https://orcid.org/0000-0002-8296-297X
Figures
Abstract
Genetic structuring in populations is the result of both historical and contemporary environmental factors driving genetic drift, natural selection and gene flow, as well as purely genetic factors, such as mutation and recombination. In Atlantic salmon (Salmo salar), re-colonisation of rivers after the last Ice Age was shown to be an important factor in shaping contemporary population structure, though the observed structure was more complex than was predicted through founder effects. Thus, other, perhaps more contemporary factors may also play a role. Here, we investigated the influence of the time since deglaciation, distance to the sea, population connectivity, temperature, water quality, waterbody modifications, and environmental protections on spatial structuring of genetic diversity, based on microsatellite data (33 loci) collected from 48 Scottish S. salar populations. The results confirmed that recently deglaciated areas are less genetically diverse and more differentiated. Modified waterbodies also exhibit less genetic diversity and greater differentiation, although this effect differs between rivers draining on the east and west coasts of Scotland. Distance to the sea also had a non-negligible effect, while the other considered factors did not have a significant effect.
Citation: Cowell F, Gaggiotti OE, Cauwelier E (2025) Post-pleistocene colonisation rather than the contemporary environment has most influenced the current population structure of Scottish Atlantic salmon (Salmo salar). PLoS One 20(10): e0333164. https://doi.org/10.1371/journal.pone.0333164
Editor: Sayyed Mohammad Hadi Alavi, University of Tehran, IRAN, ISLAMIC REPUBLIC OF
Received: August 26, 2024; Accepted: September 10, 2025; Published: October 1, 2025
Copyright: © 2025 Cowell et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Genetic relationships between populations are shaped by processes that either increase differentiation, such as natural selection, genetic drift, founder effects and mutation or decrease it, such as gene flow through migration. Understanding how these processes have influenced contemporary populations is of great interest to those involved in conservation biology as intraspecific diversity is an important component of biodiversity, especially at high latitudes [1].
At high latitudes, palearctic fishes, including anadromous Salmonids, have colonised and re-colonised rivers following the respective retreat and expansion of ice sheets [2]. Indeed, the importance of ice sheets on anadromous Salmonid populations has been reported in the Baltic Sea [3], Iceland [4] the British Isles [5,6], Norway and the areas of Russia surrounding the Barents and White seas [7]. This observed population genetic structure has been attributed to strong founder effects upon colonisation of newly ice-free rivers [8]. Furthermore, anadromous Salmonids, such as the Atlantic salmon (Salmo salar) display high levels of philopatry which may result in relatively low levels of gene flow, allowing mutations to accumulate, genetic drift to randomly fix alleles and natural selection to lead to adaptation to the unique local environment [9–11].
Initial colonisation and associated establishment of populations will have been further influenced by factors related to deglaciation, such as isostatic rebound, which would have affected river accessibility [12,13] and population connectivity. Isostatic rebound resulted in relative sea levels decreasing [13] resulting in some of the lower stretches of contemporary rivers being saltwater areas that were not colonisable by S. salar until long after the ice sheets melted. Concurrently, higher reaches of rivers, with lower carrying capacity, might have become available as the ice sheet retreated, expanding the species range upstream. It is therefore uncertain how the gradual expansion within a river affected the relationship between distance to sea and differentiation, as well as population connectivity.
Despite Cauwelier et al. [5] identifying the melting of ice sheets as having an important influence on genetic structuring on Scottish S. salar populations, ecological niche modelling could not explain all the differences between populations. Contemporary factors, such as climate change, waterbody modification, etc. are, therefore, of interest, as there have been many anthropogenic impacts on watercourses since the time of deglaciation and it is possible that these have resulted in genetic changes in S. salar populations.
The depredation of fishes by human beings is well known to have begun in prehistory but this has intensified during the late modern period [14]. River temperatures documented in Scotland during the 20th and 21st century have risen, likely a result of anthropogenic climate change [15]. The water quality itself has been impacted greatly as well and the riverscape has been changed by humans to allow for hydroelectricity generation, irrigation of farmland and the prevention of flooding [16]. All these factors have prompted environmental designations to be put in place in order to protect rivers and the organisms that live there.
There exist a number of reasons why one might expect the structure of S. salar to be impacted by human activities. In some cases, local adaptation to thermal regimes has been shown to result in greater genetic diversity within and differentiation between S. salar populations [17]. However, the reported effect of temperature may be indirect rather than causal, as it affects other elements of the ecosystem, such as bacterial diversity [18]. Poor water quality has been shown to restrict distribution [19] and reduce abundance [20] in anadromous fishes, the former having the potential to reduce gene flow, while the latter could increase genetic drift. Further direct and indirect effects of anthropogenic activity, such as modifications to waterways, e.g., dams and water abstraction have been shown to reduce genetic diversity and increase genetic differentiation [21–23]. Meanwhile, population interbreeding due to stocking and/or farm escapees associated with aquaculture also impact genetic structuring leading to less genetic differentiation between populations [24–27].
Lastly, efforts to protect S. salar populations are expected to also have an impact on diversity and structuring, whether through stocking and/or habitat conservation/restoration. In Scotland many rivers are designated as SACs (Special Area of Conservation) with this designation often being given specifically to protect S. salar habitat. The impact of these designations on S. salar populations is, however, poorly understood. One might expect an increase in population size in response to these measures which would, in turn, result in reduced levels of genetic drift.
All the above factors, both historical and contemporary, have the potential to alter levels of gene flow and genetic drift. These evolutionary forces can, in turn, influence the genetic composition of extant populations. However, the possible relationship between genetic differentiation and environmental factors, both historical and contemporary, has not been considered concurrently. Here, we investigated the effect of both historical and contemporary factors on Scottish S. salar population structure. Data for 48 S. salar populations from across Scotland were used to calculate measures of genetic differentiation and diversity. The impacts of the time since deglaciation, distance to the sea, population connectivity, temperature, water quality, waterbody modification, and environmental protections were assessed through general linear modelling using both a frequentist and a Bayesian approach.
Methods
Sample collection and screening
Due to computational demands required for the Bayesian approach, the full dataset described in Cauwelier et al. [5] was not used. 48 sites were randomly selected to represent all genetic groups identified by STRUCTURE analysis in Cauwelier et al. [5], with numbers proportional to the geographic area of the group (Fig 1). This dataset was extended by screening for an additional sixteen microsatellites, in order to increase genome coverage and potentially enhance resolution.
Numbers relate to site numbers detailed in Table 1 and the colours refer to the genetic cluster to which each site belongs, as found by Cauwelier et al. [5]. Source of map of Europe: www.freeworldmaps.net, map of Scotland: https://osdatahub.os.uk/downloads/open/BoundaryLine and salmon rivers: https://marinescotland.atkinsgeospatial.com/nmpi/default.aspx?layers=843.
Between 2003 and 2010, 48 sites across 41 Scottish rivers were electrofished for juvenile S. salar (Fig 1). The care and use of experimental animals complied with Scottish Government animal welfare laws, guidelines and policies as approved by Marine Scotland Science AWERB under Home Office project licence PP3525229. Exact field methodology may have varied slightly from river to river but, in brief, fish anaesthetics, such as tricaine methanesulphonate (MS-222), were mixed with a set volume of water in a bucket dedicated for this purpose. Fin clips were taken from these anaesthetised fish and were then allowed to recover in a bucket of aerated water, before being released and clips stored in 99% ethanol. DNA was extracted using the method described by Knox et al. [28] and initially amplified for three multiplexes encompassing fourteen microsatellites ([29]- excluding SsaD486) plus SsaD48, SsaD71 [30] and SP1608 [31]. Full details on PCR conditions and fragment analysis are described in Cauwelier et al. [8]. Screening for an additional sixteen microsatellites (EST107, EST19,EST28, EST68, MHC1, MHC2, Sleel53, Sleen82, Ssa405, Ssa407, Ssa412, Ssa98, Ssleer15, SsOSL25, Ssosl85 and SsSP2215) was carried out at the Norwegian Institute for Marine Research, using the protocol described in Harvey et al. [32]. The resulting dataset consisted of 1,044 fish genotyped at 33 microsatellite loci.
Genetic analysis
Tests for deviation from Hardy-Weinberg (HW) equilibrium were conducted using the Markov chain method in the Genepop package [34] in R version 4.0.3 [35]. This was run with 2,000 iterations of 500 batches with a dememorization length of 10,000. The same package was used to test for linkage disequilibrium (LD), with 1,000 iterations over 100 batches and a dememorization of 10,000. Handling multiple tests to avoid false positives was done by following Waples [36], using the cumulative binomial distribution to identify if the number of positive tests significantly exceeded those expected due to chance alone.
Rarefied allelic richness was calculated using hierfstat [37]. Population-specific differentiation was based on Fst values, which were calculated using GESTE [38]. GESTE is a Bayesian programme that estimates population-specific Fst values measuring genetic divergence between each population and the metapopulation as a whole.
Environmental data
Environmental data were synthesised from several sources. Data on water quality and protection status were obtained from NatureScot and the Scottish Environmental Protection Agency (SEPA) [39]. Data on the water quality in rivers used the Water Framework Directive (WFD) classification of a waterbody as either High, Good, Moderate, Poor or Bad, which was based on many factors relating to chemistry, biology, specific pollutants and hydromorphology. The WFD also details the extent to which a river has been modified by human activities, a heavily modified river being one which has been permanently physically altered in a way which has substantially changed its character, and as such, is designated under article 4(3) of the WFD. NatureScot provided information on the protection status of rivers and surrounding areas. The designated protection for each site was recorded and, if the reason for this designation included the need to protect the spawning and nursery habitat of S. salar, then this information was also recorded. The qualitative descriptors of habitat protection were coded as three binary dummy variables. The first variable showed if S. salar were specially protected or not in that section of river, the second detailed if the section of river was protected for a reason unrelated to S. salar or not, and the third detailed if the river, as well as a broader area outwith the riparian zone, was protected or not.
Water temperature data came from the Scotland River Temperature Monitoring Network, through the Marine Scotland website [40]. These are modelled data that have been inferred from strategic monitoring of temperature at specific sites throughout Scotland and the consideration of influential landscape features [41]. The figures used were the predictions of the maximum daily river temperature for the hottest year in the last 20 years.
The time since glaciation was also considered, using maps of the deglaciation process made by Cauwelier et al. [5]. Distance to the sea and elevation data (OS terrain 50 layer) were calculated using ArcMap 10.6 [42]. Pairwise distances were calculated by measuring the shortest swimmable distance from the mouth of the rivers using Free Map Tools [43]. This was then added to the distance to the sea measures of the two rivers. This pairwise distance matrix was then converted into a population-specific measure of connectivity, which was calculated as the mean of all pairwise distances between a population and all other populations.
In order to limit issues arising from multicollinearity, a correlation matrix was calculated between all the environmental variables. One variable of each pair was removed in cases of a correlation coefficient greater than 0.8. Furthermore, variables with a variance inflation factor (VIF) greater than 5 were also removed. Distance to the sea was used in place of elevation for this reason.
Modelling
Rarefied allelic richness was analysed by using General Linear Models (GLMs), with all environmental variables initially included. A procedure of backwards elimination was employed, using a p-to-remove criterion, until only a single independent variable was included in the remaining model. The best model was then chosen as the one with the lowest corrected Akaike’s Information Criterion (AICc) score. All GLMs assumed a normal distribution, as this best fitted the model assumptions. As many of the variables were categorical, the drop1 function in R was used, as this added and removed each term of the model in turn, allowing the effect of that factor on the R2 to be realised and the significance to be determined.
GESTE was used to investigate the association between population-specific Fst and environmental factors following the procedure described by Gaggiotti et al. [44]. This involved using default settings except for using ten pilot runs of 1,000 iterations, with an additional burn in period of 5,000,000 iterations. A thinning interval of 50 was then used during the main run. An initial exploratory run with all factors (full model) was used to identify the five factors with the highest marginal posterior probability across all models, which were then used in a second run. The five models with the highest posterior probabilities were reported.
For comparison with the GESTE connectivity effect, a pairwise Fst matrix was generated using Genalex 6.5 [45,46]. In an effort to detect any pattern of isolation-by-distance, this matrix was linearised (Fst/(1-Fst)), as suggested by Rousset [47] and was regressed against a geographical distance matrix with a classic mantel test using the R package ade4 [48] with 9999 repeats.
Two different spatial scales were analysed for each modelling method. Initially, all rivers were included, which encompassed seven genetic regions [8]. A hierarchical structure was then considered, whereby data were split into rivers draining along the east coast and west coast and analysed separately. Carrying out separate analyses for the two geographic regions was done in order to avoid confounding effects by factors that could not be accounted for by statistical analyses, given that the anthropogenic effects were likely to differ between the two coasts. In particular, protection of waterbody for a reason other than S. salar, as well as protection of the surrounding land, only applied to the west coast. Moreover, there are geomorphological differences between the east and west coasts which have long been recognised [49]. The three sites on the north coast of Scotland were categorised as being in the west coast group, as they form part of a genetic grouping extending west [5].
Results
Genetic diversity and differentiation
As Hardy-Weinberg equilibrium deviation and positive tests for LD were not significantly greater than the number expected due to chance alone, all 33 microsatellite loci were analysed. Mean rarefied allelic richness across the 48 Scottish S. salar populations equalled 5.42 (SD = 0.26). Population-specific Fst generated from GESTE, when all sites were considered, ranged from 0.002 in the River Don (population 10 in Table 1 and Fig 1) to 0.084 in the River Leven (population 39) with a mean Fst of 0.021 (SD = 0.019). Full diversity and differentiation results can be found in Table 1.
General linear modelling
The difference between the AICc score of the simplest model and the next lowest score was not > 2 but it was nonetheless chosen as the best model for its simplicity. Variation in rarefied allelic richness at all sites was best explained by a model considering time since deglaciation (Fig 2 and Table 2). In the case of the east coast, only waterbody modification was included in the best model. Genetic diversity on the west coast meanwhile was best predicted by the sole factor of time since deglaciation. The adjusted R2 for the east coast model was more than twice as high when compared to the all-sites or west coast models.
GESTE
When predicting genetic differentiation in a model which included all sites, the marginal posterior probabilities of each factor, when GESTE considered all nine (Table 3), was greatest for time since deglaciation. This was also true when considering the east and west coasts separately. However, additionally, on the east coast, waterbody modification and distance to the sea also had relatively high probabilities, and the former had a higher posterior probability than the latter. In all cases, connectivity had a negligible effect; a result corroborated by the classic mantel test between all sites (r = 0.084, p = 0.084).
Subsequent analysis of the effect of the five best environmental factors on population-specific Fst across all sites also revealed the predominant influence of time since deglaciation (Table 4) but also found that distance to the sea was an influential factor. Population-specific Fst was found to be higher in the more recently deglaciated rivers and lower in sites more distant from the sea (Fig 3). This was also the case when the east coast was considered separately where, additionally, waterbody modification was important (Fig 3, Table 4).
Model numbers relate to the numbers in Table 4.
Indeed, unmodified waterbodies exhibited much lower genetic differentiation than those that have been modified. On the other hand, the results from the west coast dataset showed that the null model was best. The marginal posterior probability on the west coast for waterbody modification, distance to the sea and time since deglaciation were 0.099, 0.1 and 0.598, respectively. In this same order, the east coast results were 0.172, 0.152 and 0.567.
Discussion
The aim of this study was to explore the major historical and contemporary factors contributing to genetic differentiation and diversity in S. salar. Overall, the largest factor associated with the observed diversity/differentiation was timing since deglaciation, which was also reported in Cauwelier et al. [5]. However, contemporary factors also seemed to be playing a role in shaping genetic variation, particularly when considering rivers draining on the east coast of Scotland, where waterbody modification had an effect on the same order of magnitude as time since deglaciation. Additionally, an effect of distance to sea was also detected but the effect size was very small.
Post-pleistocene colonisation
Previous studies have shown that the founder effects associated with recolonisation of newly ice-free rivers [3–6,8] are still the predominant determinant of contemporary genetic structuring in S. salar populations. These results, showing low genetic diversity and high differentiation in the most recently deglaciated rivers, are akin to the stepping-stone model [50] and central-marginal hypothesis [51]. This result is likely to be independent of the type of molecular markers used, as SNP based analysis [52] resulted in very similar groupings to when microsatellites were used [5].
Waterbody modification
A large effect of waterbody modification on genetic diversity and differentiation was only observed in the east coast analysis. The fact that genetic diversity on the east coast is significantly reduced at sites that have been heavily modified suggests, unsurprisingly, that there have been impediments to gene flow and stronger genetic drift as a result of these modifications. The non-negligible posterior probability of time since deglaciation and waterbody modification provided by GESTE further shows these two factors had an impact on the pattern of genetic differentiation seen today. Sites that have been deglaciated recently have been the major contributor to genetic differentiation, but waterbody modifications have likely reduced gene flow between populations and increased levels of genetic drift, increasing differentiation and reducing diversity further. This highlights how anthropogenic activities can have important effects on ecological timescales.
There were differences in the nature of the modifications between East and West river drainages, with hydro schemes dominating the modifications along the east coast. Modifications caused by hydro schemes vary in their nature, including, for example, both the storage of water behind dams and general water abstraction from run-of-river schemes. Meanwhile modifications for flood prevention and agriculture were more frequently found along the west coast. Indeed, hydro schemes were responsible for five of the six alterations on the east coast sites, whereas this is only true for four out of the seven modified sites on the west coast. The absence of an impact of waterbody modifications on diversity/differentiation along the west coast could suggest that these cases of flood prevention and irrigation have not significantly altered the genetic composition of these populations. Furthermore, the results provide another example that hydro schemes can have a marked impact on genetic structuring [21–23].
Distance to the sea
Distance to the sea had a non-negligible posterior probability (all sites and east coast) in the GESTE analysis. However, the slopes (Fig 3) were very close to zero, which would suggest that the distance from the sea was not biologically meaningful. In both cases, the influence of the distance to the sea was accompanied by the time since deglaciation, which was unsurprising, given that colonisation might majorly have been in an upstream direction, as the ice sheet was receding. However, the results suggest that genetic differentiation decreased with increasing distance to the sea; the opposite of what previous studies have found [17-53].
The unexpected negative relationship between differentiation and distance to sea could be due to outlier sites. These outlier sites were characterised by high population-specific Fst close to the mouth of the river or low Fst at large distances from the sea. Coincidentally, the outlier sites were among the first (low Fst) or the most recent (high Fst) to become ice free.
Overall, the existence of such outliers suggests that the influence of the distance to the sea on genetic differentiation is overridden by the effect of time since deglaciation. When the outliers were removed, the relationship between differentiation and distance to the sea became positive. Although not statistically significant (r = 0.237, p = 0.159), this is more in line with the previous findings [17-53].
Other factors
The other factors considered in the model which are influenced by anthropogenic activities (water quality, temperature and the various environmental protections) appeared to have had a minor impact on genetic structuring in these populations.
Contrasting results have been reported on the effect of temperature on genetic differentiation. Dionne et al. [17] showed that increases in genetic differentiation were associated with higher temperatures in some S. salar populations. In contrast, anadromous rainbow (steelhead) trout (Oncorhynchus mykiss) showed reduced differentiation with increased temperature [54]. This would suggest that more research is needed into the effect of temperature on population structuring in various species of Salmonids.
There were other environmental variables left unincluded due to a paucity of data and/or the level of complexity, such as those relating to hydrology and geography. For example, populations in close proximity to one another but located across geological boundaries have been shown to be more differentiated than populations within them, which might be linked to accuracy of homing [55].
Considerations for the west coast
Similar to the analysis of all sites and the east coast sites, the results from the west coast analysis showed an effect of timing since deglaciation on genetic diversity, with lower diversity found in the most recently deglaciated rivers. It is, however, notable that the R2 more than doubles if west coast sites are not included. It might be that the limited variability of the various factors considered played a role. For example, distance to sea varied from 0.113 to 58.5 km (mean = 14.9 and SD = 16.5) along the west coast, whilst this was between 2.75 and 121 km (mean = 42.2 and SD = 36.3) along the east coast. However, the low R2, together with the finding that the null model was the best when looking at population-specific Fst, might also suggest that, despite including many factors in this study, some other, more prominent factors, were missed.
One factor that was not considered in the model was the presence of aquaculture. Indeed, fish farms are almost exclusively located on the west coast of Scotland. Given that Gilbey et al. [56] have detected introgression from farmed fish in wild west coast populations, it cannot be ruled out that aquaculture has had a significant impact on the populations in this study. Glover et al. [27] showed that, in Norway, the presence and growth of the aquaculture industry has led to an erosion of wild S. salar population structure over time, with contemporary populations being less differentiated than their historical counterparts. However, the impact of aquaculture on wild populations varies across regions [25,56] and is very complex to measure and model [57,58]. As such, it would be very difficult to disentangle and robustly attribute the impact of aquaculture from the other local environmental characteristics of the rivers in question.
Conclusions
In summary, the results from this study have reinforced the findings of Cauwelier et al. [5] but put the influence of post-Ice Age colonisation in the context of the other contemporary forces considered and differences in the type of anthropogenic perturbations between east and west coast river-drainage systems in Scotland. In this regard, the most important contemporary factor is waterbody modification affecting rivers draining in the east. However, other river characteristics and anthropogenic factors not considered in this study could also be important and these should be explored in more detail so as to account for the genetic variation not explained in this study.
Supporting information
S1 Table. Posterior means, modes and highest posterior density intervals (HPDI) for the best GESTE models.
https://doi.org/10.1371/journal.pone.0333164.s001
(DOCX)
S1 File. Genetic data: Microsatellite data in GENALEX format.
https://doi.org/10.1371/journal.pone.0333164.s002
(XLSX)
Acknowledgments
Thanks to Widar Wennevik for helpful comments and discussion. Thanks also to John Gilbey for assisting in the selection of the sample sites and to Synne Winterthun for her work on screening and QC-ing the samples for the additional microsatellites. This work was made possible due to the partial sample screening carried out under the Research Council of Norway project 280308 (SeaSalar).
References
- 1. Pamilo P, Savolainen O. Post-glacial colonization, drift, local selection and conservation value of populations: a northern perspective. Hereditas. 1999;130:229–38.
- 2. McDowall RM. Why are so many boreal freshwater fishes anadromous? Confronting ‘conventional wisdom’. Fish Fisher. 2008;9(2):208–13.
- 3. Säisä M, Koljonen ML, Gross R, Nilsson J, Tahtinen J, Koskiniemi J, et al. Population genetic structure and postglacial colonization of Atlantic salmon (Salmo salar) in the Baltic Sea area based on microsatellite DNA variation. Canadian J Fish Aquatic Sci. 2005;62:1887–904.
- 4. Olafsson K, Pampoulie C, Hjorleifsdottir S, Gudjonsson S, Hreggvidsson GO. Present-day genetic structure of Atlantic salmon (Salmo salar) in Icelandic rivers and ice-cap retreat models. PLoS One. 2014;9(2):e86809. pmid:24498283
- 5. Cauwelier E, Verspoor E, Coulson MW, Armstrong A, Knox D, Stradmeyer L, et al. Ice sheets and genetics: insights into the phylogeography of Scottish Atlantic salmon, Salmo salar L. J Biogeogr. 2018;45:51–63.
- 6. Finnegan AK, Griffiths AM, King RA, Machado-Schiaffino G, Porcher J-P, Garcia-Vazquez E, et al. Use of multiple markers demonstrates a cryptic western refugium and postglacial colonisation routes of Atlantic salmon (Salmo salar L.) in Northwest Europe. Heredity (Edinb). 2013;111(1):34–43. pmid:23512011
- 7. Wennevik V, Quintela M, Skaala Ø, Verspoor E, Prusov S, Glover KA. Population genetic analysis reveals a geographically limited transition zone between two genetically distinct Atlantic salmon lineages in Norway. Ecol Evol. 2019;9(12):6901–21. pmid:31380023
- 8. Bernatchez L, Wilson CC. Comparative phylogeography of nearctic and palearctic fishes. Mol Ecol. 1998;7:431–52.
- 9. Garcia de Leaniz C, Fleming IA, Einum S, Verspoor E, Jordan WC, Consuegra S, et al. A critical review of adaptive genetic variation in Atlantic salmon: implications for conservation. Biol Rev Camb Philos Soc. 2007;82(2):173–211. pmid:17437557
- 10. Landry C, Bernatchez L. Comparative analysis of population structure across environments and geographical scales at major histocompatibility complex and microsatellite loci in Atlantic salmon (Salmo salar). Mol Ecol. 2001;10(10):2525–39. pmid:11742552
- 11. Verspoor E, Beardmore JA, Consuegra S, Garcia de Leaniz C, Hindar K, Jordan W, et al. Population structure in the Atlantic salmon: insights from 40 years of research into genetic protein variation. J Fish Biol. 2005;67:3–54.
- 12. Shennan I, Bradley S, Milne G, Brooks A, Bassett S, Hamilton S. Relative sea-level changes, glacial isostatic modelling and ice-sheet reconstructions from the British Isles since the last glacial maximum. J Quater Sci. 2006;21:585–99.
- 13. Shennan I, Hamilton S, Hillier C, Hunter A, Woodall R, Bradley S, et al. Relative sea-level observation in western Scotland since the last glacial maximum for testing models of glacial isostatic land movements and ice-sheet reconstructions. J Quater Sci. 2006;21:601–13.
- 14.
Wilson EO. The diversity of life. new ed. London: Penguin Books; 2001.
- 15. Pohle I, Helliwell R, Aube C, Gibbs S, Spencer M, Spezia L. Citizen science evidence from the past century shows that Scottish rivers are warming. Sci Total Environ. 2019;659:53–65. pmid:30594861
- 16.
Freethy R. The natural history of rivers. Suffolk: Terence Dalton Limited; 1986.
- 17. Dionne M, Caron F, Dodson JJ, Bernatchez L. Landscape genetics and hierarchical genetic structure in Atlantic salmon: the interaction of gene flow and local adaptation. Mol Ecol. 2008;17(10):2382–96. pmid:18430145
- 18. Dionne M, Miller KM, Dodson JJ, Caron F, Bernatchez L. Clinal variation in MHC diversity with temperature: evidence for the role of host-pathogen interaction on local adaptation in Atlantic salmon. Evolution. 2007;61(9):2154–64. pmid:17767587
- 19.
Maes J, Stevens M, Breine J. Poor water quality constrains the distribution and movements of twaite shad Alosa fallax fallax (Lacépède, 1803) in the watershed of river Scheldt. In: Durfour S, Prévost E, Rochard E, Williot P, editors. Fish and Diadromy in Europe (ecology, management, conservation). Developments in Hydrobiology, vol 200, 602. Dordrecht: Springer; 2008. p. 155–9.
- 20. Zdasiuk BJ, Chen CY, McCormick SD, Nislow KH, Singley JG, Kelly JT. Evaluating acid-aluminium stress in streams of the Northeastern U.S. at watershed, fish community and physiological scales. Ecol Ind. 2002;144:109480.
- 21. Deiner K, Garza JC, Coey R, Girman DJ. Population structure and genetic diversity of trout (Oncorhynchus mykiss) above and below natural and man-made barriers in the Russian River, California. Conserv Genet. 2007;8:437–54.
- 22. Samad-zada F, van Poorten BT, Harris S, Godbout L, Russello MA. Genome-wide analysis reveals demographic and life-history patterns associated with habitat modification in landlocked deep-spawning sockeye salmon (Oncorhynchus nerka). Ecol Evol. 2021;11:13186–205.
- 23. Yamamoto S, Morita K, Koizumi I, Maekawa K. Genetic differentiation of white-spotted charr (Salvelinus leucomaenis) populations after habitat fragmentation: spatial-temporal changes in gene frequencies. Conserv Genet. 2004;5:529–38.
- 24.
Cross TF, McGinnity P, Coughlan J, Dillane E, Ferguson A, Koljonen M, et al. Stocking and Ranching. In: Verspoor E, Stradmeyer L, Nielsen JL, editors. The Atlantic Salmon: Genetics, Conservation and Management. Oxford: Blackwell Publishing; 2007. p. 325–56.
- 25. Diserud OH, Fiske P, Karlsson S, Glover KA, Naesje T, Aronsen T, et al. Natural and anthropogenic drivers of escaped farmed salmon occurrence and introgression into wild Norwegian Atlantic salmon populations. ICES J Marine Sci. 2022;79:1363–79.
- 26.
Ferguson A, Fleming IA, Hindar K, Skaala Ø, Mcginnity P, Cross T, et al. Farm Escapes. In: Verspoor E, Stradmeyer L, Nielsen JL, editors. The Atlantic Salmon: Genetics, Conservation and Management. Oxford: Blackwell Publishing; 2007. p. 357–98.
- 27. Glover KA, Quintela M, Wennevik V, Besnier F, Sørvik AGE, Skaala Ø. Three decades of farmed escapees in the wild: a spatio-temporal analysis of Atlantic salmon population genetic structure throughout Norway. PLoS One. 2012;7(8):e43129. pmid:22916215
- 28. Knox D, Lehmann K, Reddin DG, Verspoor E. Genotyping of archival Atlantic salmon scales from northern Quebec and West Greenland using novel PCR primers for degraded mtDNA. J Fish Biol. 2002;60:266–70.
- 29. Ellis JS, Gilbey J, Armstrong A, Balstad T, Cauwelier E, Cherbonnel C, et al. Microsatellite standardization and evaluation of genotyping error in a large multi-partner research programme for conservation of Atlantic salmon (Salmo salar L.). Genetica. 2011;139(3):353–67. pmid:21279823
- 30. King TL, Eackles MS, Letcher BH. Microsatellite DNA markers for the study of Atlantic salmon (Salmo salar) kinship, population structure, and mixed‐fishery analyses. Mol Ecol Notes. 2005;5:130–2.
- 31. Paterson S, Piertney SB, Knox D, Gilbey J, Verspoor E. Characterization and PCR multiplexing of novel highly variable tetranucleotide Atlantic salmon (Salmo salar L.) microsatellites. Mol Ecol Notes. 2004;4:160–2.
- 32. Harvey AC, Quintela M, Glover KA, Karlsen Ø, Nilsen R, Skaala Ø, et al. Inferring Atlantic salmon post-smolt migration patterns using genetic assignment. R Soc Open Sci. 2019;6(10):190426. pmid:31824688
- 33.
QGIS Development Team QGIS Geographic Information System. Open Source Geospatial Foundation Project; 2022. Available from: http://qgis.osgeo.org
- 34. Rousset F. genepop’007: a complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour. 2008;8(1):103–6. pmid:21585727
- 35.
R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation of Statistical Computing; 2022.
- 36. Waples RS. Testing for Hardy-Weinberg proportions: have we lost the plot?. J Hered. 2015;106(1):1–19. pmid:25425676
- 37.
Goudet J, Jombart T, Kamvar ZN, Archer E, Hardy O. hierfstat: Estimation and tests of hierarchical F-statistics. R package version 0.5-10. 2021.
- 38. Foll M, Gaggiotti O. Identifying the environmental factors that determine the genetic structure of populations. Genetics. 2006;174(2):875–91. pmid:16951078
- 39.
Nature Scot and SEPA Data Analysis [cited 10.09.2021]. Available from: https://www.environment.gov.scot/data/data-analysis/
- 40.
Marine Scotland Scotland River Temperature Monitoring Network (SRTMN) – Predictions of river temperature and sensitivity to climate change [cited 09.09.2021] Available from: https://www.gov.scot/publications/scotland-river-temperature-monitoring-network-srtmn/pages/outputs-and-tools/
- 41. Jackson FL, Fryer RJ, Hannah DM, Millar CP, Malcolm IA. A spatio-temporal statistical model of maximum daily river temperatures to inform the management of Scotland’s Atlantic salmon rivers under climate change. Sci Total Environ. 2018;612:1543–58. pmid:28915548
- 42.
ESRI ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research Institute; 2011.
- 43.
Free Map Tools, Measure a Distance [cited 12. 01. 2022] Available from: https://www.freemaptools.com/measure-distance.htm
- 44. Gaggiotti OE, Bekkevold D, Jørgensen HBH, Foll M, Carvalho GR, Andre C, et al. Disentangling the effects of evolutionary, demographic, and environmental factors influencing genetic structure of natural populations: Atlantic herring as a case study. Evolution. 2009;63(11):2939–51. pmid:19624724
- 45. Peakall R, Smouse PE. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes. 2006;6:288–95.
- 46. Peakall R, Smouse PE. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research – an update. Bioinformatics. 2012;28:2537–9.
- 47. Rousset F. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics. 1997;145(4):1219–28. pmid:9093870
- 48. Dray S, Dufour AB. The ade4 package: implementing the duality diagram for ecologists. J Stat Soft. 2007;22:1–20.
- 49. Ballantyne CK. The scenery of Scotland revisited: retrospective assessment of a classic geomorphological text. Scottish Geog J. 2021;137:228–50.
- 50. Le Corre V, Kremer A. Cumulative effects of founding events during colonisation on genetic diversity and differentiation in an island and stepping-stone model. J Evolution Biol. 1998;11:495–512.
- 51. Eckert CG, Samis KE, Lougheed SC. Genetic variation across species’ geographical ranges: the central-marginal hypothesis and beyond. Mol Ecol. 2008;17(5):1170–88. pmid:18302683
- 52. Gilbey J, Cauwelier E, Coulson MW, Stradmeyer L, Sampayo JN, Armstrong A, et al. Accuracy of Assignment of Atlantic Salmon (Salmo salar L.) to Rivers and Regions in Scotland and Northeast England Based on Single Nucleotide Polymorphism (SNP) Markers. PLoS One. 2016;11(10):e0164327. pmid:27723810
- 53. Harris LN, Moore JS, Galpern P, Tallman RF, Taylor EB. Geographic influences on fine-scale, hierarchical population structure in northern Canadian populations of anadromous Arctic char (Salvelinus alpinus). Environ Biol Fish. 2014;97:1233–52.
- 54. Hand BK, Muhlfeld CC, Wade AA, Kovach RP, Whited DC, Narum SR, et al. Climate variables explain neutral and adaptive variation within salmonid metapopulations: the importance of replication in landscape genetics. Mol Ecol. 2016;25(3):689–705. pmid:26677031
- 55. Dallaire X, Normandeau É, Mainguy J, Tremblay J-É, Bernatchez L, Moore J-S. Genomic data support management of anadromous Arctic Char fisheries in Nunavik by highlighting neutral and putatively adaptive genetic variation. Evol Appl. 2021;14(7):1880–97. pmid:34295370
- 56. Gilbey J, Sampayo J, Cauwelier E, Malcolm I, Millidine K, Jackson F, et al. A national assessment of the influence of farmed salmon escapes on the genetic integrity of wild Scottish Atlantic salmon populations. Scottish Marine Freshw Sci. 2021;12:1–70.
- 57. Mahlum S, Vollset KW, Barlaup BT, Skoglund H, Velle G. Salmon on the lam: drivers of escaped farmed fish abundance in rivers. J Appl Ecol. 2020;58:550–61.
- 58. Sylvester EVA, Wringe BF, Duffy SJ, Hamilton LC, Fleming IA, Castellani M, et al. Estimating the relative fitness of escaped farmed salmon offspring in the wild and modelling the consequences of invasion for wild populations. Evol Appl. 2019;12(4):705–17. pmid:30976304