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

Data Sources and Data Processing

Ages of Adult Salmon

Sea-age- and return-date-stratified data on both the numbers and weights of salmon catches, from both the net and coble fisheries within the North Esk estuary and for (groups of) rod-fishery beats (lengths of river owned by particular proprietors) within the North Esk, were available, from records provided by the proprietors and owners of all Scottish salmon fisheries from 1952 to the present day (see http://www.gov.scot/Topics/marine/Publications/stats/SalmonSeaTroutCatches [Dec-2014]).

The different sea-age phenotypes of Atlantic salmon are associated with different marine life-histories (see Malcolm et al. 2010 for a review). This contention is supported by stable isotope studies demonstrating differences in the marine diet among groups of Atlantic salmon returning to UK rivers (MacKenzie et al. 2012), differences in the marine growth performance of Atlantic salmon (Bacon et al. 2009) and by observed variations in temporal patterns of Atlantic salmon abundance (Youngson et al. 2002).

Scale analysis (Bacon et al. 2009; Bacon et al. 2011) yields strong confirmation of the high accuracy of the estimated sea-ages reported by the netsmen from their net and coble catches (based on visual assessment of body size). However, the Atlantic salmon sea-ages provided by the rod- and net- data-sets for the North Esk have differing accuracies and, importantly in the case of the rod-fisheries only, bias. A more restricted scale-ageing study of Scottish rod-caught fish (MacLean et al. 1996) confirms that rod-angler-reported sea-ages from the rod and line fisheries are appreciably inaccurate (much less accurate than the nets). . Accordingly, in this study, the angler-reported sea-ages for rod-caught salmon were corrected, based on data provided in MacLean et al. (1996), to form our default (grilse-error corrected) data-set (details of the correction procedure are given in Supplementary Material, part A).

Adult Counter

In 1980 (the starting date for the present study) a resistivity counter was installed at the Logie weir on the North Esk, some 6 km from the sea and some 1 Km upstream of the Morphie Dyke weir (Fig.1). The weir diverts water into the Kinnaber lade, which is used to provide estimates of smolt numbers and ages (see below). Over 90% of the wetted river-area available to spawning salmon is above the Logie counter: although a few redds have been recorded between the counter and the head of tide, the great majority of the population is likely to spawn above the Logie counter. The Logie counter provides year-round counts of both upstream and downstream adult salmon movements, and the difference between upstream and downstream counts is used to enumerate total immigration. Sea trout are generally smaller than salmon in this river, and consequently these counts are not appreciably confounded by sea-trout. The counter results show that Atlantic salmon migrate into the North Esk in substantial numbers from February to November.

Commercial In-River Fishery

The net and coble fishery traditionally operated from February to August inclusive and provided scale samples from which the sea- and freshwater-ages of returning adults were obtained. This information was subsequently used to allocate sea-ages to fish passing over the counter (see below). Age estimates outside these months are based on smaller and less comprehensive samples. Since 1999, the start of the netting season has been delayed until May, somewhat altering the representativeness of the net and coble data in recent years.

Estimating numbers of grilse and MSW fish emigrating past counter

The estuarine net and coble fishery is neither age- nor size-selective, (due to the small mesh-size used and its method of deployment) so the salmon escaping this fishery were assumed to contain the same fraction of grilse as the net and coble catch ( F _N). If, over some period (say a month), the rod fishery in the part of the river between the estuary and the counter catches a total of T _RB fish of whom an (accurately evaluated, scale-read) fraction F _RB are grilse, then, provided natural mortality is small compared to fishery mortality, the fraction ( F _C) of the grilse in the resistivity counter's total for the same period ( T _C) is given by

A short derivation of this formula is provided in Supplementary Material, part D. This relationship, using the rod-catch grilse proportion from below the counter (corrected according to the prescription detailed in Supplementary Material, part A) was used to determine the grilse to MSW ratio in the overall (nett, total flux) monthly population flow measured by the Logie counter.

Estimating Pre-Estuary abundance

The numbers of salmon (by sub-stock) entering the estuary were assumed to be equal to the sums of the catches plus the fish passing the counter (thus any fish spawning below the counter were hence omitted from this pre-estuary abundance ( PEA) estimate). Assuming that natural mortality (and, in recent years, mortality associated with catch-and-release by rods) is small relative to direct fishing mortality. If so, then subtraction of the monthly aggregated sea-aged rod kill ( i.e. numbers caught, killed and retained, adjusted for grilse-reporting-error) above the Logie counter from the sea-aged nett monthly immigration past the counter yields the sea-age and run-time stratified spawner counts which were one of the key data inputs to this investigation. The other central input was the sea-age and run-time stratified abundance of individuals returning to the estuary (the pre-estuary abundance, hereafter PEA), which was obtained by adding the monthly aggregate sea-age stratified estuarine net and coble catches to the monthly rod kill below, and the monthly aggregated immigration past, the Logie counter. In this study PEA was used to provide more precise single-catchment numbers than 'pre-coastal home-water fishery abundance' ( PCHWFA), which is used in national and global assessments, but whose 'coastal catch' sub-components cannot be accurately ascribed back to particular rivers.

Smolt trap and estimates of smolt numbers

Annual smolt number estimates were obtained from the North Esk since the early 1970's. A capture-mark-recapture programme at Kinnaber Mill, close to the tidal limit of the North Esk (Fig.1) operated from late March to June or July, until several days in a row when no smolts were observed (Sheerer 1992). The Kinnaber Mill lade receives a portion of the North Esk flow and typically takes 5% to 10% of total emigrants. Smolts are caught and marked at a total-trap across the Kinnaber Mill lade (the trap is 100% efficient except on the rare occasions it spills). Known proportions of those marked smolts are transported back upriver and the proportion of such released-and-marked fish subsequently re-captured in the lade allows estimation of sampling efficiency and thus total numbers (Shearer 1992; Bjorkstedt 2005). The resulting estimates of total smolt output formed the third input to our analysis, and were sub-dividable into smolt (freshwater-winter) age-classes by scale-readings from samples of Kinnaber smolts.