude reaches a critical level at even shallower areas, the wave is breaking at a certain ratio of wave height to water depth (H/d) depending on the steepness (H/L) of the waves. Breaking waves transform large amounts of wave energy into turbulent kinetic energy, leading to decreased wave heights. Hence the structured inhomogeneous morphology at the Danube was found to lead to varying conditions regarding wave effects, which furthermore depend on the average bank water depth and thus on discharge conditions.

Even constructed groyne fields at the Danube have diverse morphologies leading to completely different trends when comparing wave height to Danube discharge (Fig. 9). The differences in the trends for bank types 1 and 2 can be explained by the differing morphology of the groyne fields. The groyne field of bank type 1 is characterized by aggradation processes and therefore a steep transition zone between fairway and groyne field is present, followed by an almost horizontal platform (see inset in Fig. 9). At low discharges, the waves break far away from the banks leading to small waves near the shore, whereas at high discharges, the waves are not breaking but increasing their height and steepness over a long distance. At bank type 2, a constant gentle gradient is found between fairway and shore (see inset in Fig. 9). Consequently, when discharge decreases, the waves show a tendency to get steeper and higher near the river banks without breaking, which leads to the inverse trend compared to bank type 1 (Fig. 9). For bank type 3 the gradients can be described similar to bank type 2 but with even gentler slopes (see inset in Fig. 2). Due to that fact, a similar trend but no significant correlation (the discharge range during observations at bank type 3 was too small for obtaining an adequate dataset) between wave height and discharge was observed at bank type 3.

As the Danube is a moving stream, the relative depth Froude numbers Fr hr (relation between the relative ships’ speed (v r-ship) and the celerity of a progressive wave) was compared to the wave height. As expected, only the subcritical vessels show a clear (more than linear) relationship between maximum wave height and relative Froude number (Fig. 10), as found in previous work (Schoellhamer, 1996; Rapaglia et al., 2011). The residual dispersion of experimental data can partially be explained by the other significant parameter (ratio between ship and channel cross-section), not explicitly considered here but somehow represented by the type of vessel: larger vessels of type 3 tend to produce higher waves than vessels of type 2. However, it can be seen that for displacement waves, the correlation of maximum height and relative Froude number is much stricter than with absolute vessel speed (compare Figs. 7, 10).

As drawdown can be seen as the parameter having the greatest impact on the early stages of YOY fish fauna (Schludermann et al., 2013), a morphological analysis was accomplished, directly linking shore morphology to ship induced drawdown. Mainly depending on the surface gradient and discharge, the shift in water’s edge can be remarkable for shore types with gentle slopes.

This is of substantial importance, as the shallow shore-near areas are normally associated with optimal instream nursery habitats, regarding feeding and growth (e.g., Wolter & Arlinghaus, 2003; Kucera-Hirzinger et al., 2009), and are temporarily lost due to passing vessels. This information was used for the study reach wide drawdown analysis, described above, leading to a habitat assessment map presented in Schludermann et al. (2013). The map shows that over 99% of shoreline areas within the 3-km river stretch are probably highly affected by vessel-induced drawdown. All potential inshore nursery habitats were situated in zones with a maximum retrogression of up to 4 m at this discharge (Schludermann et al., 2013).

Management implications

Some of the relations and results achieved in this study have the potential to be useful for planning or enhancing river restoration projects at navigable rivers.

For the relationship between wave height and ship velocity (Fig. 7), similar regression gradients were found for the trend lines of all bank types. This allows for an estimation of a mean wave height reduction as a consequence of reduced ship speeds. When calculating an overall trend, a speed reduction of 5 km h−1 would result in a wave height decrease of about 14 cm for passenger ships (ship type 1) leading to a reduction of water level retrogression of up to 24.5 m depending on discharge and morphology. As observed above, the relationship of wave height with relative Froude number is even stricter than with absolute vessel speed. It may then be worthwhile, to consider the relative vessel speed in possible new regulations.

Equally, the relationship between the distance to shore and wave height can be integrated into the planning process of restoration projects. According to the linear regression performed on the data (Fig. 8), a relocation of the shipping fairway of 50 m would result in a mean decrease of 8 cm in wave height for the analyzed type of vessels. Especially in combination with a reduction of vessel speed, this may lead to a substantial ecological enhancement if optimal instream nursery habitats are affected.

Also morphological characteristics must be considered within state-of-the-art restoration projects in navigable rivers. As described differing morphology is leading to a rising wave height with rising discharges at bank type 1 and a decreasing wave height with rising discharges at bank type 2 (Fig. 9). For discharges slightly above mean discharge, which correspond to typical flow conditions during spawning periods in spring, the trend lines (Fig. 9) for the different bank types show values being relatively close to each other. However, for all other discharges the differences are much larger, emphasizing the influence of morphology on wave heights. The analysis shows the importance of morphology in relation to the effects of vessel-induced waves, in analogy to the discharge-dependent changes in wave height (Fig. 9). It highlights the importance of careful planning within restoration projects and the possibilities arising when involving such considerations in adapting parts of a river system to riverine biota such as providing special morphologic features to create wave protected areas or reconnecting side arm systems for providing areas without wave impact.

Conclusions

Extensive measurements were performed to describe the wave characteristics of the main ship types at the Danube River in Austria and the interaction of the vessel-induced waves with three different bank types. Passenger ships were found to generate displacement waves characterized by a pronounced first primary wave with a large drawdown (trough) followed by a large crest and a series of smaller oscillatory waves; they also produced the biggest wave height (up to 0.65 m) among all types. High-speed passenger ships induced waves of high frequency with a wave period ranging between 3 and 6 s, but smaller maximum wave heights. Bulk carriers were found to have similar characteristics as passenger vessels, with a pronounced drawdown (trough) of the first primary wave but followed by a comparatively damped crest (oscillatory system). The largest wave events had comparable heights but the wave heights of the 75% quartile events were substantially smaller than waves originating from ordinary passenger and high-speed passenger ships.

Moreover, relative Froude number was found to be a main parameter controlling wave height. A ship velocity reduction of 5 km h−1 for passenger ships leads to a decrease of about 14 cm in primary wave height. A relocation of the shipping fairway of 50 m would result in a mean decrease of 8 cm in wave height of high-speed passenger vessels. The influence of morphology of the different bank types on wave height was shown to be substantially high. Especially, the bank water depth (resulting from morphology and discharge) influences wave breaking and wave damping and thus controls the primary wave height near the shores.

As drawdown is one of the most important parameters influencing the survival of YOY fish fauna, a complete map of gradients was calculated for the whole study area based on detailed morphological analyses and based on slope calculations along cross-sections aligned perpendicular to the rivers centreline (presented in Schludermann et al., 2013).

The presented results highlight the importance of considerations regarding ship-induced waves in river restoration projects. Sophisticated planning of retreat areas such as well connected side arm systems or runners with offshore morphological structures causing waves to break and thereby reduce wave energy could lower negative effects on sensitive nursery areas and potentially decrease mortality rate of YOY fishes. Also the consideration of wave impacts in shipping management was shown to be of high importance regarding a decrease of wave height. The results presented serve as a basis for applied biotic studies and basic research.