Scottish Marine and Freshwater Science Vol 6 No 12: The demography of a phenotypically mixed Atlantic salmon (Salmo salar) population as discerned for an eastern Scottish river

Introduction

Fishery managers frequently treat stocks as if they are homogeneous and fully mixing, an assumption that could have deleterious consequences if untrue. A number of wild Atlantic salmon (Salmo salar) stocks are thought to be below full reproductive capacity ( ICES 2015), raising concerns from fisheries managers and conservationists ( AST 2014). Consequently the need for local monitoring to facilitate management of remaining populations is accordingly very high (Chaput 2012). Achieving successful management is hampered by the species' complex life-cycle, which can vary very locally, within-catchments, and also by ongoing fishery exploitation.

In Scotland the Atlantic salmon fishery contributes in excess of £60m per annum to the rural economy (Riddington et al. 2002). Although historically dominated by commercial net fisheries, this contribution comes increasingly from the sporting rod-fishery (Radford and Gibson 2004), to which the larger `multi-sea-winter' ( MSW) fish are of greater value than smaller one-sea-winter fish (grilse, 1 SW). These Scottish rod-fisheries (increasingly catch-and-release) benefit substantially from the fact that fresh-run, and thus `catchable' (Thorley et al. 2007; Bacon et al. 2011), salmon are available in rivers somewhere in Scotland virtually throughout the year (Youngson et al. 2002; Stewart et al. 2002). By contrast, many other European countries have runs restricted to just a few months ( e.g. May to July in Sweden, June to August in Norway, June to September in Iceland).

Scottish stocks of Atlantic salmon have changed appreciably over the last half century (Youngson et al. 2003; Vøllestad et al. 2009), with low catches of early-running MSW salmon being viewed with particular concern by fishery managers. Less attention has been focused on grilse, whose stocks and catches seem to have remained more nearly stable or increased, and on late-running MSW fish, whose stocks varied more than grilse (Anon 2009, 2013). Although the ultimate causes of these changes are matters of debate (Vøllestad et al. 2009; Bacon et al. 2009), the desire to preserve stocks of early-running MSW salmon has led to a number of local, national and international actions aimed at reducing fishery pressures, particularly on early-running MSW fish.

Scottish MSW salmon run throughout the year, but predominantly in the early part of the season, whereas grilse dominate the run in the later months. MSW returners are the main source of individually desirable target fish, but grilse are essential to maintaining numbers of potential target fish in late season. Thus the sea-age composition of catchment stocks is of central interest to Scottish fishery owners.

Early-running salmon on the Scottish east coast are broadly associated with upland spawning habitats, whilst late-run components mainly use lowland ones (eg Laughton and Smith 1992; for a discussion see Bacon et al. 2012). Although grilse in the large catchments of eastern Scotland are mainly associated with lowland spawning, a small number, and proportion, of grilse spawners are regularly found in surveys of upland sub-catchments. Such 'upland' and 'lowland' designations probably represent extremes of a continuum, and thus no clear habitat boundaries or spawning zones can be reliably defined.

The aim of this paper is to assess whether treating phenotypically distinguishable putative sub-stocks as a single fully-mixing stock might impair appropriate management decisions. We recognise that such phenotypes have both genetic and environmental influences, which are sufficiently complex that current data cannot fully reflect them; hence any modelling will be incomplete. However, we describe a pragmatic way of defining sub-stock phenotypes, which assumes independence of stocks and provides a sharp contrast to the assumption of 'fully-mixing' for single-stock models. Reality probably lies between these extremes, involving complex interactions between stocks ( e.g. though: sub-stock mixing; assortative-matings; and heritabilities of phenotypes; all of unknown degrees). Nevertheless we consider it useful to contrast the merits of these different sub-stock paradigms for management purposes, as this highlights the strengths, weaknesses and risks associated with different approaches in the absence of perfect data and models. In order to develop an internally consistent set of demographic descriptions for the phenotypically differentiated sub-stocks (by sea-age and run-timing) which co-exist within many Scottish river catchments, one first needs to quantitatively integrate data from various life-cycle stages. Here we combined data from the monitoring of various salmon life-stages (smolt emigration; fishery removals; adult fish counter) on the intensively studied Scottish river North Esk.

Three subsidiary points were also addressed. The first was the potential impact of systematic sea-age reporting bias (`grilse error') in the salmon rod-catch data. The second was the demographic importance of inter-phenotype (sub-stock) breeding, which is known to occur, but whose impact on catchment population dynamics is currently unknown. The third was the impact of a number of changes in fisheries practice and management which have been made during the study period and that had the capacity (or indeed intent) of differentially influencing particular sub-stock components.