The geomorphic and ecological effectiveness of habitat rehabilitation works: Continuous measurement of scour and fill around large logs in sand-bed streams
Introduction
Ideally, stream rehabilitation projects should provide a degraded river with the appropriate elements for a healthy system; including appropriate: flow, sediment, coarse-wood, vegetation, etc. One would then let these elements interact to produce an ecologically superior state that would be closer to some target state. Unfortunately, most stream rehabilitation projects do not have the luxury of this laissez-faire approach. Rehabilitation projects take place in a constrained environment in which resources are limited. A typical river rehabilitation project in Australia, for example, will be attempting to optimise the environmental benefit from a limited amount of (very expensive) environmental water provided from a dam, combined with various structural measures, such as adding coarse-wood for habitat. These modifications will be imposed on a stream system that is still adjusting to altered historical flow and sediment regimes. Managers want to increase their confidence that any interventions will work before they commit to them. The result is that engineers, ecologists and geomorphologists are often asked to predict the magnitude of the physical response to an intervention before it is implemented. What flow will turn-over the gravel-bed, for example. Will a particular meandering channel infill with sediment? Here we concern ourselves with a similar question: will a log placed in a river bed initiate scour that provides deep-water and cover that allow for fish migration?
This paper provides a case-study of a small part of a major rehabilitation project on the Snowy River in south eastern Australia. This 500 km long river (catchment area of 15,500 km2) rises on the slopes of Mount Kosciusko in south-east NSW, and has its lower reaches in Victoria. Low and moderate flows in the Snowy River have been dramatically reduced by inter-basin water transfers, and the river is now the subject of one of the largest rehabilitation project in Australia (Erskine et al., 1999, Anonymous, 2000). The Victorian Government is intending to improve fish passage in the sand-bed of the lower Snowy River, East Gippsland (Fig. 1), by building structures that will optimise the benefits of environmental flows by promoting scour of the sand-bed to enhance fish passage and fish habitat (ID&A, 1998, Gippel et al., 2002). Wood is an important component of instream physical diversity (Gippel, 1995), particularly in sand-bed streams (Mutz, 2000, Webb and Erskine, 2003, Erskine and Webb, 2003). Human pressures have reduced instream physical diversity. Channelisation, riparian zone degradation, sedimentation (sand-slugs) and desnagging have been widespread throughout Australia, reducing levels of habitat diversity (Rutherfurd, 2000, Erskine and Webb, 2003). It is common for wood to be artificially restored to degraded streams to encourage bed scour, and so provide deeper, slow flowing areas (Shields et al., 2004, Brooks et al., 2004), that are considered important attributes of the physical habitat for fish (Crook and Robertson, 1999, Bond and Lake, 2003).
Response of biota to habitat manipulations has been mixed. Whilst some trials have shown positive responses to habitat manipulations (Shields et al., 1998, Shields et al., 2003, Brooks et al., 2004) others have shown modest or negative physical and biological consequences of the interventions (Jonga et al., 1997, Thompson, 2003, Champoux et al., 2003). The analysis of physical changes (such as pool depth) in such studies in sand-bed streams, however, have typically relied on single measurements of scour taken months or years apart (Shields et al., 1995, Shields et al., 1998, Till and Davy, 2000, Lintermans, 2001). Shields et al. (1995) concluded that “Repeated visual observation of pools adjacent to groins show that scour and deposition patterns at a given structure fluctuated through time. Several structures exhibited cycles of deposition: deep scour holes… would alternate with sandbars” (p.499). This reflects the conclusion of Stalnaker et al. (1996) who warned that “the temporal dynamics of habitat quantity are a major influence determining fish population responses in riverine environments”.
Perhaps the key geomorphic question is not whether flow will produce scour around the log, but will the scour still be there when it is needed by the fish? Geomorphic and engineering models have been concerned with maximum scour, rather than with the more subtle patterns of scour and fill that interact with the life cycle of fish. For example, the main body of literature on local scour in streams focuses on maximum scour around bridge piers, as this is the process that will threaten the structure (Raudkivi, 1991, Melville, 1997). Subtle patterns of fill are not considered. The same can be said for the preoccupation of geomorphologists with channel-forming discharges. Geomorphic effectiveness is not the same thing as ecological effectiveness.
Numerous trials report that instream structures can function as expected, producing scour pools and increases in instream physical diversity. After construction of groin-extensions in an incised, sand-bed stream, Shields et al. (1998) report average maximum scour depths in the following 4 years as 0.84, 0.72, 0.60 and 0.79 m, measured during summer baseflows. Similar increases in pool depths were observed at a series of weir extensions in a relatively stable mixed-bed river (gravelly-sand), where average maximum depths of scour of 1.29, 1.41 and 1.29 m were measured in the three summers following installation. Lintermans (2001) reported the formation and maintenance of scour pools with an average depth of 1.1 m and 0.8 m in the 2 years following construction of groin structures. Shields et al. (2004) noted that scour associated with woody debris structures resulted in baseflows depths approximately twice as deep as pre-construction conditions. Additionally, several flume studies have investigated scour around large wood in streams (Beschta, 1983, Cherry and Beschta, 1989, Marsh, 2001, Wallerstein et al., 2001), and a number of predictive models have been developed and applied (Abbe and Montgomery, 1996, Marsh, 2001, Wallerstein et al., 2001, Buffington et al., 2002). Until recently, predictive models of scour around instream structures have largely been based on empirical relationships between scour depths and hydraulics, sediment characteristics and shape characteristics of the instream structure (e.g. Raudkivi, 1991, Melville, 1997; but see Wallerstein, 2003 for a recent example of a mechanistic model). These empirical models can be applied to instream structures through the use of analogy (i.e. scour around a log is analogous to scour around a submarine pipeline or bridge pier). However, such models tend not to provide accurate predictions in the field (Johnson, 1995, Buffington et al., 2002). Overall, none of the existing work on scour around habitat structures in streams considers the full pattern of live-bed scour and fill over the rise and fall of several freshes, and over seasons. To begin addressing this knowledge gap, we have developed a novel field instrument, a pressure transducer that is able to continuously measure the real-time fluctuations in sand levels below natural logs in the sand-bed of the Snowy River. We then use these results to test whether the scour around the logs occurs at the time of the year when fish would need the scour holes for migration.
A number of different scour mechanisms have been identified around logs (Beschta, 1986, Abbe and Montgomery, 1996, Webb and Erskine, 2003). Beschta (1986) distinguishes three main mechanisms of wood-forced bed scour: flow over the log (a plunge pool), flow under the log, and scour around the base of the log. This last case is analogous to bridge pier scour, and can be described as horseshoe vortex scour. Formative mechanisms are: the submerged jet (Beschta, 1983) where pools are scoured by flowing water deflected downward by the debris piece; and vortex shedding, in which vortices are shed downstream of an instream object producing instantaneous velocities up to 1.6 times greater than the mean velocity (Sumer et al., 1988) (see Fig. 4). The most detailed assessments of scour around wood come from flume studies, where scour pools in uniform sediment are produced by a range of steady flows (Cherry and Beschta, 1989, Marsh, 2001). As for erosion around a bridge-pier, erosion around a log will be a function of the ratio of the average undisturbed stream velocity to the critical velocity for bed entrainment, termed the flow intensity; the depth of flow relative to the diameter of the log, termed the flow shallowness; and the median size of sediment relative to the log diameter, termed the sediment coarseness (Melville and Chiew, 1999).
Beschta (1983) explored the effect of flow rate, cylinder diameter and elevation above the bed on the magnitude and extent of pool scour around fully spanning cylinders in a flume experiment involving steady-state, clear water, and a mobile-bed. The amount and extent of scour resulted from a “complex interaction between cylinder diameter, cylinder position above the bed and water discharge” (P. 8.75). However, with fixed log-size, orientation and bed material; maximum depth of scour increases with increasing discharge (Beschta, 1983, Kuhnle et al., 1999, Marsh, 2001). In our trial, the log diameter and log elevation do not change, therefore, discharge should be the major variable. Depth of scour under live-bed conditions is about ± 10% of the scour under clear-water conditions (Raudkivi, 1991), and the time taken to reach that scour is also shorter under live-bed conditions (Melville and Chiew, 1999). To conclude, then, we would expect the depth of a scour hole to exhibit a strong relationship with discharge, producing a clockwise hysteresis loop: scouring on the rising limb of the hydrograph, and filling on the falling limb, as the rate of transport through the pool declines (Fig. 2). This is the classical pattern of scour for stream beds described by Leopold et al. (1964). Superimposed on this clockwise loop could be fluctuations related to the movement of sediment in pulses at various spatial and temporal scales (Graf, 1988).
Section snippets
Continuous monitoring of bed level
Many real-time monitoring techniques for bed elevation have been developed. Elaborations of scour chains have been used extensively in gravel-bed streams (Laronne et al., 1994). Chains were not appropriate for this study because they are too invasive, and can measure bed fluctuations through only one flow event. More sophisticated devices have been used in studies of bridge scour. Arrays of sensor rods can be attached to the bridge, where a data-logger tracks the rise or fall of some device on
Results
Patterns of scour and fill were monitored continuously for almost 1 year (mid-January to mid-December 2004). The Snowy River is a heavily regulated stream (Erskine et al., 1999). Flows in the Snowy were unusually low in 2004, when compared with 22 years of record (1969–1990) for the Jarrahmond gauge. Flow was less than the mean daily baseflow (MDBF) of 10.6 m3s− 1 for over 60% of the study period. Flow was also less than the Q75 and Q90 flows for more than 39% and 22% of the time, respectively.
Discussion and conclusions
Many stream rehabilitation projects are forced to artificially add flow and structural elements to streams in an effort to improve conditions for organisms. The Snowy River is a good example, where managers intend to introduce environmental flows and structural changes to the stream that improve habitat for target fish species. The pressure transducer method of measurement allows us, for the first time, to continuously monitor real-time fluctuations in a sand-bed stream over many months. The
Acknowledgments
This research was supported by the Cooperative Research Centre for Catchment Hydrology (CRCCH). Funding for the Snowy River Monitoring was provided by the Victorian Department of Sustainability and Environment. An initial review of real-time monitoring methods was undertaken by Nada Dashlooty, as a CRCCH Vacation Studentship Project, under the direction of Dr Tony Ladson. Dan Borg was supported by a joint scholarship from the CRCCH and Land and Water Australia (Riparian Program).
References (52)
- et al.
Bed load transport in an obstruction-formed pool in a forest, gravelbed stream
Geomorphology
(2004) - et al.
Scour chain employment in gravel bed rivers
Catena
(1994) The importance of high-resolution monitoring in erosion and deposition dynamics studies: examples from estuarine and fluvial systems
Geomorphology
(2005)- et al.
Distribution, recruitment and geomorphic significance of large woody debris in an alluvial forest stream: Tonghi Creek, southeastern Australia
Geomorphology
(2003) - et al.
Large woody debris jams, channel hydraulics and habitat formation in large rivers
Regulated Rivers : Research & Management
(1996) Heads of Agreement: The agreed outcomes from the Snowy Water Inquiry
(2000)The effects of large organic debris upon channel morphology: a flume study
Morphological features of small streams: significance and function
Water Resources Bulletin
(1986)- et al.
Characterizing fish-habitat associations in streams as the first step in ecological restoration
Austral Ecology
(2003) - et al.
Continuous monitoring of pool habitats in sand bed streams: A new method using a pressure-transducer