NOAA’s Dr. Sean Hayes reviews the results of studies of salmon predation in the San Joaquin River
At the April 19th meeting of the State Water Resources Control Board, Board members heard an informational update from Dr. Sean Hayes with NOAA’s Southwest Fisheries Science Center on the results of the latest studies of predation on salmonids in the Delta, and specifically the San Joaquin River.
Dr. Sean Hayes began by saying he would be discussing the work being done in the San Joaquin, greater Delta and Central Valley watershed that is focusing on salmon survival issues and their challenges in the Central Valley.
“California is obviously many things, and we have a lot to think about in managing California and balancing resources,” he said. “The challenge is managing the water to support all of this infrastructure in California while at the same time, protecting and preserving natural resources that are of societal concern. This is particularly hard with salmon in this case, because we’ve altered roughly 65 million acres of their habitat and less than 15% of it is unchanged; a colleague of mine recently corrected me that we’re more like 4% is unchanged or altered.”
The heart of California’s circulatory system, in terms of both fish and water, is the Delta. Prior to the reclamation efforts in the 1800s, the Delta was basically a large wetland habitat for fish and wildlife. The Delta’s wetlands were then reclaimed in the late 1800s and early 1900s for the purposes of agriculture and urbanization.
“In addition to the physical changes, we also now have to manage it in an aggressively freshwater state and maintain the X2 line to ensure that there isn’t seawater intrusion for any of the communities or farms in the area, and also to ensure freshwater transport to the Clifton Court forebay,” he said. “All of this provides a situation where it is not natural habitat for salmon in this situation, and the fish have many challenges as a result.”
“Ironically, while many of these challenges seem obvious, we’re tasked with proving it because of the potential stakeholder conflicts and challenges on this particular issue,” Dr. Hayes said. “So we’ve been tasked first with the issue of establishing whether or not fish are dying in the river versus in the ocean. The fish are dying in the river, so we really have these simple questions of identifying where, when, and what are the causes.”
He presented a graph of results from an acoustic telemetry study conducted from 2007 to 2011 on chinook salmon that involved putting acoustic tags in fish were released from the hatchery in Battle Creek near Redding. He explained that the left axis shows survival with the top of the axis being 100% survival, and the y axis shows the percentage of fish that survive as they are moving through the Delta to the Golden Gate and ultimately the ocean.
“There is a general trend that in the years of 2007 through 2010, roughly 3% of the hatchery fish, once released from the hatchery, actually made it to the Golden Gate,” he said. “We had a wetter year in 2011 where roughly 16% made it to the river, and this 16% – while sad – is on par with what we expect in a fairly healthy river system, compared to where the studies have been done farther north along the coast in large river systems. The 3% survival is not so good.”
Dr. Hayes noted that this study subsequently has been done with all other hatchery stocks in the Sacramento River as well as wild stocks of fall and spring-run Chinook in the Sacramento.
“The most effective strategy for the hatchery fish seems to be swim fast and get out of the river,” he said. “Compiled with that is the fact that if they don’t, roughly between 84 and 90% of them will die within the first two weeks at liberty after released from the hatchery. And there is a relationship here where in one wet year, survival was 5 times better, suggesting that water does appear to contribute to greater survival to the ocean.”
Since 2011, fish returning as adults either in the harvest or to spawn in the hatchery have been tracked using coded wire tag technology to measure marine survival. “In the dry years, in this example, roughly a million fish were released and only 28,000 made it to the ocean; in a wet year, about 160,000 made it to the ocean,” he said. “However, all the mortality historically was accounted for as marine mortality. If you factor out all the in-river mortality, for one of the year classes, it was estimated that marine survival was .8%. If you factor out all the fish that died in the river that year, the marine survival is actually 20%, so 95% of the fish died in the first two weeks after they were released from the hatchery; then during the next two years, the fish had a one in five chance of surviving, so in many respects they went to the ocean for safety, compared to the conditions that we have in the marine environment.”
So, the answer to the initial question, do fish die in the river is yes, and more than we would really like to see biologically, said Dr. Hayes, noting that through the large acoustic studies, it seems to be due to a combination of lower flows and warmer water conditions. As to where they are dying in the river, he said that upstream habitat can be quite poor, depending on the stock of salmon being tracked. “In the case of winter run, they seem to do very well in the highest portions of the river, but other stocks we’ve tracked, not so well. The lower mainstem of the Sacramento River seems to have good survival; but this is literally downtown Sacramento and the reality is there’s no structure or habitat; it’s just a travel corridor, so survival is good put growth is poor. Delta survival is poor for most of the stocks as is Bay survival. The why are the fish dying is really the $100 million question that we’re tasked with addressing.”
Studying mortality is a math formula that should be as simple as being able to understand and identify what is the source of mortality, and then measuring how frequently that source occurs. “It gets more complicated when we actually have multiple potential sources,” he said. “In the case of the Central Valley, we’re considering predators, water diversions, temperature, contaminants – all of these things can contribute; the frequency by which they contribute can also be influenced directly or indirectly by flow, available growth rate of potential habitat, turbidity, and other variables. So growth is definitely a function of food source and habitat to historical habitat, which lacking.”
Dr. Hayes said he was tasked to study the effects of predation. “When I first got involved with studying predation by bass and other predators in the Central Valley, I was asked the question, if bass have been present in the Central Valley for 130 years, why are they a central problem now?,” he said.
“It’s a very fair question. The reality is the question that the Delta is an unstable environment, ecologically or biologically, and every year, over time, there have been constant introductions of new invasive species which continue to perturb or alter the food web in the system. In case of the Asian clams that were introduced in the lower portion and eastern part of the bay from ballast water in the 80s, they resulted in two side effects: One, clams directly competed with juvenile striped bass for food, forcing juvenile striped bass into a more vociferous lifestyle at an earlier age; in addition, they increased water clarity which made small juvenile fish like juvenile salmon, which rely on avoidance from visual predators, more difficult. Aquatic plants, such as Egeria or hyacinth have contributed; In the case of the water hyacinth, its introduction and subsequent explosion in the Delta has come with a transition from 35 to 75% of the fish community in the Delta fish community to the centrarchidae which includes the black basses and some fish in the last 20 years.”
So just how bad is the predator issue? “Someone did a very elegant model and basically figured out how many pounds of fish the striped bass population needs to eat to survive every year, and that estimate was roughly on order of 25 million kilograms of fish that striped bass need to eat – crayfish, etc. to meet their energetic requirements every year,” said Dr. Hayes. “So working with juvenile salmon and having a rough estimate of the biomass of all the juvenile salmon in the Central Valley, I did a very conservatively high estimate and came up with a back of the envelope estimate of roughly 240,000 kilograms of juvenile salmon, which means that if striped bass were to eat every single salmon in the Central Valley, it would meet 1% of their diet. This isn’t an accusation of the bass; it’s just saying they could easily account for all missing salmon in the Central Valley.”
In order to study and evaluate predation, they worked with a set of reaches in the lower San Joaquin with the priority focus being the Head of Old River. “We collaborated with the Department of Fish and Wildlife, and used multiple electro-fishers to go in and basically estimate the abundance of predators in the reach,” he said. “Then we identified what was there, and we evaluated and found that there were four particular predators: striped bass, largemouth bass, channel cats and white catfish.”
They then conducted genetic stomach content analysis of the frequency of salmon in their stomachs, which was a basic test for a positive presence or absence of salmon. “It was a test for the genetic signature of a salmon having been eaten roughly in the last 24 hours,” he said. “This is just a binary test; a positive test doesn’t tell you whether it was one salmon in the stomach or a hundred salmon in the stomach.”
Overall, the frequency of finding salmon in the stomachs of predators in general was pretty low, Dr. Hayes said, presenting a graph of the results. “On the x axis is the various predators that we sampled: striped bass, largemouth bass, channel cats, and white cats respectively; then what we have is the frequency of detection: with striped bass, roughly 3-5% of the stomachs were positive for chinook salmon; for largemouth bass, it was 3% of the stomachs were positive; channel cats were surprisingly high at 26%, but it was a small sample of channel cats, so there is considerable variance in that; and then roughly 3% of the white catfish had salmon detected in their stomachs.”
In terms of figuring out how many salmon are being eaten; Dr. Hayes acknowledged these are relatively low numbers. “When we thought about this a little more deeply, we considered that when a prey item is rarely detected in the stomach – on the order of only 3 to 5% – the odds are in that two or three bass in the salmon in their stomachs, that they had more than one is probably very low, so we developed a conservative estimate of a positive detection = 1 salmon in the stomach as a very conservative low estimate of predation rate to equal the actual rate.”
He then presented a graph depicting salmon mortality in the reach above the Head of Old River. “This is the estimate of striped bass that were in the 1km reach that we sampled around the Head of Old River in 2014 and in 2015,” he said. “The x axis is the range of population density, so between 400 and 600 striped bass were in that 1km reach that day. In 2015, it was more like 950 to 1500 striped bass were in that reach that one day, and then based on the stomach content analysis, we were able to estimate the number of juvenile salmon that were likely eaten by just striped bass in the one reach on that one day. In 2014, it’s roughly between 35 and 60 fish. The central tendency is the dark gray line and the gray areas are the confidence intervals around that. In 2015, we estimated roughly 80 to 110 juvenile salmon were eaten in that 1km of river in that one day by that one species. We have similar data for largemouth bass and the numbers are comparable, so again another large impact.”
Dr. Hayes noted that he is showing the results for the Head of Old River because it is the focus; similar studies down in other reaches with more subtle habitat features had much lower numbers – 1 to 20 fish per day being consumed. “This is the upper estimate of what predation potential is,” he said.
The catfish had similar numbers, but since catfish are harder to catch with the methods used for sampling, they don’t have accurate estimates of the catfish population. “Our suspicion is that the white catfish population probably rivals the bass numbers in terms of population densities; the channel cats, they seem to be the most voracious of the predators, but their numbers are probably smaller than the other three predators.”
While that study gave raw estimates of mortality and consumption, the real question in terms of management is how do all these other variables come into play, and how do they interact with each other, he said. To test this, they adapted NMFS’s marine hydroacoustic surveys to conduct an acoustic sonar survey that would simultaneously map the bathymetry of the habitat, as well as count all the individual fish targets in the habitat. They then used multiple methods to identify the size targets of the fish to determine the predators and the juvenile fish.
They were then tasked with conducting a density manipulation experiment. “We were asked to basically identify if we were to remove predators from this environment, would it increase salmon survival? And so we effectively did that,” he said. “We surveyed three reaches: a control reach, a removal reach, and an addition reach; we sent juvenile salmon through all of those reaches before we manipulated the densities of any of the reaches. We then went in and removed most of the predators from the removal reach and then increased predator density in the addition reach, so that we had a strong experimental effect. We sent more juvenile salmon through the reaches to evaluate relative survival of both acoustically tagged fish as well as an alternative method.”
Dr. Hayes then described the alternative method that was used. “It is very creative but it’s the reality of what we needed to do to do the research,” he said. “We sent a live salmon smolt tethered to a detection device to evaluate whether or not it was predated upon while drifting through the reach. The device is a buoy system with a GPS tracker, a Go Pro camera, and a hook timer; the salmon is basically tied to the end of it. It is drifted through our experimental reaches and then evaluated about whether it survived or not. We did this thousands of times through our various experimental reaches.”
Dr. Hayes emphasized that this study was not meant to quantify absolute predation rate. “Obviously, this fish is far more susceptible than if it were free swimming,” he said. “The purpose is to quantify relative predation rate under varying features like environmental conditions, time of day, temperature, habitat features, etc.”
Dr. Hayes presented a slide with two pictures; the picture on the left is from a series of deployments of 20 tethers along a reach in the San Joaquin. The tether drifts from south to north or from the bottom to the top of the picture; each colored line denotes the individual path that a specific tether buoy took drifting through it, and the red dots denote where in the habitat a predation event occurred. “What you notice is that the majority of the predation events happen in a very small portion of the habitat, showing you that predation is not random,” he said.
The picture on the right is from a hydroacoustic data survey of the same reach. Dr. Hayes noted that the blue dots show a cluster of predators in that habitat feature, while at the same time there are other blue dots where there are clusters of predators but predation didn’t really seem to happen. “This provides a different way of thinking about this situation, which is that predation is non-random and actually highly focused in small percentages of the habitat,” he said. “Our goal then is to identify what are those characteristics where salmon are particularly susceptible to predation. It seems like they probably make it through 80% of the habitat completely unscathed, so what our goal is to do is to identify those features in the other 20% of the habitat that make them so susceptible to predation.”
“By drifting these tethers through, we’re able to look at all of these things in concert, such as habitat features, temperature, and dissolved oxygen,” he said. “Because the tether is drifting, we’re able to evaluate the effects of pumping, tide, flow, OMR – all of those things on the movement rate of the animal through the environment as a covariate on the susceptibility of predation as well as predator density, turbidity, light, time of day. We’re working towards developing a predictive model that could then be used to look across the landscape and identify where mortality is likely to be happening in high frequency areas. We don’t necessarily know what you’d do with that, but it would give you a new way of thinking about it.”
The other thing to consider is that it may not all just be predation, he said. “When we started getting catfish as a predator, our initial concern was when we were doing the tether work late in the season in May in 2014, and again in 2015, during drought conditions, the water got so hot that as soon as put the juvenile salmon in the river, they just died,” he said. 
“So the reciprocal question was, maybe it’s not the predators, maybe they are just eating dead salmon off the bottom of the river, so that’s something we’re working to consider and account for in our studies. We’re doing that by using swim flumes and exposing fish to the river temperatures, and then exercising them and evaluating how their swimming ability or survival capacity changes after exposure to the river with time. We can basically exercise them and challenge them in that.”
Dr. Hayes said that our current management tools for salmon recovery have emphasized flow management and hatcheries, and he acknowledged we’ll probably always have to do that. “But working with just those two features, we’ve been struggling for a long time, and we’re still struggling with recovery, so what we really need are additional knobs that we can turn with the goal of getting us towards a more stable state of natural production; ideally something that we could fix once, and then have to put less effort into it in the future.”
So the question was posed that if we were to just get rid of all the bass, would that stop the predation in this situation? “Unfortunately, ecology is far more complex than that,” Dr. Hayes said, then giving an analogous metaphor to explain why he is concerned predator control might not work in a situation like the Delta. Drawing on the cartoon character Homer Simpson, he explained: “Let’s consider that Springfield has a series of restaurants that you can eat at, and most of them serve broccoli. One of them serves doughnuts. Homer’s not very happy about eating at the broccoli restaurants, but he’s getting by and he can survive. However, the doughnut restaurant has a line out the door. If you were to remove half of the Homers from this situation, even the doughnut one, you’re still going to have a big line out the door at the doughnut restaurant in Springfield.”
“The situation I am really proposing that this analogy of getting rid of half of the bass or predators in the Central Valley is such that unless we manage the hotspots where the salmon are really getting predated upon, we won’t really be in a situation where we are reducing predation just through raw population control,” he said. “It’s really until we turn the doughnut restaurant into a broccoli restaurant that Homers will redistribute to other areas and continue feeding.”
Dr. Hayes said from a management perspective, he doesn’t have any confidence that predator control would be a successful way to enhance salmon recovery. “The reality is the predators are probably providing a stabilizing situation in the Delta and salmon are irrelevant to them as a diet source; they could eat all of them, they could all go away tomorrow and they’d still be eating something else,” he said. “Most of what they are eating is each other as well; in our genetic content studies, we find that striped bass eat largemouth bass, largemouth bass eat striped bass, and probably some cannibalism. In addition, the majority of the Delta fish community is a lot of small non-native fishes at this point, and those non-native fishes compete with juvenile salmon for food resources, and if we remove the predators from this situation, their populations might explode in such a way that it had deleterious effects on the salmon and smelt that we’re trying to recover, so those are things we want to be very cautious about coming in with a drastic approach.”
“My advice is that salmon aren’t contributing to the Delta food web; I recommend leaving it alone,” he said. “However, strategic hotspot control that focuses on these areas where we know mortality is an issue is a tool that we could use. How that tool would be implemented waits until we finish the analysis of what this model is; it may be that we allow free fishing between buoy marked areas where we know predators are a problem, or we understand the habitat features that are causing it and we reengineer the habitat so it’s less likely to be a problem, but all of these things are far more effective than trying to cut the bass population in half.”
As for the timeline for the model, the first model which is calibrated just for the Lower San Joaquin will be ready for release to the public at the end of the year. They have funding to calibrate the model to the greater Delta based on field studies in 2017, and they are hoping to find funding to calibrate the model to the Sacramento Basin with field efforts in 2017 and 2018.
“The real challenge for salmon and our other native species is that we changed the habitat from a way that gave them the advantage to giving an advantage to the non-native species, and until we really focus on floodplains and habitat restoration, that’s going to continue be a challenge,” he said. “Fish die. In the salmon lifecycle, for every 4000 eggs that are laid, a stable healthy population only gets 2 adults back, so we have to accept that, but we can manage when and where it happens. The idea of habitat restoration where we can give juvenile fish growth opportunity in floodplains and wetlands which they don’t currently have is probably where our stable state solution requires.”
“There will always be high mortality in those habitats; that has always been the case,” he said. “Before we introduced the bass, it was birds. People often ask me what is the idea of a high growth/low risk habitat in California. We have them, they are called hatcheries, and they haven’t resulted in solution a, so we’re still struggling with that one.”
“That being said, there really are creative opportunities, if we can provide some protection to the fish as well as provide them a life history diversity,” Dr. Hayes said, presenting a picture with a wild smolt from the Mill Creek Basin in the northern Sacramento Valley on the top, and a picture of a fish the same age from the Sutter Basin who spent time on a rice field paddie, which is a surrogate for lost wetland habitat. “We’re never going back to the way things were, but through creative use of existing habitat features there may be alternatives that could allow us to find balance between society’s economic and human resource needs while preserving resources.”
Sign up for daily email service and you’ll never miss a post!
Sign up for daily emails and get all the Notebook’s aggregated and original water news content delivered to your email box by 9AM. Breaking news alerts, too. Sign me up!
