- Email: [email protected]
The role of morphology on the displacement of particles in a step–pool river system
Available online at www.sciencedirect.com
Geomorphology 99 (2008) 270 – 279 www.elsevier.com/locate/geomorph
The role of morphology on the displacement of particles in a step–pool river system Hélène Lamarre ⁎, André G. Roy Département de Géographie, Université de Montréal, C.P. 6128, Succ. Centre-Ville, Montréal, Québec, Canada Received 15 February 2007; received in revised form 14 November 2007; accepted 16 November 2007 Available online 22 November 2007
Abstract Although studies of sediment transport in steep and coarse-grained channels have been more numerous in recent years, the dynamics of sediment transport in step–pool river systems remain poorly understood. This paper investigates displacements of individual clasts through Spruce Creek (Québec, Canada), a classic step–pool channel, and the effects of the channel morphology on the path length of the clasts. Passive integrated transponder tags (PIT) were used to track the displacement of 196 individual particles over a range of discharges including the bankfull stage. Clasts were tracked after five sequences of flood events. The results showed that the distance distributions match a two-parameter Gamma model. Equal mobility transport occurs for the particle size investigated during each sequence of flood events. Mean travel distance of the clasts can be estimated from excess stream power, and the mobility of the clasts is more than an order of magnitude less than the model reported in riffle–pool channels. The dominant morphological length scale of the bed also controls the path length of the clasts. These results confirm some preliminary observations on sediment transport in step–pool channels. © 2007 Elsevier B.V. All rights reserved. Keywords: Gravel-bed rivers; Step–pool; Passive integrated transponders; Radio frequency identification; Sediment transport; Particle mobility; Distance of movement; Bed morphology; Steep channels
1. Introduction Step–pool channels occur in a wide range of bioclimatic environments (Bowman, 1977), from arid (Wohl and Grodek, 1994) to humid forested areas (Heede, 1972). They are commonly found in steep headwater mountain streams, where the channel width to depth ratio is small, bed material is heterogeneous, and slopes exceed 4 to 7% (Chin, 1989; Grant et al., 1990; Grant and Mizuyama, 1991). The longitudinal profile of step–pool channel is characterized by steps viewed as congested zones where clasts have clustered (Church and Jones, 1982; Church et al., 1998). The steps are generally composed of an accumulation of cobbles and boulders that are transverse or oblique to the channel (Zimmermann and Church, 2001; Chin and Wohl, 2005) and their occurrence depends on the local availability of keystones that remain immobile on the ⁎ Corresponding author. Tel.: +1 514 343 8035; fax: +1 514 343 8008. E-mail address: [email protected] (H. Lamarre). 0169-555X/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2007.11.005
contemporary flow regime (Church and Zimmermann, 2007). Steps define breaks of slope and alternate with pools, containing finer bed material. The geometry of step–pools is defined by step spacing (L) and step height (H) and their ratio has been found to fall in the range 0.06 ≤ H/L ≤ 0.20, which corresponds to the range of gradients where step–pools are found (Chin and Wohl, 2005; Church and Zimmermann, 2007). Step spacing has also been correlated with overall stream gradient (S) in the form H/LαSβ, in which 0.42 ≤ β ≤ 0.68 (Abrahams et al., 1995). Chartrand and Whiting (2000) have shown that the step spacing was 0.6 time the channel width (W), while Bowman (1977) reported a spacing of 1.4W. Chin (1989) obtained a spacing of 2.7W for step–pools in Santa Monica Mountains (Colorado). Data from numerous locations have shown that step–pool spacing is usually less than one to four channel widths. The step–pool morphology serves a fundamental role in river systems controlling hydraulic resistance (Wohl and Grodek, 1994; Abrahams et al., 1995) and energy dissipation (Hayward,
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
1980; Whittaker and Jaeggi, 1982). They are characterized by relative submergence of the large clasts (Y /d84) that is b1.2 (Bathurst, 1978), where Y is the flow depth and d84 the 84th percentile bed material size. At high flow stage, water flows over and in between the large roughness elements that form the steps and plunge into the pools below, promoting a tumbling highly turbulent flow (Peterson and Mohanty, 1960; Whittaker, 1987). The available energy is dissipated by eddies (Wohl and Thompson, 2000). Therefore, the potential energy available for erosion and sediment transport at high stage is reduced by the steps (Chin, 2003). In their recently published review of empirical research on step–pool channels, Chin and Wohl (2005) pointed out some of the common observations on bedload transport in these systems: transported sediment is derived from localized sites on adjacent hillslopes and in the channel; bedload entrainment and transport are spatially and temporally discontinuous (Ergenzinger and Schmidt, 1990); sediment are preferentially stored in
271
and mobilized from pools (Schmidt and Ergenzinger, 1992; Marion and Weirich, 1999); bedload transport during lower flows tends to be characterized by equal-mobility transport (Blizard and Wohl, 1998; Marion and Weirich, 2003) where larger grains are as easily entrained as smaller ones because they are more exposed to lift and drag forces (Parker et al., 1982). In spite of these findings, there is still a paucity of field data on the displacement of clasts in step–pool channels. Such data are a good indicator of bedload transport response of the stream to a given water discharge and to sediment supply conditions, and are critical for understanding of the development of channel morphology (see Haschenburger and Church, 1998; Sear et al., 2000). Morphological features of step–pool channels should play a significant role on the distance of displacement of the bed material. At this time, it is still not possible to predict sediment transport rates in step–pool systems as a function of flow and inchannel sediment storage (Church and Zimmermann, 2007). It
Fig. 1. Location (A) and upstream view (B) of Spruce Ck step–pool channel.
272
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
Table 1 Number of particles in each clast size b axis (mm)
Number of tagged clasts
0–50 50–100 100–150 150–200 200–250 Total
4 124 60 6 2 196
160 mm. Removing the boulders (N256 mm = 30% of the measured particles), which can be transported only during exceptional flood, from the distribution, the d50 of the surface bed material is reduced to 90 mm. 2.2. Sediment tracking
2.1. Field site
We implanted passive integrated transponder (PIT) tags into clasts of various sizes and shapes to quantify clast movements. A PIT is a glass encapsulated pulse transponder, using a radio frequency system, with a unique identification code encrypted. A detailed description of the tags and of its application to sediment transport investigation is given in Nichols (2004), Lamarre et al. (2005) and Allan et al. (2006). PITs were installed in the laboratory within holes drilled into indigene clasts sampled on the stream bed (n = 196). The number of clasts tagged within each size fraction is given in Table 1. The size distribution of these particles is close to that of the particles composing the bed surface (Fig. 2), although two limitations prevented a perfect match. Firstly, particles with a b axis (intermediate axis) less than 40 mm were too small to host a PIT tag. Secondly, particles larger than 256 mm were difficult to bring back to the laboratory because of their excessive weight. In Spruce Ck, the d50 of the tagged clasts (86 mm) is half the d50 of the surface bed material (160 mm), but it matches the d50 of the distribution if the clasts N256 mm are not included. Our strategy aimed at focusing on the clasts that could be mobilized during low to moderate floods that were likely to occur during the survey period.
The study was conducted in Spruce Creek (alt. 520 m), a classic step–pool channel located on Mount Sutton in the Eastern Townships (Québec) (Fig. 1). The 50-m study reach is located in the middle part of the basin and drains a forested area of about 3.2 km2. The length of the study site was defined by the time needed to characterize both the channel morphology and the sediment transport characteristics. The channel bed is composed of clasts ranging from pebbles to boulders, essentially coming from bedrock fragments and a till layer. A number of clasts in the channel are in the order of 1 m in diameter. These boulders assure the stability of the step morphology, probably remaining immobile under contemporary flood regime. Mobile clasts during the survey are derived from the channel bed, the lower portion of the channel bank, and local slides on hillslopes along the reach. Average morphological characteristics of the reach were extracted from detailed topographic maps of the bed. The maps were produced using a Trimble robotic total station (model DR-5600). The total station collects x, y and z coordinates electronically and stores the observation directly in a portable computer (Trimble, 2005). More than seven topographic points per square metre were measured systematically along transects and at breaks of slope in order to extract the morphological characteristics of the channel. At the bankfull stage, the average channel width is 6.0 m, flow depth is 0.45 m, and slope is 14%. The surface bed material was characterized over the reach with a Wolman count performed on 1200 randomly selected particles. The median size (d50) of the bed surface material is
Fig. 2. Bed material (black) and tagged clasts (white) size distribution: (A) including clasts N256 mm, (B) excluding clasts N256 mm.
is to be expected that the movement of individual particles may be controlled by the large morphological features such as the steps and the pools. This control could be similar to that of pools and bars in lowland streams as suggested by Pyrce and Ashmore (2003) who have shown that the mode of the path length distribution coincides with known pool-bar spacing in gravelbed streams. The objectives of this paper are to (i) describe the movement of the clasts to compare with other preliminary observations on bedload transport in step–pool channels and (ii) quantify the effect of the step and pool spacing on the mode displacement distance of the clasts. This study is based on a unique set of particle displacement data measured on a four year period in a step–pool channel. 2. Methods
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
273
Table 2 Characteristics of the survey periods Tagged particles recovery dates
Recovery rates (%)a
Number of floodb
Qmax (m3 s− 1)c
Qmax / Qdbk
Q1 (m3 s− 1)e
Q1 / Qbk
Duration of flow N ω0 (hours)f
(m)
Mobile particles (%)
1 December 2003 25 May 2004g 27 September 2004 30 May 2006 3 October 2006 30 May 2007
Initial seeding 92 83 62 62 57
1 1 5 1 2
2.04 1.20 3.40 2.11 1.36
1.0 0.6 1.6 1.0 0.7
2.04 1.20 1.07 2.11 1.36
1.0 0.6 0.5 1.0 0.7
120 96 336 96 144
3.84 1.53 5.01 8.31 8.29
62 32 70 86 78
a
Recovery rates include all the clasts that were not found within the sampling section. The number of flood events is estimated selecting the peaks of discharge N1 m3 s− 1 (=0.5Qbk) in which case, at least 25% of the tagged clasts were mobile. c Qmax is the largest peak flow. d Qbk is the bankfull flow. e Q1 is the first peak flow. f Number of hours where specific stream power is Nω0. g Results of the first event disregarded in the analysis. b
The clasts were seeded at low flow on the channel bed on 1 December 2003. Because of the heterogeneity and roughness of the bed, we could not place the particles systematically along cross-sections. The tagged clasts were seeded in a way to mimic the position of similar clasts. We attempted to respect inasmuch as possible the natural organization of the bed and the degree of imbrication of the particles. It was difficult, however, to replace the tagged clasts to maintain the exact structure of the bed surface material. Therefore, most tagged particles appeared less constrained than the clasts originally composing the water worked surface. We supposed that after the first transport event, the clasts were reorganized in a more natural way, while they occupy a more stable position. The results of the first displacement of the clasts were then disregarded from further analysis. Note that at the upstream end of the reach, most of the tracers were seeded on one side of the channel. The other side was characterized by bed material much finer than the size distribution of the tagged particles. The location of each particle within the channel was measured with the robotic total station. The tracking of particles was carried out at low flow after five sediment transport episodes: on 25 May 2004 (removed
from the analysis), 27 September 2004, 30 May 2006, 3 October 2006, and 30 May 2007. Buried tagged particles could be detected up to a depth of 0.25 m within the substrate (Lamarre et al., 2005). The characteristics of the flood events were estimated from discharges measured by the Centre d'expertise hydrique du Québec in the neighboring Yamaska-est watershed (Table 2). Hydrograph of the average daily discharge of Spruce Ck was reconstructed using a correlation approach between the Yamaska sud-est discharge series and punctual discharge measurements made in Spruce Ck. High discharges happened before the spring surveys and discharges remained below bankfull stage (Qbk) until the autumn surveys (Fig. 3). The number of flood events between the surveys was estimated from the discharge peaks above 1 m3 s− 1 (0.5Qbk) and when at least 25% of the tagged clasts were mobile. Several constraints such as the distance of the site from the university or the winter conditions made it impossible to survey the clast positions after each flood event. The recovery rates of tagged particles were 83% on 27 September 2004, and 62% for the two subsequent surveys (Table 2). The recovery rate falls to 57% on 30 May 2007. This
Fig. 3. Hydrograph of the average daily discharges of Spruce Ck between August 2003 and June 2007. Discharges are estimated from data measured by the Centre d'expertise hydrique du Québec in the neighboring Yamaska sud-est watershed. Dates of the recovery of the tagged particles are presented in relation to flood events. Thresholds of bankfull flow (Qbk) and 0.5Qbk are superimposed to the series.
274
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
during the recovery process (Lamarre et al., 2005). The percentages of mobile tagged particles increase with discharge. The path length traveled by a tagged particle was estimated as the straight line distance between the initial location and the point of final deposition. Given that the channel has a small width and is relatively straight, we assumed that the difference between the straight line and the real distances traveled by the clasts is negligible. We also identified the number of tagged particles deposited in each morphological unit at the time of the surveys. The results obtained on the transfer of tagged clasts between each of the morphological units are influenced by the particles moving downstream during the experiment and the fact that particles are not being replaced at the upstream end of the reach after each transport event. 2.3. Step–pool configuration The locations of the particles were superimposed on the detailed topographic map of the bed. Steps were distinguished from pools using field observations in conjunction with the features seen on the topographic map. The location of the morphological units was estimated using the zero-crossing technique along longitudinal profiles (Richards, 1976) and corroborated with a visual identification. The average distance (Ls) between the step and the pool was estimated from the top of the crest to the lowest point of the pool located immediately downstream. 3. Results 3.1. Sediment transport Fig. 4 shows the distribution of the distances of clast movements (Di) scaled by the mean distance of the event () (see Table 2). The distributions of distance do not include the clasts that remain immobile. The distributions are uni-modal and tend to be regular. The mode of the scaled individual distances is
Fig. 4. Displacement distance distribution of the clasts. Clasts that did not move are excluded: (A) 25 May 2004 to 27 September 2004, (B) 27 September 2004 to 30 May 2006, (C) 39 May 2006 to 3 October 2006, and (D) 3 October 2006 to 30 May 2007. Individual distance Di is scaled by, the mean distance of the event. Classes of the 0.25 are used to produce the histograms. Each class is specific to an event.
may be due to the burial of the clasts and to the duration of the experiment. Some of the tagged clasts have exited the surveyed section, thus affecting the recovery rates. Using the tagged particles, we characterized the mobility, the distance of displacement and the morphological unit where the clasts have deposited. The mobility represents the percentage of tagged particles that were transported between each survey. A tagged particle found outside of a 0.3 m radius of its previous location was considered as being mobilized. This threshold is associated to the sampling error of the location of the PITs
Fig. 5. Relation between the displacement distances of the clasts scaled by the total time for which the flow was larger than that corresponding to ω0 and excess specific stream power. Specific stream power is estimated for the first peak flow during flood events. Clasts that did not move are excluded.
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
275
Fig. 6. Individual displacement distances in relation to grain size. Clasts that did not move are excluded: (A) 25 May 04 to 27 September 2004, (B) 27 September 2004 to 30 May 2006, (C) 30 May 2006 to 3 October 2006, (D) 3 October 2006 to 30 May 2007. Grain size b axis is scaled with the d50 of the bed surface material that does not include clasts N256 mm.
Fig. 7. Subdivision of reaches into distinct morphological units from (A) longitudinal topographic profile and (B) topographic map of the bed. Slope has been removed to emphasize the changes in bed elevation.
276
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
0.25 in each distribution. The observed distances of movement are fitted with a two-parameter Gamma distribution (Hassan et al., 1991). Chi-square tests (α = 0.05) indicate that Gamma models fit the data well, except for the first event (p = 0.002). In particular, the model underestimates the frequency for short distances. The correlation between the path length of the clasts and flow conditions is presented in Fig. 5 on an event basis. Specific stream power (ω) was used to quantify the flow conditions of the reach. In this study, specific stream power is based on the bed slope of the reach and on the first peak discharge observed in an episode of sediment transport (Table 2). This value was selected to reduce the effect of duration of the period between each survey. The excess stream power (ω − ω0) was estimated using Bagnold's (1980) method where: 12Y 1:5 log x0 c290d50 d50 (see Hassan et al., 1992). In Bagnold's original formula, d50 represents the median size of the subsurface material. Here, this value was replaced by the d50 of the bed surface, because the size of subsurface bed material is not known. It may have reduced the estimate of excess stream power values, expecting that the size of subsurface is smaller than at the bed surface. Also, step–pool channels are characterized by large-scale structural arrangements that may influence the entrainment and transport. The approach does not consider those features. The average distance traveled by the clasts is divided by the total time for which the discharge was higher than that corresponding to ω0 (< Dv>). The empirical relation between < Dv> and the excess stream power is:¼ 0:0002ðx x0 Þ1:13 R2 ¼ 0:82; pN0:01 : The linear regression models (method of least squares) between the size of the clasts scaled with the d50 of the surface bed material and the Di / < Dv> were used to test the equalmobility transport (EMT) hypothesis in the step–pool channel (Fig. 6). In the linear regression between scaled distance and grain size, the slope b1 should be near zero if EMT occurs and negative if size-selective transport (SST) occurs. The models for all individual transport events investigated in Spruce Ck do have a slope b1 that is not significantly different than 0 (α = 0.05) even if it is negative. 3.2. Effect of bed morphology on sediment transport The morphological units of the reach were reported on the map, with steps and pools corresponding respectively to elevations higher and lower than the mean bed level (Fig. 7). The channel is subdivided into ten morphological units. The steps are relatively regularly spaced along the reach. Ls is 5 m, which corresponds to 0.7W. The modal distance traveled by the clasts was extracted from Fig. 8 in absolute value and scaled with the average distance between steps and pools measured from the topographic
Fig. 8. Displacement distance distributions of the clasts. Clasts that did not move are excluded: (A) 25 May 2004 to 27 September 2004, (B) 27 September 2004 to 30 May 2006, (C) 30 May 2006 to 3 October 2006, and (D) 3 October 2006 to 30 May 2007. Individual distances Di are scaled by Ls, the average distance between steps and pools.
map. For each event, the modes are close to Ls, suggesting a significant control of the channel morphology on sediment transport. Fig. 9 shows the percentage of clasts deposited in step and pool morphological units after each transport event. The analysis is based on mobile clasts only. Results are presented both for all clasts and for different clast sizes. Classes of b axis/d50 of the bed surface material are used to produce the histograms. The clasts were equally distributed in steps and in pools at the beginning of the experiment except for the largest clasts. There is no sign that one of the morphological units was more effective in trapping the clasts after each of the transport events.
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
277
Fig. 9. Morphological units effective in trapping clasts after each event. Black bars are step units and white bars are pools. (A) All clasts, (B) 0–0.75, (C) 0.75–1.50, (D) 1.50–2.25, and (E) N2.25. Tagged clasts sizes of are scaled to the d50 of the bed material excluding clasts N256 mm.
The differences between the proportions are not significantly different from zero (α = 0.05) for all event and class sizes.
been reported in other studies on gravel-bed rivers (see Hassan et al., 1991; Hassan and Church, 1992; Schmidt and Ergenzinger, 1992).
4. Discussion and conclusions 4.1. Sediment transport in a step–pool channel The Gamma model was found to describe the observed distances of particle displacement in Spruce Ck. Even if the model seems strictly valid in a physical sense when all the particles move the same number of steps during a single flood, a constraint not likely to achieved in natural floods (Hassan et al., 1991), in general it fits our displacement data. This is mostly due to the displacement of the clasts over short distances at low to moderate flood events. The model underestimated the frequencies for short distances for the event 25 May 2004–27 September 2004. Because of the small magnitude flood event (see Table 2), more tagged clasts than predicted were transported on short distances. For each of the other transport event sequence, the goodness of fit of the model suggests that sediment transport is a random process in step–pool channels. A similar observation has
Fig. 10. Excess stream power in relation to scaled averaged distance of travel distance.
278
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
Although the relation is established with only four transport events, the excess stream power appears to predict the average path length of the clasts at low to moderate flow in Spruce Ck. However, the observations yield distance of displacements lower than the average values observed in riffle–pool systems (Fig. 10). Low values in Spruce Ck may represent the structural constraints to entrainment and the stability of the step–pool channel greater than in riffle–pool rivers. Displacement distances of clasts in Spruce Ck were also found to be lower that those measured in Lainbach River (Bavaria), a step–pool channel with a slope of 2% (Gintz et al., 1996). Small events yielded values higher than those obtained from the riffle–pool and the Spruce Ck relations. The relatively high values obtained for small events in Lainbach River may be due to the fact that after each event, the tagged clasts were carried back at the upstream end of the surveyed reach. This approach may have enhanced the mobility of the clasts. Moreover, one must note that one of the investigated flood event in Lainbach River has reached discharge level of 165 m3 s− 1, which is close to six times the mean annual maximum discharge. This event changed the bed morphology and slope, and created new boundary conditions for sediment transport. Such morphological changes were not observed in Spruce Ck under maximum discharge of 3.5 m3 s− 1. Regression models between grain sizes and particle travel distances were used to describe how the variability in bed material size affects bedload entrainment. The results obtained in Spruce Ck indicate that the relation is rather obscure and that SST does not occur for individual sediment transport events although step–pool systems exhibit a wide range of bed material sizes. There is a general trend toward EMT under the conditions surveyed in Spruce Ck and for clasts ranging in size between 40 to 256 mm. Entrainment does appear to be significantly affected by neighboring clasts and by the structural characteristics of the bed. The results confirm those of Blizard and Wohl (1998) who have concluded that EMT most adequately describe sand and gravel transport in East St. Louis Ck, a step–pool system. It also matches those obtained in Toots Ck (Arkansas, USA) a typical step–pool channel where EMT, measured by linking grain size and hydraulic measurements, was found to be more evident than SST during near-bankfull flow events (Marion and Weirich, 2003). The significance of the results in Spruce Ck must be interpreted with caution. The large sample, both in the number of clasts and in their size, to describe the statistical characteristics of particles displacement is needed to access evidence of EMT in step–pool channels. 4.2. The role of step–pool morphology on sediment transport Spatial distribution of the mobile clasts can be modified by regularities in the long profile of the river (Schmidt and Ergenzinger, 1992). Short displacement distances observed here, generally well described by the Gamma distribution, may reflect the effect of the reach-scale morphological features on the movement of clasts. In Spruce Ck, the mode of displacement distances of the mobilized fraction of the bed material during the survey seems closely associated with the average distance between the steps and the downstream pools (Ls). Pools may be acting as erosion sites and
steps are the downstream depositional site as suggested by Pyrce and Ashmore (2003) for streams on gentle slopes. Because of the low and moderate magnitude of flood events investigated here, this interpretation remains speculative. Moreover, no trapping morphological unit was identified in Spruce Ck while the evidence in the literature appears to be contradictory to this effect. In steep channels, Sawada et al. (1983) have shown that sediments were more likely to be deposited where large boulders accumulate (such as the steps). On the contrary, Schmidt and Ergenzinger (1992) reported sedimentation in the pool morphological unit and deficits in the steps in the Lainbach step–pool channel. The results may be affected by the number of clasts tracked during study, the magnitude of the floods, and the number of morphological unit investigated in the reaches (Ergenzinger and Schmidt, 1990). Field investigation of the role of channel morphology on the average individual particle displacements in step–pool channels must be pursued in order to obtain a general model of clasts deposition. Acknowledgements The authors thank the National Sciences and Engineering Research Council of Canada, the Fonds québecois de la recherche sur la nature et les technologies, and the Canadian Foundation for Innovation for their financial support. This research is conducted as part of the program of the Canada Research Chair in fluvial dynamics. We express our gratitude to Vincent Cardin-Tremblay, Bruce MacVicar and Julie Thérien for their help during the data collection. John Laronne has provided many useful comments on the first version of the manuscript. Thanks are also due to Mike Church, Hervé Piégay, and an anonymous reviewer for their constructive comments. References Abrahams, A.D., Li, G., Atkinson, J.F., 1995. Step–pool streams: adjustment to maximum flow resistance. Water Resour. Res. 31 (10), 2593–2602. Allan, J.C., Hart, R., Tranquili, J.V., 2006. The use of passive integrated transponder (PIT) tags to trace cobble transport in a mixed sand-and-gravel beach on the high-energy Oregon coast, USA. Mar. Geol. 232, 63–86. Bagnold, R.A., 1980. An empirical correlation of bedload transport rates in flumes and natural rivers. Proc. R. Soc. Lond. A372, 453–473. Bathurst, J.C., 1978. Flow resistance of large-scale roughness. Am. Soc. Civ. Eng. 104 (HY12), 1587–1603. Blizard, C.R., Wohl, E.E., 1998. Relationships between hydraulic variables and bedload transport in a subalpine channel, Colorado Rocky Mountains, U.S.A. Geomorphology 22, 359–371. Bowman, D., 1977. Stepped-bed morphology in arid gravelly channels. Geol. Soc. Amer. Bull. 88, 291–298. Chartrand, S.M., Whiting, P.J., 2000. Alluvial architecture in headwater streams with special emphasis on step–pool topography. Earth Surf. Process. Landf. 25, 583–600. Chin, A., 1989. Step pools in stream channels. Prog. Phys. Geogr. 13, 390–407. Chin, A., 2003. The geomorphic significance of step–pools in mountain streams. Geomorphology 55 (1–4), 125–137. Chin, A., Wohl, E.E., 2005. Toward a theory for step–pool stream channels. Prog. Phys. Geogr. 29 (3), 275–296. Church, M., Jones, D., 1982. Channel bars in gravel-bed rivers. In: Hey, R.D., Bathurst, J.D., Thorne, C.R. (Eds.), Gravel-bed Rivers. John Wiley and Sons, Chichester, UK, pp. 291–334. Church, M., Zimmermann, A., 2007. Form and stability of step–pool channels: Research Progress. Water Resour. Res. 43 (3) (Art. No. W03415).
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279 Church, M., Hassan, M.A., Wolcott, J.F., 1998. Stabilizing self-organized structures in gravel-bed stream channels: field and experimental observations. Water Resour. Res. 34 (11), 3169–3179. Ergenzinger, P., Schmidt, K.-H., 1990. Stochastic elements of bed load transport in step–pool mountain river. In: Sinniger, R.O., Monbaron, M. (Eds.), Hydrology in mountain regions II: artificial reservoirs, water and slopes. International Association of Hydrological Sciences, Louvain, Belgium, pp. 39–46. Gintz, D., Hassan, M.A., Schmidt, K.-H., 1996. Frequency and magnitude of bedload transport in a mountain river. Earth Surf. Process. Landf. 21, 433–445. Grant, G.E., Mizuyama, T., 1991. Origin of step–pool sequences in high gradient streams: a flume experiment. Japan–U.S. Workshop on Snow Avalanche, Landslide, Debris Flow Prediction and Control. Niigata, Tokyo, Japan, pp. 523–532. Grant, G.E., Swanson, F.J., Wolman, M.G., 1990. Pattern and origin of steppedbed morphology in high-gradient streams, western Cascades, Oregon. Geol. Soc. Amer. Bull. 102, 340–352. Haschenburger, J.K., Church, M., 1998. Bed material transport estimated from the virtual velocity of sediment. Earth Surf. Process. Landf. 23 (9), 791–808. Hassan, M.A., Church, M., 1992. The movement of individual grains on the streambed. In: Billi, P., Hey, R.D., Thorne, C.R., Tacconi, P. (Eds.), Dynamics of Gravel-bed Rivers. John Wiley and Sons, Chichester, UK, pp. 159–175. Hassan, M.A., Church, M., Schick, A.P., 1991. Distance movement of coarse particles in gravel bed streams. Water Resour. Res. 27 (4), 503–511. Hassan, M.A., Church, M., Ashworth, P.J., 1992. Virtual rate and mean distance of travel of individual clasts in gravel-bed channels. Earth Surf. Process. Landf. 17, 617–627. Hayward, J.A., 1980. Hydrology and stream sediment from the Torlesse Stream Catchment. Special publication 17, Lincoln Papers in Water Resources. Tussock Grassland and Mountain Lands Institute, Lincoln, New Zealand. Heede, B.H., 1972. Influences of a forest on the hydraulic geometry of two mountain streams. Water Res. Bull. 8, 523–530. Lamarre, H., MacVicar, B., Roy, A.G., 2005. Using passive integrated transponder (PIT) tags to investigate sediment transport in gravel-bed rivers. J. Sediment. Res. 75 (4), 736–741. Marion, D.A., Weirich, F., 1999. Fine-grained bed patch response to nearbankfull flows in a step–pool channel. In: Olsen, D.S., Pontyondy, J.P. (Eds.), Proceedings of American Water Resources Association Specialty Conference: Wildland Hydrology, pp. 93–100.
279
Marion, D.A., Weirich, F., 2003. Equal-mobility bed load transport in small, step–pool channel in the Ouachita Mountains. Geomorphology 55, 139–154. Nichols, M.H., 2004. A radio frequency identification system for monitoring coarse sediment particle displacement. Appl. Eng. Agric. 20 (6), 783–787. Parker, G., Kligeman, P.C., McLean, D.G., 1982. Bedload and size distribution in paved gravel-bed streams. J. Hydraul. Div. 108 (4), 544–571. Peterson, D.F., Mohanty, P.K., 1960. Flume studies of flow in steep and rough channels. Am. Soc. Civ. Eng. 86, 55–76. Pyrce, R.S., Ashmore, P.E., 2003. The relation between particle path length distributions and channel morphology in gravel-bed streams: a synthesis. Geomorphology 56, 167–187. Richards, K.S., 1976. The morphology of riffle–pool sequences. Earth Surf. Process. Landf. 1, 71–88. Sawada, M., Ashida, L., Takahashi, T., 1983. Relationship between channel pattern and sediment transport in steep gravel-bed river. Z. Geomorphol. 46, 55–66. Schmidt, K.H., Ergenzinger, P., 1992. Bedload entrainment, travel lengths, step lengths, rest periods-studied with passive (irons, magnetic) and active (radio) tracer techniques. Earth Surf. Process. Landf. 17, 147–165. Sear, D.A., Lee, M.W.E., Oakey, R.J., Carling, P.A., Collins, M.B., 2000. Coarse sediment tracing technology in littoral and fluvial environments: a review. In: Foster, I.D.L. (Ed.), Tracers in Geomorphology. Wiley, Chichester, UK, pp. 21–56. Trimble, 2005. Trimble 5600DR Total Station: Key Features. 10 pp. www.trimble. com. Whittaker, J.G., 1987. Sediment transport in step–pool streams. In: Thorne, C.R., Bathurst, J.C., Hey, R.D. (Eds.), Sediment Transport in Gravel-bed Rivers. John Wiley and Sons, Chichester, UK, pp. 545–579. Whittaker, J.G., Jaeggi, M.N.R., 1982. Origin of step–pool systems in mountain streams. Proc. Am. Soc. Civ. Eng., J. Hydraul. Div. 108, 758–773. Wohl, E.E., Grodek, T., 1994. Channel bed-steps along Nahal Yael, Negev desert, Israel. Geomorphology 9, 117–126. Wohl, E.E., Thompson, D.M., 2000. Velocity characteristics along a small step– pool channel. Earth Surf. Process. Landf. 25, 353–367. Zimmermann, A., Church, M., 2001. Channel morphology, gradient profiles and bed stresses during flood in a step–pool channel. Geomorphology 40, 311–327.
Geomorphology 99 (2008) 270 – 279 www.elsevier.com/locate/geomorph
The role of morphology on the displacement of particles in a step–pool river system Hélène Lamarre ⁎, André G. Roy Département de Géographie, Université de Montréal, C.P. 6128, Succ. Centre-Ville, Montréal, Québec, Canada Received 15 February 2007; received in revised form 14 November 2007; accepted 16 November 2007 Available online 22 November 2007
Abstract Although studies of sediment transport in steep and coarse-grained channels have been more numerous in recent years, the dynamics of sediment transport in step–pool river systems remain poorly understood. This paper investigates displacements of individual clasts through Spruce Creek (Québec, Canada), a classic step–pool channel, and the effects of the channel morphology on the path length of the clasts. Passive integrated transponder tags (PIT) were used to track the displacement of 196 individual particles over a range of discharges including the bankfull stage. Clasts were tracked after five sequences of flood events. The results showed that the distance distributions match a two-parameter Gamma model. Equal mobility transport occurs for the particle size investigated during each sequence of flood events. Mean travel distance of the clasts can be estimated from excess stream power, and the mobility of the clasts is more than an order of magnitude less than the model reported in riffle–pool channels. The dominant morphological length scale of the bed also controls the path length of the clasts. These results confirm some preliminary observations on sediment transport in step–pool channels. © 2007 Elsevier B.V. All rights reserved. Keywords: Gravel-bed rivers; Step–pool; Passive integrated transponders; Radio frequency identification; Sediment transport; Particle mobility; Distance of movement; Bed morphology; Steep channels
1. Introduction Step–pool channels occur in a wide range of bioclimatic environments (Bowman, 1977), from arid (Wohl and Grodek, 1994) to humid forested areas (Heede, 1972). They are commonly found in steep headwater mountain streams, where the channel width to depth ratio is small, bed material is heterogeneous, and slopes exceed 4 to 7% (Chin, 1989; Grant et al., 1990; Grant and Mizuyama, 1991). The longitudinal profile of step–pool channel is characterized by steps viewed as congested zones where clasts have clustered (Church and Jones, 1982; Church et al., 1998). The steps are generally composed of an accumulation of cobbles and boulders that are transverse or oblique to the channel (Zimmermann and Church, 2001; Chin and Wohl, 2005) and their occurrence depends on the local availability of keystones that remain immobile on the ⁎ Corresponding author. Tel.: +1 514 343 8035; fax: +1 514 343 8008. E-mail address: [email protected] (H. Lamarre). 0169-555X/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2007.11.005
contemporary flow regime (Church and Zimmermann, 2007). Steps define breaks of slope and alternate with pools, containing finer bed material. The geometry of step–pools is defined by step spacing (L) and step height (H) and their ratio has been found to fall in the range 0.06 ≤ H/L ≤ 0.20, which corresponds to the range of gradients where step–pools are found (Chin and Wohl, 2005; Church and Zimmermann, 2007). Step spacing has also been correlated with overall stream gradient (S) in the form H/LαSβ, in which 0.42 ≤ β ≤ 0.68 (Abrahams et al., 1995). Chartrand and Whiting (2000) have shown that the step spacing was 0.6 time the channel width (W), while Bowman (1977) reported a spacing of 1.4W. Chin (1989) obtained a spacing of 2.7W for step–pools in Santa Monica Mountains (Colorado). Data from numerous locations have shown that step–pool spacing is usually less than one to four channel widths. The step–pool morphology serves a fundamental role in river systems controlling hydraulic resistance (Wohl and Grodek, 1994; Abrahams et al., 1995) and energy dissipation (Hayward,
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
1980; Whittaker and Jaeggi, 1982). They are characterized by relative submergence of the large clasts (Y /d84) that is b1.2 (Bathurst, 1978), where Y is the flow depth and d84 the 84th percentile bed material size. At high flow stage, water flows over and in between the large roughness elements that form the steps and plunge into the pools below, promoting a tumbling highly turbulent flow (Peterson and Mohanty, 1960; Whittaker, 1987). The available energy is dissipated by eddies (Wohl and Thompson, 2000). Therefore, the potential energy available for erosion and sediment transport at high stage is reduced by the steps (Chin, 2003). In their recently published review of empirical research on step–pool channels, Chin and Wohl (2005) pointed out some of the common observations on bedload transport in these systems: transported sediment is derived from localized sites on adjacent hillslopes and in the channel; bedload entrainment and transport are spatially and temporally discontinuous (Ergenzinger and Schmidt, 1990); sediment are preferentially stored in
271
and mobilized from pools (Schmidt and Ergenzinger, 1992; Marion and Weirich, 1999); bedload transport during lower flows tends to be characterized by equal-mobility transport (Blizard and Wohl, 1998; Marion and Weirich, 2003) where larger grains are as easily entrained as smaller ones because they are more exposed to lift and drag forces (Parker et al., 1982). In spite of these findings, there is still a paucity of field data on the displacement of clasts in step–pool channels. Such data are a good indicator of bedload transport response of the stream to a given water discharge and to sediment supply conditions, and are critical for understanding of the development of channel morphology (see Haschenburger and Church, 1998; Sear et al., 2000). Morphological features of step–pool channels should play a significant role on the distance of displacement of the bed material. At this time, it is still not possible to predict sediment transport rates in step–pool systems as a function of flow and inchannel sediment storage (Church and Zimmermann, 2007). It
Fig. 1. Location (A) and upstream view (B) of Spruce Ck step–pool channel.
272
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
Table 1 Number of particles in each clast size b axis (mm)
Number of tagged clasts
0–50 50–100 100–150 150–200 200–250 Total
4 124 60 6 2 196
160 mm. Removing the boulders (N256 mm = 30% of the measured particles), which can be transported only during exceptional flood, from the distribution, the d50 of the surface bed material is reduced to 90 mm. 2.2. Sediment tracking
2.1. Field site
We implanted passive integrated transponder (PIT) tags into clasts of various sizes and shapes to quantify clast movements. A PIT is a glass encapsulated pulse transponder, using a radio frequency system, with a unique identification code encrypted. A detailed description of the tags and of its application to sediment transport investigation is given in Nichols (2004), Lamarre et al. (2005) and Allan et al. (2006). PITs were installed in the laboratory within holes drilled into indigene clasts sampled on the stream bed (n = 196). The number of clasts tagged within each size fraction is given in Table 1. The size distribution of these particles is close to that of the particles composing the bed surface (Fig. 2), although two limitations prevented a perfect match. Firstly, particles with a b axis (intermediate axis) less than 40 mm were too small to host a PIT tag. Secondly, particles larger than 256 mm were difficult to bring back to the laboratory because of their excessive weight. In Spruce Ck, the d50 of the tagged clasts (86 mm) is half the d50 of the surface bed material (160 mm), but it matches the d50 of the distribution if the clasts N256 mm are not included. Our strategy aimed at focusing on the clasts that could be mobilized during low to moderate floods that were likely to occur during the survey period.
The study was conducted in Spruce Creek (alt. 520 m), a classic step–pool channel located on Mount Sutton in the Eastern Townships (Québec) (Fig. 1). The 50-m study reach is located in the middle part of the basin and drains a forested area of about 3.2 km2. The length of the study site was defined by the time needed to characterize both the channel morphology and the sediment transport characteristics. The channel bed is composed of clasts ranging from pebbles to boulders, essentially coming from bedrock fragments and a till layer. A number of clasts in the channel are in the order of 1 m in diameter. These boulders assure the stability of the step morphology, probably remaining immobile under contemporary flood regime. Mobile clasts during the survey are derived from the channel bed, the lower portion of the channel bank, and local slides on hillslopes along the reach. Average morphological characteristics of the reach were extracted from detailed topographic maps of the bed. The maps were produced using a Trimble robotic total station (model DR-5600). The total station collects x, y and z coordinates electronically and stores the observation directly in a portable computer (Trimble, 2005). More than seven topographic points per square metre were measured systematically along transects and at breaks of slope in order to extract the morphological characteristics of the channel. At the bankfull stage, the average channel width is 6.0 m, flow depth is 0.45 m, and slope is 14%. The surface bed material was characterized over the reach with a Wolman count performed on 1200 randomly selected particles. The median size (d50) of the bed surface material is
Fig. 2. Bed material (black) and tagged clasts (white) size distribution: (A) including clasts N256 mm, (B) excluding clasts N256 mm.
is to be expected that the movement of individual particles may be controlled by the large morphological features such as the steps and the pools. This control could be similar to that of pools and bars in lowland streams as suggested by Pyrce and Ashmore (2003) who have shown that the mode of the path length distribution coincides with known pool-bar spacing in gravelbed streams. The objectives of this paper are to (i) describe the movement of the clasts to compare with other preliminary observations on bedload transport in step–pool channels and (ii) quantify the effect of the step and pool spacing on the mode displacement distance of the clasts. This study is based on a unique set of particle displacement data measured on a four year period in a step–pool channel. 2. Methods
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
273
Table 2 Characteristics of the survey periods Tagged particles recovery dates
Recovery rates (%)a
Number of floodb
Qmax (m3 s− 1)c
Qmax / Qdbk
Q1 (m3 s− 1)e
Q1 / Qbk
Duration of flow N ω0 (hours)f
Mobile particles (%)
1 December 2003 25 May 2004g 27 September 2004 30 May 2006 3 October 2006 30 May 2007
Initial seeding 92 83 62 62 57
1 1 5 1 2
2.04 1.20 3.40 2.11 1.36
1.0 0.6 1.6 1.0 0.7
2.04 1.20 1.07 2.11 1.36
1.0 0.6 0.5 1.0 0.7
120 96 336 96 144
3.84 1.53 5.01 8.31 8.29
62 32 70 86 78
a
Recovery rates include all the clasts that were not found within the sampling section. The number of flood events is estimated selecting the peaks of discharge N1 m3 s− 1 (=0.5Qbk) in which case, at least 25% of the tagged clasts were mobile. c Qmax is the largest peak flow. d Qbk is the bankfull flow. e Q1 is the first peak flow. f Number of hours where specific stream power is Nω0. g Results of the first event disregarded in the analysis. b
The clasts were seeded at low flow on the channel bed on 1 December 2003. Because of the heterogeneity and roughness of the bed, we could not place the particles systematically along cross-sections. The tagged clasts were seeded in a way to mimic the position of similar clasts. We attempted to respect inasmuch as possible the natural organization of the bed and the degree of imbrication of the particles. It was difficult, however, to replace the tagged clasts to maintain the exact structure of the bed surface material. Therefore, most tagged particles appeared less constrained than the clasts originally composing the water worked surface. We supposed that after the first transport event, the clasts were reorganized in a more natural way, while they occupy a more stable position. The results of the first displacement of the clasts were then disregarded from further analysis. Note that at the upstream end of the reach, most of the tracers were seeded on one side of the channel. The other side was characterized by bed material much finer than the size distribution of the tagged particles. The location of each particle within the channel was measured with the robotic total station. The tracking of particles was carried out at low flow after five sediment transport episodes: on 25 May 2004 (removed
from the analysis), 27 September 2004, 30 May 2006, 3 October 2006, and 30 May 2007. Buried tagged particles could be detected up to a depth of 0.25 m within the substrate (Lamarre et al., 2005). The characteristics of the flood events were estimated from discharges measured by the Centre d'expertise hydrique du Québec in the neighboring Yamaska-est watershed (Table 2). Hydrograph of the average daily discharge of Spruce Ck was reconstructed using a correlation approach between the Yamaska sud-est discharge series and punctual discharge measurements made in Spruce Ck. High discharges happened before the spring surveys and discharges remained below bankfull stage (Qbk) until the autumn surveys (Fig. 3). The number of flood events between the surveys was estimated from the discharge peaks above 1 m3 s− 1 (0.5Qbk) and when at least 25% of the tagged clasts were mobile. Several constraints such as the distance of the site from the university or the winter conditions made it impossible to survey the clast positions after each flood event. The recovery rates of tagged particles were 83% on 27 September 2004, and 62% for the two subsequent surveys (Table 2). The recovery rate falls to 57% on 30 May 2007. This
Fig. 3. Hydrograph of the average daily discharges of Spruce Ck between August 2003 and June 2007. Discharges are estimated from data measured by the Centre d'expertise hydrique du Québec in the neighboring Yamaska sud-est watershed. Dates of the recovery of the tagged particles are presented in relation to flood events. Thresholds of bankfull flow (Qbk) and 0.5Qbk are superimposed to the series.
274
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
during the recovery process (Lamarre et al., 2005). The percentages of mobile tagged particles increase with discharge. The path length traveled by a tagged particle was estimated as the straight line distance between the initial location and the point of final deposition. Given that the channel has a small width and is relatively straight, we assumed that the difference between the straight line and the real distances traveled by the clasts is negligible. We also identified the number of tagged particles deposited in each morphological unit at the time of the surveys. The results obtained on the transfer of tagged clasts between each of the morphological units are influenced by the particles moving downstream during the experiment and the fact that particles are not being replaced at the upstream end of the reach after each transport event. 2.3. Step–pool configuration The locations of the particles were superimposed on the detailed topographic map of the bed. Steps were distinguished from pools using field observations in conjunction with the features seen on the topographic map. The location of the morphological units was estimated using the zero-crossing technique along longitudinal profiles (Richards, 1976) and corroborated with a visual identification. The average distance (Ls) between the step and the pool was estimated from the top of the crest to the lowest point of the pool located immediately downstream. 3. Results 3.1. Sediment transport Fig. 4 shows the distribution of the distances of clast movements (Di) scaled by the mean distance of the event (
Fig. 4. Displacement distance distribution of the clasts. Clasts that did not move are excluded: (A) 25 May 2004 to 27 September 2004, (B) 27 September 2004 to 30 May 2006, (C) 39 May 2006 to 3 October 2006, and (D) 3 October 2006 to 30 May 2007. Individual distance Di is scaled by
may be due to the burial of the clasts and to the duration of the experiment. Some of the tagged clasts have exited the surveyed section, thus affecting the recovery rates. Using the tagged particles, we characterized the mobility, the distance of displacement and the morphological unit where the clasts have deposited. The mobility represents the percentage of tagged particles that were transported between each survey. A tagged particle found outside of a 0.3 m radius of its previous location was considered as being mobilized. This threshold is associated to the sampling error of the location of the PITs
Fig. 5. Relation between the displacement distances of the clasts scaled by the total time for which the flow was larger than that corresponding to ω0 and excess specific stream power. Specific stream power is estimated for the first peak flow during flood events. Clasts that did not move are excluded.
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
275
Fig. 6. Individual displacement distances in relation to grain size. Clasts that did not move are excluded: (A) 25 May 04 to 27 September 2004, (B) 27 September 2004 to 30 May 2006, (C) 30 May 2006 to 3 October 2006, (D) 3 October 2006 to 30 May 2007. Grain size b axis is scaled with the d50 of the bed surface material that does not include clasts N256 mm.
Fig. 7. Subdivision of reaches into distinct morphological units from (A) longitudinal topographic profile and (B) topographic map of the bed. Slope has been removed to emphasize the changes in bed elevation.
276
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
0.25 in each distribution. The observed distances of movement are fitted with a two-parameter Gamma distribution (Hassan et al., 1991). Chi-square tests (α = 0.05) indicate that Gamma models fit the data well, except for the first event (p = 0.002). In particular, the model underestimates the frequency for short distances. The correlation between the path length of the clasts and flow conditions is presented in Fig. 5 on an event basis. Specific stream power (ω) was used to quantify the flow conditions of the reach. In this study, specific stream power is based on the bed slope of the reach and on the first peak discharge observed in an episode of sediment transport (Table 2). This value was selected to reduce the effect of duration of the period between each survey. The excess stream power (ω − ω0) was estimated using Bagnold's (1980) method where: 12Y 1:5 log x0 c290d50 d50 (see Hassan et al., 1992). In Bagnold's original formula, d50 represents the median size of the subsurface material. Here, this value was replaced by the d50 of the bed surface, because the size of subsurface bed material is not known. It may have reduced the estimate of excess stream power values, expecting that the size of subsurface is smaller than at the bed surface. Also, step–pool channels are characterized by large-scale structural arrangements that may influence the entrainment and transport. The approach does not consider those features. The average distance traveled by the clasts is divided by the total time for which the discharge was higher than that corresponding to ω0 (< Dv>). The empirical relation between < Dv> and the excess stream power is:
Fig. 8. Displacement distance distributions of the clasts. Clasts that did not move are excluded: (A) 25 May 2004 to 27 September 2004, (B) 27 September 2004 to 30 May 2006, (C) 30 May 2006 to 3 October 2006, and (D) 3 October 2006 to 30 May 2007. Individual distances Di are scaled by Ls, the average distance between steps and pools.
map. For each event, the modes are close to Ls, suggesting a significant control of the channel morphology on sediment transport. Fig. 9 shows the percentage of clasts deposited in step and pool morphological units after each transport event. The analysis is based on mobile clasts only. Results are presented both for all clasts and for different clast sizes. Classes of b axis/d50 of the bed surface material are used to produce the histograms. The clasts were equally distributed in steps and in pools at the beginning of the experiment except for the largest clasts. There is no sign that one of the morphological units was more effective in trapping the clasts after each of the transport events.
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
277
Fig. 9. Morphological units effective in trapping clasts after each event. Black bars are step units and white bars are pools. (A) All clasts, (B) 0–0.75, (C) 0.75–1.50, (D) 1.50–2.25, and (E) N2.25. Tagged clasts sizes of are scaled to the d50 of the bed material excluding clasts N256 mm.
The differences between the proportions are not significantly different from zero (α = 0.05) for all event and class sizes.
been reported in other studies on gravel-bed rivers (see Hassan et al., 1991; Hassan and Church, 1992; Schmidt and Ergenzinger, 1992).
4. Discussion and conclusions 4.1. Sediment transport in a step–pool channel The Gamma model was found to describe the observed distances of particle displacement in Spruce Ck. Even if the model seems strictly valid in a physical sense when all the particles move the same number of steps during a single flood, a constraint not likely to achieved in natural floods (Hassan et al., 1991), in general it fits our displacement data. This is mostly due to the displacement of the clasts over short distances at low to moderate flood events. The model underestimated the frequencies for short distances for the event 25 May 2004–27 September 2004. Because of the small magnitude flood event (see Table 2), more tagged clasts than predicted were transported on short distances. For each of the other transport event sequence, the goodness of fit of the model suggests that sediment transport is a random process in step–pool channels. A similar observation has
Fig. 10. Excess stream power in relation to scaled averaged distance of travel distance.
278
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279
Although the relation is established with only four transport events, the excess stream power appears to predict the average path length of the clasts at low to moderate flow in Spruce Ck. However, the observations yield distance of displacements lower than the average values observed in riffle–pool systems (Fig. 10). Low values in Spruce Ck may represent the structural constraints to entrainment and the stability of the step–pool channel greater than in riffle–pool rivers. Displacement distances of clasts in Spruce Ck were also found to be lower that those measured in Lainbach River (Bavaria), a step–pool channel with a slope of 2% (Gintz et al., 1996). Small events yielded values higher than those obtained from the riffle–pool and the Spruce Ck relations. The relatively high values obtained for small events in Lainbach River may be due to the fact that after each event, the tagged clasts were carried back at the upstream end of the surveyed reach. This approach may have enhanced the mobility of the clasts. Moreover, one must note that one of the investigated flood event in Lainbach River has reached discharge level of 165 m3 s− 1, which is close to six times the mean annual maximum discharge. This event changed the bed morphology and slope, and created new boundary conditions for sediment transport. Such morphological changes were not observed in Spruce Ck under maximum discharge of 3.5 m3 s− 1. Regression models between grain sizes and particle travel distances were used to describe how the variability in bed material size affects bedload entrainment. The results obtained in Spruce Ck indicate that the relation is rather obscure and that SST does not occur for individual sediment transport events although step–pool systems exhibit a wide range of bed material sizes. There is a general trend toward EMT under the conditions surveyed in Spruce Ck and for clasts ranging in size between 40 to 256 mm. Entrainment does appear to be significantly affected by neighboring clasts and by the structural characteristics of the bed. The results confirm those of Blizard and Wohl (1998) who have concluded that EMT most adequately describe sand and gravel transport in East St. Louis Ck, a step–pool system. It also matches those obtained in Toots Ck (Arkansas, USA) a typical step–pool channel where EMT, measured by linking grain size and hydraulic measurements, was found to be more evident than SST during near-bankfull flow events (Marion and Weirich, 2003). The significance of the results in Spruce Ck must be interpreted with caution. The large sample, both in the number of clasts and in their size, to describe the statistical characteristics of particles displacement is needed to access evidence of EMT in step–pool channels. 4.2. The role of step–pool morphology on sediment transport Spatial distribution of the mobile clasts can be modified by regularities in the long profile of the river (Schmidt and Ergenzinger, 1992). Short displacement distances observed here, generally well described by the Gamma distribution, may reflect the effect of the reach-scale morphological features on the movement of clasts. In Spruce Ck, the mode of displacement distances of the mobilized fraction of the bed material during the survey seems closely associated with the average distance between the steps and the downstream pools (Ls). Pools may be acting as erosion sites and
steps are the downstream depositional site as suggested by Pyrce and Ashmore (2003) for streams on gentle slopes. Because of the low and moderate magnitude of flood events investigated here, this interpretation remains speculative. Moreover, no trapping morphological unit was identified in Spruce Ck while the evidence in the literature appears to be contradictory to this effect. In steep channels, Sawada et al. (1983) have shown that sediments were more likely to be deposited where large boulders accumulate (such as the steps). On the contrary, Schmidt and Ergenzinger (1992) reported sedimentation in the pool morphological unit and deficits in the steps in the Lainbach step–pool channel. The results may be affected by the number of clasts tracked during study, the magnitude of the floods, and the number of morphological unit investigated in the reaches (Ergenzinger and Schmidt, 1990). Field investigation of the role of channel morphology on the average individual particle displacements in step–pool channels must be pursued in order to obtain a general model of clasts deposition. Acknowledgements The authors thank the National Sciences and Engineering Research Council of Canada, the Fonds québecois de la recherche sur la nature et les technologies, and the Canadian Foundation for Innovation for their financial support. This research is conducted as part of the program of the Canada Research Chair in fluvial dynamics. We express our gratitude to Vincent Cardin-Tremblay, Bruce MacVicar and Julie Thérien for their help during the data collection. John Laronne has provided many useful comments on the first version of the manuscript. Thanks are also due to Mike Church, Hervé Piégay, and an anonymous reviewer for their constructive comments. References Abrahams, A.D., Li, G., Atkinson, J.F., 1995. Step–pool streams: adjustment to maximum flow resistance. Water Resour. Res. 31 (10), 2593–2602. Allan, J.C., Hart, R., Tranquili, J.V., 2006. The use of passive integrated transponder (PIT) tags to trace cobble transport in a mixed sand-and-gravel beach on the high-energy Oregon coast, USA. Mar. Geol. 232, 63–86. Bagnold, R.A., 1980. An empirical correlation of bedload transport rates in flumes and natural rivers. Proc. R. Soc. Lond. A372, 453–473. Bathurst, J.C., 1978. Flow resistance of large-scale roughness. Am. Soc. Civ. Eng. 104 (HY12), 1587–1603. Blizard, C.R., Wohl, E.E., 1998. Relationships between hydraulic variables and bedload transport in a subalpine channel, Colorado Rocky Mountains, U.S.A. Geomorphology 22, 359–371. Bowman, D., 1977. Stepped-bed morphology in arid gravelly channels. Geol. Soc. Amer. Bull. 88, 291–298. Chartrand, S.M., Whiting, P.J., 2000. Alluvial architecture in headwater streams with special emphasis on step–pool topography. Earth Surf. Process. Landf. 25, 583–600. Chin, A., 1989. Step pools in stream channels. Prog. Phys. Geogr. 13, 390–407. Chin, A., 2003. The geomorphic significance of step–pools in mountain streams. Geomorphology 55 (1–4), 125–137. Chin, A., Wohl, E.E., 2005. Toward a theory for step–pool stream channels. Prog. Phys. Geogr. 29 (3), 275–296. Church, M., Jones, D., 1982. Channel bars in gravel-bed rivers. In: Hey, R.D., Bathurst, J.D., Thorne, C.R. (Eds.), Gravel-bed Rivers. John Wiley and Sons, Chichester, UK, pp. 291–334. Church, M., Zimmermann, A., 2007. Form and stability of step–pool channels: Research Progress. Water Resour. Res. 43 (3) (Art. No. W03415).
H. Lamarre, A.G. Roy / Geomorphology 99 (2008) 270–279 Church, M., Hassan, M.A., Wolcott, J.F., 1998. Stabilizing self-organized structures in gravel-bed stream channels: field and experimental observations. Water Resour. Res. 34 (11), 3169–3179. Ergenzinger, P., Schmidt, K.-H., 1990. Stochastic elements of bed load transport in step–pool mountain river. In: Sinniger, R.O., Monbaron, M. (Eds.), Hydrology in mountain regions II: artificial reservoirs, water and slopes. International Association of Hydrological Sciences, Louvain, Belgium, pp. 39–46. Gintz, D., Hassan, M.A., Schmidt, K.-H., 1996. Frequency and magnitude of bedload transport in a mountain river. Earth Surf. Process. Landf. 21, 433–445. Grant, G.E., Mizuyama, T., 1991. Origin of step–pool sequences in high gradient streams: a flume experiment. Japan–U.S. Workshop on Snow Avalanche, Landslide, Debris Flow Prediction and Control. Niigata, Tokyo, Japan, pp. 523–532. Grant, G.E., Swanson, F.J., Wolman, M.G., 1990. Pattern and origin of steppedbed morphology in high-gradient streams, western Cascades, Oregon. Geol. Soc. Amer. Bull. 102, 340–352. Haschenburger, J.K., Church, M., 1998. Bed material transport estimated from the virtual velocity of sediment. Earth Surf. Process. Landf. 23 (9), 791–808. Hassan, M.A., Church, M., 1992. The movement of individual grains on the streambed. In: Billi, P., Hey, R.D., Thorne, C.R., Tacconi, P. (Eds.), Dynamics of Gravel-bed Rivers. John Wiley and Sons, Chichester, UK, pp. 159–175. Hassan, M.A., Church, M., Schick, A.P., 1991. Distance movement of coarse particles in gravel bed streams. Water Resour. Res. 27 (4), 503–511. Hassan, M.A., Church, M., Ashworth, P.J., 1992. Virtual rate and mean distance of travel of individual clasts in gravel-bed channels. Earth Surf. Process. Landf. 17, 617–627. Hayward, J.A., 1980. Hydrology and stream sediment from the Torlesse Stream Catchment. Special publication 17, Lincoln Papers in Water Resources. Tussock Grassland and Mountain Lands Institute, Lincoln, New Zealand. Heede, B.H., 1972. Influences of a forest on the hydraulic geometry of two mountain streams. Water Res. Bull. 8, 523–530. Lamarre, H., MacVicar, B., Roy, A.G., 2005. Using passive integrated transponder (PIT) tags to investigate sediment transport in gravel-bed rivers. J. Sediment. Res. 75 (4), 736–741. Marion, D.A., Weirich, F., 1999. Fine-grained bed patch response to nearbankfull flows in a step–pool channel. In: Olsen, D.S., Pontyondy, J.P. (Eds.), Proceedings of American Water Resources Association Specialty Conference: Wildland Hydrology, pp. 93–100.
279
Marion, D.A., Weirich, F., 2003. Equal-mobility bed load transport in small, step–pool channel in the Ouachita Mountains. Geomorphology 55, 139–154. Nichols, M.H., 2004. A radio frequency identification system for monitoring coarse sediment particle displacement. Appl. Eng. Agric. 20 (6), 783–787. Parker, G., Kligeman, P.C., McLean, D.G., 1982. Bedload and size distribution in paved gravel-bed streams. J. Hydraul. Div. 108 (4), 544–571. Peterson, D.F., Mohanty, P.K., 1960. Flume studies of flow in steep and rough channels. Am. Soc. Civ. Eng. 86, 55–76. Pyrce, R.S., Ashmore, P.E., 2003. The relation between particle path length distributions and channel morphology in gravel-bed streams: a synthesis. Geomorphology 56, 167–187. Richards, K.S., 1976. The morphology of riffle–pool sequences. Earth Surf. Process. Landf. 1, 71–88. Sawada, M., Ashida, L., Takahashi, T., 1983. Relationship between channel pattern and sediment transport in steep gravel-bed river. Z. Geomorphol. 46, 55–66. Schmidt, K.H., Ergenzinger, P., 1992. Bedload entrainment, travel lengths, step lengths, rest periods-studied with passive (irons, magnetic) and active (radio) tracer techniques. Earth Surf. Process. Landf. 17, 147–165. Sear, D.A., Lee, M.W.E., Oakey, R.J., Carling, P.A., Collins, M.B., 2000. Coarse sediment tracing technology in littoral and fluvial environments: a review. In: Foster, I.D.L. (Ed.), Tracers in Geomorphology. Wiley, Chichester, UK, pp. 21–56. Trimble, 2005. Trimble 5600DR Total Station: Key Features. 10 pp. www.trimble. com. Whittaker, J.G., 1987. Sediment transport in step–pool streams. In: Thorne, C.R., Bathurst, J.C., Hey, R.D. (Eds.), Sediment Transport in Gravel-bed Rivers. John Wiley and Sons, Chichester, UK, pp. 545–579. Whittaker, J.G., Jaeggi, M.N.R., 1982. Origin of step–pool systems in mountain streams. Proc. Am. Soc. Civ. Eng., J. Hydraul. Div. 108, 758–773. Wohl, E.E., Grodek, T., 1994. Channel bed-steps along Nahal Yael, Negev desert, Israel. Geomorphology 9, 117–126. Wohl, E.E., Thompson, D.M., 2000. Velocity characteristics along a small step– pool channel. Earth Surf. Process. Landf. 25, 353–367. Zimmermann, A., Church, M., 2001. Channel morphology, gradient profiles and bed stresses during flood in a step–pool channel. Geomorphology 40, 311–327.