Sea lion and fisheries interaction work in the Columbia River is carried out by two federal agencies, three states, and CRITFC. There is a tremendous amount of coordination and collaboration among organizations. A list of entities and a brief detail of work they carry out includes:
- NOAA Fisheries is the regulatory agency for sea lions and the Northwest Science Center at Point Adams, OR assesses upriver spring chinook mortality between the estuary and Bonneville Dam (Rub et al. 2019).
- The US Army Corps of Engineers (USACE) produces sea lion abundance and predation estimates within ½ mile of Bonneville Dam (Tidwell et al. 2019) and funds dam-based hazing that is carried out by the USDA.
- The Oregon Department of Fish and Wildlife (ODFW), Washington Department of Fish and Wildlife (WDFW), and Idaho Department of Fish and Game (IDFG) jointly hold a letter of authority under section 120 to lethally remove California sea lions at Bonneville Dam (Wright et al. 2018). ODFW is also carrying out similar removals at Willamette Falls along with estimating sea lion abundance and predation and conducts a branding program and estimates abundance at the East Mooring Basin in Astoria, OR (ODFW 2017).
- CRITFC conducts boat based hazing near Bonneville Dam, and estimates sea lion abundance, and distribution between Bonneville Dam and Astoria weekly from March 1 through mid-May (Hatch et al. 2019). In the future, we expect to be involved in lethal removal of sea lions at Bonneville Dam.
The goal of all agencies represented is to limit the impact of sea lion predation on ESA listed salmon, sturgeon and Pacific lamprey. All the above research and management activities are coordinated among the groups to reduce overlap and maximize the outcomes using limited resources. There are specific areas where CRITFC collaborates with other sea lion interaction work on individual projects and that includes:
- NOAA Fisheries: Participation in Marine Mammal Protection Act section 120 Pinniped Fishery Interaction Task Forces for sea lion removals at Bonneville Dam and Willamette Falls. Coordinated acoustic telemetry and mortality analysis with NW Science Center (Rub et. al. 2019).
- US Army Corps: Coordinate hazing activities, functional response analysis, assistance from USACE observers in locating accelerometer tagged sea lions.
- The States (ODFW, WDFW, IDFG): Coordinated applications for sea lion removal (Section 120, PL 115-329) at Bonneville Dam and Willamette Falls, and hazing efforts, we relied on the states trapping program to tag sea lion with acoustic, GPS, and accelerometer tags. We jointly tracked tagged animals.
- We provide all groups with our weekly tandem boat counts, and Phoca Rock and East Mooring Basin haul-out counts.
Since the passing of PL 115-329, the states and tribes have collaborated to craft a joint application for permanent removal of sea lions upstream of RM 112 in the Columbia River and all tributaries with anadromous species. This application will go to NOAA Fisheries for processing, which requires an application, an EIS, NEPA compliance, NOAA convenes a Task Force, etc. The entire process takes at least 1 year to complete. Note that PL 115-329 was signed in December 2018 about 1 month after the review process started for this project. Consequently, not all plans are fully developed and many components of the removal program, some that generated questions from the ISRP, will be stipulated by NOAA Fisheries (i.e. carcass deposal and culling techniques). However, we have discussed the notion that during the spring, CRITFC and WDFW will take over the removal program at Bonneville Dam and ODFW will focus on removals at Willamette Falls to maximize resources and we all work together to remove Steller sea lions at Bonneville Dam during the summer and fall period.
Specific Questions:
1. More detail is needed about the statistical methods used for the conditional Lincoln- Peterson estimators of sea lion abundance from the tandem boat surveys. How would abundance estimates differ if a sequence of additional observation boats were used (e.g., 2, 3, 4 or more) for a reach? Might drones be used?
We will answer these questions in reverse order so the follow up questions don’t get lost in the lengthy abundance methodology response.
Adding More Boats
Adding more boats to the abundance and distribution surveys would likely decrease the variance in abundance estimates proportionally. Additional boats would provide more robust estimates and would be beneficial especially at high sea lion abundance, but at lower abundance the added precision would be in fractions of sea lions and the overall estimate would not be meaningfully different. However, there are practical, logistical, and financial tradeoffs to the gain in precision. These surveys take most daylight hours to run the 140 river miles between Bonneville Dam and Astoria, OR and the survey crews stay the night in Astoria. Each boat has 2 people on board and an additional person uses a truck to pull the boat trailer from Bonneville Dam to Astoria for each survey. The current survey using two boats utilizes the following resources: 2 boats, 4 people on the boats, 2 drivers with trucks pulling trailers, and a third driver and vehicle to shuttle the truck drivers from Astoria back to Portland, for a total of 2 boats, 7 people, and 3 vehicles. Adding one more boat would require an additional boat, truck and trailer, and 3 more people. Though reducing the variance on the abundance estimate would be nice, in practical terms we would need additional resources or reallocate resources from activities that we consider higher priority (i.e. removal efforts).
Drones
Deploying a drone(s) for abundance surveys is something we have thought about, though not really seriously considered to date. When large numbers of sea lions are in the river, animal distribution is generally clumped at confluences of major tributaries. Relatively accurate sea lion counts could be made during those times with properly equipped drones. Logistical considerations for drones include the capability of flying 140 miles, ability to operate in all weather conditions including rain, freezing rain, snow, and wind speeds up to 30mph which is generally the upper limit for our boat operation. There are also operational limits flying in the proximity of Portland International Airport. Another consideration is the amount of time required to review video images to obtain sea lion counts, many times the review period in the lab far exceeds the amount of time acquiring the images.
At some point in the future this seems worth exploring and could ultimately result in more accurate and precise estimates at a lower cost. However, some distribution data would be lost if light drones were used in specific areas and didn’t cover the entire 140 mile area. Currently, BPA has a moratorium on purchasing drones until regulations are better defined. In short, deploying drones for the abundance surveys may be an excellent solution but restrictions on use and purchasing could put practical use a few years out into the future.
Tandem Boat Survey Methodology
A statistical abundance estimation model was applied to the tandem boat survey data using a conditional Lincoln-Petersen estimator that assumes each boat has its own probability of detecting an animal if the animal is present (p1 for boat 1 and p2 for boat 2). Further, the model assumes that each statistical zone is potentially unique in how the observation process occurs, i.e., that p1 in Zone 1 is not assumed to be the same as p1 in Zone 2 (nor p2). Further still, we formulate the probability of detection such that we estimate the probability of detecting an animal, and also the rate at which that probability declines as the number of pinnipeds increases.
The method in its most general form assumes the following structure in a given year:
p1iw= probability that boat 1 will see an animal given the animal is present in zone i in week w.
p2iw= probability that boat 2 will see an animal given the animal is present in zone i in week w.
ai= intercept parameter to predict the probability that boat 1 will see an animal (all zones).
bi= intercept parameter to predict the probability that boat 2 will see an animal (all zones).
g= density decay parameter for the probability that either boat will see an animal in any zone.
q1iwis the probability that boat 1 will not see an animal in zone i in week w.
q2iwis the probability that boat 2 will not see an animal in zone i in week w.
x11iw= number of animals seen by both boats in zone i in week w.
x10iw= number of animals seen by boat 1 but not boat 2 in zone i in week w.
x01iw= number of animals seen by boat 2 but not boat 1 in zone i in week w.
x00iw= the number of animals not seen by either boat in zone i in week w.
n1iw= x11iw+ x10iwis all animals seen by boat 1 in zone i in week w.
n2iw= x11iw+ x01iwis all animals seen by boat 2 in zone i in week w.
m2iw= x11iwis the animals seen by both boats in zone i in week w.
riw= n1iw+ n2iw– m2iwis the number of distinct animals seen in zone i in week w.
Note that p1i and p2iare derived from the aand bparameters, as well as the abundance. Since the abundance changes from week to week, the values of p1i and p2i also change from week to week. p1i andp2iare calculated using the equations p1iw= 1/(1+exp(-ai-griw)) and p2iw= 1/(1+exp(-bi-griw)). The equations predict that detection probabilities decline from maximum intercept values of 1/(1+exp(-ai)) and 1/(1+exp(-bi)) to a minimum value at infinite abundance, at which point the fraction of animals that an observer can see would converge to zero. Since the statistical method is using empirical observations of tandem observers to infer the abundance not observed, we can’t scale the detection probabilities with the actual abundance. Instead, we use riwas a surrogate for abundance. Since riwis the distinct number of animals seen, it differs from total abundance only by the number not seen by either observer, i.e.: x00iw.
Since tandem surveys were not conducted every week in each zone, total in-river abundance estimates across all zones was not possible for all weeks, i.e., there is no estimate for the abundance in some zones in some weeks. Furthermore, in some time periods and zones, only one boat was used. In the case where only one boat conducted a survey, we used the detection probability estimates obtained from the tandem survey to scale the observation of a single boat to obtain an abundance estimate.
We assumed that the intercept parameter for the probability of boats seeing an animal given that and animal is present in a zone does not differ among years. We used a conditional multinomial method where we use the assumption that all probabilities must add to one, so that p* = 1- (1-p1)(1-p2) (dropping the subscripts iand wfrom this and subsequent equations) is the probability that an animal is seen at least once. We calculated each week and zone likelihood using the conditional multinomial likelihood equation:
L(
, 
, 
|a,b,g)= 
where, the yearly likelihood is the product across all zones and weeks. 2012, 2013, 2014, 2015, and 2016 each had distinct likelihoods. The sum of the negative logarithm of the likelihoods across all years was minimized to obtain estimates for the four ai and bi parameters and the g parameter – a total of 9 estimated parameters.
Since p* is a function of p1and p2, we only need to obtain the values of p1and p2to maximize the likelihood of the data give the parameters. The predicted value of Niwis obtained using the equation Niw= riw/ p*. Thus by obtaining estimates of p1and p2, we have an estimate of N for each sample period because r is the observed data.
We assumed that theaiandbido not differ across years. In 2013, we used the 2012 estimates of p1and p2as prior probabilities in a Bayesian estimation because there was insufficient information in Zones 2, 3 and 4 surveys to get an estimate. The 2014 estimate did not use a prior probability. The estimates of p1and p2came entirely from the 2012, 2013, and 2014 tandem surveys, with Zones 3 and 4 having 5 tandem boat survey samples spanning the ten-week survey period between 2013 and 2014. In 2015, we treated all years as having distinct detection probabilities for each zone, but the differences in detection probabilities was predicted by the density, with griw predicting the yearly deviate from the intercept maximum detection probability for each zone and year, and all years having the same intercept parameter for each zone. The same method was applied in 2016.
We minimized the negative log of the likelihood equation above by implementing the likelihood in AD Model Builder, a freely available open source non-linear optimization software. We further investigated variation in detection probability and abundance using Markov Chain Monte Carlo (MCMC) simulations. We used a Metropolis Hastings algorithm to generate random samples from the posterior densities the intercept parameters of p1 and p2, and the density decay parameter, and we used random draws from those densities to reconstruct a random sample of the abundances in each statistical period. We sampled 10,000 random samples from a chain of 1,000,000 samples after discarding the first 100,000 “burn-in” samples.
1. 2. The estimation of abundance, distribution, and predation of sea lions is described as a continuation of the previous measurements with tandem boat observation and accelerometers. Additional information should be provided to describe how these two measurements or their analyses will be changed or improved by the proposed research. How will these results be synthesized and reported? How will they evaluate their methods with comparisons to measures by other research groups or modifications of their protocols? More detail is needed about the functional responses fit to the predation data.
Abundance, Distribution, and Predation
At this time, we are proposing to continue these two measures to estimate abundance, distribution and predation. The ODFW estimates sea lion abundance and predation at Willamette Falls and abundance the East Mooring Basin in Astoria, Oregon. The USACE estimates sea lion abundance and predation at Bonneville Dam. We are the only agency that estimates abundance on the Columbia River between Bonneville Dam and Astoria. Through an experiment in 2012 and refinement to our statistical methods through 2015, we feel confident in our tandem boat surveys to provide minimum estimates of sea lion abundance and distribution within the four zones of the lower Columbia River from Bonneville Dam to Astoria. This is a vast area and we feel that we are likely underestimating abundance due to the area two boats can cover in a single day, but it is the only measure of abundance and distribution available. If additional resources become available, we would be pleased to expand our surveys to cover more of the river or at least estimate sea lions we’d be missing with our current surveys.
There is very little overlap in estimates generated among multiple groups collecting similar data in similar areas, instead each group’s work adds to the other. For example, the USACE estimates sea lion abundance and predation with ½ of Bonneville Dam. CRITFC estimates sea lion abundance and distribution on the Columbia River from just beyond the USACE observation area at Bonneville Dam to Astoria, OR, and ODFW records sea lion haul outs at the East Mooring Basin in Astoria. All of these groups are collecting complimentary data throughout the lower river and there isn’t a comparison of one estimate relative to another. Our abundance and distribution estimates represent the best available data between Bonneville Dam and the East Mooring Basin.
We do not intend to further change the survey approach but plan to continue conducting the counts through the next review period. The analysis of these data is primarily status and trends viewed in various temporal scales. We built a way to collect these abundance and distribution data and now as more data becomes available we will be searching for spatial and temporal patterns. We envision that analysis will include comparison of abundance among and within river sections and time likely using analysis of variance. Covariates such as eulachon (Thaleichthys pacificus) abundance, which Rub et al. (2019) found to correlate with annual California sea lion abundance in the estuary and sea surface temperature off of the California coast that links to male California sea lion northern distribution (Laake et al. 2017), will likely be included in the analysis. These results will be reported in our annual report to BPA.
The only measures of salmonid predation by sea lions occurs at Bonneville Dam by the USACE (Tidwell 2019), Willamette Falls by ODFW (ODFW 2018), and NOAA Fisheries has estimates of unexplained mortality for the lower Columbia River (Rub et al. 2019). Our study with accelerometers is promising but very experimental. Much work will be required before the technique could be deployed. In addition, Rub et al. (2019) provides multiple years of sea lion predation data for the lower Columbia River. If that work continues, developing individual sea lion predation rates is less important and to a degree duplicative. Further, assuming that we successfully obtain a sea lion removal permit, the accelerometer research will rank lowest in priority for our limited budget and will be postponed.
Functional Response Methodology
We compared three functional response forms: the Type 2 and Type 3 functional responses without predator interference, and the Type 2 with predator interference. All three take on the general mathematical form of F(N,P,a,b,h,x) = aNx/(1+ahNx+bP), where b=0,x=1 for the Type 2, b=0,x=2 for the Type 3, and b>0 for the interference model. These forms predict the number of prey captured and consumed per predator per unit time (which we evaluate on a daily basis, so per day). The parameter a describes the predator search efficiency is dimensionless, and describes the fraction of the prey pool searched per day. N is the abundance of prey in the prey pool, P is the number of predators in the prey pool, h is the fraction of a day spent handling and digesting prey upon successful capture, and b is the proportional change in search efficiency lost to interference (Beddington 1975).
The functional response predicts kills per predator per unit time, which can have profound consequences on population level predation patterns. We seek to identify the predation characteristics of California sea lions on salmonids, but particularly with respect to spring chinook since they comprise the largest fraction of CSL diet. By fitting functional response formulas to empirical data, we seek to distinguish between Type 2 and Type 3 functional forms, and evidence for interference mechanisms. The fundamental difference between Type 2 and 3 forms is that the Type 2 form predicts that the search efficiency (a) does not change with the abundance of prey, whereas the Type 3 mechanism expresses the search efficiency as a linearly increasing quantity, which implies that efficiency is low at low abundances and higher at higher abundances. This predicts that the number of kills per prey will decline at lower abundances, which can occur for any number of reasons, but translates to a predator loss of interest in the focal prey species when they are less abundant. This can express itself in predation on alternate prey, or simply emigration from the predation area. The mechanism is often referred to as “prey switching” (Murdoch 1969, Smout et al. 2014). The reason for testing the alternative Type 2 and 3 assumptions is that the Type 3 form predicts that predation rates will decrease as prey abundances decrease, whereas the Type 2 form predicts that the predation rate will increase, also known as depensation, which predicts potentially catastrophic losses to predation as prey abundance declines. Since CSL presence is relatively recent, (i.e. last 15 years, Madson et al. 2016), and since we have documented an increase in abundance of CSLs at Bonneville dam (Madson et al. 2015), it is of profound interest to characterize the form of the functional response.
Disclaimer: Raw observer data were provided by USACE (Kyle Tidwell, pers. comm. 2017). Analyses and methods to parse and aggregate data in this report do not reflect opinions or views of the USACE pinniped monitoring program. Interpretations and summaries of data presented in this report are exploratory in nature and should be considered preliminary findings. Summary kill rates, and the prey and pinniped abundances reported here may differ from those reported by USACE because of the resolution of data aggregation, and reporting objectives used in the current study.
Raw pinniped predation observations were collected by USACE biologists and volunteers from 2002 through 2016. Presently, we do not include the 2004 and 2005 observations because of unresolvable inconsistencies in data that stem from methodological differences employed during these years. Since kills were not partitioned by pinniped species until 2008, as such, the 2002, 2003, 2006, and 2007 kills were all attributed to CSLs in this analysis. Given the low level of Stellar sea lion (SSL) abundance at Bonneville Dam over the years 2002-2007 (Tackley et. al. 2008a), a small number of SSL salmonid kills are likely included in the analysis.
The 2016 USACE report aggregated Spring Chinook salmon kills across the period Jan. 1 - May 31 (Madson et al, 2015). The report details the total number of observed salmonid kills, and the maximum (or average) number of individual pinnipeds observed throughout the season. Observed kills align with estimated average prey abundances, doing so provides only 14 kill rates to examine functional response forms. For the purposes of estimating functional response forms, we require a contrast of kills across a wider range of prey abundances. Thus, we re-aggregated the USACE pinniped monitoring data in three ways to achieve a more finite sampling interval for years up to 2015. These three approaches included: 1. A monthly aggregation of observations into kills observed in 3 different predation periods (March, April, and May) for each observation year, 2. An aggregation of observations at specific CSL abundances (with no distinction for month or year), and 3. An aggregation of observed kills at specific prey ranges (also without distinction for month or year). The kill rates (observed chinook and steelhead consumed per day per CSL) are shown in Figures 2-4.
The monthly aggregation (MA), predator aggregation (PA), and salmonid aggregation (SA) data offer different perspectives of how prey and predator abundances potentially affect the rate at which salmonids are captured by CSLs. The MA data provide approximately the same number of samples for each year and month, but the range of potential CSLs and salmonids can vary considerable throughout the month. As a result, there is no apparent kill rate above a prey abundance of 5000, since potentially higher prey abundances are included with lower abundance ranges. The apparent pattern is that the lower prey abundances are generally associated with the earlier time period and abundances increase later in the season. The PA data (Figure 3) display a pattern of higher kills rates when more pinnipeds are present. There is also a pattern of lower prey abundances when fewer CSLs are present. The SA data aggregation (Figure 4) is interesting because it imposes a greater range of prey abundances into the contrast of kills rates. In doing so, it provides evidence for higher kill rates. The raw data are shown in Figure 5.
The average number of kills observed per day at each CSL abundance in the PA aggregation is shown in Figure 6. This shows in increase in the number of kills as the number of CSLs increases.
We fit several alternative functional response forms to these data. The basic form is that the observed kill rate for the ith sample is Ki , where Ki = 
+ 
, where
is a normal random deviate with mean of 0 and standard deviation 
. 
has power x=1 for a Type 2 functional response, x=2 to distinguish as a Type 3. N and P are the true prey and predator abundances. We treat these models in two distinct ways with regard to observation and process uncertainties. The first case is where we assume N and P are observed without error. In this case we treat observed N and P as the true N and P, and predicted Ki differs from true Ki only by the process error term 
. In the case where we assume observation uncertainties, we assume that Ki is predicted by unobserved true quantities ni or pi instead of the data, and we introduce an observation error term such that
or 
, where 
is a random normal deviate with mean 0 and standard deviation 
. In this case predicted Ki differs from true Ki by the differences coming from N or P observation errors as well as the process errors. Including observation errors along with process errors make this a State-Space model (or latent variable model). The various model forms are shown in Table 1.
Table 1Summary of model structures.
|
Model
|
Type
|
Interference
|
Latent variable
|
|
|
1, 2
|
2 or 3
|
|
|
|
|
3
|
2
|
Yes
|
|
|
|
4, 5
|
2 or 3
|
|
Prey
|
|
|
6, 7
|
2 or 3
|
|
CSL
|
|
|
8
|
2
|
Yes
|
Prey
|
|
|
9
|
2
|
Yes
|
CSL
|
|
Models 1-9 were fitted to each of the MA, PA, and SA data aggregations. All model fits provided estimates of a, h, and , and the latent variable models 4-9 additionally estimated and a vector of either pi or ni for the unobserved CSL or prey abundances. Models were fit using Template Model Builder (a freely available c++ library and R package) and its internal algorithms for obtaining the Laplace approximation for the latent variable estimates, which integrates across the uncertainty in observation error space. The model fitting provided estimates of parameter values, a likelihood objective function value, estimates of uncertainty in fitted parameter values, and estimates of the standard deviations of process and observation uncertainties. We calculated the AIC values of the 9 models for each data aggregation in order to select the best fitting models.
3. How will culling of sea lions be evaluated? What criteria will be used to assess whether it is effective at reducing sea lion abundance and predation on adult salmonids? What factors will be considered when determining how many animals will be culled?
For context we will highlight some of the major changes to section 120 resulting from PL 115-329. First, the lower river tribes and by delegation CRITFC are eligible to receive removal permits. Second, area-based management replaces individual sea lion management regarding removal eligibility. Prior to PL 115-329, California sea lions (CSL) had to be individually identifiable, and deemed to have a significant negative impact on ESA listed salmonids. Steller sea lions (SSL) were not eligible. Under that framework, sea lions were trapped, branded, released, and observations of their behavior were made at Bonneville Dam. A removal list of individual sea lions was developed using the criteria of being observed at Bonneville Dam for 5 days, seen eating a salmon, and subjected to hazing. After these sea lions on the removal list were recaptured, they could be lethally removed. Third, the new area-based management states that sea lions of both species between Columbia RM 112 and McNary Dam and all salmon bearing tributaries are individually identifiable and having a significant negative impact on ESA listed fish populations.
The States and tribes are developing a joint application for lethal removal of sea lions under the new section 120 statute. NOAA is ultimately responsible for setting evaluation criteria for the program but we will evaluate the success of the sea lion culling with respect to expected benefits of the program including:
1) Allow the eligible entities to reduce predation on ESA listed salmon/steelhead and sturgeon by Steller sea lions, which are not currently authorized for lethal removal under existing permits;
2) Improve the efficiency of the currently authorized removal programs at Willamette Falls and Bonneville Dam by eliminating the need to mark and repeatedly handle animals and document their repeated presence in the area; and,
3) Prevent sea lions from self- or socially-habituating to tributary locations that are outside of the current geographic scope for authorizations at Bonneville Dam and Willamette Falls, but where sea lions have either been observed feeding on salmonids and/or lamprey and/or sturgeon or may forage in such areas in future.
To more effectively manage pinniped fish interactions, we propose to manage proactively by not allowing CSL and SSL to habituate within the geographical scope of this application. This approach is consistent with the intent of recent amendments to the MMPA, specifically subsection 120(f). It is anticipated that this will result in more efficient management by reducing the number of sea lions removed over time, and increasing annual spawner escapement (Schakner et al. 2016).
Past work by ODFW evaluated sea lion removals in terms of salmonids saved and we would continue with a similar analysis. The number of salmon saved has been estimated using a bioenergetics model (Brown et al. 2012) to calculate the median daily individual salmonid biomass requirement for CSL x the average number of days individual CSL spent at Bonneville Dam, based on observations by the USACE. For example, from Brown et al. (2012):
“The median estimated daily individual salmonid biomass requirement for California sea lions was 14.2 kg (95% confidence interval was 7.8 to 27.1 kg/day), which translated into a median of 3 Chinook/day (95% confidence interval was 2 to 6 Chinook/day). The median estimated seasonal salmonid requirement for each sea lion was 57 salmonids (95% confidence interval was 6 to 216 salmonids/season). The predicted number of salmonids that would have been required from 2008 to 2012 by the 54 California sea lions that have been removed due to any reason ranged from 3,742 to 13,483 fish.”
For every California sea lion removed, escapement above Bonneville Dam by salmonids increases by an estimated 57 fish per season. Since habituated sea lions have shown high fidelity to Bonneville Dam, these savings accrue over multiple years. In addition, removal of habituated animals is believed to reduce opportunities for new, naive animals to be recruited into the Bonneville Dam "population", since at least some naive animals are thought to follow habituated animals upriver from the Columbia River mouth haul-outs. Similar calculations could be made using data from the functional response analysis in lieu of bioenergetics, which would yield results at the higher end of the range. In addition, we will develop similar metrics for Steller sea lions that tend to reside within the Columbia River longer than California sea lions and are much larger animals requiring more biomass for energy.
To realize the goals that we stated above, we plan to remove as many animals as possible in the Bonneville Dam tailrace.The maximum number of individual sea lions observed at Bonneville Dam during the spring season is about 150 animals, so we are unlikely to remove over 150 animals annually and will be far below maximum take limits.
The new statute sets removal limits of 10% of Potential Biological Removal (PBR) that is calculated as part of the periodic status reports on Steller and California sea lions (NOAA 2015). PBR is effectively the maximum number of sea lions that could be removed annually while allowing the population to reach or maintain its optimal sustainable population. For the Steller sea lions PBR is 2,378 (Allen and Angliss 2012) and for California sea lions it is 9,200 (NOAA 2015). Therefore, the maximum removal limits, by statute are 237 for Steller sea lions and 920 for California sea lions.
Additional factors that could governor our take and are being considered include trap efficiency, processing capacity, animal availability, and restrictions in the take authorization from NOAA Fisheries.
4. 4. Although lethal removal has been controversial to date, it is likely going to get much more so with increased culling. Is there a CRITFC or NOAA public relations plan in place to address a public response to the culling program?
No, CRITFC does not have a specific public relations plan in place to address potential response to the culling program. However, our team and the CRITFC Public Relations group have been addressing this issue since 2004 and are well briefed on sea lion / salmon interactions. We routinely give a dozen interviews a year on sea lions, their impacts, and arguments supporting removals. It is our experience that press coverage has consistently been very favorable to the sea lion management program.
We are not aware of NOAA public relations plan regarding the culling program.
5. 5. What are the culling techniques and what do they plan to do with the carcasses? Will the meat, hides, and bones be used? If so, is there concern about possible contaminants in the meat?
We are proposing in our section 120 removal application to capture animals by trapping (Figure 1) or darting. If suitable relocations are available (determined by NOAA) for the captured sea lions, the animals will be moved there. If no suitable relocation areas are available, the animals will be chemically euthanized, per statutory requirement.
In our application we will request tribal cultural use of carcasses, however, preliminary responses from NOAA indicate that cultural use will not be allowed. Ultimately, NOAA will make the determination on carcass deposition. Under the current state authorization, all carcasses are required to be rendered. There is concern about contaminants in the meat, and we plan to have toxicology analysis performed to determine if it is safe for consumption. In addition, PL 115-329 requires that animals be chemically euthanized. This could further complicate consuming meat, although we understand that drugs exist that would not contaminant meat are available, but again their use would need to be approved by NOAA.
3. 6. A brief description of how adaptive management occurs is needed.
Decision making for this project has relied on a combination of approaches including adaptive management. The collective goal of CRITFC and our Agency partners is to vastly reduce the salmon predation loss to sea lions and to better understand what the impacts are in the lower Columbia River.
Sea Lion Adaptive Management Framework
(see Figure below)
Goal
Limit salmonid predation in the lower Columbia River by California and Steller sea lions.
PLAN
Work plan to achieve the goals of the project:
- · Shift project resources to sea lion removal in the Bonneville Dam area.
- · Haze sea lions.
- · Estimate sea lion abundance, distribution, and predation in the lower Columbia River.
IMPLEMENT
CRITFC will implement work plan through the following actions:
· Sea lion removal
o Submit application to NOAA to remove problem sea lions at Bonneville Dam.
o Implement removal based on NOAA’s LOA terms and conditions.
o Coordinate with Co-Managers (USACE, States, NOAA).
· Removal plan.
· Division of labor and resources.
· In-season and post-season coordination
· Haze sea lions
o Haze sea lions at the level required for the removal program.
· Estimate sea lion abundance, distribution, and predation
o Conduct tandem boat runs to estimate sea lion abundance and distribution in the lower Columbia River.
o Estimate predation.
· Pursue accelerometer tag technology to estimate sea lion predation remotely.
LEARN AND ADAPT
· Sea lion removal
o Monitoring and Evaluation with Co-Managers for decision making.
· Effects of the program on sea lion abundance.
· Effects of the program on salmonid abundance.
· Collaborative findings reported to NOAA annually.
· Haze sea lions
o Evaluate effectiveness.
· Estimate Sea lion abundance, distribution, and predation in the lower Columbia River.
o Use data generated to inform Co-Manager decision making on:
· Sea lion population and distribution trends.
· Salmonid predation trends.
REPORTING
- · Annual sea lion Report to BPA – The project reports accomplishments for the year and lessons learned to BPA and the region.
- · Annual LOA Report to NOAA – The Co-Manager report submits accomplishments and evaluation of sea lion removal and RM&E activities to the management agency.
- · Annual BiOp Report to BPA for implementation RPAs 49 and 69.
- · Annual Sea Lion Statement of Work – The annual sea lion SOW is adjusted as necessary to support changes in direction, priority, or focus of the project.
References
Allen, B.M., and R.P. Angliss. 2012. Alaska marine mammal stock assessments, 2013. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-AFSC-277, 294 p.
Beddington, J.R. (1975) Mutual interference between parasites or predators and its effect on searching efficiency. Journal of Animal Ecology, 44:331–340.
B.M. Allen, and R.P. Angliss. 2012. Steller sea lions. NOAA-TM-AFSC-245 https://www.afsc.noaa.gov/nmml/PDF/sars/ak2012slst-e.pdf
Brown, R., S. Jeffries, D. Hatch, B. Wright, and S. Jonkers. 2011. Field Report: 2011 Pinniped management activities at and below Bonneville Dam. Oregon Department of Fish and Wildlife, 7118 NE Vandenberg Ave. Corvallis, OR 97330.
Brown, R., S. Jeffries, D. Hatch, B. Wright, and S. Jonker. 2012. Field report: 2012 Pinniped research and management activities at and below Bonneville Dam. Report to NOAA. https://www.westcoast.fisheries.noaa.gov/publications/protected_species/marine_mammals/pinnipeds/sea_lion_removals/2012_states_field_rpt.pdf
Hatch, D.R., R. Lessard, and J. Whiteaker. 2019. Sea lion monitoring and non-lethal hazing. Annual Report to BPA Project # 2008-004-00. https://www.researchgate.net/publication/331873563_Sea_Lion_Monitoring_and_Non-Lethal_Hazing
Laake, J.L., M.S. Lowry, R.L. Delong, S.R. Melin, and J.V. Carretta. 2017. Population growth and status of California sea lions. The Journal of Wildlife Management. DOI: 10.1002/jwmg.21405.
Madson, P.L., B. K. Van Der Leeuw, K.M. Gibbons, and T.H. Van Hevelingen. 2016. Pinniped predation on adult salmonids and other fish in the Bonneville Dam tailrace, 2016. U.S. Army Corps of Engineers, Bonneville Lock and Dam, Cascade Locks, Oregon. 97014.
Murdoch, W. W. (1969). Switching in general predators: experiments on predatorspecificity and stability of prey populations. Ecological Monographs 39 (4) 335:354.
NOAA 2015. California sea lion status. https://www.fisheries.noaa.gov/webdam/download/76143366
ODFW. 2017. Request for marine mammal protection act section 120 authorization to remove California sea lions from the Willamette River. https://tinyurl.com/Section120Application
Rub, Wargo, A.M., N.A. Som, M.J. Henderson, B.P. Sandford, D.M. Van Doornik, D.J. Teel, M.J. Tennis, O.P Langness, B.K. van der Leeuw, and D.D. Huff. 2019. Changes in adult Chinook salmon (Oncorhynchus tshawytscha) survival within the lower Columbia River amid increasing pinniped abundance. Canadian Journal of Fisheries and Aquatic Sciences.
Schakner, Z.A., M.G. Buhnerkempe, M.J. Tennis, R.J. Stansell, B.K. van der Leeuw, J.O. Lloyd-Smith, and D.T. Blumstein. 2016. Epidemiological models to control the spread of information in marine mammals. https://doi.org/10.1098/rspb.2016.2037
Smout, S., A. Rindorf, P.S. Hammond, J. Harwood, and J. Matthiopoulos. 2014. Modelling prey consumption and switching by UK grey seals. ICES Journal of Marine Science, Volume 71, 1:81-89.
Tackley, S.C., R.J. Stansell, and K.M. Gibbons. 2008a. Pinniped predation on adult salmonids and other fish in the Bonneville Dam tailrace, 2005-2007. U.S. Army Corps of Engineers, Bonneville Lock and Dam, Cascade Locks, Oregon. 97014.
Tidwell, K.S., B.A. Carrothers, K.N. Bayley, L.N. Magill, and B.K. van der Leeuw 2019. Evaluation of Pinniped Predation on Adult Salmonids and other Fish in the Bonneville Dam Tailrace, 2018. U.S. Army Corps of Engineers, Portland District, Fisheries Field Unit. Cascade Locks, OR. 65pp.
Wright, B., S. Jeffries, and D. Hatch 2018. Field report: 2018 Pinniped research and management activities at and below Bonneville Dam. Report to NOAA. https://www.westcoast.fisheries.noaa.gov/publications/protected_species/marine_mammals/pinnipeds/sea_lion_removals/2012_states_field_rpt.pdf
Help
Mitigation work
Reporting
Funding
Search
Log in