Long term monitoring of Lake Ontario's Bay of Quinte provides the opportunity to examine the impact of dreissenid invasion on the zooplankton community. Weekly or biweekly zooplankton samples have been collected from 1975 to 2008 at 3 stations: Belleville (B), Hay Bay (HB), and Conway (C) along a trophic and depth gradient down the bay. Rotifers have been collected since 2000. Biomass estimates based on measured zooplankton lengths started in 1995. Archived seasonal composite samples prior to 1995 were reanalysed and biomass recalculated from length-weight equations to allow for comparable data in trend analysis. Mean May 1– October 6 zooplankton biomass was low from 1975 into the early 1980s and peaked between 1982–1983 and 1991. Biomass fell during the cold summer of 1992 associated with the Pinatubo eruption and was low after the invasion of dreissenid mussels. From 1979–1991 (after phosphorus control and prior to dreissenid invasion), biomass averaged 265, 253 and 84 mg m−3 at B, HB and C, respectively. Seasonal biomass of most zooplankton groups, as well as total biomass, was significantly lower at all stations after the dreissenid invasion. Cladocerans still dominated zooplankton biomass after the invasion, averaging 56% to 80% of the total. Cyclopoid numbers and biomass fell dramatically.
After the invasion of Cercopagis pengoi in 1999, calanoid and cyclopoid biomass at HB and C decreased by approximately 50%. Bosmina biomass did not change. Seasonal mean rotifer biomass over the 2000 to 2008 period was 1.9% to 4.4% of total zooplankton biomass.
Reductions in zooplankton following dreissenid and C. pengoi invasions are thought to be caused by both direct predation of microzooplankton (e.g. nauplii) by mussels and copepods by C. pengoi (at C and HB), and competition between zooplankton and dreissenids for food resources.
The Bay of Quinte is a narrow embayment 64 km long and 254 km2, on the north-eastern shore of Lake Ontario (Johnson and Hurley, 1986; Figure 1). The Bay was officially listed as an Area of Concern in the 1987 amendment of the Great Lakes Water Quality Agreement (International Joint Commission, 1988). However, concerns over cultural eutrophication started in the 1960s, and led to the development of “Project Quinte” in 1972 (Minns et al., 1986). This multi-agency, multi-year research program was largely developed to evaluate ecological responses to phosphorus reductions implemented in 1978. This monitoring has continued to the present (2008), providing one of the longest running and most complete limnological data sets focusing on water quality, zooplankton, phytoplankton and fish.
The Remedial Action Plan developed for Quinte listed zooplankton as one of its impaired beneficial uses (Bay of Quinte RAP Coordinating Committee, 1993). The zooplankton community changes in structure from the shallow, eutrophic habitat of the upper bay, through less eutrophic conditions in the middle section, to a meso-oligotrophic environment in the mouth of the bay. Zooplankton community structure and biomass are controlled not only by the physical environment but also by the quality and quantity of available food and level and type of predation. This is determined by nutrient conditions, the structure of the fish community which is the primary source of mortality in this system, and the introduction of new species which either compete for food or alter predation on or within the zooplankton community. A few strongly invasive species colonized the Bay in the 1990s. Zebra Mussels (Dreissena polymorpha) and Quagga Mussels (D. rostriformis bugensis) were established in the lower bay by 1993, and D. polymorpha was found in the upper bay by 1994 (Dermott et al., 2003). Cercopagis pengoi, a predatory cladoceran, arrived in Lake Ontario in 1998 (MacIsaac et al., 1999) and were found in the lower bay in 1999.
At the beginning of Project Quinte, it was postulated that phosphorus reductions, and resulting lowered algal abundance would lead to a drop in zooplankton standing crop and production, and changes in species composition (Cooley et al., 1986). However, monitoring results comparing two years (1975–1976) prior to phosphorus control with the five years following phosphorus control (1979–1983) indicated that these changes did not occur. In the 1960s and early 1970s, small-bodied cladocerans such as Bosmina, Eubosmina and Chydorus dominated the upper Bay, and this dominance continued throughout the Bay in the early post-phosphorus control period (Cooley et al., 1986). Cyclopoid copepods, particularly juvenile nauplii and copepodids, increased in relative abundance down the length of the bay toward Lake Ontario. As a follow-up to Cooley et al. (1986), the purpose of this paper is to provide an overview of trends in zooplankton community composition, abundance, and biomass from 1975 to 2008, with emphasis on the changes that occurred following invasion by dreissenid mussels and C. pengoi. Our goal was not to examine interactions with planktivorous fish or with changes in the algal community in the Bay of Quinte; those analyses will be discussed in future publications.
A suite of physical (temperature, oxygen, light attenuation), chemical (total phosphorus, soluble reactive phosphorus, nitrate+nitrite, total Kjedhal nitrogen, conductivity) and biological (benthos, zooplankton, phytoplankton and as of 2000 rotifers) variables were collected as part of the monitoring program in the Bay of Quinte (Minns et al., 1986). Here we examine the zooplankton and rotifer data, touching on other relevant data. At least three primary stations, Belleville (B), Hay Bay (HB), and Conway (C) which fall along a trophic gradient from the upper to lower Bay (Figure 1), have been sampled every one or two weeks through much of the ice-free season, generally from early to mid-May until early October: a total of 17–23 (1975–1982) or 9–12 (1982–2008) samples per station-year. Station depth increases from 5 m at B, to 11 m at HB and 32 m at C. This means that thermal stratification occurs throughout the summer at C, sporadically at HB and not at B.
Zooplankton sample collection
Discrete samples for zooplankton were collected through the water column with a 41-L Schindler-Patalas trap fitted with 64-μm mesh. Three depths were sampled at B (1 m, 2 m and 3 m), five at HB (1 m, 2 m, 3 m, 6 m and 9 m), and eight at C (1 m, 3 m, 5 m, 8 m, 10 m, 15 m, 20 m and 25 m). Samples were preserved with 4% sugar, buffered formalin. A single composite sample was constructed for each station-date by combining 50% of the sample from each depth.
Rotifer sampling began in 2000 by collecting 1 L of water from each depth indicated above using a Van Dorn sampler. For each station-date, water was pooled and filtered through 20-μm mesh. Rotifers were narcotized using carbonated water and preserved as above. Each year, a seasonal composite sample for each station was made by combining 50% of the sample from each date.
Zooplankton were enumerated using a weighted counting procedure where enumeration of dominant species ceased before that of rarer species (Cooley et al., 1986). The count was considered complete when a minimum of 400 individuals had been identified. Loose eggs were counted in all aliquots examined. A minimum of 200 organisms were counted in a sample dominated by only one to two species with extremely few other species present, as can occur in the early spring. For numerically dominant species, a minimum 50 individuals were measured. The exotic predatory cladoceran C. pengoi was found at B and C in 1999, and at all three stations by 2000. After its arrival, each of the discrete depth samples was filtered through 400-μm mesh, and all C. pengoi enumerated. This species cannot be sub-sampled because individuals form clumps as their caudal spines become entangled.
Seasonally-weighted mean (SWM) abundance and biomass of zooplankton were calculated over the May 1 to October 6 sampling season, weighting for time between each sample, including the start (May 1) and end (October 6) dates, in order to standardize the time period and adequately account for missing samples etc. Starting in 1995, zooplankton lengths were measured on each date using a SummaSketch III digitizing system and dissecting microscope equipped with a camera lucida. A program written by Russ Hopcroft (University of Alaska) recorded the lengths and calculated the biomass using length-weight relationships from Johannsson et al. (2000), except for Bosmina and Eubosmina where a relationship developed for Lake Erie zooplankton was employed. This is W =10.715*L2.12, where weight (W) is in μg and length (L) is in mm. C. pengoi weights were estimated using equations in Grigorovich (2000). Cladoceran taxa not listed in Johannsson et al. (2000) were rare and most closely resembled Daphnia, thus the Daphnia equation was applied. Starting in 1995, biomass was calculated for each species on each station-date by multiplying the abundance and estimated mean individual weight for that date.
Prior to 1995, a single individual mean weight was initially assigned to each species based on average measured dry weights of individuals caught in Lakes Superior, Huron and Ontario in the early 1970s. These weights were originally determined by J.B. Wilson and N.H.F. Watson (Fisheries and Oceans Canada, Burlington, ON, Canada, unpublished data) and Cooley et al. (1986). Biomass was initially estimated as abundance times these weights. However, it was necessary to standardize our methodology across all years to examine temporal trends and to assess the impact of this change in methodology on biomass estimates in the Bay. We collated and analysed archived samples across the season for each station-year from 1975–1994 in order to get species lengths and estimates of species annual mean weights. SWM biomass of each species was then calculated as the product of the station-year specific weight and SWM abundance. For rare taxa, for which there were few or no measurements during these years, mean weights were assigned depending on available data in this order: (1) the average for the 1975–1994 period, (2) the average for the 1995–2008 period, and (3) a mean weight determined for a similar Great Lakes’ species (mean GL) determined by Wilson and Watson; that is, so rare that it did not occur in any of the collated samples, nor was it found after 1994.
Enumeration and identification of rotifers were similarly performed using a weighted counting method and limits similar to that of the zooplankton. In cases where samples contained few rotifers, a maximum of 25% of the sample by volume was analysed. The biovolume (mm3) for each individual was calculated by using formulae of Ruttner-Kolisko (in McCaulley, 1984). Biovolume equals wet weight, assuming a density of 1. Formula used for Polyarthra species are given in Johannsson et al. (2000). Dry weights were calculated by multiplying wet weight by 0.1, or in the case of Asplanchna sp. 0.036 (Dumont et al., 1975). Mean weights of rotifer taxa from Quinte were compared to those from Lake Erie (1993, 1994) (Graham et al., 1996; O. Johannsson and J. LeBlanc, Fisheries and Oceans Canada, Burlington, ON, Canada, unpublished data).
Unless otherwise indicated, comparisons were considered significantly different at p < 0.05. The pre-phosphorus control (pre-P) period from 1975 and 1976 was excluded from statistical analyses as the effect of phosphorus control was described by Cooley et al. (1986). The years 1992 and 1993 were also omitted from the analyses because of their unusually low zooplankton biomass associated with the cold summer of the Pinatubo eruption (Newhall and Punongbayan, 1996) and slow return of copepod abundance. 1994 was also omitted because it was a transition year for dreissenid colonization. For each zooplankton group (total zooplankton, cladocerans, cyclopoids and calanoids), the remaining 27 year data set was log-transformed (no zero values present) and split into the following time stanzas: pre-dreissenid (pre-DM): 1979 to 1991 and post-dreissenid post-DM): 1995 to 2008.
For each group, differences in SWM biomass were examined using general linear models (GLM) in SystatV11, with station, stanza and station*stanza as factors. Within each stanza, differences among stations were examined using ANOVA and Bonferroni post-hoc tests (p < 0.025). At each station, differences among the time stanzas were examined using t-tests. Cladoceran mean length, an index of fish predation, was also compared between stanzas at each site using t-tests.
In order to test for impacts associated with the C. pengoi invasion, the post-DM data were divided into pre-CP (1995–1998) and post-CP (1999–2008) periods. Differences in biomass of each zooplankton group and dominant taxa (veligers, Bosmina spp., Eubosmina coregoni, Daphnia galeata mendotae, Daphnia retrocurva, Ceriodaphnia sp., Chydorus sp., cyclopoid nauplii and copepodids and calanoid nauplii) were assessed using t-tests on the transformed data.
Total zooplankton densities and biomass
Zooplankton biomass peaked at all three stations between 1981–83 and 1991 and appeared to be lower and more variable thereafter, although means at HB have been relatively stable since 2004 (Figure 2A– C). High values were again observed in 1994 and 2001 at B, and in 1996 and 2008 at C. Biomass reached unusually low levels at all three stations in 1992 (and again in 1993 at C and HB). The eruption of Mount Pinatubo in the Philippines in June 1991 (Newhall and Punongbayan, 1996) cooled the earth and was reflected in unusually low summer temperatures in the Bay of Quinte in 1992. The June to August mean water temperature at B was the lowest observed during the study: 20.3°C compared with the 1975–2008 mean of 22.3°C (excluding 1992).
GLM analyses revealed that zooplankton densities and biomass decreased significantly between the pre-DM and post-DM time stanzas (Figure 3A). The patterns were similar across stations as indicated by a lack of significant interaction terms (stanza*station) in the GLM models. Mean total densities in the pre-DM stanza averaged 207 ± 16, 275 ± 28 and 92 ± 7 individuals L−1 at B, HB and C, respectively - values which were all higher than the corresponding values in the post-DM stanza (146 ± 9, 167 ± 9 and 55 ± 6 individuals L−1). In the pre-DM period, mean total biomass (±1 S.E.) was higher at B (265 ± 21 mg m−3) and HB (253 ± 19 mg m−3) compared to C (84 ± 5 mg m−3). In the post-DM stanza, B (166 ± 20 mg m−3) and HB (192 ± 13 mg m−3)were again higher than C (48 ± 5 mg m−3). Declines in total biomass were 37% at B, 24% at HB, and 42% at C (Figure 3A).
Compared to the pre-DM period, cladoceran mean length increased in the post-DM period at C (0.356 ± 0.011 mm to 0.425 ± 0.009 mm) and HB (0.401 ± 0.014 mm to 0.450 ± 0.013 mm), but remained unchanged at B (0.439 ± 0.018 mm to 0.434 ± 0.017 mm). Within the post-DM period, cladoceran mean length decreased post-CP at HB (0.503 ± 0.025 to 0.423 ± 0.033 mm) but remained unchanged at B and C.
Zooplankton group biomass
With the exception of cladocerans and calanoids at HB, biomass of each zooplankton group declined after the dreissenid invasion (Figures 3B– D). The patterns were similar across stations as indicated by a lack of significant interaction terms (stanza*station) in the GLM models. Cladocerans were dominant at all stations (Figure 2), and values were higher at B and HB relative to C in both time stanzas (Figure 3B). In the pre-DM stanza, percent cladocerans (by biomass) was highest at B (79%), intermediate at HB (66%) and lowest at C (52%). In the post-DM stanza, the proportion of cladocerans remained relatively stable, with both B (80%) and HB (71%) higher than C (56%). Based on t-tests, dominant cladoceran taxa also showed significant differences following dreissenid invasion, but these were often station-specific (Table 1). For example, Bosmina spp. showed a decline only at C, but E. coregoni dropped at all three stations following invasion.
In both stanzas, cyclopoid biomass was higher at HB relative to B and C. The proportion of cyclopoids in the pre-DM stanza was highest at C (42.7%), intermediate at HB (31.8%) and lowest at B (19.3%). These values dropped at all stations in the post-DM period, with corresponding values of 21.4%, 19.6% and 12.5%. Calanoid biomass tended to be highest at HB in both stanzas. Percentages of calanoids were highest at C, intermediate at HB and lowest at B in both time stanzas, and ranged from 1.2% to 5.2%.
Following dreissenid invasion, the proportion of veliger biomass in the zooplankton community largely replaced the proportion lost by cyclopoids. Veligers were more dominant at C (17.5%) than at HB (5.9%), and B (6.2%). Veliger biomass showed no significant differences among stations.
Only a few Bythotrephes have been found at Quinte, and these have been restricted to the lower bay in recent years. C. pengoi has only rarely been found at B and HB, and was usually abundant at C for a short summer period. Between 1999 and 2008, mean SWM biomass of C. pengoi averaged 0.01 ± 0.00, 0.04 ± 0.02 and 0.58 ± 0.12 mg m−3 at B, HB and C, respectively. Changes within the post-DM period were further examined to determine possible effects of C. pengoi invasion. At C, biomass of total cyclopoids, total calanoids, cyclopoid nauplii, cyclopoid copepodids and D. retrocurva declined significantly in the post-CP period (1999–2008) compared to the pre-CP, post-DM period (1995–1998) (Figure 3E– H). Total cladocerans, veligers, calanoid nauplii and the cladocerans Bosmina spp., E. coregoni, D. galeata mendotae, Chydorus and Ceriodaphnia remained unchanged. At HB, where C. pengoi were much less abundant, total biomass, total cyclopoids, total calanoids, cyclopoid nauplii and cyclopoid copepodids also declined over this period, and veligers increased. At B, where C. pengoi were rarely found, the only change was an increase in Ceriodaphnia.
Species composition and mean sizes
The following zooplankton taxa were dominant at all stations (>5% of biomass): the cladocerans Bosmina spp., D. galeata mendotae, D. retrocurva and E. coregoni, cyclopoid copepodids and dreissenid veligers (Table 2, in bold). The cyclopoid Diacyclops thomasi was a dominant species at C.
Individual dry weights of many zooplankton taxa and species, estimated from mean length data and length-weight regressions, were lower than the mean GL weights determined from the early 1970s (Table 2). Mean weights of the dominant taxa are based on hundreds or thousands of measurements, whereas values for the rare taxa (indicated by an *) may be based on only a few measurements. Daphnia and Ceriodaphnia taxa, Diaphanosoma birgei, Holopedium gibberum, and Sida crystallina were notably smaller in Quinte than in the open waters of the Great Lakes in the early 1970s. Species individual dry weights tended to be more similar for the copepod taxa, although both cyclopoid nauplii and copepodids were generally smaller in Quinte. Over the 1995 to 2008 period, SWM biomass calculated using species dry weights determined from length measurements were 38 to 45% lower than biomass estimates calculated using the 1970s mean Great Lakes weights.
Seasonal mean dry rotifer biomass (±S.E.) over the 2000 to 2008 period was 2.87 ± 0.66 mg m−3 at B, 2.83 ± 0.92 mg m−3 at HB, and 1.69 ± 0.54 mg m−3 at C (Figure 4). Mean rotifer biomass was only 1.7, 1.5 and 3.4% of zooplankton + rotifer SWM biomass at the three stations, respectively. The dominant rotifer taxa (>5% of biomass) were Asplanchna priodonta, Asplanchna sp., Keratella cochlearis, Polyarthra dolichoptera, Polyarthra major, Polyarthra vulgaris, Trichocera cylindrica and Trichocera multicrinis (Table 3). Estimated measured weights for a number of rotifer taxa were smaller than the Lake Erie means based on the same length-weight equations, including Asplanchna sp., P. dolichoptera, P. major and T. multicrinis. A few others, including A. priodonta and T. cylindrica were larger.
Biomass values in this study were calculated from length-weight relationships. Such estimates of zooplankton biomass are preferable to estimates based on abundance and a set species weight because they capture more variation in the populations associated with growth patterns and levels and types of predation. These estimates based on length measurements in the Bay of Quinte were 38 to 45% lower than the initial estimates based on weights of offshore Lake Superior, Huron and Ontario animals from the early 1970s. The length-weight equations used at Quinte are widely recognized relationships from the literature and were selected because they best represented actual length-weight data determined for Daphnia retrocurva, Diacyclops thomasi (narrow cyclopoid) and Mesocyclops edax (wide cyclopoid) from Lakes Ontario and Erie from the late 1980s and early 1990s (Johannsson, unpubl. data). The 1970s measured mean species weights may be higher that the means observed at Quinte due to differences in predation amongst the different regions. Nearshore, nursery areas, such as the Bay of Quinte, tend to have higher levels of planktivory than more offshore regions, which would select for smaller body sizes (Hall et al., 2003). Importantly, the method has now been standardized across all years, allowing trend analyses.
Zooplankton biomass has significantly declined in the Bay of Quinte as of the mid-1990s, with the largest change seen in cyclopoid copepods. Declines in zooplankton biomass may be driven by a number of causes: unusually cold weather (to witness, 1992), increased predation, declines in food supplies and alteration of habitat. Reduced phosphorus loading to the Bay in the late 1970s did not result in lower zooplankton biomass or production (Cooley et al. 1986). Rather the zooplankton increased in biomass and reached peak levels during the 1980s (Figure 2). Any effect of reduced nutrients was overshadowed, likely by a decline in planktivory associated with changes to the fish community (Johannsson and Nicholls, 2002).
A number of observations suggest that dreissenids were implicated in the decline of zooplankton in the mid-1990s. Dreissenids are ecological engineers, causing change that ripples through food webs. Dreissenids were observed through out the Bay starting in the early 1990s (Dermott et al., 2003). Veliger larvae of dreissenids contributed 7.0, 2.6 and 6.8% of total zooplankton biomass at B, HB and C by 1995, indicating a strong presence in these regions by this time. Veligers have continued to be a significant part of the zooplankton communities up until the present time (2008). Declines in chlorophyll a, (43 to 55% across the three stations) and increases in water clarity occurred at all sites as of 1995 (Bedford and Gerlofsma, 2010) and rapid expansion of macrophytes were observed in the upper and middle Bay (Leisti et al., 2006). Similar changes have been observed at other sites of invasion (Dahl et al., 1995; Fahnenstiel et al., 1995). Such changes could also be caused by decreases in phosphorus loading with consequent declines in total phosphorus levels (TP); however, TP levels were lower in 1995 but did not decline across the 1990s except perhaps at C (Nicholls, 2010; E.S. Millard and M. Burley, Fisheries and Oceans, Burlington, unpubl. data). Chlorophyll a/TP ratios, an indication of dreissenid grazing often observed after the dreissenid invasion (Nicholls et al., 1999; Hall et al., 2003), plummeted between 1995 and 1996 at B and C, and between 1994 and 1996 at HB (Figure 5). Levels of the cyanobacterium Microcystis increased as of 1995 in the upper Bay (Nicholls, 2002). Phytoplankton biomass declined in 1995 at all stations, increased for a few years in the late 1990s and remained low again as of 2000 (Nicholls, 2010). Nicholls et al. (in press) consider the ecosystem changes that occurred in the upper Bay in 1995 to be a regime shift. Six of the top ten drivers which defined this regime shift were phytoplankton groups. All these changes suggest that dreissenids were altering the pelagic ecosystem throughout the Bay as of 1995/1996.
Dreissenid-mediated declines in zooplankton could occur through changes in phytoplankton/seston quantity and quality and/or changes in predation associated with the increases in water clarity which improves detection by predators. Lack of decrease in cladoceran mean length between the pre- and post-DM periods indicates that the decline in zooplankton was not due to increases in fish predation. The observed reductions in chlorophyll/phytoplankton suggest that food limitation may be one of the causes of zooplankton decline, as in other systems (MacIssac et al., 1995; Jack and Thorp, 2000; Nicholls et al., 2002; Thorpe and Casper, 2003). Mussels consume particles in a wide size range including microzooplankton such as protozoa, ciliates, rotifers, veligers and nauplii (MacIssac et al., 1995; Miller and Watzin, 2007) as well as phytoplankton. This could explain the low rotifer biomass in the Bay (1.7–3.4% of total zooplankton biomass): normally rotifers constitute slightly <30% of total zooplankton biomass in shallow eutrophic lakes in Europe where dreissenids are not a major driver (data from 50-μm mesh samples; Gyllström et al., 2005). Omnivorous copepod populations were likely affected directly through predation on nauplii and indirectly through competition for prey (rotifers, ciliates, small cladocerans, copepod juveniles and phytoplankton).
That dreissenids could impact the zooplankton in the Bay of Quinte is further suggested by the documented impacts of dreissenid invasions on zooplankton in other areas. Zooplankton, particularly copepod nauplii, and rotifers significantly declined in the shallow, unstratified western basin of Lake Erie following mussel establishment in the late 1980s (MacIsaac et al., 1995). In the eastern basin of Lake Erie (Outer Long Point Bay), biomass declined by 60% between 1984–1987 (prior to dreissenids) and 1993–1994 (post-dreissenids) (calculated from Table 3; Johannsson et al., 1999). Densities of cladocerans and copepods in Lake St. Clair, another shallow system somewhat comparable to the upper Bay of Quinte, fell by about 50% following the arrival of dreissenids, and rotifers dropped by 86% (David et al., 2009). In the Hudson River, total zooplankton biomass declined by more than 70% following invasion (Pace et al., 1998), and rotifers and nauplii were particularly impacted. Declines in zooplankton upon dreissenid invasion do not always occur (Idrisi et al., 2001), but the general experience has been a marked decline. Unfortunately, any decrease in rotifers following invasion cannot be assessed due to the lack of pre-invasion data.
The chlorophyll a/TP ratio increased again at B and HB between 2002 and 2006/7 suggesting some relaxation of the impact of dreissenids (Figure 5). Round goby (Neogobius melanostomus) invaded the Bay in the early 2000s and reached average catches >1 per gillnet set averaged across the three stations between 2002 and 2005 (Hoyle, 2010). Young gobies are produced throughout the season and consume zooplankton, while adult gobies consume benthos including dreissenids (Jude et al., 1995; Ray and Corkum, 1997). Thus a series of interactions will occur amongst gobies, dreissenids and zooplankton. Zooplankton biomass at B was low between 2002 and 2005 suggesting that gobies may play a noticeable role along with deissenids in regulating zooplankton biomass in the shallow upper Bay.
C. pengoi was present for a short period each summer at C. During that summer period, the abundance of copepod nauplii and copepodids declined (Benoit et al., 2002). Similar seasonal patterns were observed in nearshore Lake Ontario (Warner et al., 2006). C. pengoi was very rare at B and no zooplankton declines were observed between the pre- (1995–1998) and post- (1999–2008) C. pengoi periods. However, in the middle (HB) and lower (C) Bay total cyclopoid and total calanoid biomass declined in synchrony (Figure 2) with the C. pengoi invasion. It is unknown how much of this decline was due to this invasive cladoceran. Although C. pengoi biomass was an order of magnitude lower at HB compared to C, it is possible that their effects were greater at HB due to the lack of a hypolimnetic refuge for zooplankton prey. C. pengoi at C were most abundant in the epilimnion and metalimnion. Contrary to expectations based on observations in the literature (Benoit et al., 2002; Laxon et al., 2003; Warner et al., 2006), small species, like Bosmina, did not decline. Either they are less susceptible to C. pengoi predation and/or other factors are important in the copepod declines. There were no increases in cladoceran mean length between the pre-DM and post-DM period which might have supported an increase in the importance of invertebrate predation. To balance this observation, it must be remembered that cyclopoids themselves are ominvores/invertebrate predators and the loss of one predator may partly offset the impact of the other.
Although zooplankton populations/communities are governed by a number of drivers, a dominant structuring force in the Bay of Quinte since 1995 has been Dreissena spp. The continued decline of copepods in the middle and lower Bay may be due to the invasion of C. pengoi. The trends in chlorophyll a/TP ratios suggest that the effect of dreissenid mussel grazing has declined, possibly because of predation by round goby. This needs further examination.
Foremost, we acknowledge our indebtedness to the Fisheries and Oceans Canada field and laboratory crews who have made this work possible: J. Gerlofsma, R. Bonnell, A. Bedford, M. Burley, H. Niblock, M. Fizpatrick, B. Timmins, T. Hollister, K. Bonnell and others too countless to name. Thanks to C. Tudorancea, W. Geiling, D. Geiling and J. Moore, among others, for zooplankton and rotifer sample enumeration, taxonomy and data analysis. We thank J. Cooley, J. Moore and K. Minns who started and ran the zooplankton data collection and analysis prior to us, and to the Directors of GLLFAS (DFO) whose continuous support of the program since its inception helped to ensure its viability. Total phosphorus data were provided by Scott Millard and Michele Burley, Fisheries and Oceans, Burlington. We acknowledge the support of the Bay of Quinte RAP Restoration council and other leads on Project Quinte over the years, including M. Munawar, R. Randall, R. Dermott, K. Minns, K. Nicholls, M, Johnson, J, Christie, D. Hurley and others. Thanks also to Environment Canada's Technical Operations Services for assistance with moorings, and Ontario Ministry of Natural Resources for use of their Glenora lab. Financial support was provided by the Great Lakes Action Plan.