C.N. Gibbins, H.J. Moir, N. Hakanpaa and H.M. Jackson

Department of Geography and Environment, University of Aberdeen,

Aberdeen, Scotland. AB24 3UF.

Tel. +44 (0) 1224 272328, Facsimile +44 (0) 1224 272331,

E-mail: c.gibbins@abdn.ac.uk

Abstract: Juvenile salmonid fish defend small territories and feed primarily on invertebrates drifting downstream in the water column. This study compares patterns of invertebrate drift in a relatively pristine stream and a stream modified to aid agricultural drainage. Invertebrate drift rates in the modified stream were consistently greater than those in the natural stream, despite the numbers of invertebrates on the bed of each stream not being significantly different. This suggests that invertebrates in the modified stream are more prone to drift. As expected, drift rates varied over time, largely in parallel with discharge. However, in both streams, relative drift rates at individual nets were not constant: locations switched from yielding a large biomass of drifting invertebrates to yielding little biomass and vice versa. The implication of this is that relative territory ‘quality’ is fluid and, to procure maximum food resources, fish may have to defend different locations at different times.

 

Keywords: Invertebrate drift, salmonid fish, stream modification, Scotland

Salmonid fish (the Salmonidae) are an important resource in many countries. Flow regulation and channel modification affect the ability of river systems to support salmonid populations (Parrish et al., 1998) and there are concerns that this might have economic implications for local areas. Each salmonid life stage has specific hydraulic habitat requirements (Gibson, 1993; Bardonnet and Bagliniére, 2000) and the direct impacts of channel modification on the suitability of rivers and streams for fish has been documented widely (Carnie, 1994; Nakamura and Tsukisaka, 1994). However, the indirect impacts of channel modification on salmonids, via, for example, changes to their food resource, have rarely been assessed.

Juvenile salmonids feed primarily on invertebrates drifting downstream in the water column (Elliott, 1970, 1973). Individuals defend territories that have suitable hydraulic conditions and which provide sufficient drifting invertebrates to yield a net energy gain (energetic benefits derived from food intake minus the costs of swimming and maintaining position; Fausch, 1984, 1993). Patterns of invertebrate drift are known to be dependent on local hydraulics, particularly channel velocities (Brittain and Eikeland, 1988). As channel modification such as straightening, widening and overdeepening can affect instream hydraulics, it is also likely to affect patterns of invertebrate drift. This, in turn, may affect the relative suitability of different locations for juvenile fish. This paper reports the preliminary findings of a study comparing patterns of invertebrate drift in two streams in North-east Scotland, one natural, one canalised (Fig. 1). The paper addresses three questions: (i) How does the total drift flux compare in the two streams? (ii) How do patterns of drift within and between sites vary in the two streams? and (iii) What are the potential implications of (i) & (ii) for salmonid territorial behaviour?

Benthic and drifting invertebrates were sampled from the Girnock Burn and the Newmill Burn. The Girnock is an important Atlantic salmon, Salmo salar, spawning stream, although stocks have declined in recent years (Moir et al., 1998; Moir, 1999). The stream channel in the Girnock has not been modified, other than locally as a result of small bridges and fords. The Newmill Burn is highly canalised. It is surrounded by agricultural land and has been extensively straightened and deepened to aid agricultural drainage. The Newmill Burn supports high densities of sea trout, Salmo trutta, and smaller numbers of Atlantic salmon.

Two sites were studied in each stream: the Gauge site and Hampshire’s Bridge in the Girnock and the Gauge site and Claypots in Newmill Burn (Fig. 1). Newmill sites averaged 1.5 m wide, Girnock sites 8-12 m wide. Invertebrate samples were collected on 11 occasions between February and June 2000. Five drift nets were located at each site over a channel length of circa15 m. These consisted of a base-plate fitted permanently into the stream bed (Fig. 2). Onto these, net heads were fitted and removed as required. Net positions were determined from preliminary surveys of site hydraulics. At each site, single nets were located at points with the maximum and minimum water velocity recorded during the preliminary surveys; three more nets were then positioned to sample drift in locations with a range of intermediate velocities. Vertical adjustment of the net heads allowed them to be positioned to sample the part of the water column where velocities approximated mean column velocity (i.e. at 0.6 of the water depth). Nets can affect hydraulic conditions in their immediate vicinity while the drift delivery rate to a net will be influenced by a net located immediately upstream. Net positions were also determined by the need to avoid these effects. Consequently nets were not placed directly in line (with respect to dominant flow path) while the lateral distance between adjacent nets was always greater than 1m.

The larger a drift net, the greater the likelihood that water velocity will vary across its width. As a key part of the work was to quantify the relationship between water velocity and drift, it was important that velocity measured at a particular net at a particular time was representative of the velocity across the width of the net as a whole. The preliminary trials showed that velocity was relatively homogenous across 100 mm wide patches of channel in the two streams. Trials also indicated that nets with an aperture of 100 mm X 100 mm collected enough invertebrates per unit time to make statistical comparison between samples possible. The small size of the nets also minimised hydraulic effects and limited the extent to which nets affected drift rates into downstream ones. Net mesh size was 0.5 mm. This allowed fine material (silts and organic debris) to pass through, so avoiding clogging which can reduce capture efficiency and cause backflushing of invertebrates. However, it also meant that invertebrates less than 0.5 mm (i.e. microinvertebrates) were lost.

For consistency, all drift sampling commenced between 0900 and 1000hrs. Nets were left in position for 2-4 hours, always during periods of stable flow. Mean column velocity was measured at each net on each sampling occasion using an electromagnetic velocity probe. Discharge was monitored automatically at one site in each catchment (gauge sites; Fig.1). Discharge was estimated for the ungauged sites using proportional catchment area relative to respective stream gauge sites. At the end of each sampling period, net heads were removed from their locator poles and transported to the laboratory. Net contents were then washed into white plastic trays and invertebrates picked out and preserved in formaldehyde.

Five replicate surber samples of the benthos were collected from random locations at each site on each drift sampling occasion. The surber sampler had a 300 mm X 300 mm base frame and net mesh of 0.5 mm. Bed material within the base area was disturbed to a depth of approximately 100mm using a trowel. Net contents were emptied into bottles in the field and, on arrival back at the laboratory, invertebrates were preserved in formaldehyde.

In total, 220 drift and 220 benthic samples were collected over the 11 sampling occasions. Data presented here are a subsample consisting of 140 drift and 140 benthic samples from 7 occasions. The number of invertebrates present in each sample was counted. Invertebrate samples were then oven dried at 60^oC for 12 hours and dry weight of invertebrates measured using precision scales (accurate to 0.00001g) to give an estimate of biomass. All drift values (number and biomass of invertebrates) were converted to rates per hour; benthic densities were converted to number and biomass per square metre (m²).

There were no consistent differences in benthic invertebrate densities between the two streams (Two-way ANOVA on square root transformed data: P=0.2003, df_1,85, F=1.6, F crit.=3.9 [Fig.3a]). However, the drift flux in the Newmill Burn was consistently greater than that in the Girnock (Fig. 3b). Per unit velocity, there were more invertebrates drifting through the Newmill sites than the Girnock sites. This was true whether drift was quantified according to abundance or biomass of invertebrates. ANOVA results for benthic data indicate that differences in drift rates between the two streams cannot be explained by differences in benthic densities. The relationship between drift rate and water velocity differed between the two streams. In the Girnock Burn, there was a significant linear relationship between velocity and drift (r² 0.553, P<0.0001, Fig. 4) whereas there was no significant relationship in the Newmill Burn (P>0.05, Fig. 5). This suggests that drift in the unmodified Girnock varies predictably as a function of velocity; in contrast, in the canalised Newmill Burn high drift rates are not necessarily associated with high velocities.

At each site drift rates varied markedly over time (Fig. 6), broadly in parallel with discharge. Temporal variation in the Girnock was less marked than in the Newmill Burn. There was also noticeable intra-site variation in drift rates, with variation often greatest at high discharges. Relative drift rates through individual nets often differed over time. For example, net 5 at Newmill Gauge had by far the greatest drift rate on 23/03 but had the second lowest on 30/3. Similarly, net 2 at Girnock Hampshire’s Bridge had the highest drift rate on 31/03 whereas on other dates it had average values. Conversely, some nets ranked consistently: net 5 at Girnock Gauge always had consistently low drift rates.

A number of factors influence the stream positions used by individual salmonid fish. Position choice varies with season (Heggenes, 1994), time of day (Fraser et al., 1995; Kadri et al., 1997) and life-history strategy (Valdimarsson and Metcalfe, 1999). In addition, individual fish use different hydraulic habitats (water depths and velocities) as they mature and grow larger (Gibson, 1993). When feeding, fish tend to hold station in an area of relatively low water velocity (where they use little energy maintaining position) and occasionally make darting forays into areas of higher water velocity (Stradmeyer and Thorpe, 1987) that deliver larger numbers of drifting invertebrates per unit time (Fausch, 1984). Such ‘good quality’ locations are defended by fish; less profitable locations are occupied by smaller and less aggressive individuals (Giannico and Healey, 1999). Overall, the distribution of fish within a site is determined by the relative suitability of different locations and the dominance hierarchies that determine which fish procure which territories. By altering instream hydraulics, channel modification could affect the hydraulic suitability of, and the food resources available in, specific territories.

Drift rates in the Newmill Burn, a canalised stream, were consistently greater than in the Girnock Burn. One possible explanation for this is simply that there were more benthic invertebrates (the source of drift) in the Newmill Burn than in the Girnock. However, there were no significant differences in the numbers or biomass of invertebrates on the bed of the two streams. This suggests that a greater proportion of the benthic invertebrates in the Newmill Burn are prone to drift. There are two components to invertebrate drift: (i) voluntary or active drift and (ii) involuntary or passive drift (reviewed by Brittain and Eikeland, 1988).

Voluntary drift is thought to result from animals redistributing either to seek out sources of food or to avoid predators. Involuntary drift results from accidental dislodgement, often as a consequence of increases in discharge and channel velocities. Further research is needed to assess the cause of the increased drift rates in the Newmill Burn. It may be that canalisation increases velocities more rapidly during the rising limb of the hydrograph, such that a larger proportion of benthic invertebrates are dislodged. As part of the canalisation works, large (boulder-sized) items of bed material were removed from the burn. In natural channels these roughness elements help create hydraulic heterogeneity and the flow refuge areas that invertebrates use to avoid extreme hydraulic forces (Winterbottom et al., 1997). The relatively homogeneous bed of the Newmill Burn (Moir, 1999) may reduce the availability of flow refuge areas. Again, this may contribute to increased rates of dislodgement and hence drift.

As would be expected, drift rates at each of the four sites varied over time. There was also considerable variation in drift rates between individual streambed locations (i.e. between nets) within sites. Experimental studies have shown that fish select locations preferentially that provide greater food delivery rates (Fausch, 1984, 1993; Giannico and Healey, 1999). However, data for both the Newmill and Girnock Burns indicate that the relative quality of streambed locations (indexed by drift rate) varies over time, most likely as a function of discharge and its influence on local hydraulics. Also, for sites in both streams, the disparity between ‘good’ and ‘poor’ locations differed over time; on some dates/at some discharges, drift rates were similar at all net locations whereas at other times there were marked disparities. In the Newmill Burn, there was no predictable relationship between velocity and drift: locations that had a high velocity did not necessarily provide large numbers of drifting invertebrates. Exploration of the possible reasons for this requires further analysis using drift-velocity data for individual nets which, in turn, requires further sample processing.

Although preliminary, data for the Newmill and Girnock Burns raise three points of relevance to salmonid fish. First, while channel modification in the Newmill may have impacted hydraulic habitat heterogeneity, the stream’s gross invertebrate biomass appears to be sufficient to support salmonid populations; indeed, salmonids occur at greater densities in the Newmill than in the Girnock (Youngson, unpublished data). Second, drift rates, the primary source of salmonid food, were much greater in the Newmill Burn than the Girnock. More specifically, each Newmill net yielded greater drift biomass (and therefore energy) per unit velocity than each Girnock net. Consequently, even the relatively ‘food poor’ locations that theory and experimental studies suggest will be occupied by less dominant fish, may yield a net energy gain. A key question for further work is whether energetic benefits gained from increased food resources in the Newmill are offset by the higher energy requirements for swimming and holding station in this canalised stream. Finally, data for individual nets suggest that relative territory quality is fluid – locations can switch from providing relatively large to relatively little drift biomass. The implication of this is that, in order to maximise energy consumption, fish may have to occupy different streambed locations at different times in response to variations in drift rate.

The work was funded by a small grant from University of Aberdeen. Authors are grateful to Ian Malcolm and Jens Petry for flow gauging in the Newmill Burn and to Alan Youngson for help with collection of drift samples and arranging access to field sites.

Bardonnet, A. and J-L. Bagliniere (2000). Freshwater habitat of Atlantic salmon. Can. J. Fish. Aquat. Sci., 57, 497-506.

Brittain, J.E and T-J. Eikeland (1988). Invertebrate drift: A review. Hydrobiologia, 166, 77-93.

Carnie, C.G. (1994). Salmon in the river basin. In C. Kirby and W.R. White (eds) Integrated river basin Development. Wiley. 389-397.

Elliott, J.M. (1970). Diel changes in invertebrate drift and the food of trout. J. Fish Biol. 2, 161-165.

Elliott, J.M. (1973). The food of brown and rainbow trout in relation to the abundance of drifting invertebrates in a mountain stream. Oceologia 12, 329-347.

Fausch, K.D. (1984). Profitable stream positions for salmonids: relating specific growth rate to net energy gain. Can. J. Zool., 62, 441-451.

Fausch, K.D. (1993). Experimental analysis of microhabitat selection by juvenile steelhead and coho salmon in a British Columbia stream. Can. J. Fish. Aquat. Sci. 50, 1198-1207.

Fraser , N.H.C., J Heggenes, N.B. Metcalfe and J.E. Thorpe (1995). Low summer temperatures cause juvenile Atlantic salmon to become nocturnal. Can. J. Zool., 73, 446-451.

Giannico, G.R. and Healey, M.C. (1999). Ideal free distribution theory as a tool to examine juvenile coho salmon habitat choice under different conditions of food abundance and cover. Can. J. Fish. Aquat. Sci. 56, 2362-2373.

Gibson, R.J. (1993). The Atlantic salmon in freshwater: spawning, rearing and reproduction. Reviews in Fish Biology and Fisheries, 3, 39-73.

Heggenes, J. (1994). Physical habitat selection by brown trout and yound Atlantic salmon in spatially and temporally heterogeneous streams: implications for hydraulic modelling. Proc. First International Symposium on habitat hydraulics. Norwegian Institute of Technology, Trondheim, Norway, June 1994. 12-30.

Kadri, S., N.B. Metcalfe, F.A. Huntingford amd J.E. Thorpe. (1997). Daily feeding rhythms in Atlantic salmon II: size-related variation infeeding patterns of post-smolts unders constant environmental conditions. J.Fish Biol., 50, 273-279.

Moir, H. J. (1999). ‘The characterisation of Atlantic salmon spawning habitat in the River Dee and River Don catchments, north-east Scotland.’ PhD thesis. Department of Geography, University of Aberdeen.

Moir. H. J., Soulsby, C., and Youngson, A. (1998). ‘Hydraulic and sedimentary characteristics of habitat utilized by Atlantic salmon for spawning in the Girnock Burn, Scotland.’ Fisheries Management and Ecology, 5: 241-254.

Nakamura, S and M. Tsukisaka (1994). Field investigations on artificial habitat in a Japanese canalised stream. Proc. First International Symposium on habitat hydraulics. Norwegian Institute of Technology, Trondheim, Norway, June 1994. 305-315.

Parrish, D,L. R.J. Behnke, S.R. Gephard, A.D. McCormick ans G.H. Reeves. (1998). Why aren’t there more Atlantic salmon Salmo salar? Can. J. Fish. Aquat. Sci. 55 (Suppl. 1): 281-287.

Stradmeyer, L. and J.E. Thorpe (1987). Feeding behaviour of wild Atlantic salmon, Salmo salar, parr in mid- to late-summer in a Scottish river. Aquaculture and Fisheries Management, 18, 33-49.

Valdimarsson, S.K and Metcalfe, N.B. (1999). Effect of time of day, time of year and life history strategy on time budgeting in juvenile Atlantic salmon, Salmo salar. Can. J. Fish. Aquat. Sci. 56, 2397-2403.

Winterbottom, J.H., S.E. Orton, A.G. Hildrew and J. Lancaster (1997). Field experiments on flow refugia in streams. Freshw. Biol. 37, 569-580.