There are three key projects in this category, examining the impacts of bird, fish and seal predation on juvenile chinook and coho salmon. During 2014, these studies were focussed in Cowichan Bay, and occurred simultaneously with the juvenile salmon sampling studies. In 2015, we focussed primarily on seal and fish predation, and the field study locations were expanded out of Cowichan Bay to include Puntledge and Big Qualicum Rivers. For 2016-2018, the seal predation study has been expanded to allow for validation using additional diet data from non-estuary sites, as well as utilizing an independent method to reconstruct diets. In 2017, focussed studies to assess levels of predation in the freshwater environment of Cowichan River were also carried out. These studies examined the impacts of mergansers, brown trout, otters and raccoons on juvenile chinook as they make their way down the river before heading out to sea.

1.    Bird Predation in Cowichan Bay

Team: Joy Wade (contractor, Fundy Aquatic Services Inc.) and local volunteers

The movie above shows mergansers feeding in Cowichan River.

In order to determine the sources of early marine mortality for Chinook salmon in the Cowichan River system, all sources of predation, including from birds, must be considered.  To date, there has been little effort to characterize the abundance and diversity of bird species at the time of smolt outmigration from the Cowichan River in either the river or the estuary.  The purpose of this project was to collect preliminary data on the abundance and diversity of the non-passerine birds in this area through a citizen science based project.

In addition, scat have been collected from known piscivorous birds for DNA analysis at a future date.  It is believed that such an analysis will be able to determine which species of salmon are being consumed if any.

This project was piloted in 2014. A total of 31 different species were recorded (including unidentified categories) in Cowichan Bay estuary (Table 1).These species have been classed according to their feeding habits (Table 2); the majority are non-piscivorous. The number of individuals enumerated varies with the observation location but it can be generalized that the Great Blue Heron, Common Merganser and gull (sp.) were the most numerous piscivorous species counted.

Wade Pic 1
A group of mergansers in Cowichan Bay estuary
Bird scat
Collecting bird scat for genomic analyses of bird die

Despite the short time span of this survey a high species diversity was recorded in the Cowichan Bay estuary. In addition, the researchers were able to observe the seasonal succession of species. For example, the American Wigeon is reported by local birders to be present throughout the winter, we were able to capture the departure of the American Wigeon and the arrival of the Common Merganser within the time frame of this study. Additionally, It is possible with the data collected to look at the breakdown between males, females and ducklings in the Cowichan River. Males and females were both present in the river until the end of May, at which time the males disappear entirely. Shortly after, the females are seen with ducklings. In the Maritimes it has been documented that broods are raised upstream in rivers containing Atlantic salmon and trout then move to the estuary to feed (White, 1936, 1937, 1957; Anderson, 1985; Anderson et al, 1985). The same progression was observed in the Cowichan Bay estuary. This is not new information and has been documented in other areas. These examples remind us of the ever changing avifauna in the wild and the need to understand what brings them to that particular area at that time.

Potential Predators
The potential impacts of bird predators on juvenile salmon populations is well documented (Mace, 1983; Wood, 1987a; Feltham, 1995; Krohn and Blackwell, 1986; Kennedy and Greer, 1988). They found very few species of piscivorous birds in Cowichan River and Cowichan Bay estuary. Common Merganser, gulls (sp.) and Great Blue Heron were those species present in highest abundance. Other piscivorous species including as the Double-crested Cormorant, Belted Kingfisher, Bald Eagle and Osprey were present in such low abundances that their potential predation on outmigrating salmon smolt is considered negligible for the observation period of this study.

b.      Seal Predation

Team: Ben Nelson, Austen Thomas, Hassen Allegue and Andrew Trites (UBC)

The long-term goals of this research are to 1) estimate the numbers of chinook and coho smolts consumed in the Strait of Georgia by harbour seals, 2) evaluate their impact on salmon recovery, and 3) propose ways to mitigate the impact of seals.

Since the 1970s, native stocks of Chinook and coho salmon have declined throughout the Strait of Georgia (SOG), despite sizable reductions in harvest from commercial and recreational sources. Coinciding with the decline of these species are rapid increases in populations of pinnipeds native to the SOG, including Steller sea lions, California sea lions and harbour seals.

Salmonids are a significant dietary component of seals and sea lions in the SOG, particularly in estuaries where adult salmon return to their natal steams. There is also direct scientific evidence of seals preying on out-migrating juvenile salmon during the spring. Such predation in combination with habitat loss, fishery removals, declining prey abundance and climate regime shifts may explain the dramatic declines in salmon abundance in the SOG. However, the trends in pinniped abundance have led some scientists to speculate that predation on salmon by marine mammals may be particularly significant and may be impeding recovery. High rates of predation by harbour seals on salmon smolts may explain the decline and lack of recovery of coho and Chinook salmon in the Salish Sea. However, rates of predation and the relative spatial and temporal vulnerabilities of smolts to predation by seals are unknown.


2015- Predation estimates using scat samples

This team assessed the impact of seals on salmonids in Cowichan Bay by counting the number of predators in the Cowichan Bay, quantifying predation events from visual observations, and determining diets from morphological and genetic analysis of fecal samples. Data collected for 2012-2014 provides the following estimates of the percentage of chinook and coho juveniles lost to seal predation in the Strait of Georgia:

-Chinook: Mean = 40% (95% CI: 32-45%)

-Coho: Mean = 47% (95% CI: 44-52%)

These data have been used in a model developed by PhD student Ben Nelson to test whether pinnipeds are inhibiting the recovery of commercially and recreationally important stocks of Chinook and coho salmon in the Strait of Georgia. Given these estimates, it appears that harbour seal predation is likely responsible for a significant amount of natural mortality in the early marine stage for both Chinook and coho salmon in the SOG. Ben notes that seal predation appears to be highest on juvenile fish between 115-145mm in length, but also cautioned that the consumption estimates and mortality could be biased due to high emphasis on “estuary seals” because of the locations that scats were sampled.

He suggested that any future sampling should focus on NE Strait and open water/rocky reef habitat as most of the scat sampling sites were in estuaries which could bias the results. A more stratified and extensive sampling plan was suggested and has been implemented for 2016-2017 (see below). These samples should allow for validation of Ben’s model.

2015- Predation estimates using an experimental PIT tagging methodology

Additionally, the US and Canada have both pursued a technological solution to quantifying interactions between harbour seals and salmonids exhibiting poor marine survival.  In B.C., a development project was undertaken during 2014 to create a head-mounted PIT tag scanner for harbour seals that could be used to quantify the number of PIT tagged salmon smolts consumed by pinnipeds. This study aimed to track harbour seals to estimate rates of predation on coho smolts, and will identify when and where predation is occurring.

Over 3 years the team developed a satellite-linked PIT tag scanner that is designed to quantify the number of PIT tagged fish ingested by harbour seals. These RFID tags are suitable to be affixed to the heads of harbour seals, and the internal PIT tag scanner is activated when seals attempt to capture fish (head-strikes), thereby detecting the presence of PIT tagged fish in the mouths of the seals. PIT tag detections are logged by the instrument and then transmitted via the ARGOS satellite network, providing the number consumed and the complete PIT tag IDs.

A  second  tag  was also developed which incorporates a  GPS  and 3D accelerometers  and which allows for reconstruction of the fine‐scale  movements  of individuals, to determine the locations and the ways in which seals are consuming the smolts. The data collected by the seals are being used to create maps of spatial predation risk needed to identify predation ‘hot spots’ during the critical period of smolt outmigration. The aim of this study is to test whether harbour seals could be responsible for high salmon mortality in the Salish Sea, and will provide resource managers with important information needed to design mitigation strategies for seal predation.

Half duplex PIT tags were implanted in 36,900 coho smolts at the Big Qualicum hatchery more than one month  prior  to  the  release  of  381,800  total  coho  smolts  in early May 2015 (9.7%  tagging  rate).  20  harbour  seals  were captured prior to the fish release and outfitted with RFID tags (the PIT tag scanners) in addition to back mounted  GPS/3D  accelerometers  (used  for  reconstructing  seal  movements).  Seals were tagged  in  the mouth of the Big Qualicum River (9 animals), in addition to nearby rocky haulout sites (11 animals). The purpose of this  design  was  to  determine  a  coho  smolt  predation  rate  for  the  animals  that  likely  specialize  on  out migrating smolts, versus those seals that likely feed in other areas and represent the larger portion of the seal population. Coho smolts were released from Big Q hatchery on Monday May 4th.

The group received detections from 4 of the 20  seals,  and  each  transmitted  multiple  PIT  tag  IDs.  This indicated that these seals were consuming Big Q coho smolts. All of the seals with PIT tag detections were animals tagged in the estuary, and haulout observations indicate these animals stayed in the estuary during smolt outmigration. Corrected counts of harbour seals in the area indicate that the number of seals using the river was ~96 individuals.

Based on the timing of PIT tag detections the researchers found that seal predation on smolts occurs primarily at dusk and midnight. During peak outmigration, the per seal smolt predation rates suggested that seals consume ~1kg coho smolts per night. This equals ~½ of their daily energy needs, indicating that smolt consumption is only part of the daily foraging pattern.

Extending these data only to the ~96 seals using the river, seals consumed ~6.19% of the outmigrating smolt population.  Note that this estimate is substantially lower than the estimate provided by Ben Nelson’s method, above. The two estimates may be reconciled as follows: 6% of the juvenile salmon entering the Strait (from hatcheries) may be eaten by estuary specialists (as in this study). Another 34% of mortality of smolts may occur from high numbers of seals feeding throughout the Strait of Georgia, giving a total of 40% overall seal-related mortality.

Integration of an accelerometer trigger into the RFID tags extended battery life of the tags from 2 weeks to 2-6 months.  PIT tag detections continued for over 3 weeks after the release date. Future work will explore alternative ways of doing the population level data expansion.

The backpack GPS unit stored the  seal  location  information  and  began  transmitting  the locations (from the whole deployment) on July 1 2015. This was done to prevent collisions between satellite transmissions of the head mounted tag and the back mounted tag. The location information and PIT tag detections are being used to  determine  how  many  smolts  were  consumed,  and  when  and  where  that predation  primarily  occurred.

According to preliminary spatial distribution data, 16 of 20 harbour seals seemed to use a relatively limited area around their main haul-out sites whereas 4 harbour seals traveled long distance (>100 km) and transited between multiple haul-out sites. The seals that consumed PIT tag smolts seemed to forage mainly in Big Qualicum estuary while traveling back and forth between their respectively main haul-out sites and the estuary. The seals that did not consume PIT tag smolts have different distribution patterns than seals that consumed PIT tag smolts and Big Qualicum estuary was not their main foraging spot.

2016-2017.  Seal Predation Completion Study

Given the results of the seal predation analyses for 2012-2014, it appeared that high rates of predation by harbour seals on salmon smolts could help explain the decline and lack of recovery of coho and Chinook salmon in the Salish Sea. As noted above, this hypothesis was tested using models to estimate predation rates from the frequency of occurrence of smolt remains contained in the faeces of harbour seals in estuaries. It was also tested by studies of direct predation near Big Qualicum River by capturing 20 adult harbour seals and equipping them with biologging tags. Preliminary results indicate that 20% of harbour seals in the Strait of Georgia may be eating over half of the smolts that enter these coastal waters. However, these high rates of predation require validation using additional diet data from non-estuary sites, and an independent method to reconstruct diets. Foraging behaviours of the few seals they recorded exploiting smolts also needs further investigation to determine whether these individuals used specific foraging strategies targeting smolts, or whether they were behaviorally no different from the seals that did not consume smolts.

In these additional studies, the researchers wished to assess behavioural differences using the 3D accelerometry data recorded by the 20 tagged seals, and to address major uncertainties in their estimates of predation by collecting additional seal scats in 2016 from non-estuary sites. They will also compare their 2016 estimates of diet derived from DNA metabarcoding and hard-part frequency-of-occurrence with a third method―biomass reconstruction. This research will complete their study and ensure robust estimates of predation and a fuller understanding of the predatory behaviour of seals ― information that is needed to design mitigation strategies.

The long term goals of this research are to 1) estimate the numbers of chinook and coho smolts consumed in the Strait of Georgia by harbour seals, 2) evaluate their impact on salmon recovery, and 3) propose ways to mitigate the impact of seals.

The short term objectives of the 2016-2017 project were to 1) address major uncertainties in their estimates of predation in terms of where scats were collected and the methods they used to reconstruct diets, and to 2) complete their analysis of the foraging behaviours of seals that eat smolts and those that do not to identify ways in which predation risk might be reduced. These studies will ensure robust estimates of predation and will provide a fuller understanding of the predatory behaviour of seals ― information needed to design mitigation strategies.

The 2016-2018 work plan included the following:

  • Collect harbour seal scats every two weeks for 8 months from 5 sites (April – November 2016)
  • Undertake DNA analysis and measure size of salmon otoliths recovered from 2016 harbour seal scats
  • Compare 2016 diet findings with samples collected in previous years from estuaries to determine if diets vary significantly by region in the Strait of Georgia
  • Recalculate estimates of mortality on salmon smolts caused by harbour seals in the Strait of Georgia
  • Compare biomass reconstruction estimates of diet for 210 samples collected in 2016 with estimates obtained using two other methods―frequency of occurrence and DNA metabarcoding
  • Assess the biases of the three diet reconstruction methods (from simulations and data), and recalculate (if necessary) estimates of mortality on salmon smolts caused by harbour seals in the Strait of Georgia
  • Process the acceleration data collected by the 20 tagged seals using specialized animal behaviour software to determine and compare time-activity budgets between seals that ate smolts and those that did not
  • Calculate the foraging efficiencies of the two groups of seals to assess whether targeting smolts is energetically beneficial or inconsequential
  • Assess whether seals that feed on smolts employ specialized foraging strategies or are behaviourally not different from the seals that do not consume smolts―and identify mitigation strategies that could be employed to reduce predation risk on salmon smolts.

Results to date suggest the following:

  • Site selection influences consumption estimates of juvenile and adult salmonids
  • Estuaries should not be used as proxies for non-estuary sites when assessing consumption of specific prey by harbour seals  in the Strait of Georgia (particularly salmonids in the fall) because of the higher proportion of salmonids in the diets of seals foraging in estuaries
  • Annual and seasonal variability influence estimates
  • Salmon diet in Cowichan Bay seal samples follow an annual pattern but is highly variable year to year and between seasons
  • Long-term monitoring is critical for assessing impacts of predation, particularly for species with large degrees of natural variability in production such as salmonids

c.       Fish Predation

Team: Dick Beamish (DFO-PBS, retired), Joy Wade (contractor), Dave Preikshot (Madrone Environmental), Lana Fitzpatrick (DFO-PBS)

Pacific salmon experience heavy and highly variable losses in the ocean, with natural mortality rates generally exceeding 90-95% during their marine life. Most of this mortality is thought to occur in coastal marine ecosystems during two critical periods: an early predation-based mortality that occurs during the first few weeks following ocean entry, and a starvation-based mortality period during the first fall and winter at sea. During 2014, a number of studies were carried out in Cowichan Bay to determine whether fish predation of juvenile Chinook and coho accounts for a significant level of mortality in this early marine period.  Sampling occurred primarily at night-time, at a variety of depths, and using a variety of fishing gear including gillnet, trawl and purse seine surveys. This study was replicated in 2015 and expanded to also sample in the vicinity of the Big Qualicum River.

Fishing for predators using gillnets
Fishing for predators using a small trawl

A gillnetting survey was carried out weekly from April 26 to June 19th 2015 in the estuaries of Cowichan Bay and Big Qualicum River.

The PIs:
• Confirmed the absence of fish predators in Cowichan Bay during smolt outmigration.
• Determined that there were few fish predators in Big Qualicum estuary during smolt outmigration.
• Confirmed that predators can be captured with the fishing gear used.

The absence of fish predators in the surface and bottom gillnet sets in Cowichan Bay estuary were consistent with the study conducted in 2014. Similar results were found for Big Qualicum estuary. Because it was possible that large abundances of predators arrived after they had finished fishing, they made several sets in the early morning, late afternoon and evening. Large abundances of Pacific salmon juveniles were present in the vicinity of the gillnets indicating that the potential food source was there. Seals were observed actively feeding at both locations.