Juvenile salmon (Oncorhynchus spp.) use of reservoir food webs is understudied. We examined the feeding behavior of subyearling Chinook salmon (O. tshawytscha) and its relation to growth by estimating the functional response of juvenile salmon to changes in the density of Daphnia, an important component of reservoir food webs. We then estimated salmon growth across a broad range of water temperatures and daily rations of two primary prey, Daphnia and juvenile American shad (Alosa sapidissima) using a bioenergetics model. Laboratory feeding experiments yielded a Type-II functional response curve: C = 29.858 P *(4.271 + P)-1 indicating that salmon consumption (C) of Daphnia was not affected until Daphnia densities (P) were < 30 · L-1. Past field studies documented Daphnia densities in lower Columbia River reservoirs of < 3 · L-1 in July but as high as 40 · L-1 in August. Bioenergetics modeling indicated that subyearlings could not achieve positive growth above 22°C regardless of prey type or consumption rate. When feeding on Daphnia, subyearlings could not achieve positive growth above 20°C (water temperatures they commonly encounter in the lower Columbia River during summer). At 16–18°C, subyearlings had to consume about 27,000 Daphnia · day-1 to achieve positive growth. However, when feeding on juvenile American shad, subyearlings had to consume 20 shad · day-1 at 16–18°C, or at least 25 shad · day-1 at 20°C to achieve positive growth. Using empirical consumption rates and water temperatures from summer 2013, subyearlings exhibited negative growth during July (-0.23 to -0.29 g · d-1) and August (-0.05 to -0.07 g · d-1). By switching prey from Daphnia to juvenile shad which have a higher energy density, subyearlings can partially compensate for the effects of higher water temperatures they experience in the lower Columbia River during summer. However, achieving positive growth as piscivores requires subyearlings to feed at higher consumption rates than they exhibited empirically. While our results indicate compromised growth in reservoir habitats, the long-term repercussions to salmon populations in the Columbia River Basin are unknown.
Citation: Haskell CA, Beauchamp DA, Bollens SM (2017) Linking functional response and bioenergetics to estimate juvenile salmon growth in a reservoir food web. PLoS ONE 12(10): e0185933. https://doi.org/10.1371/journal.pone.0185933
Editor: Jeffrey Buckel, North Carolina State University, UNITED STATES
Received: March 13, 2017; Accepted: September 21, 2017; Published: October 11, 2017
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by a National Science Foundation GK-12 Grant to Stephen Bollens and administered as a fellowship to Craig Haskell by Washington State University. The U.S. Geological Survey provided funds through the YES (Youth and Education in Science).
Competing interests: The authors have declared that no competing interests exist.
Freshwater food webs are important for juvenile salmon (Oncorhynchus spp.) as they migrate seaward. In the Columbia River Basin (CRB), main-stem food webs have been drastically altered from river impoundment, the introduction of non-native species, and supplementation with hatchery conspecifics . The effects of these food web changes are mostly unstudied but are likely important for juvenile salmon growth and their ultimate survival. Juvenile salmon survival is directly influenced by growth, with larger individuals less susceptible to predation at critical early life history stages . Greater size achieved during early life stages can confer a survival advantage through the lifespan of a population [3,4]. The growth of juvenile salmon during this time is a function of water temperature, consumption, and prey quality [5,6]. Therefore, examination of these factors and their ability to constrain life-stage specific growth is warranted.
The rearing capacity of lower Columbia River food webs for juvenile salmon remains largely unstudied. The CRB has more than 200 dams, many of which have elaborate collection and bypass systems to aid juvenile salmon downstream migration , and researchers have focused mostly on downstream salmon passage at main-stem dams and its relation to survival rather than food web interactions. If indeed main-stem food webs are important for juvenile salmon, they are probably particularly important for fall Chinook salmon (O. tshawytscha) juveniles that exhibit an extended residence in main-stem habitats relative to other juvenile salmon life histories . For example, juvenile spring Chinook salmon rear in tributary habitats and emigrate as yearlings on the ascending limb of the hydrograph. Juvenile fall Chinook salmon, also known as subyearlings, are more reliant on main-stem habitats and food webs because they migrate more slowly, at a smaller size, and at lower flows relative to other juvenile salmon . Therefore, reservoir food webs in the lower CRB, while utilized by all juvenile salmon to some degree, are perhaps most important to subyearlings.
Run-of-the-river reservoirs in the lower Columbia River exhibit a high degree of spatiotemporal variability in the potential prey base for subyearlings. The prey base fluctuates on an annual and seasonal basis [10,11], but also longitudinally from upstream (tailrace) to downstream (forebay) . Therefore, it is challenging to generalize the diet of subyearlings and other planktivores in the lower Columbia River due to the prevalence of run-of-the-river-reservoirs and the prey base they support. Thus, although examination of the rearing capacity of food webs for juvenile salmon is warranted, models need to be spatially and temporally explicit.
There is concern that reservoir food webs in the Columbia River are unable to meet the energetic demands of subyearlings given the large numbers of hatchery salmon and non-native juvenile American shad (Alosa sapidissima) that also rear there . During summer, subyearlings in the lower Columbia River primarily consume Daphnia [12–14], but Daphnia decline dramatically in early August which is about the time young shad begin feeding. In 2015, nearly 140 million hatchery reared juvenile salmon were released into the CRB, but there are many more juvenile shad. In John Day Reservoir alone, as many as 112 million juvenile shad are produced annually, with many more originating in other Snake and Columbia River reservoirs and the Columbia River Estuary . Juvenile shad consume up to 80% of the available zooplankton production in John Day Reservoir during the time that subyearlings are also migrating seaward . It is unknown if reductions in Daphnia forage influence the consumption rate of subyearlings.
Functional response models can provide a fundamental framework for studying predator-prey interactions within food webs. The functional response model describes predator consumption as a function of prey density. Three primary functional response models have been identified. A Type-I response exhibits a linear increase in consumption with an increase in prey density. Type-I responses are characteristic of filter feeders and a spatially homogenous prey base. A Type—II response increases linearly initially but reaches an asymptote as handling time begins to limit predator consumption. Type-II responses are indicative of planktivores that forage in patchy or prey limited environments . A Type-III or sigmoid response is exhibited by organisms that switch between prey types. There are no functional response models for juvenile Chinook salmon.
Although fine-scale growth is difficult to measure empirically given the transient nature of juvenile salmon in large rivers, it can be estimated using bioenergetics if consumption is known. The use of bioenergetics models in fisheries has increased due to the capacity of these models to address complex questions while reducing the expense and complications of field sampling . Bioenergetics models are energy balance equations that can be used to estimate the growth or consumption of a fish when one of these parameters is known or measured empirically. Although the most common use of bioenergetics models is to estimate consumption using empirical growth estimates, the model can be rearranged to estimate growth using empirically derived consumption estimates . Consumption and associated evacuation rates can be estimated empirically by sampling fish stomach contents at regular intervals over a 24-h period . These empirical consumption estimates can then be used as inputs to estimate fish growth using bioenergetics models. Empirical consumption estimates for subyearlings migrating through lower Columbia River reservoirs during July and August have been recently estimated and provide an opportunity to relate consumption to growth . These estimates generally indicated high consumption rates, but the relation between consumption, growth, and the high water temperatures at which subyearlings emigrate has not been evaluated.
Water temperature is also an important determinant of juvenile salmon growth [21,22]. Increased water temperatures at the lower end of a species thermal range increase growth, however increased temperatures at the upper end of the thermal range can decrease growth. Subyearlings can exhibit decreased growth at water temperatures exceeding 21°C . However, increases in prey consumption or prey energy density can counteract the negative energetic effects of higher water temperatures to some extent [24,25]. Subyearlings in the lower CRB migrate seaward through reservoirs when water temperatures are approaching their critical thermal maxima (20.9°C; ). During summer, subyearlings in the lower CRB switch from consuming Daphnia to juvenile American shad (which have a higher energy density) after Daphnia diminish . Here, we investigate if subyearlings can compensate for the deleterious effects of high water temperature on growth by switching from planktivory to piscivory during summer.
We estimated the consumption and growth of subyearlings as they migrate through the lower Columbia River during July and August—a time when they are subject to feeding competition from juvenile American shad, a dynamic prey base, and water temperatures approaching the upper end of their thermal tolerance. We focused our laboratory experiments on the seasonal period (July and August) when subyearlings are preying on Daphnia in the lower Columbia River. We first conducted laboratory functional response trials to estimate the feeding response of subyearling Chinook salmon to changes in Daphnia density. Next, we used empirical Daphnia densities from previous field studies to predict subyearling consumption using our functional response curve. Finally, we used a bioenergetics model to estimate growth changes associated with increasing water temperatures and a switch from planktivory to piscivory.
Materials and methods
Empirical data for this study were collected from John Day and McNary reservoirs, the third and fourth upstream reservoirs on the lower Columbia River. Together, John Day (river kilometer (rkm) 348) and McNary (rkm 471) dams impound a 221-km stretch of the lower Columbia River from Rufus, Oregon upstream past the Snake and Columbia River confluence to Pasco, Washington (Fig 1). Upstream of Pasco, the Columbia River is unimpounded for 70 km to Priest Rapids Dam. This stretch of river, commonly referred to as the Hanford Reach, is home to the largest naturally reproducing stock of fall Chinook salmon in the Columbia Basin . Therefore, McNary and John Day reservoirs are important downstream rearing areas for both Hanford Reach fall Chinook juveniles and Endangered Species Act (ESA)-listed fall Chinook originating in the Snake River . We chose John Day and McNary reservoirs because of their importance for juvenile salmon rearing, and because a relatively large amount of empirical data on Daphnia densities, water temperature, and subyearling consumption exist from past studies.
The Western Fisheries Research Center is exempt from IACUC approval because fish are exempt from the U.S. Department of Agriculture definition of animals under the Animal Welfare Act. Because our laboratory specifically receives no funds from PHS agencies (The National Institute of Health, Centers for Disease Control and Prevention, and the Food and Drug Administration), we are not required to obtain IACUC approval. Fish for this experiment were obtained from the Willard National Fish Hatchery operated by U.S. Fish and Wildlife Service. The U.S. Geological Survey is required to obtain state permits for the transport of fish to our facility for experimentation. State of Washington permits require that we dispose (euthanize) fish after the conclusion of experiments, however, protocols for euthanizing fish are not stipulated. Although not required by state permitting, we euthanized fish by administering a lethal dose of MS-222.
Functional response laboratory trials
To examine the relationship between subyearling Chinook salmon consumption and Daphnia density, we used 61 cm (24 inch) cylindrical trial tanks ranging from 53.1 to 61.0 L. Five hundred ‘Tule’ strain juvenile fall Chinook were obtained from the Little White Salmon fish hatchery in Willard, Washington on April 10, 2014. A random sample of forty fish had a mean fork length (FL) of 71.7 mm (range = 63–86 mm) and a mean weight of 3.8 g (range = 2.4–7.2 g). Before trials, salmon were held in large cylindrical tanks supplied with aerated 7°C well water and fed 5% of their body weight per day with BioClark’s Fry 1.5 mm feed. Trials were run at temperatures similar to those experienced by subyearlings in John Day Reservoir during field consumption trials conducted by Haskell et al. (; 20°C). About 10 days before running trials, single juvenile salmon were moved to trial tanks and brought up to trial temperature at a rate of 1–2°C a day and fed frozen zooplankton collected from Lower Granite Reservoir, Washington for 3 days to acclimate them to the prey used in trials. Fish were starved for the final 48 h leading up to the trials.
Prey for the trials were live Daphnia pulex. Daphnia were obtained from Sachs Systems Aquaculture and then cultured in 21 L aquaria with Biofilm filters. Water temperatures in the aquaria were ambient with air temperature held constant at 21.1°C (70°F). Photoperiod mimicked conditions in John Day Reservoir during mid-August, when daylight was from 0600 to 2010 hr. We fed Daphnia an equal-parts mixture of dried Chlorella, Spirulina algae, and activated dry yeast. To each aquarium, 0.0625 g of the mixture was added when increasing water clarity generally indicated the need for more feed. Daphnia were seeded to other aquaria as required to increase densities. We used Seminole Rams-Horn snails (Planorbella duryi) to help reduce metabolites and subsequent fouling of aquarium water. Snails were periodically culled to maintain about 20 per tank. Daphnia were culled during water changes to keep cultures from overpopulating and crashing.
Individual functional response trials were run for 10 min, beginning from the time of first Daphnia capture [29,30] as monitored by an overhead video camera connected to a monitor (Speco technologies model CVC-320WP). Light intensity within trial tanks was constant, but ranged from 11.8 to 24.8 lx across tanks. We introduced Daphnia to achieve densities ranging from 0.1 to 161.4 · L-1. We wanted to populate the functional response curve with an ecologically relevant range of Daphnia densities from John Day Reservoir  but also some higher densities to determine at what densities the curve would ‘flatten out’. We defined the ecologically relevant range of Daphnia densities as those < 40 · L-1. For low-density trials (< 5 · L-1), individual live Daphnia were counted at the beginning of the trial. For higher density trials, we used the number of uneaten Daphnia filtered from the trial tank and preserved in ethanol at the end of the trial to determine trial densities. Initial trials were run at densities based on the availability of cultured Daphnia. After initial trials, we ran trials at higher or lower density to cover the range of densities reported for John Day Reservoir when subyearlings are present. Daphnia were filtered through a 500 micron mesh net to remove smaller Daphnia that were not in the size range consumed by subyearling Chinook salmon .
During trials, prey density was maintained by adding 10, 20, or 30 Daphnia via a feeder hose based on the real-time observations of consumption rate using the camera system. The feeder hose was connected to a standard funnel located behind an office partition to minimize shadows that could affect salmon behavior during the trial. At the conclusion of a trial, fish were immediately euthanized in MS-222 and frozen. Fish guts were later examined and the number of Daphnia in the stomach was counted. Consumption rate was determined by counting the number of Daphnia in individual stomachs and dividing by the trial time. At the conclusion of a trial, tank water was filtered through a 63 micron mesh to collect uneaten Daphnia. In cases where the total number of uneaten Daphnia was large (> 1,300), the sample was diluted to a standard dilution and five 1-ml aliquots of at least 50 individuals were counted. The mean of the aliquots was used to estimate the total number of uneaten Daphnia .
We compared linear and non-linear regression models to evaluate whether our data indicated a Type-I or Type-II functional response, respectively, and then used the appropriate model as a predictor of consumption rate based on prey density. Linear regression models were developed using PROC REG and nonlinear regression models were developed using PROC NLIN (SAS 2002).
The nonlinear equation form was: (2) where β0 is the maximum consumption rate and β1 is the prey density at which consumption reaches half of its maximum . We used our resulting functional response curve to predict subyearling consumption from empirical Daphnia densities.
Functional response model evaluation
We developed two models from the laboratory data. The first model we developed from the entire range of Daphnia densities and subyearling consumption rates that we measured. The second model we developed from the ecologically relevant range of Daphnia densities (< 40 · L-1) in McNary and John Day reservoirs as determined from the literature. We compared our non-linear (Type-II functional response) and linear (Type-I functional response) models using corrected Akaike’s Information Criteria (AICc; ). Here, (3) where K = the number of parameters +1, N = the number of trials, and RSS = the error sum of squares. The more appropriate model was selected based on the lower AICc score with AICc differences > 10 indicating little support for model similarity. Models were also selected by how they performed within the range of ecologically realistic Daphnia densities from the literature. We used Cook’s D statistic to examine the influence of individual points on the original linear model. We did not test a Type-III response based on the general appearance of the data.
Empirical Daphnia densities
To determine empirical Daphnia densities in John Day Reservoir, we consulted previous studies. In June-August of 1980–1982, Rondorf et al.  collected zooplankton using a horizontally towed paired Miller sampler with 153 μm-mesh in McNary Reservoir. Using the same sampling gear in 1994–1996, Haskell et al.  collected weekly zooplankton samples in McNary Reservoir during June and July and every other week in John Day Reservoir during August–November. This collection was intended to coincide with the migration of subyearling Chinook salmon and juvenile American shad. Gilbreath et al.  collected monthly zooplankton in John Day Reservoir from April 1994 to September 1995 using vertical plankton tows with a Wisconsin style 80 μm-mesh net. The net was towed from either 5m to the surface (upper reservoir) or from 15 m to the surface (mid and lower reservoir). Lastly, Emerson et al.  collected zooplankton monthly in the forebay of John Day Reservoir from July 2009- June 2011 using a 0.5 m, 73 μm-mesh net towed vertically from 0.5 m off the bottom to the surface. We used all of these data to better understand the range of Daphnia densities in lower Columbia River reservoirs at the time subyearling Chinook salmon are migrating and feeding.
We estimated stomach capacity using empirical Daphnia densities from various sources in McNary and John Day reservoirs, predicted consumption from our functional response curve, empirical Daphnia weights estimated from length to weight relations [16,35], and a relation between stomach capacity in grams and fish mass derived for sockeye salmon ranging from 2–350 g : (4)
Here, CP = capacity (g) and M = subyearling Chinook salmon mass (g). We estimated time to satiation (h) using empirical Daphnia densities from various field studies, predicted consumption rates from our functional response curve, and evacuation rates from previously conducted consumption field trials . Evacuation rates for subyearlings ranged from 0.58 · h−1 at 20.78°C in July to 0.51 h−1 at 22.02°C in August. At ambient temps of 20–22°C, the empirical estimates of evacuation rates were 0.51–0.58 · h-1.
Subyearling Chinook salmon bioenergetics model runs
Bioenergetics models are energy balance models which can be used to estimate the required consumption to achieve a certain growth rate . The model can also be rearranged to estimate growth given observed consumption over a specific time interval where: (5)
Species-specific physiological parameters exist and are packaged in the ‘Wisconsin model’ but often rely on parameters that are borrowed from closely related species given the difficulty in deriving species-specific parameters in the laboratory . The Chinook salmon bioenergetics model was originally developed by Stewart and Ibarra  using consumption parameters borrowed from a coho salmon (O. kisutch) model with a temperature dependent Cmax curve peaking around 16.5°C. We incorporated newly developed temperature dependent consumption parameters specifically for subyearling Chinook salmon into the existing bioenergetics model (Table 1). These parameters suggest a temperature-dependent Cmax curve about 4°C higher than the previous model . We used % Cmax and growth efficiency (GE) as relative measures of feeding performance. Cmax is the maximum possible consumption for a fish of a given weight at temperatures where consumption is maximized. For subyearlings: (6) where W = weight (g). (7) where C = specific consumption (g · g · d-1). Growth efficiency (GE) is defined as: (8) where G = specific growth (g · g · d-1) from the bioenergetics output.
We used the bioenergetics model to run two sets of growth simulations for subyearling Chinook salmon. The first set was two 4-day simulations using empirical prey percentages, water temperatures, and empirical consumption rates estimated from 24-h field trials conducted on July 21–24 and August 11–14, 2013 as described in . In this study, subyearlings predominately consumed Daphnia in July and juvenile American shad in August. Other studies have also noted the importance of Daphnia to juvenile salmon in the lower Columbia River during this time [12,38].
For the second set of simulations, we wanted to more broadly understand the effect of varying prey consumption, prey type, and water temperature on subyearling growth beyond the two 4-day scenarios that we originally modeled. Therefore, we ran multiple 20-d scenarios, varying the consumption rates on the two primary prey items of subyearlings, Daphnia and juvenile American shad, and by varying water temperature within the range experienced by subyearlings migrating through John Day Reservoir during June–August . For these model runs, we used a starting weight of 20 g for subyearlings and assumed 100% by weight of the diet consisted of Daphnia in our planktivory simulations and 100% juvenile American shad in our piscivory simulations. In the planktivory simulations, we varied water temperature from 16–24°C in 2°C increments and varied Daphnia consumption from 2,000 to 32,000 Daphnia · day-1 in 5,000 Daphnia · day-1- increments. For the piscivory simulations, we used the same temperature ranges and 20-d durations, but varied consumption from 10 to 40 shad · day-1 at 5 shad · day-1 -increments. Thirty five scenarios (5 temperatures * 7 Daphnia consumption rates) were run for the planktivory simulations and 35 scenarios (5 temperatures * 7 shad consumption rates) were run for the piscivory simulations. For individual scenarios, water temperature and consumption rate were held constant over the 20-d model runs. Prey energy densities were collected from various sources (Table 2), and we assumed 10% of prey were indigestible . Predator energy density was estimated based on a 78.8% moisture content converted to a species specific energy density for juvenile Chinook salmon of 4,730 J · g-1 . We used a wet weight of 0.083 g for juvenile American shad prey (N = 25, mean FL = 24.4 mm, range FL = 11–40 mm) . We used the midpoint of the energy density range reported for larval fish by Hanson et al. (1997) for shad energy density inputs although some of those fish had probably transitioned to the juvenile stage.
Twelve juvenile Chinook stomachs were examined from each date.
The estimation of ‘average’ Daphnia wet weight was an important consideration for the estimation of subyearling growth, % Cmax, and stomach capacity (g) calculations. To estimate the wet weight of ambient (in the river) Daphnia, we first used a mean length from empirical estimates converted to a species-specific dry weight following Culver et al. . We assumed that all Daphnia, both consumed and ambient, were D. retrocurva- the primary Daphnia species in the lower Columbia River during July and August [10,11]. We then assumed that the dry weight of Daphnia was 10% of wet weight . However, kokanee salmon and other planktivores have a demonstrated ability to reduce the water content of Daphnia by about 50% after consuming them, thereby increasing the number of Daphnia they can consume . Therefore, we assumed ambient Daphnia dry weight was 7% of consumed wet weight and consumed Daphnia had an energy density of 1,620 J/g .
During sampling in the early 1980’s, Rondorf et al.  reported a mean length for ambient Daphnia of 1.05 mm in July and 1.14 mm in August in contrast to Haskell et al.  who found a mean length for ambient Daphnia ranging from of 0.84 mm in the mid-1990’s during July and August. Differences in mean Daphnia length between the two sampling periods could be due to increased planktivory by juvenile American shad and the mysid, Neomysis mercedis . Neomysis was not present during sampling in the early 1980’s and shad numbers were much lower than they are today. Nevertheless, Rondorf et al.  reported a 0.36 mm difference between the size of ambient Daphnia and those consumed by subyearlings. Therefore, we assumed a similar difference between the mean Daphnia length reported by Haskell et al. , but compensated for size selection by assuming subyearlings were consuming Daphnia that were 1.2 mm on average.
Functional response model selection
Across the entire range of Daphnia densities that we evaluated and a subset of ecologically relevant densities, AICC scores provided strong evidence for the non-linear model over the linear model. Our non-linear model across all densities was: C = 29.858 P * (4.271 + P)-1, while for the ecologically relevant range it was: C = 28.561 P * (3.530 + P)-1. In both cases, a Type-II functional response was supported over the linear Type-I response based on AICC scores (Fig 2 and Table 3). Also, both the Type-I and Type-II ecologically relevant density models were superior to their respective Type-I and Type-II model over the entire range of Daphnia densities based on lower AICC scores. Both non-linear models had asymptotes that were about 30 · L-1, which is about the maximum Daphnia reported in the literature. In fitting the original linear model, we had one point (Density = 161.1 · L-1, Consumption = 48.2 · min-1) that exhibited a high degree of influence (Cook’s D value = 3.7). Another point (Density = 97.5 · L-1, Consumption = 20.5 · min-1) had a Cook’s D value of 0.523. Although these points had ecologically irrelevant (high) densities compared to empirical densities from the literature (Table 4), we retained these data points because they had no effect on the overall outcome of model selection based on AICC score when removed.
The two data sets are from all the data and an ecologically relevant subset (Daphnia densities < 40 · L-1).
Empirical Daphnia densities
Our comparisons of Daphnia densities from the literature to consumption rates from our Type—II functional response curve predicted subyearling consumption ranging from 0.29 to 26.95 Daphnia · min-1. Previous studies of Daphnia densities in McNary and John Day reservoirs indicated increasing Daphnia densities through June and July, with peaks in early August followed by sharp declines starting in mid-August. Specific Daphnia densities ranged from 0.1 to 2.6 · L-1 in July and from 0.1 to 39.5 · L-1 in August. Daphnia densities in July and August, corresponded with Daphnia consumption predicted from our functional response curve ranging from 0.29 to 24.37 Daphnia · min-1. Based on our Type-II functional response, the greatest Daphnia densities reported by Gilbreath et al. (; 39.54 · L-1), subyearling consumption would support a feeding rate of 26.95 Daphnia · min-1. In July through September of 1968 and 1969, subyearlings predominately consumed similar amounts of Daphnia but at densities that were lower. In 1968, subyearlings exhibited an average of 1 Daphnia stomach-1 in July, 259 Daphnia · stomach-1 in August, and 971 Daphnia · stomach-1 in September. In 1969, subyearlings exhibited an average of 3 Daphnia · stomach-1 in July, 647 Daphnia · stomach-1 in August, and 961 Daphnia · stomach-1 in September. However, during this study, Daphnia densities peaked in August at 2.5 · L-1 in 1968 and 1.7 · L-1 in 1969 .
Subyearling bioenergetics growth simulations
Our bioenergetics runs using empirical data collected by Haskell et al.  in July and August, 2013 predicted negative growth under both simulations (Table 5). Increases in growth rate from July to August, although still negative, were associated with decreases in daily ration, about a 1.5°C increase in water temperature, and a change from planktivory to piscivory. In July when subyearlings were feeding on Daphnia, subyearling growth ranged from -0.23 to -0.29 g · d-1 and when subyearlings switched to juvenile American shad prey, their growth ranged from -0.06 to -0.10 g · d-1.
Initial start weight and daily ration (D) values were obtained from Haskell et al. .
Growth potential for subyearlings varied considerably across the range of temperatures, prey densities, and prey types that could be experienced during the summer in these reservoirs. Subyearlings did not exhibit positive growth at water temperatures above 22°C, irrespective of prey type or amount (Fig 3). When feeding on Daphnia, subyearlings needed to consume about 27,000 Daphnia · day-1 (representing 90% Cmax) to achieve positive growth at 16°C, but could not achieve positive growth when feeding on Daphnia at 20°C. When feeding on juvenile shad, subyearlings needed to consume 20 shad · day-1 to exhibit positive growth at 16°C (Fig 3). At 20°C, subyearlings needed to consume about 25 shad · day-1 (representing 60% Cmax) to achieve positive growth. At 24°C, subyearlings could not consume enough shad to achieve positive growth (Fig 3).
Our stomach capacity and time to satiation data indicate that during July, subyearlings would need to feed considerably longer to achieve satiation because Daphnia densities were lower than in August. We estimated a maximum stomach capacity of 1.54 g for a 20 g subyearling. With this stomach capacity, subyearlings could consume a maximum of 17,380 Daphnia (1.2 mm). Using the range of empirical Daphnia densities from the literature, subyearlings feeding in July could achieve satiation in 28.7 to 831.4 h (Table 4). Based on August Daphnia densities as reported in the literature, subyearlings could achieve satiation in as little as 12.9 h at densities approaching 20 · L-1. However, a nearly two-fold increase of Daphnia density from 20.0 to 39.5 · L-1 was not associated with a reduction in satiation time.
Our feeding trials indicated that subyearling consumption of Daphnia increased with prey density and exhibited a Type-II response. Although there are no functional response models for juvenile Chinook salmon, other juvenile salmon models have been developed using laboratory feeding trials. Koski and Johnson  estimated the functional response of kokanee salmon (O. nerka) fingerlings to Daphnia density under varying light conditions and compared them to natural conditions in Blue Mesa Reservoir, Colorado. While kokanee had a linear response at low light levels typical of crepuscular periods, they exhibited a Type-II response at higher light levels. Koski and Johnson  reported a linear response for kokanee feeding at 0.1 lx but a Type-II response at 15 lx and 30 lx- light levels that occur during crepuscular periods when feeding is highest in Blue Mesa Reservoir. In our trials, light was held constant but varied slightly across tanks from 11.8–24.8 lx. Given the similarity of our light levels to Koski and Johnson , we surmise that light differences had little effect on subyearling feeding rate in our trials. In the wild, subyearling consumption of Daphnia is highest during late afternoon . Subyearlings can probably forage on Daphnia at a higher rate than on copepods given the less evasive nature of Daphnia. Juvenile chum (O. keta) and pink salmon (O. gorbuscha) had far lower consumption rates feeding on copepods .
Subyearlings in lower Columbia River reservoirs feed at a high rate but are unlikely to reach satiation in July because of the relatively small size of Daphnia in the lower CRB. With consumption of 1.2 mm-sized Daphnia, a 20 g subyearling could reach satiation by consuming 17,380 Daphnia. Using an empirical consumption estimate (which accounts for evacuation rate) of 11.0% body weight, subyearlings feeding on Daphnia in July could reach satiation in 16.7 h, an unlikely scenario given that subyearlings feed little at night . For comparison, a 20 g kokanee in Blue Mesa Reservoir could reach its stomach capacity by consuming 9,386 Daphnia owing to the larger size of the Daphnia there (1.84 mm; ). Although size selective planktivory can account for smaller-sized zooplankton , rivers also favor small bodied zooplankton taxa which have shorter generation times [47,48].
Empirical data collected in John Day Reservoir in 2013 indicated that when feeding almost exclusively on D. retrocurva (97% of total prey) during the time of maximum consumption (1700 hr), 20.1 g subyearlings consume 11% of their body weight per day . During peak Daphnia density in August of about 20 to 40 · L-1, a 20.0 g subyearling Chinook salmon would need to forage from 12.9 to 23.3 h to reach satiation when consuming Daphnia. Time to satiation was 12.9 h at a Daphnia density of 18.5 · L-1, but 23.3 h at a Daphnia density of 19.0 · L-1. This difference in satiation time despite only a small Daphnia density difference was due to larger Daphnia in 1982 (1.5 mm) relative to 1994 (1.2 mm). When Daphnia densities were highest (39.5 · L-1), as observed by Gilbreath et al. , time to satiation was still 21.1 h due to subyearlings feeding beyond the asymptote on the flat part of the Type-II functional response curve. Due to the relatively high Daphnia densities observed in some studies during August, subyearlings can feed at a high level that approaches satiation, however, when Daphnia densities are low as in July, subyearlings cannot reach satiation.
Juvenile American shad are potential competitors of subyearlings, because their presence overlaps in main-stem reservoirs and they consume Daphnia . Our functional response curve indicates that at Daphnia densities greater than 30 · L-1, reductions in Daphnia density have little effect on subyearling consumption. It is not until Daphnia density decreases below 20 · L-1 that subyearlings exhibit reduced consumption. Our data indicate that prey-limiting (Daphnia) conditions are more likely in July, when Daphnia densities are low, than in August. During July, most young shad have yet to develop beyond the larval stage . However by August, subyearlings have switched from feeding on Daphnia to juvenile shad. We conclude that while Daphnia are an important prey item of both subyearlings and shad, they are utilized only briefly by juvenile shad in late July and early August when shad are about 20–40 mm. Juvenile shad mostly consume copepods in August and September .
We used functional response to predict subyearling consumption using changes in Daphnia density. Given the temporally and spatially dynamic nature of predator-prey interactions in lower Columbia River reservoirs, additional functional curves for other important predators and prey would be useful. For example, subyearlings prefer D. retrocurva over the invasive calanoid copepod, P. forebsi during late summer . However, despite the numerical dominance of P. forebsi, subyearlings switch to piscivory on juvenile shad instead of consuming P. forebsi after the mid-summer decline of Daphnia . Subyearlings may find it more energetically beneficial to switch from Daphnia to juvenile American shad rather than P. forebsi. Juvenile shad reduce Daphnia production otherwise available to subyearlings and may be able to exploit Daphnia at higher rate than juvenile salmon. A comparative functional response approach could be used to understand if nonnative juvenile shad consume Daphnia at a higher rate than juvenile salmon- a potential measure of competition [51–53], or whether subyearlings consume native Daphnia at a higher rate than nonnative copepods (e.g., P. forebsi) as suggested by others [50,54]. Additional functional response curves could be developed for the period after Daphnia decline and subyearlings begin feeding on juvenile shad and shad themselves begin feeding on copepods .
Our bioenergetics modeling predicted that subyearlings would need to consume as many as 32,000 Daphnia to grow. Similarly large predictions for other salmonids are not unprecedented. Kokanee need to consume 6 D. pulex per second (21,600 · h-1) to reach observed growth in Blue Mesa Reservoir . Examination of subyearling stomach contents from the CRB indicated individual stomachs containing as many as 2,878 Daphnia (N = 16, median = 1,180) . Lastly, we observed consumption rates of nearly 50 · min-1 in our functional response trials. At this rate, subyearlings could consume 32,000 Daphnia in about 11 h. We believe that the upper range of Daphnia consumption we modeled is feasible despite the resulting growth constraints.
Using bioenergetics, we predicted subyearling growth across a range of prey and temperature scenarios that subyearlings could experience while migrating seaward through the lower Columbia River. Our simulations broadly indicated that while the energy density of juvenile American shad prey is greater than Daphnia, subyearlings would need to consume about 35 shad · day-1 to exhibit positive growth at 22°C or higher. Although we did not conduct functional response experiments using juvenile shad prey which would have allowed us estimate maximum consumption rates as we did with Daphnia, others have reported subyearlings consuming up to 20 larval and juvenile Pacific sand lances (Ammodytes hexapterus) per day . While subyearlings derive an energetic benefit with a shift from planktivory to piscivory in late summer, their growth is limited by the constraints of migrating under high water temperatures. The trends we demonstrated are consistent with decreasing optimal temperatures and thermal tolerances as daily rations decline . However, because increased Daphnia densities are associated with higher water temperatures, subyearlings face a trade-off between increased foraging opportunity and their thermal tolerance .
The data supporting our bioenergetics model runs  have some uncertainties. Uncertainty surrounding bioenergetics parameters, especially for salmonids, has been conducted using both individual parameter perturbation and Monte Carlo simulations  indicating that the mass-dependent parameters for Cmax (CA or CB) or respiration (RA or RB) are typically the most sensitive. The Chinook salmon bioenergetics model has been corroborated with predictions of consumption or growth typically within 5–10% of measured responses in lab studies  and when compared to independent field estimates . These studies indicate that most of the uncertainty results from population-specific inputs to the model rather than the physiological parameters. For our model runs, these inputs were daily ration (consumption), prey and predator energy densities, and thermal experience. We used energy densities from the literature and modeled growth using empirical water temperatures and also using a broad range of temperatures subyearlings could experience. We addressed the primary source of uncertainty by developing a functional response curve with a 95% confidence interval.
Our second set of bioenergetics simulations had subyearlings consuming 100% Daphnia and 100% juvenile shad. This Daphnia percentage was supported by empirical data from the forebay of John Day Reservoir where subyearlings consumed about 98% Daphnia . Although we believe that subyearlings predominately consume Daphnia as supported in the literature from reservoir habitats, it is possible that their diet is more variable in upstream portions of the reservoir and over longer time frames than the 3-day period that was sampled. The empirical data also indicated that subyearlings consume close to 100% juvenile shad by weight (71% by number; ). These data were similarly collected over a 3-day period in the John Day Reservoir forebay during August 2013. Although we used simplified diet scenarios in our simulations, we believe that they were representative of the two important prey items of later migrating subyearlings.
Subyearlings move though John Day Reservoir fairly quickly but compromised growth opportunity there could have lasting effects. The mean travel times (2006–2015) of subyearlings through John Day Reservoir ranged from 3.1 to 5.3 days  with only small differences between wild (Mean = 4.6, SE = 0.2) and hatchery origin (Mean = 4.2, SE = 0.2) subyearlings. Over the 10 years, subyearlings exhibited a mean travel time of 3.1 days (SE = 0.02) in July and 4.0 d (SE = 0.15) in August. Based on these travel times along with the consumption and growth data presented here, we infer that the “average” subyearling could lose 3.4% of its body weight in July and 1.2% of its body weight in August while migrating through John Day Reservoir. Although not widely demonstrated in field studies, young salmon can temporarily withstand periods of negative growth resulting from unfavorable temperatures . However, the long term consequences of these periods on survival and smolt-to-adult- ratios (SAR’s) are unknown.
Although the topic of how much energy loss equates to mortality or debilitation of fish has not yet received adequate attention in the literature, a suggestive convergence in relative energy loss-mortality has begun to emerge for juvenile and adult salmonids. Mortality has been associated with a decline to about 40–45% of initial energy content [25,62]. When compared to the 1.2–3.4% loss in body mass, which is analogous to energy loss, juvenile Chinook migrating through lower CRB reservoirs are not in immediate danger of starvation. However, compromised growth could have serious consequences for the survival of early marine life stages due to strong size-selective mortality reported for yearling and subyearling Chinook [3,4,63,64].
Human mediated habitat alterations have delayed the seasonal timing of subyearling migration through John Day Reservoir. Whereas historically, subyearlings migrated through the lower Columbia River in June and July , most now migrate starting in July, with Snake River subyearlings migrating seaward into August. Prior to 2005, later-migrating ESA-listed Snake River subyearlings were nearly all collected and transported downstream, leaving fewer fish in the river. Changes brought forth in the Biological Opinion for operation of the Federal Columbia River Power System created more spill to aid juvenile migrants during July and August . Despite shad-mediated Daphnia reductions, subyearlings are feeding at a relatively high rate in the lower Columbia River during July and August . By switching from lower energy density Daphnia to higher energy density fish prey, subyearlings can partially compensate for the metabolic costs of higher water temperature. However, empirical data suggests that they are not growing despite this switch and our modeling indicates that subyearlings need to consume 35 shad · day-1 to achieve positive growth at 22°C.
Water temperatures of 22°C during July and August are prevalent not only in John Day Reservoir, but throughout the lower Columbia River . One management strategy for fish during this time is increasing spill to aid downstream migration. Our data suggest that water temperatures above 20°C coupled with reservoir prey taxa of lower energetic quality provide little growth opportunity for subyearlings during this time. While our results indicate compromised growth in reservoir habitats, the long-term repercussions to salmon populations in the Columbia River Basin are unknown. The topic of how much energy loss equates to mortality or debilitation of fish has not yet received adequate attention.
This work was supported by a National Science Foundation GK-12 Grant to Stephen Bollens and administered as a fellowship to Craig Haskell by Washington State University. The U.S. Geological Survey provided funds through the YES (Youth and Education in Science) program and provided lab space to conduct functional response trials. We thank Kenneth Tiffan and Joe Warren of the U.S. Geological Survey for assistance with functional response trials. William Connor of the U.S. Fish and Wildlife Service retrieved and summarized PIT-tag data from PTAGIS. We also thank Doug Sweet, Ohio Division of Wildlife, for advice on culturing Daphnia and snails. We acknowledge the constructive and helpful reviews of Andrew Hafs, Kenneth Tiffan, John Harrison, Brian Tissot, and an anonymous reviewer. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
- 1. Naiman RJ, Alldredge JR, Beauchamp DA, Bisson PA, Congleton J, Henny CJ, et al. Developing a broader scientific foundation for river restoration: Columbia River food webs. Proc Natl Acad Sci. 2012;109: 21201–21207. pmid:23197837
- 2. Connor WP, Tiffan KF. Evidence for parr growth as a factor affecting parr-to-smolt survival. Trans Am Fish Soc. 2012;141: 1207–1218.
- 3. Duffy EJ, Beauchamp DA. Rapid growth in the early marine period improves the marine survival of Chinook salmon (Oncorhynchus tshawytscha) in Puget Sound, Washington. Can J Fish Aquat Sci. 2011;68: 232–240.
- 4. Tomaro L, Teel D, Peterson W, Miller J. When is bigger better? Early marine residence of middle and upper Columbia River spring Chinook salmon. Mar Ecol Prog Ser. 2012;452: 237–252.
- 5. Hanson PC, Johnson TB, Schindler DE, Kitchell JF. Fish bioenergetics 3.0. Madison: University of Wisconsin, Sea Grant Institute, WISCU-T-97-001; 1997.
- 6. Thompson JN, Beauchamp DA. Growth of juvenile steelhead Oncorhynchus mykiss under size-selective pressure limited by seasonal bioenergetic and environmental constraints: bioenergetic constraints on o. mykiss. J Fish Biol. 2016; pmid:27397641
- 7. McConnaha WE, Williams RN, Lichatowich JA. Introduction and background of the Columbia River problem. Pages 1–28 In, Williams R. E., editor, Return to the River. Amsterdam: Elsevier; 2006.
- 8. Dauble DD, Page TL, Hanf RW Jr.. Spatial distribution of juvenile salmonids in the Hanford Reach, Columbia River. US Natl Mar Serv Fish Serv Fish Bull. 1989;87: 775–790.
- 9. Coutant CC, Whitney RR. Hydroelectric system development: effects on juvenile and adult migration. in Williams R N, editor, Return to the River- Restoring Salmon to the Columbia River. Amsterdam: Elsevier Academic Press; 2006. pp. 249–324.
- 10. Haskell CA, Tiffan KF, Rondorf DW. Food habits of juvenile American shad and dynamics of zooplankton in the lower Columbia River. Northwest Sci. 2006;80: 47–64.
- 11. Emerson JE, Bollens SM, Counihan TD. Seasonal dynamics of zooplankton in Columbia—Snake River reservoirs, with special emphasis on the invasive copepod Pseudodiaptomus forbesi. Aquat Invasions. 2015;10: 25–40.
- 12. Rondorf DW, Gray GA, Fairley RB. Feeding ecology of subyearling Chinook salmon in riverine and reservoir habitats of the Columbia River. Trans Am Fish Soc. 1990;119: 16–24.
- 13. Muir WD, Emmett RL. Food habits of migrating salmonid smolts passing Bonneville Dam in the Columbia River, 1984. Regul Rivers Res Manag. 1988;2: 1–10.
- 14. Haskell CA, Beauchamp DA, Bollens SM. Trophic interactions and consumption rates of subyearling Chinook salmon and nonnative juvenile American shad in Columbia River reservoirs. Trans Am Fish Soc. 2017;146: 291–298.
- 15. Haskell CA. From salmon to shad: Shifting sources of marine-derived nutrients in the Columbia River Basin. Ecol Freshw Fish. 2017; 1–13.
- 16. Haskell CA, Tiffan KF, Rondorf DW. The effects of juvenile American shad planktivory on zooplankton production in Columbia River food webs. Trans Am Fish Soc. 2013;142: 606–620.
- 17. Brachvogel R, Meskendahl L, Herrmann J-P, Temming A. Functional responses of juvenile herring and sprat in relation to different prey types. Mar Biol. 2013;160: 465–478.
- 18. Chipps SR, Wahl DH. Bioenergetics modeling in the 21st century: reviewing new insights and revisiting old constraints. Trans Am Fish Soc. 2008;137: 298–313.
- 19. Hartman KJ, Hayward RS. Bioenergetics. Analysis and interpretation of freshwater fisheries data. Bethesda, Maryland: American Fisheries Society; 2007. pp. 515–560.
- 20. Eggers DM. Comments on some recent methods for estimating food consumption by fish. J Fish Res Board Can. 1979;16: 1018–1021.
- 21. Brett JR, Shelbourn JE, Shoop CT. Growth rate and body composition of fingerling sockeye salmon, Oncorhynchus nerka, in relation to temperature and ration size. J Fish Res Board Can. 1969;26: 2363–2394.
- 22. Mantua N, Tohver I, Hamlet A. Climate change impacts on streamflow extremes and summertime stream temperature and their possible consequences for freshwater salmon habitat in Washington State. Clim Change. 2010;102: 187–223.
- 23. Marine KR, Cech JJ. Effects of high water temperature on growth, smoltification, and predator avoidance in juvenile Sacramento River Chinook salmon. North Am J Fish Manag. 2004;24: 198–210.
- 24. Sommer TR, Nobriga ML, Harrell WC, Batham W, Kimmerer WJ. Floodplain rearing of juvenile Chinook salmon: evidence of enhanced growth and survival. Can J Fish Aquat Sci. 2001;58: 325–333.
- 25. Beauchamp DA. Bioenergetic ontogeny: linking climate and mass-specific feeding to life-cycle growth and survival of salmon. American Fisheries Society Symposium. 2009. p. 19.
- 26. Plumb JM, Moffitt CM. Re-estimating temperature-dependent consumption parameters in bioenergetics models for juvenile Chinook salmon. Trans Am Fish Soc. 2015;144: 323–330.
- 27. Huntington C, Nehlsen W, Bowers J. A survey of healthy native stocks of anadromous salmonids in the Pacific Northwest and California. Fisheries. 1996;21: 6–14.
- 28. NMFS (National Marine Fisheries Service). Endangered Species Act. Threatened status for Snake River spring/summer Chinook salmon. Federal Register, 57:14653–14663. 78(April 22, 1992) Apr 22, 1992.
- 29. Koski ML, Johnson BM. Functional response of kokanee salmon (Oncorhynchus nerka) to Daphnia at different light levels. Can J Fish Aquat Sci. 2002;59: 707–716.
- 30. Moss JH, Beauchamp DA. Functional response of juvenile pink and chum salmon: effects of consumer size and two types of zooplankton prey. J Fish Biol. 2007;70: 610–622.
- 31. Galbreath GC, Schneider JL. Sampling zooplankton in lakes. Chapter 18 in Schneider, James C. (ed.) 2000. Manual of fisheries survey methods II: with periodic updates. Ann Arbor, Michigan: Michigan Department of Natural Resources; 2000.
- 32. Real LA. Ecological Determinants of Functional Response. Ecology. 1979;60: 481–485.
- 33. Burnham B K, Anderson A D. Model selection and inference: a practical information-theoretic approach. New York: Springer-Verlag; 2002.
- 34. Gilbreath LG, Smith DA, Ledgerwood RD, Pool SS, Grabowski SJ. Limnological investigations in John Day Reservoir including selected upper reservoir habitats, April 1994 to September 1995. Seattle, WA: NOAA Fisheries; 2000 p. 140.
- 35. Culver DA, Boucherle MM, Bean DJ, Fletcher JW. Biomass of freshwater crustacean zooplankton from length—weight regressions. Can J Fish Aquat Sci. 1985;42: 1380–1390.
- 36. Brett JR. Satiation time, appetite, and maximum food intake of sockeye salmon (Oncorhynchus nerka). J Fish Res Board Can. 1971;28: 409–415.
- 37. Stewart DJ, Ibarra M. Predation and production by salmonine fishes in Lake Michigan, 1978–88. Can J Fish Aquat Sci. 1991;48: 909–922.
- 38. Craddock DR, Blahm TH, Parente WD. Occurrence and utilization of zooplankton by juvenile Chinook salmon in the lower Columbia River. Trans Am Fish Soc. 1976;105: 72–76.
- 39. DART (Data access in real time). Interactive data resource relating to Columbia River salmon populations. Columbia Basin Research, School of Aquatic and Fisheries Sciences, University of Washington, Seattle. 2016;
- 40. Moss JH, Beauchamp DA, Cross AD, Farley EV, Murphy JM, Helle JH, et al. Bioenergetic model estimates of interannual and spatial patterns in consumption demand and growth potential of juvenile pink salmon (Oncorhynchus gorbuscha) in the Gulf of Alaska. Deep Sea Res Part II Top Stud Oceanogr. 2009;56: 2553–2559.
- 41. Trudel M, Tucker S, Morris JFT, Higgs DA, Welch DW. Indicators of energetic status in juvenile coho salmon and Chinook salmon. North Am J Fish Manag. 2005;25: 374–390.
- 42. Luecke C, Brandt D. Estimating the energy density of daphnid prey for use with rainbow trout bioenergetics models. Trans Am Fish Soc. 1993;122: 386–389.
- 43. Cummins KW, Wuycheck JC. Caloric equivalents for investigations in ecological energetics. Int Assoc Theor Appl Limnol. 1971;18.
- 44. Stockwell JD, Bonfantine KL, Johnson BM. Kokanee foraging: a Daphnia in the stomach is worth two in the lake. Trans Am Fish Soc. 1999;128: 169–174.
- 45. Haskell CA, Stanford JA. Ecology of an estuarine mysid shrimp in the Columbia River (USA). River Res Appl. 2006;22: 739–753.
- 46. Brooks JL, Dodson SI. Predation, Body Size, and Composition of Plankton. Science. 1965;150: 28–35. pmid:17829740
- 47. Pace ML, Findlay SEG, Lints D. Zooplankton in Advective Environments: The Hudson River Community and a Comparative Analysis. Can J Fish Aquat Sci. 1992;49: 1060–1069.
- 48. Gosselain V, Viroux L, Descy J-P. Can a community of small-bodied grazers control phytoplankton in rivers? Freshw Biol. 1998;39: 9–24.
- 49. Gadomski DM, Barfoot CA. Diel and distributional abundance patterns of fish embryos and larvae in the lower Columbia and Deschutes rivers. Environ Biol Fishes. 1998;51: 353–368.
- 50. Adams JB, Bollens SM, Bishop JG. Predation on the invasive copepod, Pseudodiaptomus forbesi, and native zooplankton in the Lower Columbia River: an experimental approach to quantify differences in prey-specific feeding rates. Gorokhova E, editor. PLOS ONE. 2015;10: e0144095. pmid:26618851
- 51. Dick JTA, Gallagher K, Avlijas S, Clarke HC, Lewis SE, Leung S, et al. Ecological impacts of an invasive predator explained and predicted by comparative functional responses. Biol Invasions. 2013;15: 837–846.
- 52. Guo Z, Sheath D, Amat Trigo F, Britton JR. Comparative functional responses of native and high-impacting invasive fishes: impact predictions for native prey populations. Ecol Freshw Fish. 2016;
- 53. Alexander ME, Dick JTA, Weyl OLF, Robinson TB, Richardson DM. Existing and emerging high impact invasive species are characterized by higher functional responses than natives. Biol Lett. 2014;10: 20130946–20130946. pmid:24522629
- 54. Bollens S, Breckenridge J, Cordell J, Rollwagen-Bollens G, Kalata O. Invasive copepods in the Lower Columbia River Estuary: Seasonal abundance, co-occurrence and potential competition with native copepods. Aquat Invasions. 2012;7: 101–109.
- 55. Duffy EJ, Beauchamp DA, Sweeting RM, Beamish RJ, Brennan JS. Ontogenetic diet shifts of juvenile Chinook salmon in nearshore and offshore habitats of puget sound. Trans Am Fish Soc. 2010;139: 803–823.
- 56. Schindler DE, Rogers DE, Scheuerell MD, Abrey CA. Effects of changing climate on zooplankton and juvenile sockeye salmon growth in southwestern Alaska. Ecology. 2005;86: 198–209.
- 57. Bartell SM, Breck JE, Gardner RH, Brenkert AL. Individual parameter perturbation and error analysis of fish bioenergetics models. Can J Fish Aquat Sci. 1986;43: 160–168.
- 58. Madenjian CP, O’Connor DV, Chernyak SM, Rediske RR, O’Keefe JP. Evaluation of a chinook salmon (Oncorhynchus tshawytscha) bioenergetics model. Can J Fish Aquat Sci. 2004;61: 627–635.
- 59. Brodeur RD, Francis RC, Pearcy WG. Food Consumption of Juvenile Coho (Oncorhynchus kisutch) and Chinook Salmon (O. tshawytscha) on the Continental Shelf off Washington and Oregon. Can J Fish Aquat Sci. 1992;49: 1670–1685.
- 60. PTAGIS (Pit Tag Information System). The Columbia Basin PIT tag information system. Pacific States Marine Fisheries Commission, Portland, Oregon. In: PTAGIS (Pit Tag Information System) [Internet]. 2016. www.ptagis.org/ (October 2016)
- 61. Huusko A, Maki-Petays A, Stickler M, Mykra H. Fish can shrink under harsh environmental conditions. Funct Ecol. 2011;25: 628–633.
- 62. Quinn TP. The Behavior and Ecology of Pacific Salmon and Trout. University of Washington Press; 2004.
- 63. Woodson L, Wells B, Weber P, MacFarlane R, Whitman G, Johnson R. Size, growth, and origin-dependent mortality of juvenile Chinook salmon Oncorhynchus tshawytscha during early ocean residence. Mar Ecol Prog Ser. 2013;487: 163–175.
- 64. Claiborne A, Miller J, Weitkamp L, Teel D, Emmett R. Evidence for selective mortality in marine environments: the role of fish migration size, timing, and production type. Mar Ecol Prog Ser. 2014;515: 187–202.
- 65. Mains EM, Smith JM. The distribution, size, time and current preferences of seaward migrant Chinook salmon in the Columbia and Snake rivers. Washington Department of Fisheries; 1956.
- 66. NMFS (National Marine Fisheries Service). Endangered Species Act Section 7(a)(2) Consultation Biological Opinion and Magnuson-Stevens Fishery Conservation and Management Act Essential Fish Habitat Consultation. 2008 May.