Alternative life‐history strategy contributions to effective population size in a naturally spawning salmon population

Abstract Alternative life‐history tactics are predicted to affect within‐population genetic processes but have received little attention. For example, the impact of precocious males on effective population size (N e) has not been quantified directly in Pacific salmon Oncorhynchus spp., even though they can make up a large percentage of the total male spawners. We investigated the contribution of precocial males (“jacks”) to N e in a naturally spawning population of Coho Salmon O. kisutch from the Auke Creek watershed in Juneau, Alaska. Mature adults that returned from 2009 to 2019 (~8000 individuals) were genotyped at 259 single‐nucleotide polymorphism (SNP) loci for parentage analysis. We used demographic and genetic methods to estimate the effective number of breeders per year (N b). Jack contribution to N b was assessed by comparing values of N b calculated with and without jacks and their offspring. Over a range of N b values (108–406), the average jack contribution to N b from 2009 to 2015 was 12.9% (SE = 3.8%). Jacks consistently made up over 20% of the total male spawners. The presence of jacks did not seem to influence N b/N. The linkage disequilibrium N e estimate was lower than the demographic estimate, possibly due to immigration effects on population genetic processes: based on external marks and parentage data, we estimated that immigrant spawners produced 4.5% of all returning offspring. Our results demonstrate that jacks can influence N b and N e and can make a substantial contribution to population dynamics and conservation of threatened stocks.

array of animal and plant life histories showed that age-at-maturity and adult lifespan, two life-history parameters that impact variance in reproductive success, can dramatically affect variation in N e and N b and N b /N e (Waples et al., 2013).
Despite a rich theoretical framework to incorporate life-history variation into N e and N b calculations (Charlesworth, 1994;Wang et al., 2016), alternative and cryptic life histories are often overlooked because they are difficult to study and complicate analyses (but see Johnstone et al., 2013;Perrier et al., 2014;Saura et al., 2008). Yet, life-history variation is ubiquitous across species, and different life-history strategies impact a variety of critically important population processes and parameters (Charlesworth, 1994;Den Boer, 1968). For example, studies of semelparous plants and animals have shown that variable age-at-reproduction can increase population viability (Acker et al., 2014), population growth rate (Greene et al., 2010), and N e (Watters et al., 2003). Similarly, immigrants can profoundly influence local adaptation (Lenormand, 2002) and increase population persistence over time (Hill et al., 2002).
Consequently, empirical studies where the effects of life-history variation on N b and N e are quantified over multiple generations are particularly valuable.
Fishes are known for their diverse life-history and mating strategies (Avise et al., 2002), and the family Salmonidae has particularly unique and well-studied strategies. Salmonids lay their eggs in riverbeds, after the eggs hatch, individuals rear in freshwater until they out-migrate to brackish waters or the sea for better growth opportunities, and after this most migrate back to their river of origin to spawn. Atlantic salmon (Salmo salar) and several species of Pacific salmon (Oncorhynchus spp.) exhibit alternative life-history tactics: some males reach sexual maturity at smaller size and younger age compared to typical full-sized males that mature after gaining more mass after spending more time in the ocean (Gross, 1985). In Atlantic salmon, males that mature without migrating to sea are referred to as "mature male parr." In populations of Coho, Sockeye, and Chinook Salmon, anadromous males that return to spawn after less time at sea than full-size males are referred to as "jacks." Jacks and mature male parr can outnumber full-sized males during reproduction, but their relative abundance varies greatly across species, populations, and years.
Although most genetic studies of closely related Atlantic Salmon do not explicitly consider mature male parr in N e and N b estimates because they are difficult to sample (Ferchaud et al., 2016), where their contributions have been quantified (e.g., Perrier et al., 2014;Saura et al., 2008) or inferred (Johnstone et al., 2013), they substantially increase N b and N e . Saura et al. (2008) found mature male parr decrease variance in male reproductive success, have little effect on the reproductive success of females, and increase overall N e by 2-3 times without affecting the N e /N ratio. Similarly, Perrier et al. (2014) found mature male parr increase N b by roughly twofold, even though they had lower reproductive success. In contrast, mature male parr contributions to N e have been much smaller in experimental (Garcia-Vazquez et al., 2001;Jones & Hutchings, 2001 and simulation-based (Palstra et al., 2009) studies of Atlantic Salmon. Based on these results, we predicted that jacks would at least somewhat increase Auke Creek Coho Salmon N e and N b .
These data suggest the alternative jack life history can contribute significantly to population genetic processes in wild Pacific salmon populations. Yet, their contributions to N e in natural populations are poorly understood, probably because they are more difficult to detect and count accurately due to their small size and because of their small economic value relative to full-sized individuals. To our knowledge, there has been no multi-generational study to quantify jack contribution to N b and N e using individual adult-to-adult reproductive success from all individuals in a population of Pacific salmon. It is important to understand the effect of variable life histories on N b and N e because such variability can further complicate the relationship between census size and effective population size, complicating management for long-term population viability.
Besides diverse male mating tactics, another defining feature of Pacific salmon life history is the return of most adults to their natal stream to spawn, termed "homing." Homing of adults to their natal stream facilitates adaptation to local spawning conditions, but some individuals fail to return home, instead "straying" into other spawning streams. For a given population, reproductively successful strays from other streams are effectively immigrants that add adaptive, maladaptive, or neutral genetic variation and will affect population genetic processes over time. As straying increases, local genetic processes will increasingly reflect regional processes and effective sizes (Kimura & Maruyama, 1971;Palstra & Ruzzante, 2008;Ryman et al., 2019). However, estimates of the frequency and reproductive success of strays vary greatly among salmon species, regions, and other factors, and are poorly understood and rarely quantified in wild populations.
We used a long-term and detailed demographic and genetic dataset to determine the contribution of jacks to the effective number of breeders and effective population size over multiple generations in a naturally spawning population of Pacific salmon in Southeast Alaska. To achieve this, we quantified adult-to-adult reproductive success of Auke Creek coho salmon over 11 years.
We determined the percentage of successful reproducers that had strayed into this population, based on comprehensive genetic sampling of returning adults and marking of out-migrating smolts (juveniles switching to marine habitats). Our specific objectives were to (i) quantify the contribution of jacks to N b and N e using demographic and genetic methods; (ii) compare N b and N e estimates among methods; (iii) estimate age-specific contributions to N b and N e using life tables, and (iv) quantify the reproductive success of immigrants into this population. We found that jack life history contributes substantially to N b and N e . Immigrants were among the adults that reproduced successfully and their contribution influenced N e and likely the population genetic dynamics of the population. We discuss the implications of our findings for small and declining populations of Pacific salmon.

| Study population
This study was conducted on the Coho Salmon population in the Auke Lake drainage in Juneau, Alaska, USA. Tissue samples and demographic information were collected at a permanent two-way weir located between Auke Lake and Auke Bay that is operated by the National Oceanic and Atmospheric Administration (NOAA). Counts of out-migrating smolts and returning adults have been recorded annually since 1980 and genetic samples have been collected since 2009. The weir catches all out-migrating smolts and the smolts are released downstream after their adipose fin is clipped and they receive a coded-wire tag. The weir also captures all returning adults, which can be identified as originating from Auke Creek by the absence of an adipose fin. As each adult is manually transported over the weir, an axillary process is removed and stored in 95% ethanol and each individual is assigned a sex/spawning type (female, full-size male, or jack) based on morphological characteristics. Roughly one-third of returning adults are sampled for length and age. Adults with intact adipose fins could either be adults that strayed into Auke Creek from other populations or were smolts that avoided capture and tagging at the weir; we used parentage (described below) to distinguish between these two possibilities.
This study did not require ethics approval because we exclusively analyzed archived tissue samples collected by NOAA and not live fish. This study used data from adults returned between 2009 and 2019. Auke Creek Coho Salmon eggs are laid in the fall and after hatching in the spring, individuals spend either 1 or 2 years rearing in freshwater before smolting and migrating to sea. Jacks spend 6 months and full-size females and males spend 1 full year at sea before returning to spawn (Table 1). Females and full-sized males return 3 or 4 years after being spawned (this includes time spent in the gravel, 1 or 2 years in freshwater, 1 year at sea, and returning to spawn). Jacks return 2 or 3 years after being spawned (this includes time spent in the gravel, 1 or 2 years in freshwater, 6 months at sea, and returning to spawn). Males that return after 6 months at sea, regardless of freshwater age, are referred to as jacks and are clearly smaller than males that spent a full year at sea. Age classes are ab- Note: Labels and cell shading denote durations of life stages. Eggs are laid in the gravel in the fall and hatch the following spring. All individuals spend the next 1 or 2 years rearing in freshwater. After smolting, jacks spend 6 months at sea before returning to spawn at age 2 or 3 years old. Full-size individuals spend 1 full year at sea before returning to spawn at age 3 or 4 years old.

| Genotyping and parentage assignment
Genotyping and parentage assignment methods were the same as those presented in King et al. (2023). In brief, we employed a contract lab (GTseek LLC) to extract and sequence tissue samples of all returning individuals from 2009 to 2019 using the "genotypingin-thousands" (GTseek LLC) protocol (Campbell et al., 2015). The single-nucleotide polymorphism (SNP) panel used, composed of 259 loci, was developed specifically for Coho Salmon parentage by the Columbia River Inter-Tribal Fisheries Commission (Hess et al., 2016).
We used the program FRANz (Riester et al., 2009) to assign parentage for adult offspring from each individual return year from 2013 to 2019. We constrained the set of possible parents for each year to fish that returned 2-4 years prior. Sex was not used in parentage assignment due to uncertainty in field identification. For each FRANz run we used the following parameters: N max was calculated by multiplying the number of potential parents, as enumerated at the weir, by (1.1)/2 (half of the potential parents with a 10% buffer); genotyping error rate was assumed to be 0.01 (default FRANz value); the allowed maximum number of mismatching alleles was five for dyads and seven for triads; and the minimum loci typed per individual was 150 (out of 251). Parentage assignments with posterior probability <0.9 were discarded. We expected assignment error to be small because we genotyped such a large proportion of the population and the panel had 122 loci with minor allele frequency >0.25.
We determined whether fish captured in 2013-2015 with intact adipose fins were likely strays (i.e., individuals that assigned with high confidence to no Auke Creek parents) or were unmarked locals (i.e., individuals that assigned with high confidence to at least one parent from Auke Creek). We then quantified the percent of returning adults from stray parents that spawned in 2013-2015. For additional details, see King et al. (2023).

| Demographic estimate of N b and N e
Parentage assignment allowed direct calculation of the inbreeding effective number of breeders (N bD ) in the population each year, and from this demographic information an inbreeding effective population size (N eD ) was calculated for each generation (for full notation see Table 2). N e estimates apply over a generation (4 years for Auke Creek Coho Salmon), while N b estimates are annual. First, we calculated N bD of each sex for each return year from 2013 to 2019 using population size (N), mean number of offspring (k), and variance in the number of offspring per individual (V k ) with the following equation (Crow & Denniston, 1988;Crow & Kimura, 1970): After this, we calculated the net N bD for each year [i] using the male and female effective number of breeders with the following equation (Crow & Kimura, 1970;Wright, 1931): Then N eD values for different generations were estimated using N bD estimates from the corresponding years in each generation: where X i is the proportional contribution of breeders from year i to the next generation (Ryman & Laikre, 1991).
Coho Salmon return to Auke Creek up to 4 years after being spawned, so for each generation we calculated four N eD values using overlapping N bD values with a sliding scale of 4 years over 11 years (2009-2012, 2010-2013, 2011-2014, and 2012-2015). Individuals that were not successfully genotyped were included in the calculations of sex-specific N b and were assumed to have the same mean and variance in reproductive success as those successfully genotyped of the same sex. In 2010, of the 816 individuals that returned, 109 were not assigned a sex. In each of 400 iterations, we randomly assigned the unsexed fish a sex (given a 50:50 sex ratio) and then To examine the sizes of N bD and N eD relative to the population size, we calculated two ratios. In N bD /N, N represents the number of individuals returning to spawn in a single year, whereas N in N eD /N represents the total number of individuals returning over the entire generation (which includes multiple spawning/return years).
To investigate the specific contribution of jacks to N bD , we also calculated N bD using the same dataset but removing all jack parents and their offspring (after Perrier et al., 2014;Saura et al., 2008). Jack contribution was defined in the following way: Jack contribution represents the percent difference in annual estimates of N bD when jacks are excluded from calculations. Positive values would indicate that the presence of jacks increases N bD .
We made three choices when excluding jacks from the dataset.  (Do et al., 2014) and the temporal method of Waples et al. (2007) in the program SALMONNb (N bS ). For N bLD and N bS , we analyzed individual cohorts (individuals born in the same year) for each year 2009 to 2015. Individuals were grouped into cohort years using parentage-based aging. Bias in the estimates of N bLD was corrected using the haploid number of chromosomes (Waples et al., 2016), which for Coho Salmon is 30 (Uyeno, 1972

| Demographic estimate for Auke Creek Coho
We created a multi-year, multi-generation pedigree from genotypes collected from Coho Salmon captured at the Auke Creek weir in 2009-2019. These data allowed us to assign offspring to parents, to determine the reproductive success of jacks and full-sized adults, and to determine the proportion of offspring produced by immigrants into this population from elsewhere. The mean annual jack contribution to N bD was 12.9% (SE = 3.8%), and closer inspection of their contribution to N bD shows they increased N bD via malespecific values, N bDm , with little impact on female-specific values, N bDf (Table 3). The mean annual jack contribution to N bDm was 23.9% (SE = 6.3%), and jacks increased N bDm in every year except 2012.
In contrast, the mean annual jack contribution to N bDf was 0.7% (SE = 2%). In two of 6 years, N bDf without jacks was greater than N bDf with jacks because excluding jacks decreased the female variance in reproductive success in those years which increased N bDf . In 2012, N bDf without jacks was larger than N bDf with jacks and N bDm without jacks did not change, so the overall N bD without jacks (174) was larger than N bD with jacks (170) and jack contribution was negative (−2.9%). N bD /N did not notably change when jacks were excluded because jacks affect both the numerator and denominator of this ratio.
N bDf each year ranged between 50 and 168, while N bDm ranged from 60 to 260. In all 6 years analyzed, N bDm exceeded N bDf .
Between 0.84% and 9.76% of returning adults each year did not have a fin clip, implying that they may be strays into Auke Creek.
Of these, 168 (63.9%) were likely strays (assigned with high confidence to no Auke Creek parents). Most strays were not successful at producing offspring, but strays produced 4.5% of all offspring from 2013 to 2015 that returned to Auke Creek as adults in subsequent years (through 2019).

| Using life tables to estimate N e
The composite life tables for males and females created from our demographic and reproductive success data from 2013 to 2015 revealed that the overall generation length was 3.72 years (Male: 3.61, Female: 3.84), N bLT was 564 (Male: 306, Female: 261), and N eLT was 2099 (Table 4). Full-size males contributed the most offspring, but the contribution from jacks was still substantial (jacks produced 17.3% of offspring in the table). Most of the jack contribution was from age-3 jacks, as age-2 jacks were less common (5.5% of jacks were age-2). Overall, jacks had smaller mean and variance in reproductive full-size males (Age 3: b x = 0.5 V x = 1.24, Age 4: b x = 0.5 V x = 1.58), but similar mean-adjusted variance. Age-4 females had similar reproductive success (b x = 0.55) to age-3 females (b x = 0.53), but their higher abundance resulted in a greater contribution of offspring than age-3 females.

| DISCUSS ION
Our study yielded several important insights into the influence of life-history variation on N e . First, jacks made a substantial contribution to N e , but their contribution varied widely across years. Jack

| Effect of jacks on N e
The presence of jacks in this population generally increased N b (male and overall), but the magnitude of the contribution was highly variable from year to year. Removing jacks decreased the number of males and typically increased the variance in reproductive success which lowered N bm . Reductions in N bD were primarily caused by the smaller numbers of individuals because removing jacks had F I G U R E 1 Auke Creek Coho Salmon population size including unknown sex/type adult individuals (N) and effective population size (N eD ) for four overlapping generations calculated using yearly effective number of breeders values and the proportional contribution of offspring from each year.

F I G U R E 2
Inbreeding effective number of breeders over time using the demographic method (N bD ), linkage disequilibrium method (N bLD ), and SALMONNb method (N bS ) for Auke Creek Coho Salmon from 2009 to 2015. 95% CIs are included (for N bLD and N bS ).

TA B L E 4 Female and male composite life tables for Auke
Creek Coho Salmon, created using individuals of known age and genotype that returned from 2013 to 2015.

Sex Age class
Age Note: All age-3 males are pooled together (jacks and full-size males). Abundance (N x ), proportion (N x /N), mean number of offspring produced (b x ), total number of offspring produced (B x ), and variance in number of offspring produced (V x ) are shown for each age class. For a stable population, the mean number of offspring (b' x ), total number of offspring produced (B' x ), and variance in number of offspring (V' x ) are noted with the prime symbol (').
little effect on N b /N (the difference between the two estimates was consistently <0.05). Similarly, a study on Atlantic Salmon found that including mature male parr increased N e without changing N e /N (Saura et al., 2008). Although there is no consensus on the typical range of N e /N values for natural populations of plants and animals (Waples, 2002), the N eD /N for Auke Creek Coho Salmon (0.08-0.18) is comparable to the values reported for a wide range of species. Frankham (1995) first reported that N e /N averaged ~0.11 over 100 species and a more recent review by Palstra and Ruzzante (2008) found a median N e /N of 0.14.
We found little effect of removing jacks on N bDf , which aligns with a study on Atlantic Salmon that found that mature male parr exclusion or inclusion had a negligible impact on female N e (Perrier et al., 2014). We did observe 1 year when N bDf was larger without jacks because removing jacks decreased the variance in female reproductive success and N bDm did not change after removing jacks, so overall N bD without jacks was slightly larger than with jacks. The lack of influence on N bDf suggests that spawning success of females in this population is not male limited. Females are the less abundant sex and 88% of females with at least one identified mate had one or more full-size male mates.

| Comparison across estimates
All N b estimates showed a similar temporal pattern and reflected the dynamics in the number of returning adults. While the N bLD values were slightly lower than the N bD values, they followed the same trend for the entire series, which supports the potential of N b estimates to serve as a population monitoring tool (Charlier et al., 2012;Luikart et al., 2021;Tallmon et al., 2010). Other studies have also found close relationships between demographic and LD estimates of N b , yet comparisons on this scale are still relatively rare. A study of the impact of Atlantic Salmon mature male parr on N b found that the demographic estimate (220) and LD estimate (198)  could explain why N bLD is slightly lower than N bD and why N eLD was smaller than N eD , but the influences of immigration on N b and N e are complex. Sampling individuals from divergent subpopulations or following pulses of migration can decrease N bLD estimates (Waples & England, 2011;Whiteley et al., 2017). Ryman et al. (2019) found that even one migrant per generation can cause differences among the N e estimates of commonly used methods in a modeled population with discrete generations. Identifying the source of strays and the size, structure, and migration rates of regional subpopulations of Coho Salmon in Southeast Alaska will improve our understanding of how immigration affects our various N b and N e estimates.

| Conservation and evolutionary implications
Monitoring  (Ulaski et al., 2020;Watters et al., 2003). Coho Salmon have less variability in total age than Chinook and Sockeye Salmon, potentially making them more vulnerable to longer lasting and more frequent natural and anthropogenic disturbances . Jacks, by spending less time at sea, might help compensate for low full-size male returns in periods of high marine mortality. There is accumulating evidence that lifehistory variation provides resilience and stability in natural populations (Moore et al., 2014;Munsch et al., 2022;Watters et al., 2003), but this life-history variation affects average reproductive success and its variance, further underscoring the need to monitor N b and N e in populations of conservation concern. The contribution of multiple cohorts to an individual return year reduces the effects of variable recruitment among years. Perrier et al. (2014) found that Atlantic Salmon mature male parr increased N b and allelic richness and hypothesized that the presence of mature male parr may compensate for low amounts of returning full-size males and potentially buffer declines in N e . Jacks, by spending less time at sea, might help compensate for low full-size male returns in periods of high marine mortality. In this population, jacks increased N b in most years.
Although our results show jacks tend to increase N b , the ability of jacks to bolster N e across a range of abundances depends upon complex relationships between N, N b , and the frequency of jacks relative to full-size males. Fluctuations in annual recruitment can result in jacks from larger cohorts mating among full-size males from smaller cohorts, which increases jack representation (DeFilippo et al., 2019). Harvest of full-size males may also increase the frequency of jacks (Young et al., 2020). On an individual basis, jacks tend to have lower reproductive success at higher jack frequencies (Berejikian et al., 2010), which may attenuate their contribution to From a conservation perspective, the N e and immigration data suggest it is important to consider Auke Creek Coho in a regional context. The N e values we observed of at least a few hundred per generation suggest that a loss of genetic variation is not a shortterm threat to Auke Creek Coho Salmon. We also identified that strays (immigrants) contributed 5% of returning adult offspring.
Strays will add genetic variation that can be adaptive, maladaptive, or neutral, but we expect local genetic variation will be maintained at higher levels than predicted strictly from the local N e estimates above to reflect the N e of populations in the region linked by migrants (Palstra & Ruzzante, 2008;Ryman et al., 2019). Because this Coho Salmon population has rapidly shifted its run timing over the past four decades, presumably in response to changing environmental conditions (Kovach et al., 2013), it would be helpful to know whether continued straying will aid this population's future adaptation. It is clear from our results that valuable management and evolutionary insights can be gained from using multi-generation genetic data to infer the contributions of life-history variants, such as small males and strays, frequently ignored in studies of wild populations.

CO N FLI C T O F I NTE R E S T S TATE M E NT
We declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data and scripts that support the findings of this study are openly available in Dryad at https://doi.org/10.5061/dryad.2280g b5z6. The raw data that supports the findings of this study are available in the supplementary material of King et al. 2023