Chinook and Coho salmon hybrids linked to habitat and climatic changes on Vancouver Island, British Columbia

Abstract Between 2013 and 2019, 63 presumed Chinook salmon Oncorhynchus tshawytscha sampled primarily in the Strait of Georgia (0.63% of total sample) were identified as potential Chinook–Coho (Oncorhynchus kisutch) hybrids by the presence of anomalous microsatellite genotypes. Their hybrid origin was confirmed by single nucleotide polymorphism amplification of two species‐specific amplicons. Mitochondrial DNA indicated that most of these fish resulted from the hybridization of Coho salmon females and Chinook salmon males. Although no diagnostic external features were identified, several individuals displayed an abnormal scale arrangement on the caudal peduncle. One hybrid juvenile examined for meristics exhibited a pyloric caeca count intermediate between published values for Chinook and Coho salmon. Most hybrids originated in the Cowichan River during the 2014 brood year. Their prevalence in the watershed is a naturally occurring event, likely exacerbated by prolonged low water levels which limit habitat and delay Chinook salmon spawning, in addition to the differential abundance of the parental species. This research is the first to document ongoing natural hybridization (Chinook–Coho salmon crosses) and link it to habitat and climatic changes, and includes the identification of eight F1 adults and two juvenile backcross or F2 hybrids. The potential negative impacts of hybridization, particularly in Coho salmon through potential introgression, warrant hybrid identification as an ecosystem monitoring tool within a survey program.

Hybrids are important from an ecosystem monitoring perspective given that they tend to occur under circumstances that suggest rapid environmental change; determining their frequency and the conditions of their appearance could help scientists and managers take action to minimize negative ecological outcomes.
Pacific salmon occupy freshwater habitats subject to increasing human habitation, anthropogenic modification, water extraction, and climate change (Mantua et al., 2010). These factors place salmon at risk of "reverse speciation" and introgression, as environmental change weakens ecologically mediated reproductive isolation (Owens & Samuk, 2020). Negative consequences of introgression include outbreeding depression, lower performance in parental environment for traits such as feeding and predator avoidance (Wessel et al., 2006), early life history mortality (Bartley et al., 1990), and susceptibility to infectious diseases (Hedrick et al., 1987). In summary, highly hybridized populations have an elevated risk for extirpation or extinction (Rhymer & Simberloff, 1996).
Natural hybridization between Chinook (Oncorhynchus tshawytscha) and Coho salmon (Oncorhynchus kisutch) has been documented but is rare (Bartley et al., 1990;Chevassus, 1979;Johnson & Ringler, 1981). Its infrequency has been attributed to prezygotic isolation mechanisms that prevent viable development (Bartley et al., 1990). Utter et al. (1989) found no evidence of Chinook-Coho salmon hybrids in a Chinook salmon population structure study of 86 collection sites from the Babine River in British Columbia to the Sacramento River in California. However, Chinook-Coho F1 (first-generation) hybrids have been detected in the wild (Bartley et al., 1990;Verspoor & Hammart, 1991) and viable and fertile hybrids are readily produced in laboratory settings (Argue & Dunham, 1999;Blanc & Chevassus, 1982;Foerster, 1935;Hedrick et al., 1987;Seeb et al., 1988;Smirnov, 1972). McKenzie et al. (2021) found no evidence for strong prezygotic isolation in noncompetitive heterospecific mating trials and in vitro fertilization experiments performed with Chinook and Coho salmon, although hybridization did not occur in mating trials in which a conspecific mate was present.
Few laboratory studies have addressed F2 (second generation) and backcrossed individuals resulting from the reproductive viability of F1 hybrids. Chevassus (1979) and references therein reported male fertility and moderate hatching success for F2 individuals. Devlin et al. (2001) reported excellent prehatch survival for eggs produced from an F1 male (from a Coho male by Chinook female cross) backcrossed to a Coho female. In natural environments, evidence is lacking on both the reproductive viability of hybrids beyond the F1 generation and the associated interspecific introgression that occurs between salmonids (Allendorf et al., 2001;Scribner et al., 2001).
Hybrids are often identified opportunistically in the context of other studies, based on atypical morphological characteristics or through genetic analysis. The latter is especially useful in the case of "cryptic" hybridization, which results in minimal morphological or meristic alteration (Allendorf et al., 2001;Scribner et al., 2001).
In particular, the genotyping of codominant Mendelian markers at multiple loci, including those with alleles shared between the species, enables both the early detection of hybridization in closely related species and the estimation of proportions of first and second generation, as well as backcrosses to parental species (Anderson & Thompson, 2002). Inclusion of diagnostic single locus polymorphisms (SNPs) in multilocus genomic analyses facilitate hybrid detection in studies that may involve extensive sampling of individuals (Beacham & Wallace, 2020) and the application of many such SNPs enables the estimation of genome-wide levels of introgression (Hohenlohe et al., 2011). Finally, examination of species-specific mitochondrial DNA sequences can identify the maternal species involved in hybridization events (Karlsson et al., 2013), thus providing behavioral clues of the matings.
The hybrids in this study were detected inadvertently by the presence of anomalous microsatellite genotypes during day-today genetic stock identification (GSI) analyses at the Fisheries and Oceans Canada -Molecular Genetics Laboratory (MGL) located at the Pacific Biological Station. GSI is used to assign mixed stock samples to a population of origin employing a baseline of genotypes and has been performed routinely in fisheries management and research contexts for four decades (e.g., Grant et al., 1980). In the current study, the hybrids were captured during the first summer of marine residence near the southern BC coast (Figure 1) as part of a tagging study to investigate survival of Cowichan River Chinook salmon (Pacific Salmon Foundation, 2017).
The Cowichan River drains a 930-km 2 watershed on southern Vancouver Island. Flow is partially controlled by a weir at the outlet of the Cowichan Lake. This system has been subject to extensive channel modifications, as well as surface and groundwater diversions. The Cowichan River has a recovering population of fallrun Chinook salmon and supports a historically large run of Coho salmon, the current status of which is poorly monitored. Natural spawning of fall Chinook salmon is supplemented by a hatchery operated by the Cowichan Tribes with the support of Fisheries and Oceans Canada. The Cowichan Lake Salmon Enhancement Society (CLSES), a volunteer-driven organization that rescues Coho salmon fry from becoming trapped by droughts during the summer months, also operates a small hatchery which supplements Chum salmon (Oncorhynchus keta).
The main objective of this study was to understand the biological and environmental processes behind the observed hybridization. To achieve this, we conducted a retrospective examination of the anomalous genotypes encountered in presumed Chinook salmon sampled primarily near the Cowichan River in the Strait of Georgia. We confirmed their hybrid status and determined if they had arisen through natural or hatchery spawning events. We applied molecular analyses of nuclear and mitochondrial DNA sequences to confirm the maternal species involved in hybridization, and estimated the numbers of potential parents and the proportions of first (F1) and second or higher generation (F2). We investigated if hybrids confirmed by molecular techniques exhibited consistent meristic traits that would allow rapid identification.
We suggested the Cowichan River watershed as the predominant source of these hybrids and examined habitat and climatic alterations that have occurred in recent decades and their effects on spatiotemporal overlap spawning between Chinook and Coho salmon spawning.

| ME THODS
The samples in this study originated from a variety of sources, which prompted the creation of several databases (Table 1). Aspects such as number of individuals analyzed, category (e.g., hybrids vs. pure breeds), origin (e.g., surveys vs. hatchery origin), and the quantity and type of genetic markers used (e.g., microsatellites or SNPs) were employed depending on the purpose. Although not directly related to the genetic analyses, we include two sections on morphology and hydrology to examine meristic traits that can help screening Chinook-Coho hybrids and provide environmental clues about the cospawning of the two species involved. The following organizations provided the samples: Fisheries and Oceans Canada (DFO), the BC Conservation Foundation, the Cowichan Tribes, the CLSES, and the North West Indian Fisheries Commission in the United States.

| Study subjects
Between 2013 and 2017, 62 presumed Chinook salmon sampled in coastal waters of British Columbia by a variety of surveys were identified as potential Chinook-Coho salmon hybrids (Table 2; Figure 1).
These fish exhibited unusual genotypes and "stutter patterns" or peak morphologies observed during DNA amplification of 15 microsatellite loci (Table 3), which are used routinely to perform GSI in the identification of mixed stock samples. The hybrid samples were observed from a total mixed stock sample of 10,003 fish (0.63%). The majority of the hybrid samples were juvenile fish captured by microtrolling (Duguid & Juanes, 2017) in marine waters near the Cowichan River as part of a passive integrated transponder (PIT) tagging study to investigate survival of Cowichan River Chinook salmon (Pacific Salmon Foundation, 2017). Additional samples were collected in routine juvenile surveys in the Strait of Georgia by Fisheries and Oceans Canada and in creel monitoring in the same area and elsewhere in British Columbia. Two juveniles (fry) were collected within the Cowichan drainage during a fry salvage operation by the CLSES.
An additional sample (# 63) was an adult tagged in marine waters near Cowichan Bay in September 2019 during a project to investigate pinniped predation on returning Cowichan River Chinook salmon. All hybrid samples were screened for coded wire tags (CWTs) or adipose fin clips that would indicate hatchery origin. More than 90% of Chinook salmon juveniles released from the Cowichan River hatchery in the years pertinent to this study had been fin-clipped as part of the Salish Sea Marine Survival Project (K. Pellett, unpublished data).

| Hybrid identification microsatellites
Hybrid fish were detected in microsatellite analyses of presumptive Chinook salmon with a standard set of 15 loci for the species F I G U R E 1 Locations of Chinook-Coho hybrid captures. Empty circles indicate juveniles, full circles adults. Survey details are provided in Table 2 ( Table 3). Most samples had been collected nonlethally and consisted of fin clips affixed to Whatman paper or preserved in 95% undenatured ethanol or scale samples on gummed scale books. DNA was extracted using a chelex resin protocol (Small et al., 1998)  Hybrid fish were identified by the presence of both Chinook and Coho salmon alleles, which presented anomalous peak morphologies among the Chinook salmon microsatellites used in their original screening (out of the allelic range expected for the species and presenting alleles corresponding to Coho salmon instead; see Table 3).
In addition, Coho and Chinook salmon have distinctive 2-and 4-bp "stutter" patterns, though with overlapping allele size ranges at the Oki100 locus. Many hybrids were initially recognized by the presence of one allele with the Coho 2-bp stutter pattern. Because of their interspecific nature, hybrid fish could not be assigned to a population of origin by traditional GSI methods which require a baseline of genotypes from pure species individuals.
Pure species Coho salmon were not genotyped with the Chinook panel of 15 microsatellites as a control group because this has been done multiple times during regular GSI analyses at the DFO MGL, which results in informative allelic frequencies for the set of shared six loci used in this study, but "out of range" (e.g., allelic frequencies not available) for the set of nonshared loci. As the hybrid samples were detected using the microsatellite panel (used in day-to-day operations in the MGL as the default GSI method), and given the small number of markers, we added a SNP panel to validate their hybrid identification.

| SNP species ID
The SNP panel is composed of two amplicons and five diagnostic SNPs. DNA amplicons OkiOts_120255 and Oki_RAD41030 have SNP sites fixed for alternate base pairs in Chinook and Coho salmon (Beacham & Wallace, 2020). OkiOts_120255 has a forward primer sequence of TGGAGTTGACAAAACATCCGATGTC and a reverse primer sequence of CCAGCAGACAGTCATCCTAAAAGAAA (Starks et al., 2016). The forward and reverse primers for  Table 2  The main purpose of this procedure was to inspect the MGL genetic baseline for errors in broodstock identification, potentially suggesting human-induced hybridization.

| Filial generation and family structure
A subset of six microsatellite loci that amplified well in both species and had little or no overlap in allele size ranges ( 3). We employed the software package NewHybrids (Anderson & Thompson, 2002) to determine the most probable identity (F1, F2, backcross to Chinook, or backcross to Coho salmon). It is important to clarify that although NewHybrids has the ability to be distinguished between F2s and backcrosses, it is difficult to assign an individual to either category with certainty, and the margin of error may be too large to reach a definitive conclusion without sufficient loci showing extreme differences in allele frequencies (Anderson & Thompson, 2002), which is not the case for the sister species involved in the hybridization. The analysis was conducted for two million sweeps, without the training option indicating pure individuals of each species and assigning posterior probabilities of filial generation in hybrids.
We employed Colony 2.0.6.5 (Jones & Wang, 2010) to identify sibling relationships in the absence of parental genotypes. The chosen method assumed polygamy in both parents and applied full likelihood without parental genotypes and an error rate of 0.01. An analysis of the hybrid fish alone allowed us to estimate the number of Chinook and Coho salmon (or F1 hybrid) parents involved in their production and simulate their genotypes. In addition, we investigated if the Chinook salmon parents involved in the hybridization also produced purebred progeny with conspecific mates using the 540 juvenile samples.

| Mitochondrial DNA
We employed mitochondrial DNA analyses to determine the maternal species involved in the hybridization and obtain behavioral clues of the matings. We conducted qPCR of the cytochrome c oxidase subunit 1 in Chinook salmon (Laramie et al., 2015) and of TA B L E 3 Microsatellite locus suite run for Chinook salmon in the DFO MGL and allelic size ranges found in Cowichan River Chinook salmon for each locus, as well as ranges in Coho salmon and putative Chinook-Coho salmon hybrids cytochrome b in Coho salmon qPCR assay (Pilliod & Laramie, 2016) on each hybrid fish, in addition 40 control samples of pure Coho and Chinook salmon (20 of each) were added. Amplification was conducted with Taqman probes labeled with 6FAM at the 5′ end and a minor groove binding nonfluorescent quencher at the 3′ end (MGB-NFQ, Life Technologies). These Coho and Chinook salmon assays amplify in other salmonids at a reduced sensitivity (>1/25), so it is necessary to run both assays for complete maternal species identification. Additionally, we developed a Chinook salmon cytochrome b qPCR assay to confirm the results. The qPCR panel is presented in Table 4.
Reactions for qPCR analysis consisted of 12 μl total volume.  Table S1).

| Morphology
Most hybrids were identified by genetic analyses after the physical specimens were no longer available for inspection, having been released as part of tagging studies or consumed by fishers. However, photographs were available for six juvenile hybrids captured and tagged as part of the Cowichan Chinook PIT tagging study and for the adult fish captured as part of the pinniped predation study in 2019 (Table 2). A single hybrid juvenile captured during the Cowichan PIT tagging in 2017 (#62 in Table 5) was retained and dissected in the laboratory for counts of pyloric caeca and branchiostegal rays.

| Hydrology and spawning
We looked at relationship between high and low flows and the timing of spawning in the Cowichan River for the species involved in the hybridization. Prespawning salmon (particularly those that enter the rivers in the late summer months) tend to wait until higher flow events facilitate their passage, obtain optimal temperature, and reach adequate dissolved oxygen levels (Bjorn & Reiser, 1991). This information allowed us to make inferences about the amount of habitat available and explore potential delays in the time of spawning, which provides clues on the convergence of the two species involved in the hybridization, with Chinook salmon historically spawning in the late summer and Coho salmon in the fall.
We obtained monthly discharge data from the Cowichan River

TA B L E 5 (Continued)
and spawning migration time data were provided by the South Coast Stock Assessment Section captured at the fence operated by

Fisheries and Oceans Canada. We estimated the weighted average
Julian day of Chinook salmon counts by year, the Julian day where the peak of the migration occurs, and the mean run Julian day considering the beginning and end of the run independent of the number of counts. Chinook-Coho sampled in the marine environment and in relation to total samples identified by GSI. These results are not meant to be interpreted as hybridization rates, given the variable sample size and that these surveys were not originally designed to investigate hybrid occurrence. The intention of this information is to put the observed hybridization in context and present a general idea of its frequency:

| RE SULTS
(1) independently of origin and (2) assuming that hybrids belong entirely to the cohorts originating in the Cowichan River. GSI results are probabilistic and thus the percentages presented in Table 2 may be subject to biological interpretation (e.g., the composition of the mixture employed and the baseline used in the analyses). Putting this caveat aside and considering only those surveys near the Cowichan watershed, the proportion of hybrid fish in the sample ranged between 0.57% when considering hybrids independent of the origin and 4.92% when assuming that hybrids originated in the Cowichan River.

| Hybrid identification with microsatellites
All the 63 fish that were originally detected as potential hybrids in the GSI microsatellite analyses were confirmed as being the offspring of Chinook and Coho salmon at some or all of the loci surveyed. In the 54 hybrids with complete genotypes for the subset of six loci used in subsequent analyses (Table 1, No. 3), the Chinook salmon alleles carried by the hybrid progeny were those characteristic of the Cowichan Chinook salmon population, and ranged in number from 5 to 19 among loci. Alleles unambiguously identifiable as of Coho salmon origin in the hybrids ranged from 2 to 19 among loci.

| SNP species ID
Of the 63 hybrids, 50 were successfully amplified at the speciesspecific SNP amplicons (Appendix Table S2). All 50 were heterozygous for a Chinook and Coho haplotype at both SNP loci, confirming these as the parental species involved in the hybridiza-

| Filial generation and family structure
In the NewHybrids analysis (

| Mitochondrial DNA
Mitochondrial DNA amplification was successful for 45 hybrid fish This individual did not exhibit abnormality of the caudal peduncle.
The most anterior ray of the anal fin did not reach the attachment point of the most posterior ray when folded against the body. The branchiostegal count was 16, while the pyloric caeca count was 95.
A Cowichan River origin Chinook salmon sampled lethally on the same day (141 mm NFL) had a branchiostegal count of 16 and a pyloric caeca count of 137.

| Hydrology and spawning
Mean Chinook salmon run timing by Julian day, independent of the run size, has not changed significantly from the historical averages

| Analytical considerations in hybrid identification
We have documented the repeated occurrence of natural hybridization between Chinook and Coho salmon. The frequency of hybrid occurrence varied depending on context (0.57% independently of their origin and 4.92% assuming origin in the Cowichan River). Given the opportunistic nature of this study, these values can only give us a general idea of hybrid presence, because the surveys in which hybrids they were encountered were not originally designed to target hybridization rates, lacking control variables and in-river sampling.
Despite this caveat, the hybridization observed may be an underestimate, as some hybrid fish in the surveys may have been identified as pure species Coho salmon, which were not subject to GSI.
Although the hybridization observed in this study is most likely natural, it was important to note that neither all Cowichan Chinook salmon broodstock were genotyped nor all the hatchery fish were marked or coded wire tagged (in recent years an average of 90%); therefore, human error is not entirely out of the question. However, given the reoccurrence of hybridization in multiple years, the number of parents involved, and that no hybrid or Coho salmon were genetically identified in the screened of hatchery broodstock, the observed hybridization remains very likely a nonhuman-mediated event.
Although we intended to investigate if hybrids confirmed by molecular techniques could be identified by consistent meristic traits for rapid identification, due to the small sample size and the release of animals, our findings remain preliminary. Most hybrids were identified by genetic means among nonlethally sampled presumed Chinook salmon. Some hybrids could be identified visually by a general appearance intermediate between the parental species or by a disordered arrangement of scales on the caudal peduncle and associated lateral line deflection or disappearance (Figure 2). The pyloric caeca count for the single dissected hybrid (Fish 62, Table 5) was intermediate to the ranges reported for Coho and Chinook salmon by Clemens and Wilby (1961) and consistent with a report of intermediate counts in three natural Chinook-Coho salmon hybrids from a Lake Ontario tributary (Johnson & Ringler, 1981). A similar intermediate pyloric caeca count occurred in hybrids between lake trout and brook trout (Scott, 1967). Morphological and meristic characteristics of salmonid hybrids may be intermediate to, equal to, or exceed parental values and do not always provide reliable hybrid identification (Scribner et al., 2001).

| Mechanism of hybridization
We suggest that the observed hybridization has been facilitated by prolonged low water levels in the Cowichan River, which limit habitat and promote later peak spawning of Chinook salmon, allowing the two species to cospawn.
Summer water levels in the Cowichan River watershed have declined significantly from the 1960s to the present as a result of F I G U R E 3 Hydrological and abundance trends. Weighted average Julian day of Chinook salmon counts by year at the Cowichan River fence operated by Fisheries and Oceans Canada, the Julian day where the peak migration occurs, and the mean run Julian day considering the beginning and end of the run independent of the number of counts (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018). Abundance (Ab) and spawning migration time data were provided by the DFO South Coast stock assessment section (K. Pellet, unpublished data). Hydrological data (Station 08HA002) were from the Water Survey of Canada (2019) increasing air temperatures, climatic shifts, and human pressures (Cowichan Valley Regional District, 2017;Spittlehouse, 2017). The number of days with low flows (below the 25th percentile) has increased nearly twofold from 1965 to 2015 (Pike et al., 2017). These low-flow events are extending into September and October, during the historical peak Chinook salmon migration (Damborg et al., 2015).
The main stem of the Cowichan River is heavily utilized by spawning Coho salmon in addition to Chinook salmon, which are also occasionally observed in tributaries to the lake (K. Pellett, unpublished data). The upstream Chinook salmon migration peaks at the counting fence from late September to mid-October, while spawning occurs from late October to early November. In contrast, Cowichan River Coho salmon spawn primarily from November through January (K. Pellett, unpublished data).
The  (Newman et al., 2016). Other salmonid species have shown altered migration timing in response to climate change via water temperature and habitat availability (Crozier et al., 2011;Kovach et al., 2015).
Surface water diversions for commercial and residential use have Lake in 1956 for water storage purposes. There is limited hydrological monitoring in smaller streams, but anecdotal evidence supports the possibility that a lack of water availability during spawning reduces habitat causing the two species to overlap more than they did historically. Climate change that alters the habitat available for reproduction may sponsor hybridization between species that previously shared watersheds with little interaction (Garcia de Leaniz & Verspoor, 1989;Muhlfeld et al., 2009;Young et al., 2016).
Mitochondrial DNA indicated that the majority of hybrids were the result of female Coho salmon spawning with Chinook salmon males (all but two individuals from a reciprocal cross). This finding suggested two possible scenarios: accidental fertilization in crowded spawning grounds or heterospecific choice of mate when conspecifics are not available (e.g., differential abundance). If the available spawning habitat is very limited when the species overlap, it is possible that the hybrids are primarily the result of fortuitous fertilization from a "milt cloud" formed by an abundance of jack Chinook salmon attempting sneak fertilizations with Coho salmon females. However, differential abundance is a more plausible explanation.
Chinook salmon abundance in the Cowichan River drainage has increased in recent years. The abundance continues to recover from a record low of 540 adult natural spawners in 2009; 13,975 Chinook salmon were present in 2018. Cowichan River Chinook salmon are supplemented by hatchery production, but the level of hatchery production and proportion of hatchery fish in the watershed has declined over the past decade. The opposite is true for Coho salmon.
Limited assessment work indicates that 20,000 individuals or less may now return to a system that once supported 10 times more (K. Pellet, unpublished data). Although Coho salmon could be relatively more abundant than Chinook salmon throughout the watershed, its abundance may be lower in spawning grounds traditionally shared by both species, such as the river section adjacent to the lake outlet.
A reproductive "bottleneck" may be experienced by Coho salmon females as they attempt reproduction when no conspecific males are available. This phenomenon is characteristic of the depensation or positive Allee effect and commonly observed in animal populations when their numbers dwindle (Gascoigne & Lipcius, 2004;Stephens et al., 1999). McKenzie et al. (2021) documented that females of both species preferentially mated with conspecific males, but would mate with a heterospecific male in the absence of a conspecific choice. Their study also found that the observed differences between fertilization rates of Chinook and Coho salmon eggs by males of either species were not significant, concluding there was little postzygotic reduction in embryonic viability. This finding suggests that the mechanism of hybridization is an incomplete reproductive barrier between these sister species.

| Ecological concerns of hybrid occurrence
We report the existence of eight Chinook-Coho hybrid adults as well as the first F2+ juveniles encountered in the natural environment. In the case of the F2+ juveniles, the number of loci used in this study provides only a probabilistic estimation of their generation. One of these fish had a genotype compatible with being the Hybridization that leads to introgression is not uncommon in salmonids (Allendorf et al., 2001;Scribner et al., 2001) and may have a variety of outcomes ranging from self-limiting low levels of hybridization with restricted geographical penetration as has been documented for the genus Salvelinus (Gruzdeva et al., 2018;Hagen & Taylor, 2001) to the loss of both species in a "hybrid swarm" (Forbes & Allendorf, 1991;Young et al., 2016). While hybridization can be a relevant source of new variability that could lead to improved fitness or the development of new taxa (Abbott et al., 2013;Verspoor & Hammart, 1991), it can also be responsible for extirpation and extinction (Allendorf et al., 2001;Rhymer & Simberloff, 1996).
Given the potential negative impacts of hybridization, identification of hybrid fish extends beyond scientific curiosity. Hybrid occurrence can be used as a monitoring tool of ecosystem changes and determining its origin with certainty warrants the value of a largescale genetic monitoring program. Such program can help estimate hybrid frequency and assess the conditions under which they appear. The multiple and continuous years of hybridization observed in this study, along with the findings by McKenzie et al. (2021), suggest the frequency of hybridization between the two species may be more common than previously presumed. Examining why the hybridization between the two salmonid species is not observed more frequently in natural environments could guide subsequent studies.

| CON CLUS ION
Although presumed infertile and extremely rare, 63 Chinook-Coho salmon hybrids were encountered in a variety of surveys conducted over a 4-year period in the vicinity of the Cowichan River. Among these hybrids, we report the existence of eight adults, as well as the first F2+ hybrid juveniles encountered in the natural environment.
Although a targeted research program will be required to identify both the causative factors and true prevalence of hybrids among Chinook and Coho salmon, we provide evidence that implicates freshwater habitat modification resulting from climatic and human pressures in the increased cospawning of the two species. This overlap in spawning may have both spatial and temporal components arising from reduced habitat due to extended summer low-flow periods. In addition, differential abundance between recovered levels of Chinook salmon and Coho salmon at historical lows may be leading to density-dependent hybridization. The likelihood of ongoing warming temperatures and drier summers in watersheds of southern BC combined with increasing human pressures on watersheds indicate that interspecific salmonid hybridization and introgression may be observed more frequently in the future, which warrants the value of hybrid identification as an ecosystem monitoring tool within a large-scale survey program.

ACK N OWLED G M ENTS
Many people contributed to this project far beyond the authors list. Katherine Horst, and Carrie Gummer conducted the laboratory work, and Sarah Webber editorial support. We are grateful for the contributions of the journal editor and two reviewers who helped us provide better context, quality, and projection.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Supplementary data are available in this document as an Appendix S1 and through DRYAD.