Fish Eating Birds and Salmonids in Scotland

1. Introduction

In Scotland, concern has long been expressed by proprietors of fisheries about the predation of Atlantic salmon and brown trout by sawbill ducks (goosander Mergus merganser and red-breasted merganser Mergus serrator) and the great cormorant Phalacrocorax carbo. However, the impact of fish-eating birds is difficult to assess due to the paucity of information. In Canada, attempts have been made to assess impact by culling sawbills on a large scale and then monitoring changes in populations of fish, but such studies had design problems that make the results difficult to interpret (reviewed in Marquiss & Carss 1994, Russell et al. 1996). In Scotland, an alternative, but complementary, approach of using simple mathematical models to estimate the scale of impact has been initiated (Shearer et al. 1987). The first simulation model suggested that the impact of birds on fish catches could be significant, but as more information accrued (Marquiss et al. 1991) it was found that some assumptions were invalid and more detailed data were needed to model impact.

The present study was commissioned to supply better information on which to base estimates of the impacts of fish-eating birds in Scotland. The specific aims of the project were: to estimate the abundance of goosanders and mergansers in the catchments of the Rivers Dee and North Esk throughout the year; to assess how the diet of fish-eating birds varies throughout Scotland, and between seasons and years, using the stomach contents of birds shot under licence by fishery managers; to investigate the diet and foraging ecology of goosander and merganser ducklings on the Rivers Dee and North Esk respectively; to assess whether populations of juvenile salmon compensate for losses in areas of stream in which their numbers have been depleted; to review methods used in attempts to reduce sawbill predation.

Not all the work undertaken is detailed here. For example, where appropriate we refer to technical papers that are published elsewhere (such as methods of estimating the diet of fish-eating birds, Marquiss & Carss 1997, and measuring fish movements, Armstrong et al. 1996). The duckling foraging work, conducted as a PhD study is summarised in Appendix 2, but available as a thesis (G D Alexander 1995, University of Aberdeen).

The main aim of this report is to detail those aspects relevant to assessing the impact of predation by fish-eating birds on fisheries in Scotland. In doing so it draws on previous sawbill abundance data (summarised in Marquiss et al. 1991), includes all stomach material processed by ITE to date, and reviews the pertinent literature.

The potential for bird impact is examined in a series of calculations to produce estimates of the numbers of fish of commercial interest consumed by birds. The calculations involve multiplying: the numbers of birds present at a certain time and place (birds/ha water surface) the daily food intake (g/bird/day) the numbers of fish consumed there and then (fish/g food intake)

The resulting figures for fish consumed per hectare of water surface per day can then be compared with the available estimates of the standing fish stock. By considering the extent to which fish populations might compensate, it can be seen whether bird predation might have an impact on the numbers of fish available to a fishery. This can then be assessed in terms of the reduction in predation that might be achieved by bird control measures. The chapters that follow this introduction are ordered in that sequence.

Chapter 2 summarises the population data for sawbills on the Rivers Dee and North Esk to explore how the abundance of birds varies in predictable ways. This allows speculation as to the factors (intrinsic or extrinsic to the river system) that might most influence bird abundance. The chapter culminates with estimates of density for various seasons and places, expressed in terms of birds per hectare of river surface. This is because our figures for fish consumed have to be compared with salmonid (salmon and trout) population densities which are expressed as fish per unit area of river surface. The densities are categorised by stream width because sawbill abundance, fish community characteristics, salmonid density and presumably their propensity for compensation, covary with stream width.

Chapter 3 details the composition of the diet (the proportion by mass of various fish species) and the numbers of fish of commercial interest consumed by fish-eating birds in Scotland. The accurate estimation of the proportion by mass of fish in bird diet is important for two reasons. First, the lower the proportion of salmonids in the diet, the fewer of such fish are removed per bird and the more difficult it is to prevent loss to a fishery as very large numbers of birds have to be removed to gain any increase in harvest. Second, a high proportion of other fish species in the diet means that by removing birds there is the theoretical possibility of wider fish community effects, such as a change in salmonid populations resulting from an increase in their competitors or (fish) predators. We discuss the possibility that the amount of salmonids eaten might be determined by the availability of other (more vulnerable) fish prey.

Chapter 4 combines the numbers from Chapters 2 and 3, and the best available estimates of the daily food intake of sawbills and cormorants to calculate the consumption of fish of commercial interest. The fish considered of commercial concern include all salmon and trout, but the size of salmonids consumed is crucial because it affects the likelihood of impact on a fishery. For example, the loss of salmon fry or small parr could result in the enhanced growth or survival of the remaining fish, or the enhanced recruitment of juveniles the following season (Chapter 5). Conversely the removal of large parr or smolts (fish of 90mm or more in length, Feltham 1990, Feltham & MacLean 1996) or reduced production of these, may result in fewer fish reaching the sea and fewer adult fish returning to be harvested. The removal of small trout could be beneficial to a trout fishery if it led to the increased growth of survivors and thus fewer but larger fish whereas the largest of trout taken by birds represent a direct loss to the fishery as they are no longer available to anglers. Under these circumstances the loss to the fishery would be equivalent to the proportion of the whole stock (of large 'angleable' fish) removed by birds (Carss & Marquiss 1996b). In Chapter 4 we calculate not only the numbers of all salmon and all trout consumed by birds, but also the numbers of those more important salmon (90mm or longer) and 'angleable' trout (>200mm).

Where birds consume large numbers of commercially important fish, some consideration must be given to the potential for compensation by the fish population. Chapter 5 discusses the nature of predator-prey interactions and the scope for compensation within salmonid populations subject to predation. It also describes a series of experiments that artificially depleted salmon populations on river sections of various stream widths, to investigate short-term recolonisation. Chapter 6 reviews the literature on methods of control of fish-eating bird predation and describes and evaluates case histories of shooting as a control measure for sawbills in Canada and Scotland.

Finally, Chapter 7 draws together the contents of the preceding chapters in a synthesis discussing the likelihood of bird impact on fisheries and whether or not existing methods of control are likely to be effective. Although much work has been done, there is still much more to do, so an important aspect of Chapter 7 is the clarification of the scientific approach required to proceed along a cost-effective line of investigation. The chapter concludes by assessing strategies for the integration of modelling and experiment and considers requirements for future work.