Salmonid Fisheries: Freshwater Habitat Management

By Paul Kemp

Salmonid Fisheries is a landmark publication, concentrating on river management, habitat restoration and rehabilitation, disseminating lessons learnt in relation to the intensively studied salmonids that are applicable to future interventions, not just for salmonid species but for other non-salmonid species, biota and ecosystems. The contents of this book are the product of the Atlantic Salmon Trust’s 40th Anniversary Conference, held in association with the Game and Wildlife Conservation Trust.

Drawing together carefully-edited contributions from many of the world leaders in river restoration from academia, commercial management and government agencies, this important book highlights the need to view river management from the context of the catchment and to adopt an ecosystem-based approach to restoration. The book is broadly divided into two sections which discuss first, the status of current understanding concerning the relationship between lotic habitat management, the response of salmonid fisheries and the theory of river restoration, and secondly, the application of this to habitat management and river restoration.

Salmonid Fisheries is an extremely valuable work of reference for fisheries managers, ecologists, environmental scientists, fish biologists, conservation biologists and geomorphologists. Libraries in all universities and research establishments where biological and earth sciences, and fisheries management are studied or taught should have copies of this book on their shelves.

• Contributions from a wide range of well known experts
• Published in association with the Atlantic Salmon Trust
• Habitat management is crucial for dwindling wild salmon populations
• Of great importance to aquatic ecologists and fisheries managers

• Language: English
• Publisher: Wiley
• Release date: Jun 13, 2011
• ISBN: 9781444347906

Book preview

Preface and Acknowledgements

This book is a product of the Atlantic Salmon Trust’s 40th anniversary conference held in association with the Game and Wildlife Conservation Trust at the University of Southampton in September 2007. This international conference on freshwater habitat management for salmonid fisheries brought together many of the world’s leading academic experts on river restoration and management, representatives of several government agencies and conservation groups, riparian owners and river managers and those with an interest in salmon and trout fishing. The book concentrates on river management and habitat restoration/rehabilitation for salmonid fisheries, not with the intention of ignoring non-salmonid species or indeed other biota and aquatic ecosystems functioning in general, but rather to disseminate lessons learnt in relation to the intensively studied salmonids that are applicable to future interventions. Key issues are highlighted to provoke readers into consideration of uncertainty, contradictions and differences of perspective so that future research and application is based on more holistic thinking. The intended audience includes the academic researcher (e.g. fisheries biologist and geomorphologist), undergraduate and post-graduate students, river managers and those with conservation and/or fisheries interests.

The chapters cover a broad spectrum of topics that span the academic theory of catchment-scale approaches to river restoration, local practical experience of manipulation of riparian and in-channel features, and technical descriptions of process and application. The opening section of the book comprises six chapters that discuss the status of current understanding about the relationship between lotic habitat management and the response of salmonid fisheries, and the theory of river restoration. In the opening chapter, John Armstrong compares predictive biological models that determine the effect of discharge on salmonid habitat quality by evaluating habitat type based on density and frequency of occurrence with those based on energetic consequences. The significant failings of easily applied traditional physical habitat simulation models are highlighted and recommendations for the development of energetic-based models made. Malcolm Newson, in Chapter 2, explores the relationship between the search and development of appropriate tools and metrics, the mantra of scientific guidance and the mania of opportunistic community-driven attempts to protect, restore and manage rivers. In Chapter 3, I discuss the experimental approaches adopted to quantify mechanisms that underpin juvenile salmonid response to enhanced within channel structural complexity and argue for the need to consider fish behaviour when assessing the influence of habitat manipulation. David Sear reviews the current status of in-channel river habitat management and the concepts on which it is based to inform the development of practices (Chapter 4). It is suggested that innovative new modelling techniques and advances in data acquisition can now enable ecological theory to be better integrated with river management to enhance salmonid habitat restoration. In Chapter 5, Phil Roni, George Pess and Sarah Morley highlight the need for well-designed programmes to monitor the results of restoration activities and discuss the application of methods developed based on the North American experience to other regions, including the UK and Europe. Key to all restoration projects are clearly defined objectives, while post-restoration monitoring largely remains inadequate. In Chapter 6, Larry Greenberg and Olle Calles describe the importance of maintaining vertical and longitudinal habitat connectivity for ecosystem functioning in regulated rivers, considering the requirements of multiple life stages of brown trout (Salmo trutta). The authors call for an integrated holistic approach to river restoration in which improving habitat quality and diversity are considered alongside re-establishing connectivity.

Aspects of habitat management and river restoration are discussed in detail in the second section of the book. Management of the riparian zone for wild salmonid populations is considered in Chapters 7 and 8 by Keith Nislow and David Summers. Both chapters illustrate how paradigms in fisheries management can often be based on limited supporting evidence. Johan Höjesjö and co-authors describe an experimental assessment of the influence of the presence of trout on the distribution of juvenile Atlantic salmon (Salmo salar) in Chapter 9. The importance of pools for salmon rearing and migration is considered for Canadian rivers by Bill Hooper in Chapter 10 and methods for successful restoration discussed. In Chapter 11, Keith Nislow and co-authors discuss the ‘planting’ of adult Atlantic salmon carcasses in Scottish rivers, a technique sometimes used to augment river nutrient levels as an approach to habitat management in the Pacific salmon rivers of North America. Bill Riley and Mike Pawson assess the influence of shading and flow on juvenile salmon and trout habitat use, population density and mortality in Chapter 12, while Stuart Clough and co-authors describe the advantages of employing aerial photography to assess salmonid habitat in Chapter 13. A case study is presented by Mark Sidebottom to describe cost-effective construction of fish passage easement in the New Forest (UK) in Chapter 14. This chapter illustrates the conflicts and contradictions associated with placement and maintenance of woody structure within river channels, commonly considered a positive rehabilitation technique, and concerns over reduced habitat connectivity as a result of debris dams that can impede the movement of migratory salmonids. In Chapter 15, Tim Jacklin, Simon Johnson and Edward Twiddy outline the role of, and work undertaken by, the Wild Trout Trust. The final two chapters consider the influence of river temperatures on salmonid populations. Carolyn Knight and co-authors discuss the relationship between river temperature and the annual number of returning adult Atlantic salmon to the River Thames (UK) in Chapter 16. Finally, in Chapter 17, David Solomon considers the evidence for changing river temperature regimes in southern lowland Britain, predictions for the future and what options might be availableforamelioratingtheimpactofclimatechangeonsalmonidpopulations.Of interest, and obvious concern, is the bleak prospect that predicted increases in temperature, as a result of a shifting climate, may push lowland English chalk stream salmon population to the brink, and that options for mitigation appear limited.

This book highlights the need to view river management from the context of the catchment and to adopt an ecosystem-based approach to restoration. River restoration should be, and often is not, based on meeting clearly defined objectives against which success and failure can be evaluated. The adoption of objectivebased strategies for meeting restoration goals, e.g. to enhance juvenile salmonid productivity and emigration of smolts, may be considered more appropriate than the reference-based strategies in which the definition of an appropriate reference condition may prove either difficult to achieve or misguided. Whatever objectives are defined, evaluation of success requires sufficient resources to be allocated to monitoring. The bias towards consideration of salmonid habitat and restoration is evident, and while salmonids also provide the focus of this book, increasing our understanding of the habitat and fish passage requirements of other species should form the focus of future research. Other biases include the application of practice at the scale of the river reach, most often in relation to increasing in-channel physical complexity, and to research conducted in North America involving species of Pacific salmonid. While the transfer of techniques developed in regions different to those in which they are employed may result in similar responses, in other cases they may not, and an outcome less desirable than predicted may be observed. Finally, it is apparent that sound habitat management and restoration require the adoption of an inter disciplinary approach that utilises expertise from a range of disciplines, while the views and concerns of multiple stakeholders should be obtained and considered.

As editor I should like to thank all of the contributors for their hard work in producing this book and the long list of reviewers whose comments, concerns and criticisms provided a valuable contribution to each chapter and the book as a whole. I thank Nigel Balmforth, Kate Nuttall, Tiffany Feist and Shalini Sharma of WileyBlackwell for their valuable support throughout the book’s production. Finally, I thank the members of the conference organising committee, David Solomon, Nick Sotherton, Lynn Field, Dylan Roberts, Seymour Monro, Ivor Llewelyn, Terry Langford and David Sear, who were keen that this book should be delivered as a testament to the success of the meeting.

Paul Kemp

University of Southampton

Chapter 1

Variation in Habitat Quality for Drift-Feeding Atlantic Salmon and Brown Trout in Relation to Local Water Velocity and River Discharge

John Armstrong

Summary

There is a requirement to determine the effect of water discharge on the qualities of rivers and streams for resident drift-feeding salmon and trout. Two main categories of predictive biological model have been widely considered to address this issue, both of which link to an underlying template of variation in structure of hydrology and physical topology across flows. The first approach, typified by physical habitat simulation modelling (PHabSim), ascribes values to each of the local habitat types as functions of the densities and frequency of occurrence of animals that occupy them. This approach has the advantage of being relatively easily applied but has been criticised on the basis that local fish density can be a poor indicator of patch quality and does not easily relate overall habitat quality to meaningful population parameters. The second approach ascribes values to the local habitat types in terms of the food intakes, net energy gains or fitness of animals that occupy them. This concept has found favour in being potentially more robust in structure than the PHabSim approach, but parameterisation of the models cannot be achieved by simple field observations. Here, the application of energy and fitness value models to salmon and trout is explored. Morphological differences between salmon and trout are related to patch quality in terms of energetics through linking optimal food intake models to energy budgets. Using these models, relationships are established between velocity niche width and population density and nutrient status of the stream. The trout velocity niche is narrower than that of salmon and skewed to lower velocities, particularly at low food availability. The importance of understanding community dynamics in predicting responses of fish to variations in discharge is demonstrated. Consideration is given to factors that further affect the values of patches and their availability to salmon and trout. These factors include fish size, food, among-fish variation in metabolism, diel variation in activity patterns, competition, genetic relatedness of neighbours and mortality risk through predator abundance and availability of shelter. Information on the likely capacities of populations of salmonids to respond to temporal change in the spatial distribution of patch qualities is then considered. Constraints and opportunities are compared between application of PHabSim, patch fitness value and empirical models for recommending river discharge criteria for resident drift-feeding salmon and trout.

1.1 Introduction

Human demands for water and changes in land use and climate influence water discharge (flow) regimes in rivers and affect the habitat available for Atlantic salmon, Salmo salar L., and brown trout, Salmo trutta L. Moderation of this impact through planning and implementation of effective water discharge and abstraction regimes should be based on an understanding of the requirements of the fish for water flow. An approach to addressing this issue in part has been to develop an understanding of the relationships between total discharge and local water velocities and other habitat variables throughout the system of interest, and to determine how fish and fish populations respond to changes in discharge and differ among discharge regimes. A main potential advantage of this approach is that by investigating such processes it might be possible to make predictions about the effects of discharge on fish populations on a more robust and wide basis than is possible from using reference to limited empirical observations.

Salmon and trout require different flows at different stages of their life cycles (Armstrong et al. 2003) and plans for seasonal provision of water would need to take this factor into account. The focus here is on the requirement for trout and salmon to obtain food, largely from drifting invertebrates, during their development and growth within rivers and streams. The overall aims are to outline the relevant factors, through reference to literature and derivation where appropriate, and to seek a structure for moving forward the development of models to predict optimum discharge regimes.

The first part of this chapter provides a context for interpreting local responses of salmon and trout to water velocities by considering scale of impacting factors and variation in population dynamics across time and space. Then two approaches to quantifying habitat quality are considered. First, assessment of patch (area of stream) quality on the basis of local population densities is evaluated (PHabSim). Second, the approach of assessing patch quality on the basis of the fitness value to fish occupying it is discussed. This latter approach is expanded by development of an energetics model to estimate patch quality for growth as a function of the station holding mode of occupants, specifically as a comparison of typical trout and salmon behaviour types. This model is then used to consider the potential for predicting change in patch quality across discharges, with emphasis on the importance of linkages to the wider community of animals, particularly invertebrate prey.

The chapter then considers some of the biological parameters of trout, salmon and their environments that can be expected to influence local patch quality independent of water velocity. Consideration is also given to evaluation of mortality risk in combination with growth as a component of fitness and to the abilities of salmon and trout to move in response to changes in patch qualities. Finally, some consideration is given to routes for development of practical models for predicting the consequences of change in flow rates on populations of salmon and trout.

1.1.1 Habitat requirements of salmon and trout

The numbers, growth and size distributions of Atlantic salmon, S. salar L., and brown trout, S. trutta L., are strongly influenced by their habitats. Their basic needs for growth, survival and reproduction in fresh water include food, shelter, oxygen, clean water and suitables pawning conditions (reviewed in detail by Armstrong etal. 2003). Food, primarily in the form of invertebrates but also including fish, provides energy to fuel metabolism and growth. Shelter may be afforded by factors such as low light, physical obstructions and camouflage that make the fish less accessible to predators and physical trauma, for example from suspended solids and high water flows. Oxygen is required to enable a physiological scope for maintenance, feeding, processing food and evading predators. Water must be sufficiently clean that there is no significant impedance to the physiological processes that enable growth, minimise tissue maintenance costs and facilitate other key functions, such as imprinting on and detecting odours (Sutterlin & Gray 1973). Composition of the substratum determines local water velocities and shelter. There is a vast complexity of interacting biotic and abiotic factors that influence accessibility to these basic requirements (Armstrong et al. 2003), one of which is local water velocity.

1.1.2 Integrating across scales of influence

The habitat at any particular location in a stream is influenced by processes that occur across a broad range of spatial and temporal scales (Frissell et al. 1986). For example, at a given point in time at the local scale of perhaps 10 cm, a fish may be affected by the velocity of the flow from which it must extract food in competition with others. Yet the amount of such food and the rate of the flow may be influenced by the whole-catchment land use and underlying continental geology, which may be under the influence of geomorphic processes that have occurred across millions of years and human influence over thousands of years (Armstrong et al. 1998b).

Fine-scale processes are also crucially important in understanding the effects of water flow on fish. For example, early investigations of flow limitations on trout and salmon focused on the velocities needed to displace fish completely and were conducted on fine substratum in almost laminar flow in which there were no refuges (Ottaway & Clarke 1981). Such an approach is clearly not readily transferable to rough river beds in which velocity varies dramatically over small space scales providing refuge areas.

1.1.3 Bottlenecks as functions of time and space

The effect of habitat on fish populations can depend critically on the relationship between local densities and sizes of fish and the numbers of fish of those sizes that the habitat potentially could support. This relationship can vary across time, in terms of fish development stage (Armstrong & Nislow 2006), and across space as a function of distance from spawning areas independent of other habitat features (Armstrong 2005). At critical periods, such as after emergence of fry in some populations (Elliott 1989) and after attaining a certain size in others (Rincón & Lobón-Cerviá 2002), fish may saturate available habitat required for their specific size. The population may then continue to saturate the habitat in a self-thinning process, such that individuals can grow only if numbers of fish decrease due to increased individual resource requirements with size (Grant & Kramer 1990). Alternatively, the population may fall below the carrying capacity, but individual growth rates may nevertheless relate inversely to density in some cases (Grant & Imre 2006). The cohort strength may be maximised at intermediate density and size due to size-dependent survival at some stage, for example over the winter period (Crisp 1995, Armstrong 2005).

1.1.4 Local variation in population densities and dynamics

A basic practical significance of these processes is that unless spawning is abundant and homogeneous, there is likely to be substantial spatial variation in numbers, weights, dynamics and even year class number of salmon and trout regardless of local habitat quality (Armstrong 2005). A test of this prediction by field experiment in replicated streams at two levels (clumped and dispersed) confirmed empirically the importance of even local-scale patchiness of egg distribution on population dynamics (Einum et al. 2008). This spatial variation has three important consequences for investigation of habitat for salmonids. First, it is likely that models relating suites of habitat variables to local population density, weight and biomass will result in substantial residual unexplained variation and transfer poorly between different systems (with their different spatial arrangements of habitat types for the different life stages). This expectation is realised in empirical analyses (Fausch et al. 1988). Second, it is essential that in field trials to test for any local effect of a particular habitat variable, there is a substantial number of independent replicates to overcome the expected high degree of background variation in local densities, sizes, growth rates and age class strengths. Third, models to predict the consequences of varying habitat features, such as discharge, should be designed to operate across a broad range of population levels in relation to carrying capacity.

1.2 Defining local habitat quality

1.2.1 Defining patch quality in terms of local fish density

Streams can be considered as assemblages of patches of habitat. The scale at which patches can be assigned can vary and has a large influence on measurement and understanding of physical and biological processes (Folt et al. 1998). A patch may include only the space between two adjacent boulders, for example, and constitute less than the scale of a fish territory, or it may include an area of similar habitat such as a pool and may accommodate several fish territories. For measuring density of fish it may be appropriate in small patches to record likelihood of a fish being present whereas in larger patches the density of fish in occupancy may be an appropriate measure. The density experienced by a fish depends on its distance from nearest neighbours.

An influential approach for predicting consequences of variation in discharge has been through the application of Physical Habitat Simulation (PHabSim) (Bovee 1986) and similar models (e.g. Capra et al. 1995). PHabSim comprises physical and biological components. The physical model includes data from spatially referenced measurements of instream parameters, generally including local flow velocity and water depth, and can predict how these vary across discharges. Overlain on the network of physical habitat structure are estimates of the quality of the local habitat for fishes, from so-called preference curves. The preference curves may be derived from direct measurement in the study site at one or more discharges, imported from studies of other sites, or evolved from peoples’ opinions (‘expert opinion’), which is usually of unknown bias.

The model includes three major assumptions. First, density of fish in a habitat type is a true reflection of the value of that habitat (preference). Second, preference for each particular habitat type is constant across discharges. Third, that fish freely move to best available habitats when discharge changes. A fourth assumption that is implicit in application of the models is that the output (weighted usable area) has some meaning in terms of the fish population, for example the biomass, growth and densities that can be supported by the overall habitat.

Regarding the first assumption, Atlantic salmon and brown trout compete aggressively for high quality feeding patches within streams and tend to form dominance hierarchies in which the top-ranking fish can exclude others from preferred habitat (Sloman & Armstrong 2002). At low densities there may be a direct relationship between patch quality and density (Girard et al. 2004). However, as numbers of fish increase, densities become highest in more marginal areas, due, for example, to dominant fish holding the best stations and displacing other individuals (Greenberg 1994, Bult et al. 1999, Holm 2001, Blanchet et al. 2006, Stradmeyer et al. 2008, see also Baker & Coon 1997).

Regarding the second assumption, a direct experiment in which positions and local habitat quality of Atlantic salmon parr were measured across a range of discharges provided evidence of the large error that can occur in application of PHabSim to predict changes in overall habitat quality in terms of ‘weighted usable area’ (Holm et al. 2001). It was clear in this experiment that preference curves were not independent of discharge even though density remained constant and a closely controlled environment allowed precise measures of habitat and fish positions.

The third assumption that fish can move in response to change in distribution of habitat patches can be expected to depend on how those patches are juxtaposed in space across discharges, which is not included as part of the PHabSim model. Capacity of salmon to respond to change in habitat is discussed in Section 1.5. The issue of how the output of PHabSim, the weighte duseable area, relates to population processes is unclear and has been a concern for many researchers (Rosenfeld 2003).

The simple function used in PHabSim to predict habitat suitability considers fish to be constant objects and takes no account of variations in their behaviour, physiology and ecology under different physical and biological environments. This is exemplified with reference to observations of Stradmeyer et al. (2008), who monitored behaviour and feeding of salmon and trout in response to abstraction under controlled mesocosm conditions. Decrease in discharge resulted in a change from widespread distribution throughout pool and riffle to almost exclusive use of marginal pool areas and a switch to cryptic inactive behaviour in subordinate fish, which employed a ‘sneaky’ feeding mode under these conditions (Höjesjö et al. 2005). By contrast, dominant fish increased their aggressive activity markedly (Figure 1.1). All fish used pools during low water and observations of how this habitat was used at normal discharge provided no insight into its capacity to support fish at low discharge. In this case, all fish survived during the two-day abstraction event but there would probably have been consequences for their energy balance, growth and mortality risk, for example through attacks from the dominant fish. The impacts of such abstractions would no doubt vary with their durations and intensity, and in some cases may have negligible effects on growth (Flodmark et al. 2006).

1.2.2 Defining patch quality in terms of fitness for occupants

An alternative approach to assessment of the effect of discharge on fish populations is to use the physical habitat component of PHabSim, or a similar model, and ascribe values to each habitat patch independent of fish distribution. Quality of a patch would be expected to be related to fitness of fish that occupy it and includes potential for growth and survival (Mangel & Clark 1986, 1988, Clark & Mangel 2000, Railsback & Harvey 2003). When fish are not feeding, survival potential should dominate and in these circumstances salmonids are known to use secure secluded locations such as below-gravel shelters (e.g. Gries & Juanes 1998). Favoured feeding locations are likely to enable high rates of net energy gain, both to increase growth potential and to minimise time exposed from secure shelters. There is potential for a balance between reducing immediate risk in a foraging location and reducing the amount of time exposed out of shelter and thus overall risk of the foraging bout.

Figure 1.1 Aggression as a function of dominance rank in (a) groups of eight salmon (homogeneous, n = 9) and (b) groups of four salmon and four trout (mixed, n = 8) (see Stradmeyer et al. (2008) for details). Each group of wild-caught fish was held in a section of a glass-sided stream, landscaped to provide pool and riffle habitats, and was observed during normal flow, during a two-day period of reduced discharge (dewatering) and following rewatering. Ranks of the fish, as shown by the grey scale, were determined on the basis of outcomes of aggressive encounters during the period of establishment in the stream. Dewatering caused a significant increase in aggression from the dominant fish (which was always a trout in the mixed group) as it monopolised central space in the pool. Rewatering resulted in a reduction in aggression.

c01_image001.jpg

Considering the growth component of fitness alone, the rate of net energy gain at a particular velocity can be estimated from energy budgets. Bioenergetics models have been applied to explain distributions of a number of species of fishes including Atlantic salmon and brown trout (Fausch 1984, Guensch et al. 2001, Booker et al. 2004) and to predict energy balance in trout (Rincón & Lobón-Cerviá 1993, Hayes et al. 2000). Such models can vary in complexity depending at what level predictions from first principles (Hughes & Dill 1990) are substituted by empirical data (e.g. Nislow et al. 1999). It has been argued that application of bioenergetics estimates of local patch qualities is likely to be promising in relating variations with discharge to population processes, but that further research is required to achieve this goal (Anderson et al. 2006).

An important consideration is determination of how patch quality varies depending on the species of occupant. This issue is explored using the growth component of fitness in relation to salmon and trout in the following section and provides an example of the application of energy budgets.

1.3 Variation in patch quality as a function of typical salmon and trout station-holding modes

An intriguing difference between salmon and trout is that trout must swim continuously to hold station against the flow in the water column, whereas salmon can use their large pectoral fins (Pakkasmaa & Piironen 2001) as hydrofoils to hold position on the substratum (Arnold et al. 1991). It appears from recent study of Pterygoplichthys spp. that this hydrofoil mode requires no significant expenditure of energy for fish to hold station (Blake et al. 2007). The question arises as to how this variation in swimming options can be expected to influence local habitat quality for the species as a function of water velocity.

Ignoring energetics and mortality risk, a patch value has been determined in terms of the rate of food intake of fish that occupy it. Food intake (I) from feeding on drifting invertebrates varies with local water velocity (S) because within a discharge, food availability (F) increases and capture efficiency (C) decreases with velocity. These functions may be approximately linear with the result that in Atlantic salmon the relationship between food intake and velocity can follow a symmetrical parabolic curve (Nislow et al. 1999):

(1.1) c01_image001.jpg

(1.2) c01_image001.jpg

(1.3) c01_image001.jpg

where a > 0 and c < 0, as depicted in Figure 1.2a.

Capture efficiency is likely to vary with time and place due to factors such as water clarity, time of day (Metcalfe et al. 1997) and turbulence; local food availability will vary with the overall rate of invertebrates entering the drift, which is likely to be a function of the nutrient richness of the stream, the rate at which food is removed by other foragers (Elliott 2002, Nilsson et al. 2004) or settles out, and also with time of day (e.g. Sagar & Glova 1992, Giroux et al. 2000). Therefore, for purposes of comparing station-holding modes here, arbitrary values are given to a, b, c and d (0.01, 0.1, –0.02 and 1.2 respectively) in Equations (1.1)–(1.3) to produce the type of curve observed empirically for Atlantic salmon by measurements of Nislow et al. (1999). These authors, who incorporated data from Hill & Grossman (1993), concluded that for drift foraging salmon, optimal food intake is at intermediate values of water velocity and because of the extended parabolic shape of the I–S curve, high intake (e.g. within 90% of maximum) was experienced across a broad range of velocities. This effect is shown at overall high and low total availability of food (henceforth referred to as high and low richness to distinguish from variation in food availability across velocities), differing by a factor of two, in Figure 1.2a.

Determination of patch quality in terms of food intake alone takes no account of variation with velocity in the costs of obtaining that food. Such costs need to be considered to calculate the value of a patch in terms of energy balance. The net energy change (ΔE) in a foraging location depends on gain in terms of food intake (I) and losses in terms of faecal and ureic waste (0.27I; Brett & Groves 1979) and energy expended for resting metabolism (Mr), to hold station (Ms), and to handle, digest and assimilate food (apparent specific dynamic action, ASDA; 0.14I; Brett & Groves, 1979):

(1.4) c01_image001.jpg

Further costs will be incurred (e), for example in activity for social interactions, to catch food and to evade predators. For now, e is held constant and I is varied with S according to Equation (1.3).

Ms is taken to be zero for salmon holding station on the substratum. By contrast, trout must swim actively in the water column and Ms is then an exponential function of S (Brett 1964), assuming laminar flow:

(1.5) c01_image001.jpg

Values for Mr, f and g are taken from Brett’s (1964) estimates for an 18-cm fish at 15°C, and estimates of faecal loss, ureic loss and specific dynamic action are those for a ‘carnivorous fish’ (Brett & Groves 1979). These general values are appropriate in view of the objective to illustrate the patterns of variation in ΔE with S as functions of station-holding mode rather than to simulate patch quality for a specific population in a particular stream, when appropriate adjustments can be made for fish size and temperature (Rincón & Lobón-Cerviá 1993). ΔE is somatic growth and reproductive investment when positive and drains on energy stores when negative.

Figure 1.2 Comparison of salmon-type (no swimming costs) and trout-type station holding on growth as a function of water velocity. Relationships are modelled between food intake and growth (J day−¹) and current speed (cm s−¹). A swimming cost is included for the trout-type model but otherwise the models use the same values for energy budget components and same arbitrary values for food availability and capture efficiency. (a) Variation in food intake assuming a linear increase in food supply and linear decrease in foraging efficiency with increase in current speed (Nislow et al. 1999), where a = 0.01, b = 0.1, c = 0.02 and d = 1.2. The dashed and solid lines show two levels of food richness. (b) Growth of salmon holding station at no energy cost. Optimum velocity for growth coincides with that for food intake and is independent of food richness. (c) Growth of trout and salmon swimming in the water column to hold station. Optimum velocity for maximising growth is lower than that for maximising food intake and shifts to lower velocities with decline in food availability.

c01_image001.jpg

Several interesting basic patterns emerge in deriving ΔE as functions of S using these relationships. First, in salmon-type station holding where Ms = 0, the shape of the curve relating ΔE to S is similar to that relating I and S (Figure 1.2b), whereas in trout the curve is asymmetrical, displaced downwards and scewed to the left (Figure 1.2c). This effect results in a lower optimum velocity in terms of maximising ΔE for trout than salmon and a much narrower niche width, in terms of velocities supporting growth, mainly due to inability to prosper at high velocities. Decline in food richness reduces the velocity niche breadth in both station-holding modes due to energy balance becoming negative at both high and low velocities (Figure 1.3); the optimal velocity (peak of the curve) shifts to lower velocities in trout- but not salmon-mode (Figure 1.3). Growth of trout per unit energy ingested is less than that of salmon at any given water velocity due to Ms.

Figure 1.3 Change in patch net energy quality across a 50% decrease in discharge and local water velocity from an initial value of 40 cm s−¹. If total invertebrate drift remains constant, then patch quality increases (square). If invertebrate drift halves with discharge, then the velocity-growth relationship shifts to a lower level and the patch quality decreases (triangle).

c01_image001.jpg

1.4 Predicting change in patch energy value with discharge

The relationships derived in the previous section are useful for exploring how possible it is to predict local values for ΔE at one discharge from measurements at another discharge. The key issue is whether the quantity and type of invertebrates entering the drift remain constant or vary with discharge. If drift of suitable food items is independent of discharge, then ΔE could, in principle, be predicted by following the ΔE–S relationship from within a discharge at a given overall food richness. For example, with reference to Figure 1.3, if a 50% reduction in discharge resulted in a 50% reduction in local current speed from a starting value of 40 cm s−¹, then increase in ΔE would result. However, if drift reduces with discharge, for example due to fewer invertebrates being displaced into the water column (Gibbins et al. 2007), then prediction would require a shift to the ΔE–S curve at a lower overall food richness. In the example here (Figure 1.4), if drift is exactly proportional to discharge, then prediction across a 50% decline in discharge and local velocity would require a shift from the high to the low food richness curves (which differ by a factor of two) and would result in a reduction in ΔE.

This example demonstrates that although it is possible in principle to estimate potential ΔE values and thus quantify habitat quality (with appropriate assumptions and simplifications), the predictive value of such estimates is only as good as our general understanding of variation with discharge at a community level, in this case through predicting food supply.

Figure 1.4 Components of the relationships among characteristics of habitat and fish populations that determine the response to change in water discharge. The initial local patch qualities, in terms of fitness of fish that occupy them, are determined by basic environmental variables and juxtaposition in relation to other patches. The initial fish population may be determined by initial habitat conditions depending on its relationship to carrying capacity. Change in discharge induces a change in basic environmental variables that determine local patch qualities and so the initial patches change in character to emergent patches. The capacity of the fish population to respond to the change depends on a number of factors that may constrain its distribution. Overall emergent fish population parameters depend on summation of patch qualities across several dimensions. For comparison of stable populations at different discharges (rather than response to a change in discharge) only the emergent environment, patch qualities and fish populations are required for each population. These processes are discussed in detail in the main text.

c01_image001.jpg

1.5 Other considerations

Simplified models are very useful for identifying key characteristics of complex systems, such as those influencing stream-dwelling salmonids. Furthermore, they enable testing of the basic model components in controlled environments to seek assurance that the structure is appropriate. However, in developing such conceptual models for making predictions for management application, it is important to reconsider the complexities inherent in natural systems. Some of the factors that affect patch selection, energetics and the relationship between water velocity and food intake are discussed in this section.

1.5.1 The other fitness component: mortality risk

It is difficult to assess mortality risk or the fish’s perception of it with any degree of precision. However, potential vulnerability to predators and availability of shelter affect the behaviour of fish in ways that suggests a strong contribution of mortality risk to local habitat quality. Presence of predators can cause fish to move to more marginal and less profitable habitats