Early evidence for establishment of a Chinook salmon population in a restored watershed

Abstract

As fish populations face compounding pressures under climate change, highly modified rivers are receiving increasing research and conservation attention as important sites for restoration. Across the North Pacific Ocean, Chinook salmon (Oncorhynchus tshawytscha) have experienced unprecedented declines and extirpations because of habitat loss and fragmentation and climate variability. Here, we studied a rare example of a novel salmon population developing in Putah Creek, California, a dam‐controlled stream flowing through an area of intensive agriculture where salmon were unlikely to occur prior to the rehabilitation of a more natural flow regime. We used otoliths from adult Chinook salmon carcasses recovered from Putah Creek to determine river or hatchery of origin for five spawning year classes. Our results provide evidence of successful salmon reproduction, outmigration survival, and natal homing of individuals back to Putah Creek in recent years. Although hatchery‐origin fish that strayed to Putah Creek to spawn dominated returns each year, this study documents life‐cycle completion required for the potential development of a self‐sustaining salmon population. This study demonstrates that targeted restoration and flow rehabilitation efforts can generate viable new salmon spawning habitat in dam‐controlled river systems, which could help mitigate habitat lost to dams and developments. Further, successful anadromous life‐cycle completion can occur in new spawning habitat within the first several generations, which will help inform efforts to restore or reintroduce salmon in other altered river systems.

INTRODUCTION

Freshwater and marine ecosystems are impacted globally by human alterations and a rapidly changing climate, resulting in a mounting biodiversity crisis (Albert et al., 2021; Lotze et al., 2006; Reid et al., 2019; Selig et al., 2014; Strayer & Dudgeon, 2010; Vitousek et al., 1997). Understanding how fish populations respond to these manifold changes is of increasing interest in conservation ecology, especially for threatened and imperiled fishes (Barot et al., 2019; Lynch et al., 2016; Till et al., 2019). In particular, evaluating the role of dams and other disruptive water management practices is crucial because of their significant ecosystem alterations and impacts on native fish, as well as the likelihood that demands for water and energy sources will continue to increase (Bain et al., 1988; Carlisle et al., 2011; Pringle et al., 2000). Migratory fishes like Pacific salmon (Oncorhynchus spp.) are especially sensitive to anthropogenic and climatic change, as such changes can degrade or limit access to spawning and rearing habitat, which in turn may affect productivity and the persistence of different life-history strategies (Hamilton et al., 2005; Herbold et al., 2018; Sturrock et al., 2019). Dams affect salmon directly by blocking or hindering access to spawning grounds and rearing areas, and indirectly by affecting downstream habitat quality through the alteration of natural sediment, temperature, and hydrologic regimes, all of which influence myriad ecological processes and ecological resilience (McClure et al., 2008; Nehlsen et al., 1991; Thompson et al., 2019).

The long-term persistence of salmon in human-dominated landscapes, such as the Central Valley of California, will require bolstering salmon populations, although there is disagreement regarding the appropriate role of hatcheries in achieving this goal (Brannon et al., 2004; Nuetzel et al., 2023; Ruckelshaus et al., 2002). During droughts in the Central Valley, many hatcheries in the Sacramento River basin truck or barge salmon parr and smolts for direct release into the Estuary because this results in higher rates of successful ocean entry. Yet these artificially transported fish tend to stray into non-natal river systems and hatcheries as spawning adults at much higher rates than hatchery fish released on site (Bond et al., 2017; Sturrock et al., 2019). High rates of straying may contribute to the genetic homogenization of salmon populations (Williamson & May, 2005), diminish the fitness of natural populations through introgression (Keefer & Caudill, 2014), and mask declines of natural populations (Johnson et al., 2012). Though salmon are philopatric, some degree of straying can be adaptive (Keefer & Caudill, 2014; Taylor, 1991; Waples, 1990): straying individuals counter inbreeding depression and reduce competition among kin (Hendry et al., 2003; Quinn, 1993). Further, straying provides a mechanism for salmon to establish populations in newly available habitats, such as recently deglaciated streams (Pitman et al., 2021), or to reestablish populations following the loss of a local population from a catastrophic event (Anderson & Quinn, 2007; Pess et al., 2012).

California's Central Valley is a model system for evaluating the potential for establishment or reestablishment of salmon populations following restoration. Rapid human population growth and agricultural development have generated extreme demand for freshwater and extensive ecological modifications, like dams to store water and levees to prevent floodplain activation (Hanak et al., 2007; James & Cutter, 2008; Mount, 1995). These alterations ultimately rendered up to 95% of spawning habitat inaccessible (Yoshiyama et al., 2001), resulting in severe population declines and extirpations (Katz et al., 2013; Yoshiyama et al., 1998). At present, two of the four genetically and ecologically distinct Chinook salmon runs found in the Central Valley are listed for protection under the US Endangered Species Act. Though the Central Valley fall run is not listed, poor forecasted returns have resulted in the repeated closure of the commercial salmon fishery in California, with the latest closures occurring in 2023 and 2024. There is growing interest in the region to assess how Chinook salmon populations may become established in systems that were previously unavailable because of anadromous fish passage barriers or poor-quality habitat (Allen et al., 2016; Anderson et al., 2014; Brewitt, 2016).

Lower Putah Creek, a small tributary river in the Central Valley system, is an ideal watershed for evaluating the processes by which habitat rehabilitation has promoted populations of native and anadromous fishes. Following two dam installations in the mid-20th century that linked Putah Creek to the Central Valley Water Project, salmon were not observed in Putah Creek, and the waterway experienced further intensive development and degradation (Jacinto et al., 2023; Kiernan et al., 2012). Restoration actions initiated in 2000 resulted in conditions and habitat more suitable for salmon, including a modified flow regime (Jacinto et al., 2023; Kiernan et al., 2012), and annual observations of both spawning adult salmon and outmigrating juvenile salmon using Putah Creek have been consistent since monitoring began in 2014 (Jacinto et al., 2023; Willmes et al., 2021). While the observation of Chinook salmon in Putah Creek has been hailed as a local conservation and restoration success story, many questions remain regarding the origins and life-cycle success of Chinook salmon using this restored habitat. Namely, it remains unknown whether Putah Creek is capable of supporting full life-cycle completion of salmon (i.e., salmon hatch in Putah Creek, migrate out to sea, then successfully return to spawn in Putah Creek) and whether there are any adult salmon in the annual spawning year classes that are of Putah Creek origin. Alternatively, the population may be composed exclusively of straying hatchery-origin fish each generation, and Putah Creek-origin fish are not yet present among the spawning year classes. If Putah Creek can support the successful spawning, rearing, and homing of salmon, the development of a persistent salmon population in this waterway over time could help provide spatial diversity to promote resilience in Central Valley Chinook salmon (Herbold et al., 2018; Price et al., 2021; Stier et al., 2020; Sullaway et al., 2021).

In this study, we used a combination of coded wire tag (CWT) information, strontium isotope analysis, and trace elements analysis to evaluate whether there is evidence of salmon produced in Putah Creek surviving and returning to Putah Creek to spawn following restoration and management efforts. We assess (1) if there are natal, Putah Creek-origin salmon in the spawning population and (2) whether the prevalence of natal Putah Creek-origin salmon changes over time. We hypothesized that some fraction of salmon was of Putah Creek origin and that the percentage of Putah Creek-origin fish would increase from the earliest to the most recent cohort.

METHODS

Study system

Lower Putah Creek (California, USA) is a tributary of the Sacramento River draining a total watershed area of 184 km². Historical reports (Shapovalov, 1940, 1947) suggest that Chinook salmon and steelhead trout could access Putah Creek during wet years when there was enough flow to attract migrating fishes; however, it is unclear whether there was a persistent salmon run in Putah Creek prior to extensive modification of the basin, including water extraction and diversions. In 1957, two new dams were built that fragmented the stream ecosystem (Figure 1). The larger dam (Monticello Dam) built approximately 55 km upstream of the river mouth, forms Berryessa Reservoir, an 8100-ha impoundment used for both recreation and water storage. A smaller dam ~13 km downstream (Putah Diversion Dam) diverts the majority of the catchment's water into a canal for agricultural and urban usage. Both impoundments blocked anadromous fish access to upper Putah Creek, and dam management practices provided year-round flows to only a short segment (~ 5 km) of the creek immediately downstream of the diversion dam (Kiernan et al., 2012). In the decades following impoundment, the stream ecosystem had insufficient flows to support most native fishes (Jacinto et al., 2023).

[IMAGE OMITTED. SEE PDF]

Lawsuits filed during the 1990s promoted the restoration of the Putah Creek ecosystem, culminating in a legal settlement, the Putah Creek Accord (“The Accord”), which required increased flows in Lower Putah Creek and the release of the remaining water that is not diverted for agricultural and urban use. In 2000, a modified flow regime was implemented with the intent of producing a more natural hydrograph and promoting ecological processes suitable for native fishes (Jacinto et al., 2023; Kiernan et al., 2012; Appendix S1: Figure S1). The new flow regime included increased fall flows to promote the attraction of anadromous fishes, increased spring flows to assist the rearing and outmigration of juvenile fishes, and permanent base flows to prevent the creek from drying in the ~42 km of creek downstream of the Putah Diversion Dam (Appendix S1: Figure S1). Subsequent monitoring studies found that native fishes displaced non-native fishes in many segments of Putah Creek following the implementation of the flow regime (Jacinto et al., 2023; Kiernan et al., 2012), and adult Chinook salmon began to appear in Putah Creek more consistently and in greater numbers (Jacinto et al., 2023; Willmes et al., 2021).

The reaches of Putah Creek accessible to anadromous fishes include the 42 river kilometers located downstream of Putah Diversion Dam (“Lower Putah Creek”). Lower Putah Creek flows past the cities of Winters, CA and Davis, CA, with the habitat surrounding the creek channel composed of a narrow riparian corridor, extensive agricultural land holdings, and semi-urban land use (Figure 1). Flows into Lower Putah Creek are primarily the product of water released from Putah Diversion Dam as stipulated by The Accord. The migratory path that adult Chinook salmon accessing Putah Creek must take traverses from the Pacific Ocean via San Francisco Bay, up the Sacramento River to Liberty Island, then through Yolo Bypass floodplain area, a managed floodplain of the Sacramento River, to Lower Putah Creek (outmigrating juveniles complete the reverse process). Most suitable salmon habitat is located in the most upstream reaches (~15 km) accessible to salmon.

Adult sampling

Since 2016, the adult Chinook salmon population of Lower Putah Creek has been monitored by weekly canoe-based carcass surveys conducted during the months of October–January, with peak abundances usually observed in November (Table 1). Surveys were conducted in an upstream-to-downstream direction, and all detected live and dead salmon were counted. Carcasses were detected from visual observation, and the location of each carcass was recorded using a handheld GPS. A uniquely numbered metal or plastic tag was zip-tied to the mandible of each carcass for individual identification during carcass resighting.

TABLE 1 Number of Chinook salmon carcasses recovered in Lower Putah Creek by study year, the number of spawning individuals included in the study, either through otolith microchemistry and microstructure and/or through coded wire tag (CWT) origin and age data, and the number of juvenile salmon counted in the rotary screw trap (RST) or fyke net the following spring.

Survey year	Spawner carcasses recovered	Spawners included in study	Juvenile salmon counted
2016	126	104	…
2017	208	97 (49 with otoliths, 48 with CWT)	215
2018	211	80 (47 with otoliths, 33 with CWT)	33,513
2019	1	0	26
2020	70	51	200
2021	81	75	5078

When the condition of deteriorating carcasses allowed, the fork length of the carcass was measured (in millimeters) and the presence or absence of an adipose fin was documented. Hatcheries operating in the Central Valley participate in a marking program in which a constant fraction (25%) of released juvenile fish are injected with a small CWT in their snout coded with hatchery and release information; these fish also have their adipose fin removed to indicate the potential presence of a CWT (Nandor et al., 2010). For carcasses missing an adipose fin (or presumed missing the adipose because of carcass deterioration), the snout of the carcass was removed and frozen for CWT recovery at the California Department of Fish and Wildlife Coded Wire Tag Laboratory in Sacramento, CA (Nandor et al., 2010). CWTs extracted from carcasses were thus linked to a hatchery, year, and location of release. Sagittal otoliths were removed on site and air dried in Falcon tubes. All sampling and collections were conducted in accordance with California Department of Fish and Wildlife scientific collecting permit S-183530003-18360-003.

Juvenile sampling

Downstream migrant trapping on Lower Putah Creek was conducted each spring following spawning adult carcass surveys to evaluate juvenile production (UC Davis IACUC Protocol #22677). A rotary screw trap (RST) was operated during every study year except 2020, during which a fyke net was operated instead to allow for social distancing of field staff during the COVID-19 pandemic. Constraints in the annual installation timing of the RST and variation in spawning success and juvenile production contributed to highly variable numbers of juveniles observed in Putah Creek annually (Table 1), especially in high-flow years when the RST could only be safely installed after high pulse periods, which could also have coincided with pulses of outmigrating juveniles (Miner, 2022). Incidental juvenile mortalities in traps were collected and frozen and were later sampled for otoliths to provide fish of known Putah Creek natal origin as a reference dataset for the otolith geochemical methods.

Otolith preparation

After air drying, sagittae collected from juvenile (2017, 2018, and 2021) and adult salmon (2017–2021) were weighed and photographed whole under a stereomicroscope at 0.5× magnification. The length (distance from the midpoint of the rostrum, through the primordium, to posterior edge) and width of each otolith were measured to the nearest 0.001 mm from photographs using ImagePro Plus software. Rostral tips were removed anterior to the core region, and the otolith was mounted sulcal side up in the sagittal plane on round glass slides in Crystal Bond thermoplastic resin, as preparations in this method both expose the core and preserve annual growth rings with high fidelity. Mounted otoliths were ground until the core region was exposed using an AccuStop specimen holder on 500–1200 grit wet/dry sandpaper and polished with 3-μm lapping film. The sanded side was adhered to a small glass slide using superglue; then, the otolith was flipped and sanded using the same methods until the core region was exposed and polished. Otoliths collected in 2016 from a previous study (Willmes et al., 2021) were prepared similarly to otoliths from 2017 to 2021, with the exception that 2016 otoliths were prepared using a transverse section; thus, age and growth data from 2016 were also incorporated into this study. Sagittae from juvenile salmon were extracted from individuals recovered as incidental mortalities from RSTs operating in California Central Valley systems.

Otolith age estimations

Annual ages in otoliths were determined in the sagittal plane on the dorsal lobe. Bands were counted as a sequence of winter (translucent) and summer (opaque) bands. Characteristic checks produced at hatching (hatch check), onset of exogenous feeding (exogenous feed check), and smoltification (smolt check) were also identified. Two readers independently counted annual growth rings for each otolith to estimate the annual age of each fish (Barnett-Johnson et al., 2007). When age estimates generated by the two readers were in disagreement, concert reads were used (both readers assessed the otolith together). If the two readers could not reach a consensus or the bands were too obscure to read, the otolith was excluded from the final age dataset.

Otolith strontium isotope analysis and initial natal origin classification

Naturally occurring stable isotopes of strontium (⁸⁷Sr/⁸⁶Sr) vary in their abundance due to differences in underlying geology, which results in variations in Sr isotope ratios among local watersheds (Chesson et al., 2012). Matching Sr isotope ratios observed in organisms with those mapped from water and soil samples can establish habitat use and migration of mobile animals (Hobson et al., 2010). In particular, deposited layers within an otolith can reflect the water chemistry of a fish's location during deposition, allowing for the movements and migrations of a fish to be mapped against chemical tracers specific to certain water bodies, such as Sr (Barnett-Johnson et al., 2008; Elsdon et al., 2008). Sr isoscapes for California's Central Valley, which were developed by mapping Sr isotope ratios of water samples and juvenile Chinook salmon from major tributaries and hatcheries, have been used to identify the natal origins and movement timing of Chinook salmon by measuring Sr isotope ratios in otoliths (Barnett-Johnson et al., 2008). Sr tools have been applied extensively in the Central Valley to answer connectivity, life history, and conservation-related questions in salmon populations (Barnett-Johnson et al., 2007, 2008; Cordoleani et al., 2021; Johnson et al., 2012; Phillis et al., 2018; Willmes et al., 2021).

Sectioned otoliths were mounted onto petrographic glass slides (six otoliths per slide) using double-sided tape. ⁸⁷Sr/⁸⁶Sr isotope ratios for each otolith were analyzed at the University of California, Davis Interdisciplinary Center for Plasma Mass Spectrometry using established techniques (Barnett-Johnson et al., 2008). A Nd:YAG 213-nm laser (New Wave Research UP213) coupled to a New Plasma HR multicollector inductively coupled plasma mass spectrometer (MC-ICP-MS) (Nu032) was used for Sr isotope analyses (Appendix S1: Table S1). A 40-μm laser-beam transect was pulsed at 10 Hz along the otolith from the dorsal edge, through the core, to the ventral edge at a speed of 10 μm/s. Isotope data were reduced using IsoFishR (Willmes, Ransom, et al., 2018) to yield Sr isotope ratio profiles along the transect line of the otolith. Laser run times were converted to transect lengths (in micrometers) using the known laser run speed of 10 μm/s. Following laser ablation, each otolith was photographed at 100× magnification using Image-Pro Plus software with the “live tiling” function. Laser transect orientations were reconciled with a standardized transect using the Line function to trace laser transect lines and otolith axis lines, and marking lengths (in micrometers) to known landmarks (e.g., laser start, otolith start, ventral lobe bands, core, dorsal lobe bands, otolith end, laser end) between the two lines. The natal region of the otolith was defined as the period following the exogenous feeding check, typically ~200 μm from the core along the dorsal radius (Barnett-Johnson et al., 2008) and verified using visual inspection of the otolith section and the geochemical data. Mean strontium isotope values were then calculated for the natal region as input for the natal classification model.

A classification and regression trees (CART) model was used to predict the natal origin for each fish by matching the ⁸⁷Sr/⁸⁶Sr isotope ratio of the natal region of each otolith with known ⁸⁷Sr/⁸⁶Sr isotope ratios of different Central Valley rivers and fish hatcheries (Barnett-Johnson et al., 2008; Phillis et al., 2018; Sturrock et al., 2015; Willmes et al., 2021; Willmes, Hobbs, et al., 2018). Data were split into training (75%) and test (25%) data, and the training data was resampled to 25 samples per source, with replacement. Then a bagging (bootstrap aggregating) ensemble algorithm was used (500 trees, 10-fold cross validation) to improve the stability and accuracy of the decision tree using the CARET package (Kuhn, 2008). The CART model achieved a classification accuracy of 87% (CI = 76%–94%, kappa = 85%) as evaluated on the test data. However, the ⁸⁷Sr/⁸⁶Sr ratios for the natural Feather River population is known to partially overlap with the natural Putah Creek ⁸⁷Sr/⁸⁶Sr ratios (Figure 2), making it impossible to determine Putah Creek fish with certainty using ⁸⁷Sr/⁸⁶Sr alone.

[IMAGE OMITTED. SEE PDF]

Otolith trace element analysis and refined natal origin classifications

Trace elements in otoliths (²⁴Mg, ⁴³Ca, ⁴⁴Ca, ⁵⁵Mn, ⁶⁶Zn, ⁸⁸Sr, ¹³⁷Ba) were measured for individuals classified as originating from Putah Creek or Feather River by the Sr isotope CART model in an attempt to find additional elements that differed between Putah Creek and Feather River, since the Feather River ⁸⁷Sr/⁸⁶Sr values overlap with the Putah Creek ⁸⁷Sr/⁸⁶Sr values. First, we analyzed the trace elements in the otoliths from known-origin juveniles recovered from incidental mortalities in RSTs in Putah Creek (2022: n = 3; 2018: n = 2) and Feather River (2022: n = 10; 2019: n = 5; 2015: n = 5) to characterize any differences in elemental chemistry. Juvenile and adult otoliths were analyzed for element concentrations using a 193-nm laser coupled to an Agilent 7700× Quadrupole Inductively Coupled Plasma Mass Spectrometer at the Stable Isotope Laboratory, University of California, Davis (Appendix S1: Table S2). The laser was set with a repetition rate of 10 Hz and a fluence of 1.44 J/cm³, each sample was pre-ablated at a slit size of 100 μm and a run speed of 100 μm/s, and ablation transect lines were set with a slit length of 40 μm and a run speed of 10 μm/s. For juvenile otoliths, transect lines were drawn on the ventral lobe before the core to the dorsal edge and bent to intersect the core. For adult otoliths, transect lines 1000-μm long were drawn from the core through the dorsal lobe to capture the natal freshwater residence region of the otolith. Analysis cycle time was approximately 1.0 s, allowing each element to be sampled for every second, with varying dwell times depending upon element abundance. Before and after each set of samples was ablated, three sets of four reference materials were ablated along 300-μm transects in the following order: NIST 610, NIST 612, SrNano, Macs3. Laser run times were converted to transect lengths (in micrometers) using the known laser run speed of 10 μm/s.

Data were reduced using Iolite v4 software (Paton et al., 2011). Elements were then expressed as ratios against ⁴³Ca and grouped according to their origin (e.g., Putah Creek juvenile; Feather River juvenile; unknown adult; Figure 3). Ratios of Ba/Ca and Sr/Ca are known to vary among watersheds (Sturrock et al., 2012) and show distinct differences between known Feather River and known Putah Creek juvenile samples and consequently were utilized as tracers for subsequent natal-origin classification analysis. The natal region for juveniles and adults was defined using the same criteria following the approach outlined in the strontium isotope analysis. In short, the average natal Ba/Ca and Sr/Ca values were calculated for the region of the otoliths just outside of the exogenous feeding check, generally about 200 μm from the core. Adult otoliths were classified as either Putah Creek or Feather River origin using a linear discriminant analysis, with scaled elemental ratios of Ba/Ca and Sr/Ca from juvenile otoliths of known natal origin used as the training dataset, with 91%–100% classification accuracy based on a jackknife leave-one-out approach. All analyses were conducted in R version 4.3.1 (R Core Development Team, 2023).

[IMAGE OMITTED. SEE PDF]

RESULTS

A total of 407 adult fish from five spawning year classes of Chinook salmon were assigned to river or hatchery of origin, with 326 assigned using population-specific otolith ⁸⁷Sr/⁸⁶Sr isotope ratios alone or in concert with CWT data (2016: n = 104; 2017: n = 49; 2018: n = 47; 2020: n = 51; 2021: n = 75; Table 1), and 81 assigned using CWT data alone (2017: n = 48; 2018: n = 33). There were 16 fish for which both otolith-assigned natal origins and CWT data were available, and the CART model classified 81% of them correctly to their hatchery of origin. Fish with CWT information and incorrect otolith chemistry (n = 3) assignments were corrected to their CWT-known origins.

Natal origins of five spawning year classes of Chinook salmon recovered in Putah Creek revealed that these fish originated and strayed from at least 12 sources in California's Central Valley, including seven rivers and five hatcheries (Figure 4; Appendix S1: Figure S2). The CART Sr isotope model identified a total of 13 Chinook salmon otoliths as originating from either Putah Creek or Feather River (2016: n = 1; 2017: n = 1; 2020: n = 1; 2021: n = 10). These 13 fish were subsequently analyzed for trace elements along with 25 juvenile fish of known Feather River or Putah Creek origin in order to more confidently assign the natal origin of adults using multiple markers. Linear discriminant analysis of Ba/Ca and Sr/Ca ratios assigned 11 of the 13 unknown fish as Putah Creek origin (Figure 5), with these fish clustering closely with known Putah Creek-origin juvenile fish, and two fish were uncertain in their classification as a result of low confidence in assignment (<85% single source assignment).

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

Overall, the proportion of natural-origin fish (including Putah Creek-origin fish) was greater in the four later year classes (e.g., 2016 = 0.028; 2021 = 0.186; Figure 4), though hatchery-origin fish were the most abundant origin group in every spawning year class studied. A χ² test with 20,000 simulations indicated there were differences in the number of hatchery-origin, natural-origin, and Putah Creek-origin fish among study years (χ² = 38.299, p < 0.0001; Table 2). Pairwise post hoc comparison tests with Holm correction found differences in the frequency of different natal classifications between 2016 and 2021 (p < 0.01), 2017 and 2021 (p = 0.04), and 2018 and 2021 (p = 0.04). One post hoc comparison presented null values between 2016 and 2018, owing to zero values for observed Putah Creek-origin fish in 2016 and 2018. The contributions of the different natal origins varied by brood year (calculated by subtracting the annual age of a fish from its survey year and adding one), with the largest hatchery-origin contributions originating in the 2015 (n = 91) and 2016 brood years (n = 90) and the largest Putah Creek-origin contribution from the 2019 brood year (n = 8; Appendix S1: Figure S4).

TABLE 2 Results from the χ² test evaluating differences in the number of fish per natal-origin assignment classification for all study years.

n	Group 1	Group 2	Statistic	df	Adjusted p	Significance
191	2016	2017	6.8578	2	0.194	ns
172	2016	2018	…	2	…	…
144	2016	2020	2.0479	2	0.980	ns
167	2016	2021	14.3104	2	0.007	**
173	2017	2018	2.6077	2	0.980	ns
145	2017	2020	2.8167	2	0.980	ns
168	2017	2021	10.5007	2	0.042	*
126	2018	2020	1.8854	2	0.980	ns
149	2018	2021	10.3516	2	0.042	*
121	2020	2021	4.8746	2	0.437	ns

Of the 326 otoliths with age estimations, readers were able to form a consensus on 317 otoliths (97%), with the remaining nine otoliths excluded from age structure analyses; fish without analyzed otoliths but with known ages from CWTs were included in analyses (2017: n = 48; 2018: n = 33). There were 14 fish that had known ages from CWT tags and otolith-estimated ages, and these agreed in 79% of the cases, with deviations from the age either + or − 1 year. Fish with CWT information and incorrect age assignments were corrected to their CWT-known ages. Demographics of the spawning population varied in the five spawning year classes studied. Age-2 and age-3 fish were present in near-equal proportions in the first spawning year class (Figure 6). In every subsequent year, age-three fish predominated, and proportions of age-2 and age-4 fish varied among spawning year classes. Fork lengths (in mm) were larger for fish of older age classes (Figure 7), though size discrimination was weak between the older age classes (age-3 and age-4 fish). Sex distributions of fish were roughly equal between males and females for all age classes and natal-origin categories (Appendix S1: Figure S3).

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

DISCUSSION

We analyzed the otolith microchemistry and CWTs of Chinook salmon spawning in Putah Creek, which enabled us to assign spawning salmon to their natal stream or hatchery of origin with high confidence and demonstrate that some of the individuals using Putah Creek to spawn were of Putah Creek origin—the first evidence, to our knowledge, that Chinook salmon are able to successfully complete their full life cycle using this restored stream. Previously uncertain natal classifications resulting from overlap in ⁸⁷Sr/⁸⁶Sr isotope ratios (Willmes et al., 2021) were resolved using trace element analyses of Ba/Ca and Sr/Ca, resulting in greater confidence in natal-origin classifications for spawning Chinook salmon of putative Putah Creek origin, using two strong and separate lines of evidence to identify Putah Creek-origin adults in the spawning year classes of Putah Creek.

This study examines year classes of spawning salmon present between 2 and 7 years after the onset ofsalmon monitoring at Putah Creek in 2014. Since Chinook salmon in the Central Valley are between 2 and 4 years of age at spawning, this study only captures up to three generations of salmon spawning in Putah Creek, providing an early window into the stages of salmon population reestablishment (Allen et al., 2016; Liermann et al., 2017; Pess et al., 2011). Though these early findings are promising, there are several potential obstacles to establishing a self-sustaining Chinook salmon run in Putah Creek: some of these obstacles are local in nature and may be resolved with local management actions moving forward, but other obstacles derived from the broader management of salmon in the Central Valley are likely more permanent. While the high number of annual hatchery strays may preclude the development of a locally adapted self-sustaining, natal population of salmon at Putah Creek, this study's finding of natal-origin salmon present in the spawning population within the first few generations of habitat use is a promising sign that Lower Putah Creek is capable of supporting Chinook salmon through their life cycle. The initial success observed at Lower Putah Creek suggests that other altered, dam-controlled river systems in the Central Valley and beyond may be restored to generate viable salmon spawning habitat and cultivate new populations of salmon, and adds to a growing body of literature examining how different restoration actions can contribute to the resilience and reestablishment of salmon worldwide (Beechie et al., 2023; Leinonen et al., 2020). The Putah Creek-origin fish observed in this study were age-3 or age-4 individuals (Appendix S1: Figure S3), and older, larger salmon may have greater fitness and reproductive success than younger, smaller individuals (Anderson et al., 2013; Berejikian et al., 2010; Christie et al., 2014; Williamson et al., 2010), while hatchery-origin individuals tend to spawn at younger ages (Hankin et al., 2009; Macaulay et al., 2023). The older spawning individuals in the Putah Creek salmon population could promote high rates of reproductive success, aiding juvenile production and population establishment at Putah Creek.

In each spawning year class, hatchery-origin Chinook salmon were most abundant, and natural and hatchery sources contributed variably to several years of Putah Creek spawning year classes (Appendix S1: Figure S2). Multiple source populations may result in a higher degree of genetic variation present in the Putah Creek founding population. Such genetic diversity could in turn provide more genes and traits upon which selection may act to determine which source populations will be most successful in contributing to the salmon produced at Putah Creek (Nuetzel et al., 2023). The high number of hatchery-origin Chinook salmon observed spawning in Putah Creek reflects in part the management of hatchery juvenile salmon during drought years in California, when fish are trucked or barged to estuaries and released, rather than being released directly at the hatchery (Sturrock et al., 2019). These facilitated transport methods, while effective at producing higher outmigration survival during droughts, disrupt the natural imprinting process in smolts as they migrate seaward, resulting in higher straying rates when adults return to spawn (Bond et al., 2017; Dedrick & Baskett, 2018; Sturrock et al., 2019). Since the trucking of juvenile hatchery-origin salmon is still widely practiced in the Central Valley, and despite concern about the effectiveness of the strategy (California Hatchery Scientific Review Group [California HSRG], 2012), it is likely that hatchery-origin fish will continue to be present in high numbers in future spawning year classes at Putah Creek. These consistent and extensive hatchery inputs could delay the emergence of a locally adapted self-sustaining, natal run of salmon at Putah Creek, or could mask declines or other problems in a developing population. Though the role of hatcheries in fisheries management is still a dynamic question (Brannon et al., 2004; Keefer & Caudill, 2014; Ruckelshaus et al., 2002; Williamson & May, 2005), in this study, we found that hatchery-origin fish were a useful resource for supplying large numbers of founding individuals for the early development of a novel population. In this way, straying fish of hatchery origin can provide a valuable mechanism for founding new fish populations in sites that lack nearby or ancestral founders (Nuetzel et al., 2023), in a mechanism similar to the “natural” straying that traditionally would have allowed fish to establish populations in newly available habitats or following the stochastic loss of an established population (Anderson & Quinn, 2007; Pess et al., 2012; Taylor, 1991; Waples, 1990). As many other freshwater systems worldwide work to restore fish habitat and establish new fish populations to combat continuing declines (Jin et al., 2022; Lavelle et al., 2021; Leinonen et al., 2020; Munsch et al., 2020; Polivka & Claeson, 2020; Stoffers et al., 2022), leveraging hatchery populations as founders may provide a strategic resource for promoting more rapid and effective development of nascent fish populations.

In addition to the obstacle posed by region-wide Central Valley hatchery management practices, there are smaller, more local obstacles to the development of a natal salmon population at Putah Creek, some of which may be addressed with local management actions. For example, though currently suitable spawning habitat is limited to the most upstream 15 km accessible to salmon, which inherently restricts the number of salmon that the waterway is able to support, local efforts to restore more downstream reaches impacted by channelization and historic gravel mining may provide more suitable spawning and rearing habitat throughout the accessible length of Putah Creek, which could increase the number of salmon the system is able to support. Separately, a privately operated agricultural impoundment dam located near the terminus of Putah Creek in the Yolo Bypass (Figure 1) is currently operated such that Putah Creek is inaccessible until comparatively late in the spawning run timing for Central Valley fall-run Chinook salmon. As such, the current seasonal removal of the dam in the fall selects for a reduced number of spawners that have delayed run timing. The dam is also reinstalled every spring for agricultural water pumping, which restricts the outmigration timing and success of juvenile salmon produced in Putah Creek. This may select for juveniles that develop and outmigrate earlier, as later-arriving juveniles become trapped behind the dam and are unable to continue their migrations toward the estuary (Miner, 2022). Together, the dam's selections for late-entering spawners and early-departing juveniles put high selective pressure on both life stages of Chinook salmon utilizing Putah Creek. Changing the management and installation timing of the agricultural impoundment dam could lift this selection pressure and promote better salmon success at Putah Creek.

While the presence of several Putah Creek-origin Chinook salmon documented in the 2021 spawning year class of this study was promising, these individuals were ultimately unable to spawn. All salmon from the 2021 spawning year class documented in this study died near the junction of Lower Putah Creek and the Sacramento Toe Drain, many kilometers downstream of the nearest viable spawning habitat on Putah Creek (Rabidoux et al., 2022). An atmospheric river event coincided with the timing of fall attraction flows, and high inputs of debris, organic matter, and ammonia-saturated tailwaters from the storm contributed to high amounts of runoff in Putah Creek during the arrival of spawning salmon. Warm waters and rapid decomposition of organic matter ultimately resulted in a prolonged hypoxic event that killed nearly all the migrating salmon and many other fish in the area (Rabidoux et al., 2022). It is a significant but not insurmountable setback that the spawning year class with the greatest proportion of Putah Creek-origin fish was unable to contribute to juvenile production, but the dynamic, collaborative management of the Putah Creek system in the future will play an important role in preventing setbacks like this one. Ensuring earlier migratory access to Putah Creek, providing alternate migration pathways into Putah Creek, cultivating stakeholder engagement, and developing a real-time response system for future emergency situations will likely determine the fate and resilience of the Chinook salmon population at Putah Creek.

The evidence for successful production of natal Chinook salmon at Putah Creek documented here has important management implications. Putah Creek is the sole accessible tributary on the west side of the Sacramento River entering the Yolo Bypass. Prior to evidence of a Putah Creek Chinook salmon population, all salmon migrating through the Yolo Bypass Wildlife Area and the Toe Drain were assumed to be hatchery-origin strays, and salmon that migrate past Putah Creek in the Yolo Bypass either spawn (likely unsuccessfully) and die in the Yolo Bypass or are recovered upstream of the Putah Creek mouth at a fish salvage facility (Wallace Weir) for release into the Sacramento River. The most recent severe drought year resulted in an extensive loss of connectivity between the Toe Drain and the fish salvage facility during the fall-run migration period, making Putah Creek the sole available stream for spawning Chinook salmon migrating through the Toe Drain. Current policy suggests that all salmon that migrate into the Toe Drain are destined for the Wallace Weir fish salvage facility, but incorporating Putah Creek into Chinook salmon management plans for the Toe Drain and Yolo Bypass will provide an alternative route for salmon using these waterways, especially when the fish salvage site is hydrologically disconnected during drought years.

The reestablishment of Chinook salmon in Putah Creek following community-led restoration initiatives and changes in streamflow management provides context for promoting the resilience of salmon runs in the Central Valley and the larger Pacific region, many of which are in decline or already extirpated (Irvine & Fukuwaka, 2011; Nehlsen et al., 1991). While many sweeping and expensive management actions have been taken in recent decades to abate or slow declines in salmon populations, most have been unsuccessful (Knudsen & McDonald, 2000). Lower Putah Creek provides a case study for restoring salmon spawning habitat downstream of dams in modified Central Valley systems, which could help mitigate the costs of habitat lost to extensive development, agriculture, and permanent dams. Many other dammed waterways in the Central Valley could be candidates for or are currently implementing flow rehabilitation and habitat restoration similar to that conducted at Putah Creek, including Alameda Creek, Lower Stoney Creek, Cache Creek, and Lower Clear Creek, among others. Such actions implemented at a regional scale across many altered waterways could bolster Central Valley salmon populations or provide stability to the region's salmon stocks in a changing future. It is also important to recognize that salmon recovery began to occur only following the restoration of a functional flow regime. Prior to functional flow regime implementation, this system was unsuitable and inaccessible to migrating salmon due a lack of a permanent wetted channel and flow releases that did not necessarily coincide with cues relevant to salmon. Implementing a functional flow regime in this system ensured that the stream hydrology elements that were most relevant to cold water fish populations were present in Putah Creek, which was essential to making the system suitable for attracting and supporting salmon, and was previously demonstrated in this system to promote an assemblage of native, coldwater-adapted fishes in the place of introduced, warm water-adapted fishes (Jacinto et al., 2023). Stream restoration techniques such as functional flow regimes are thus a powerful tool for promoting populations of native fishes, among other habitat restoration methods used for fisheries recovery (Geist & Hawkins, 2016; Marttila et al., 2019), and could be implemented to great effect in other systems in the Central Valley to create hydrologic conditions more suitable for salmon in altered ecosystems. Though a locally adapted, entirely self-sustaining Chinook salmon run may not emerge at Putah Creek under the current management of salmon in the Central Valley, the future of salmon in this system will be better enhanced and protected under future climate change if more of the accessible habitat located below dams is restored and managed for environmental flows, making Putah Creek the first of many successes rather than an isolated anecdote in the story of the restoration of California's salmon in the 21st century.

AUTHOR CONTRIBUTIONS

All authors contributed to study conceptualization and review and editing of the manuscript. Lauren G. Hitt, Malte Willmes, George Whitman, and Mackenzie C. Miner curated data and conducted investigation. Lauren G. Hitt, Malte Willmes, and George Whitman conducted formal analysis. Lauren G. Hitt, Nann A. Fangue, and Andrew L. Rypel acquired funding. Lauren G. Hitt, Malte Willmes, George Whitman, and Rachel C. Johnson developed methodology. Malte Willmes and Andrew L. Rypel contributed to project administration. George Whitman and Dennis E. Cocherell organized resources. Lauren G. Hitt and Malte Willmes generated data visualization. Lauren G. Hitt wrote the original draft.

ACKNOWLEDGMENTS

We thank Solano County Water Agency (Contract no. 03-00206VR), especially M. Stevenson and R. Marovich, as well as D. Jones, A. Rabidoux, R. Sanford, and the rest of the staff at SCWA. We also thank local stakeholders and landowners who permitted land access for this research, especially J. Pickerell, D. Kilkenny, J. & E. Hasbrook, J. Hobbs, and H. Wimmer, as well as the many community members who care so deeply about the Putah Creek ecosystem. We thank E.D. Chapman, G. Singer, E.E. Jacinto, J. Colby, and the many volunteers and staff who helped conduct field surveys and operate the rotary screw trap to obtain the samples used in this study. We thank J.J.G. Glessner, L. Lewis, and S. Araya for technical support in operating the mass spectrometers. We thank L. Koerber and the staff of the CDFW CWT laboratory for their assistance in recovering CWTs and hatchery-origin data for fish included in this study. We also thank P. Herrera-Thomas for assistance with creativity in spatial analysis. Open access publishing facilitated by University of Canterbury, as part of the Wiley - University of Canterbury agreement via the Council of Australian University Librarians.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT

Data (Willmes & Hitt, 2025) are available from Dryad: .

References

Albert, J. S., G. Destouni, S. M. Duke‐Sylvester, A. E. Magurran, T. Oberdorff, R. E. Reis, K. O. Winemiller, and W. J. Ripple. 2021. “Scientists' Warning to Humanity on the Freshwater Biodiversity Crisis.” Ambio 50: 85–94. [DOI: https://dx.doi.org/10.1007/s13280-020-01318-8].

Yoshiyama, R. M., E. R. Gerstung, F. W. Fisher, and P. B. Moyle. 2001. “Historical and Present Distribution of Chinook Salmon in the Central Valley Drainage of California.” California Department of Fish and Game Fish Bulletin 179(1): 71–176.