Studies to examine larval Lake Whitefish (Coregonus clupeaformis) abundance, diet composition and growth, and the abundance of their zooplankton prey were conducted during eight years (1991–1993, 1995–1996 and 2003–2005) over the course of two decades that spanned a period of major ecosystem change—primarily dreissenid mussel related impacts—on the Bay of Quinte, northeastern Lake Ontario. Larval fish were captured in shallow, nearshore waters (0.2–2.0 m) from early April to mid-May each year. Larval Lake Whitefish fed primarily on cyclopoid copepods and small-bodied cladocerans. The key finding of our studies was that prey abundance declined by 89% from the earlier (1991–1993, 1995–1996) to the later (2003–2005) sampling years. Larval fish growth during spring was significantly correlated with prey availability. Recruitment to the juvenile stage in August was correlated with spring prey availability and larval fish growth. The observed decline in larval Lake Whitefish prey in the Bay of Quinte may be contributing to poor stock performance during and following a period of significant ecosystem change.
The Bay of Quinte Lake Whitefish (Coregonus clupeaformis) spawning stock formed the basis for a commercial food fishery on Lake Ontario for many years (Christie, 1963), and has undergone severe changes in stock abundance as a result of a variety of anthropogenic stressors (Christie, 1972; Casselman et al., 1996; Hoyle et al., 2003). Although scientific and commercial harvest monitoring programs of juvenile and adult life stages are on-going, studies of earlier life history stages have not been conducted since the 1920s (Hart, 1930). McKenna and Johnson (2009) and Johnson et al. (2009) recently studied larval Lake Whitefish from early April to early May (2004–2006) in Chaumont Bay, an open-embayment in New York waters of Lake Ontario. In this paper, we investigate larval Lake Whitefish abundance, diet composition and growth, and the abundance of their zooplankton prey spanning a period of dramatic change in the Bay of Quinte ecosystem from the early 1990s to the mid 2000s.
The study described here was initially undertaken because invasion of dreissenid mussels (zebra mussel, Dreissena polymorpha and later quagga mussel, D. bugensis) into the Bay of Quinte area was imminent by the early 1990s. The mussels were expected to blanket the Lake Whitefish spawning substrate and, perhaps, impact egg or larval fish survival. In 1991 and 1992 spawning sites were identified and egg deposition and fry emergence were described (Hoyle, 1992; Hoyle 1993; Hoyle and Melkic, 1991). Three sites were established to examine larval whitefish abundance and diet, and prey availability in 1991–1993 and 1995–1996—years that spanned mussel invasion and proliferation. Cyclopoid copepods and small-bodied cladocerans dominated the larval fish diet. The early life history studies were suspended after 1996.
Following dreissenid mussel invasion in the early 1990s, the Lake Whitefish population collapsed. Hoyle (2005) attributed the collapse to the loss of the burrowing amphipod Diporeia spp.—the primary diet of juvenile and adult whitefish (Hart, 1931; Ihssen et al., 1981)—which coincided with dreissenid mussel invasion. Correlates with the whitefish population collapse included diet change, declines in body condition and growth, delayed age at maturity, changes in distribution and feeding patterns, and very poor reproductive success (Hoyle, 2005). Hoyle (2005) interpreted low juvenile abundance in August bottom trawls as an indication of poor reproductive success and hypothesized that the very low catches from 1998–2002 may have been linked to poor condition of adults and a consequence of low lipid reserves; i.e. reduced egg viability.
However, independent, long-term and annual monitoring of Bay of Quinte zooplankton showed that a 60% decline in abundance of cyclopoid copepodids was coincident with the dreissenid mussel invasion (Johannsson and Nicholls, 2002). Therefore, during 2003–2005, we repeated our larval whitefish studies, using the same sites and methodology as used in the earlier years, to investigate the possibility that declines in larval whitefish food resources may contribute to low recruitment to the juvenile stage in August. In a Lake Huron Lake Whitefish population, Freeberg et al. (1990) identified a bottleneck at the larval stage and found that prey availability, expressed as the number of zooplankton per fish (z f−1), was positively linked to larval fish growth and survival and to recruitment to a later life stage. Here, we investigate the possibility of a similar bottleneck in the Bay of Quinte whitefish population.
Our objectives were to: (1) examine trends in the abundance of larval whitefish and their zooplankton prey and, (2) examine the relationships among prey availability (z f−1), larval fish growth rate and the abundance of juvenile Lake Whitefish in August.
Lake Whitefish early life history studies were focused at three sites: Trident Point, Sherman's Point and Indian Point in the Upper, Middle and Lower Bay of Quinte, respectively (Figure 1). All sites were adjacent to shore at or near points of land and characterized at Trident Point by bedrock with crevices and pockets of gravel and sand, at Sherman's Point by bedrock overlaid with medium-sized broken rock cobble, and at Indian Point by gravel and small cobble substrate close to shore and sand offshore. In each year, we visited each site 3–12 times, on a rotational basis, from “ice-out” in late March or early April until mid-May when larval Lake Whitefish disappeared from the shallow nearshore waters. Larval fish studies were carried out from 1991–1993, 1995–1996 (hereafter also referred to as early years), and 2003–2005 (hereafter also referred to as later years).
Juvenile Lake Whitefish (age approximately 4 months) were sampled annually, in August, near Conway in the Lower Bay of Quinte (Figure 1) at a depth of about 20 m.
Larval fish sampling
Larval fish were sampled with a 0.5 m diameter tow net like that described in Loftus (1982) and Cucin and Faber (1985) and towed from a 6.1 m open aluminum boat, as close to shore as possible (water depth ranging from 0.2–2.0 m), at 1 knot (0.5 m s−1) for 5–15 min. Larval fish abundance was expressed as number m−3. One to three replicate tows were made during each visit to each site. Basic environmental data were collected during each visit including water temperature. After each tow, larval fish were identified, enumerated and a random sample of up to 100 individual Lake Whitefish were kept in ice-water for processing on the same day in the laboratory. All fish were measured to the nearest 0.1 mm and a smaller number were weighed to the nearest 0.0001 g wet weight. Three stomach samples were taken from each tow. Each sample consisted of the combined stomach contents of 10 fish and stored in 5% buffered formalin. Stomach samples were later analysed in the same manner as zooplankton samples (see below) and diet composition was expressed as proportions by number.
Juvenile fish sampling
Juvenile Lake Whitefish were sampled annually in August (age approximately 4-mth) with a ¾ Western bottom trawl. Lake Whitefish do not inhabit the Upper or Middle reaches during summer because water temperatures are too warm; at this time the juvenile fish are found in the Lower Bay (Hoyle et al., 2008). Four replicate trawls of 6 min duration and a covering a distance of 463 m (¼ nautical mile) were made during each of two visits annually (see Hoyle et al., 2008 for more detailed methodology). Juvenile fish abundance was expressed as number trawl−1 and used to assess Lake Whitefish year-class strength since 1972 because this long term index of juvenile abundance has been shown to be significantly and positively correlated with age-3 population size three years later (Ward, 2007).
Zooplankton sampling was conducted concurrently with larval fish sampling using a 73 μm mesh Wisconsin-style plankton tow with a mouth opening of 0.115 m dia. The tow net was not metered and a 100% flow efficiency was assumed. Horizontal tows were conducted at 1 knot (0.5 m s−1) for up to 360 seconds duration in earlier years of the study (1991 and 1992) and standardized to 80 seconds in all subsequent years. Samples were preserved in 5% buffered formalin. Individual zooplankton was identified and enumerated from a 2–100% fraction of each sample. Zooplankton abundance was expressed as number m−3.
We also used an independent zooplankton dataset from Fisheries and Oceans, Canada (DFO) for offshore sites in the Upper, Middle, and Lower reaches of the Bay of Quinte (Figure 1). Every year since 1975, a vertical series of Shindler-Patalas trap samples (64-μm sleeve) are collected at each site bi-weekly starting in early May (Bowen and Johannsson, 2011, in this issue for more details regarding sampling and enumeration). Preliminary analysis indicated that these offshore zooplankton data during May (number m−3) were not significantly correlated with the nearshore zooplankton abundance in April/early May. However, our expectation was that these offshore zooplankton data would reflect larval Lake Whitefish prey availability after the larval fish disappeared from nearshore waters and might be correlated with larval fish survival in August.
Zooplankton were grouped as cladocerans (primarily Bosmina longirostris and Chydorus sphaericus), calanoid copepods (excluding nauplii), calanoid nauplii, cyclopoid copepods (excluding nauplii), and cyclopoid nauplii. Zooplankton abundance was expressed as number m−3. Diet composition of larval Lake Whitefish was expressed as proportion by number of each prey taxa.
Statistical analyses were used to test for differences between earlier (1991–1993, 1995–1996) and later (2003–2005) sampling time-periods for the dependent variables larval fish and edible zooplankton abundance (based on diet composition, see below) abundance for a window of sampling dates common to both time-periods (April 1–May 10 with water temperatures ranging approximately from 3–13°C). For comparing mean abundances between time-periods, analysis of variance (ANOVA; STATISTICA 8.0) was used on log10 transformed data. All test statistics were considered significant at p < 0.05.
Mean annual juvenile Lake Whitefish abundance data were summarized for three time periods: 1991–1997, 1998–2002 (time-period of very poor recruitment reported by Hoyle, 2005), and 2003–2005. For comparing mean juvenile abundances among these three time-periods, analysis of variance was used on log10 transformed data. The Bonferoni post-hoc multiple comparison test was used to identify those pairs of means statistically different from each other (p < 0.05). Edible zooplankton taxa were grouped and an annual prey availability estimate determined as the ratio of the number edible zooplankton per individual larval fish (z f−1) following Freeberg et al. (1990):
Initial inspection of larval fish length at age (days) data revealed a general pattern of little or no change in length for the first 2–3 wks of sampling each year followed by a rapid increase in length over the remainder of sampling (mid-May). For each year, we pooled all larval fish length data and calculated instantaneous daily growth during a 3-wk period of accelerated growth (approximately wks 3–5) using the approach of Freeberg et al. (1990):
u = (logelt – logelt−1)/T;
u = instantaneous daily growth;
lt = mean length (mm) at beginning of rapid growth phase;
lt−1 = mean length (mm) 3 wks later; and
T = time interval in days (21 days).
Mean larval fish length was calculated annually for the period of time prior to the 3-wk accelerated growth phase.
Simple correlation coefficients were calculated between August juvenile (age approximately 4-mth) Lake Whitefish abundance in bottom trawls and spring larval fish and edible prey abundance, prey availability (z f−1), mean larval fish length, instantaneous daily growth, and May cyclopoid copepod (excluding nauplii) abundance. Correlation coefficients were also calculated between larval fish length and abundance and between larval fish growth and prey availability (z f−1) and mean larval fish length.
A total of 269 larval fish tows and 193 zooplankton samples were taken during eight years of sampling (1991–1993, 1995–1996, 2003–2005) at the three Bay of Quinte sites. Over 17,000 larval Lake Whitefish were caught. Larval fish were first captured immediately after ice-out and as early as March 27 but more commonly in early April. Fish were caught as late as May 19. Highest catches tended to occur in late April and the first week of May corresponding to water temperatures of about 7–11°C.
Generally, larval Lake Whitefish abundance tended to be lower (83%) in later sampling years (2003–2005; mean = 0.055 m−3, sd = 0.049) than in earlier years (1991–1993 and 1995–1996; mean = 0.325 m−3, sd = 0.722; Table 3) but a two-way ANOVA with sampling time period, i.e. early years and later years, and sampling site as factors indicated that abundance was not significantly different (F1,18= 2.601, p = 0.124). Neither site (F2,18= 0.525, p = 0.601) nor period x site interaction (F2,18= 0262., p = 0.772) terms were significant.
The early spring zooplankton community was dominated by cyclopoid copepods (67% by number) and small-bodied cladocerans (30%; primarily Bosmina longirostris and Chydorus sphaericus). Larval fish stomachs contained food only after about 2–3 wks and these same zooplankton taxa groups, excluding cyclopoid nauplii and referred to hereafter as edible zooplankton, also dominated the diet composition (Table 1). Cyclopoid copepods were most common in the diet at 61% by number followed by cladocerans at 32%. Although cyclopoid nauplii were the most common group in the zooplankton community (36%) they were infrequently eaten by larval whitefish (4%). Zooplankton abundance declined dramatically (e.g. total zooplankton declined by 78% by number) from earlier sampling years to later sampling years (Tables 2 and 3). Edible zooplanktion abundance was examined with a two-way ANOVA with sampling time period, and sampling site as factors. The results indicated that edible zooplankton abundance declined significantly (89%) from earlier (mean = 3871 m−3, sd = 4537) to later sampling years (mean = 425 m−3, sd = 208; F1,18 = 34.408, p = 0.00002). Neither site (F2,18 = 1.475, p = 0.255) nor period x site interaction (F2,18 = 1.111., p = 0.351) terms were significant.
Mean annual larval fish total length prior to the period of accelerated growth ranged from 13.6 mm in 2003 to 14.9 mm in 1993 (Table 3). Mean length was not significantly correlated with larval fish abundance (r = −0.150, p > 0.05, Table 4). Growth in length accelerated as water temperatures increased above 7–8°C in late-April. During the period of accelerated growth, instantaneous daily growth ranged from 0.008 in 2004 to 0.028 in 1991 (Table 3). Growth was highly correlated with prey availability (z f−1, r= 0.813, p < 0.05) but not with mean larval fish length prior to the period of accelerated growth (r=0.469, p > 0.05, Table 4).
Juvenile Lake Whitefish catches in August Bay of Quinte bottom trawls were variable, but generally high in the early 1990s nearly zero from 1998–2002 and relatively low thereafter (Figure 2). Mean abundance from 1991–1997 (18.95 fish trawl−1, sd = 15.40) was significantly higher than that for 1998–2002 (0.15 fish trawl−1, sd = 0.16, p = 0.001) but only weakly higher than for 2003–2005 (3.15 fish trawl−1, sd = 3.28, p = 0.084). Mean abundance for 1998–2002 was not different than for 2003–2005 (p = 0.421).
May cyclopoid copepod abundance (excluding nauplii) was much higher in early years. Mean abundance from 1991–1995 was 20,447 m−3 while that for 1996–2005 was 5802, a decline of 72%. The four highest values occurred during the five early years of our larval Lake Whitefish study (Figure 2).
For years corresponding to the larval fish studies, juvenile Lake Whitefish abundance was significantly and positively correlated with larval Lake Whitefish prey availability (z f−1, r= 0.846, p < 0.05) and instantaneous daily growth (r= 0.740, p < 0.05) and with May cyclopoid copepod abundance (r= 0.932, p < 0.05) but not with larval fish abundance (r=−0.365, p > 0.05), edible prey abundance (r= 0.567, p > 0.05) or mean length (r=0.176, p > 0.05) prior to the period of accelerated growth (Table 4).
When early life history studies were resumed in spring 2003, we were very uncertain about larval whitefish status. Since the last larval fish study in 1996, the adult population had been subjected to severe stress—the loss of their preferred prey, Diporeia—possibly to the extent that reproduction was impaired because of poor nutritional status (Hoyle, 2005). Very few young-of-year whitefish were captured in August bottom trawl surveys between 1998 and 2002 (Figure 2). In spring 2003 we confirmed that larval fish were present at the same sites previously studied. However, in August 2003 juvenile abundance in bottom trawls had increased relative to those very low catches of 1998–2002. If reproductive success had been impaired during 1998–2002 by some factor such as poor adult nutritional status, this factor appeared to be alleviated by 2003. Larval catches tended to be lower in the later years (2003–2005) although the decline in abundance was not statistically significant due to high variation in catches.
Cyclopoid copepods and small-bodied cladocerans were the major prey of larval Lake Whitefish in the present study. This result was similar to that for Hart's (1930) feeding study in the 1920s on the same Bay of Quinte Lake Whitefish population. Johnson et al. (2009) found that cyclopoids were the primary prey of whitefish larvae over a 3-year period (2004–2006) in Chaumont Bay, Lake Ontario. We observed less calanoid copepods and more cladocerans in the diet than reported by Johnson et al. (2009). Johnson et al. (2009) found that, although larval whitefish feeding was more intense during the day, chironomid larvae were an important part of their diet at night. We did not sample larval fish at night and did not observe chironomids in the larval Lake Whitefish stomachs in our Bay of Quinte study.
Perhaps the most significant finding of our study was that the major prey types available to larval Lake Whitefish in nearshore waters had declined dramatically from the years studied in the 1990s to the later sampling years (2003–2005). A major gap in our data series between 1996 and 2003 impaired our ability to pinpoint the period of major zooplankton decline. Nevertheless, the reason for the decline is unclear. The May cyclopoid copepod abundance data collected annually in offshore waters by DFO suggest that Bay of Quinte zooplankton populations were impacted by the mid-1990s (Figure 2). At least two ecological changes that could have impacted zooplankton dynamics occurred over the course of our larval fish studies in the Bay of Quinte. First, dreissenid mussel establishment impacted water quality (e.g. lower phosphorus and chlorophyll a and increased clarity) and phytoplankton communities by 1995 (Hoyle et al., 2003). Secondly, Cercopagis pengoi (a predatory cladoceran) invaded in 1999 and appeared to negatively impact small-bodied zooplankton types that year and in 2000 (Benoit et al., 2002). Bowen and Johannsson (this issue) discuss, in detail, potential causes for the decline in Bay of Quinte zooplankton. The reductions are thought to be caused by both direct predation of microzooplankton (e.g. nauplii) by mussels and C. pengoi, and competition between zooplankton and dreissenids for food resources.
Was larval Lake Whitefish prey abundance limiting whitefish recruitment to later life stages in the recent time period of our study? Larval fish prey abundance decreased significantly between earlier and later sampling years. We observed that prey availability (z f−1) was correlated with larval fish growth and to juvenile abundance later in the summer, and recent year-class strength, measured as juvenile abundance in August, remains low relative to that in the early 1990s. Freeberg et al. (1990) showed that a very abundant Lake Whitefish year-class (1984) at hatch was subsequently depleted by a prey availability (z f−1) bottleneck only a few weeks later, despite prey (zooplankton) abundance that was similar to the previous year when many fewer fry hatched. The key was the ratio of zooplankton prey to individual larval fish. In our study, the positive relationships between prey availability, larval fish growth and juvenile abundance were largely driven by very high prey abundance in earlier sampling years. We contend that if the depressed zooplankton abundance levels observed in this study continue, recruitment of a strong Lake Whitefish year-class is unlikely because z f−1 cannot be sufficiently high with low zooplankton abundance. High survival rate from larval to juvenile stages is possible at low zooplankton abundance but only at very low levels of larval fish abundance.
Without accounting for prey abundance, larval Lake Whitefish abundance was not correlated with juvenile abundance later in the summer. Mean length of larval fish, prior to accelerated growth, was also not correlated with juvenile abundance. Larval fish abundance and initial size, prior to exogenous feeding and accelerated growth, are likely related to number, quality and hatching success of eggs. We suspect that these and related factors (e.g. spawning stock size, nutritional status of adults, over-winter and early spring environmental conditions), possibly in addition to low zooplankton levels, were also likely limiting Lake Whitefish year-class strength during the 1998–2002 time period. Further consideration of these factors is beyond the scope of this study but could be the focus of future work. A laboratory study by Brown and Taylor (1992) concluded that both egg composition and prey availability significantly influence the growth and survival of larval Lake Whitefish. Claramunt et al. (2010) found that variation in larval density was best explained by larval length, spring wind intensity, and adult stock density.
Cyclopoid copepod abundance in May (DFO data) from offshore waters of the Bay of Quinte was positively correlated with juvenile abundance later in the summer. Larval Lake Whitefish moved offshore about mid-May in our study and, perhaps, zooplankton abundance there was also important for continued larval fish survival. At the very least, the same conditions that were important for high cyclopoid copepod abundance also favoured juvenile Lake Whitefish survival. This fortuitous result is useful because the zooplankton abundance data is collected annually and could be used by fisheries managers as a very early indicator of Lake Whitefish recruitment levels.
Hoyle (2005) reported a decline in adult Lake Whitefish length-at-age in the mid-1990s and attributed it to the loss of Diporeia—the major food source for juvenile and adult whitefish. The results of the present study suggest that larval whitefish prey availability may also be a critical factor influencing Lake Whitefish recruitment and population dynamics in the Bay of Quinte. Here, we report an 89% decline in larval whitefish zooplankton prey that appears to be causing decreased larval fish growth and survival to the juvenile life stage.
We gratefully acknowledge past and on-going contributions by field and laboratory crews at the Ontario Ministry of Natural Resources, Glenora Research Station in Picton, and at the Fisheries and Oceans Canada office in Burlington. Thanks to W. Geiling, D. Geiling and C. Tudorancea for zooplankton sample enumeration, taxonomy and data analysis. Thanks to C. Lake for preparing the map of the Bay of Quinte. We also thank two anonymous reviewers for their thoughtful suggestions on an earlier version of this manuscript. Funding was provided by the Ontario Ministry of Natural Resources and Fisheries and Oceans Canada.