Secchi depth, total phosphorus, soluble reactive phosphorus, silica, chlorophyll a, and zooplankton (density, biomass, and average size) were sampled as part of a lake-wide, seasonal (spring, summer, and fall) assessment of Lake Ontario in 2003 to characterize the status of the lower aquatic food web. For each parameter, spatial comparisons were performed to test for differences between habitats (nearshore and offshore) and between regions (east and west) during each season. Significant differences between habitats were found only for silica and chlorophyll a; silica was higher nearshore in fall, and chlorophyll a was higher offshore in fall. Significant differences between regions were detected in Secchi depth, epilimnetic zooplankton mean length, total phosphorus and Cercopagis pengoi density and biomass; Secchi depth and zooplankton mean length were higher in the east in spring, total phosphorus and Cercopagis pengoi biomass were higher in the west in summer, and Cercopagis pengoi biomass was higher in the east in fall. Cercopagis pengoi was present lake-wide in summer and fall, but Bythotrephes longimanus was present only in fall in the Kingston basin. Mean spring total phosphorus, soluble reactive phosphorus, chlorophyll a, and epilimnetic zooplankton density and biomass are at or near record low levels. As we move into the future, persistent low levels of these lower food web elements will continue to stress alewife populations both through reduced food resources and food quality for zooplankton, and may force these fish to seek alternative food such as Mysis.
The physical, chemical, and biological structure of the Great Lakes ecosystems, including that of Lake Ontario, changed considerably during the three decades from 1970 to 2000 (Madenjian et al. 2002; Mills et al., 2003, 2005; Bronte et al., 2003; Dobiesz et al., 2005). In Lake Ontario, water clarity increased, phosphorus concentrations declined, and several non-native species became established. Reduced phosphorus loading, mandated by the Great Lakes Water Quality Agreement of 1972, was largely responsible for the oligotrophication of the lake and declines in productivity of phytoplankton and zooplankton during this period (Johannsson 1987, 2003; Neilson and Stevens 1987; Stevens and Neilson, 1987; Johannsson et al., 1991; Johengen et al., 1994; Millard et al., 1996; Millard et al., 2003). Through the latter part of this period, several invasive, non-indigenous species (NIS) became very abundant, in particular, the zebra mussel (Dreissena polymorpha), quagga mussel (D. bugensis), spiny water flea (Bythotrephes longimanus), and fishhook water flea (Cercopagis pengoi) (Mills et al., 1993; MacIsaac et al., 1999). Addition of these species has caused changes in community structure, trophic relationships, and productivity patterns in several of the Great Lakes (Nalepa, 1991; Dermott and Kerec, 1997, Lozano et al., 2001; Johannsson et al., 2000; Benoit et al., 2002; Barbiero and Tuchman, 2004; Warner et al., 2006; Pangle et al., 2007).
Declining productivity and establishment of NIS are of particular concern relative to Lake Ontario's non-native salmonid fishery which is supported primarily by alewife (Alosa pseudoharengus), a non-native planktivore. A declining alewife population (Jones et al., 1993; O'Gorman and Stewart, 1999) led fishery managers to shift from a strategy of alewife control to concern about the ability of alewife to sustain the salmonid fishery. The alewife decline may be linked to the establishment of other NIS (i.e. dreissenid mussels and the invasive predatory cladocerans, C. pengoi and B. longimanus). For example, Mills et al. (2003) suggest that filter-feeding dreissenids have depressed zooplankton production, particularly in nearshore habitats where the ratio of dreissenids to water volume is highest. Also, Cercopagis and Bythotrephes have added competitive and predatory pressure on the zooplankton community and likely increased competition among zooplanktivores (Warner et al., 2006). Declines in zooplankton abundance and changes in community composition are also reflective of the overall decline in lake productivity as evidenced by low nutrient levels (Johannsson, 2003).
In this study, we examine the status of water clarity, total phosphorus (TP), soluble reactive phosphorus (SRP), soluble reactive silica (SRS), chlorophyll a, and zooplankton biomass, density, and mean length (ZML) in Lake Ontario in 2003. These data are compared with long-term trends and a previous large-scale spatial study to examine the possible impacts that changes in lower food web parameters may have had on the ecosystem, and to establish a baseline for future assessment. In addition, we examine the importance of season and of geography (nearshore vs. offshore; east vs. west) in describing the present patterns in these variables. In this work we depend heavily on historical, long-term databases and studies developed and executed by Environment Canada (Surveillance Program, 1969–2003), Fisheries and Oceans Canada (Bioindex Program, 1981–1995; Lake Ontario Trophic Transfer Project, 1990–1996), and the United States Environmental Protection Agency (1986–2003).
When sampling large lakes such as the Great Lakes, resources dictate that a choice must be made between temporally intensive sampling at a few representative sites and broad spatial sampling a few times during the season. Although a temporally intensive sampling program provides a much better understanding of the dynamics and productivity of a region (El-Shaarawi and Kwiatkowski, 1977; Minns, 1984; Johannsson et al., 1998), it cannot be used to evaluate spatial patterns. Due to resource limitations and the fact that a spatially intensive survey had not been performed recently, the sampling design used in our study (the Lake Ontario Lower Aquatic Food Web Assessment or LOLA) was modeled after the earlier Canadian Lake Ontario Trophic Transfer (LOTT) program (1990–1996), the most recent lake-wide survey of Lake Ontario's lower trophic levels.
Sample collection and processing
Water and zooplankton samples were collected during the day during three lake-wide cruises in spring (April 28–May 3), summer (August 10–11 and August 19–21), and fall (September 21–25), 2003 from sites along four north-south transects (Figure 1) in Lake Ontario. The cruises were planned to coincide with organism life cycles and provide a spatial component of the season cycle. Spring is the time of isothermal water temperatures and provides the initial chemical conditions for the year prior to significant uptake of nutrients by the biota. The spring survey provides a picture of the amounts of nutrients available for the biological activity that will occur during the year. Summer and fall surveys (August and September) characterize the summer and late summer zooplankton production and community structure. The 30-m bathymetric contour was used to delineate nearshore (gray area in Figure 1) and offshore habitats (nearshore n = 7, offshore n = 16 for nutrients; nearshore n = 4, offshore n = 16 for zooplankton). The lake was divided into eastern and western regions by the 77.8°W longitude line (eastern region n = 13, western region n = 10 for nutrients; eastern region n = 11, western region n = 9 for zooplankton). Zooplankton samples were lost for three nearshore stations (black triangles in Figure 1) in the spring; therefore, those stations were omitted from all zooplankton analyses.
Sampling procedure varied with season due to the location of the thermocline. An electronic bathythermograph (EBT) or conductivity-temperature-depth (CTD) profiler was used to determine thermocline depth. During spring isothermal conditions, integrated water samples were collected from 20-m depth or two meters above the bottom (for shallow stations) to the surface. In summer and fall, integrated water samples were collected from one meter above the thermocline to the surface. Water samples were collected using an integrated tube sampler (1.9-cm inside diameter Nalgene tube), lowered to the appropriate depth, and kept perpendicular to the lake bottom. Water was transferred to 2-L Nalgene containers and processed immediately on board the ship. Water was shaken to mix thoroughly and a 100 mL aliquot was taken and preserved with 1 mL 30% H2SO4for later analysis of TP. For soluble reactive phosphorus (SRP) and soluble reactive silica (SRS), water was filtered through a 0.45-micron membrane filter and 100 mL aliquots of filtrate were taken. Water samples were stored in the dark at 4°C until returned to the lab for analysis. For chlorophyll a, up to 2L of water was filtered through a Whatman GF/C filter at a pressure not exceeding 300 mmHg, and the filter was frozen for later analysis. Once samples were returned to the lab, TP concentration was determined colorimetrically using the ammonium molybdate – stannous chloride method after persulfate digestion (Philbert and Traversy, 1973). SRP and SRS were analyzed in an autoanalyzer using the ammonium molybdate – stannous chloride method heteropoly – blue method, respectively (Philbert and Traversy, 1973). Chlorophyll a concentration was determined by 90% acetone extraction (Strickland and Parsons, 1972). Secchi depth (m) and temperature (°C) were recorded at each site.
Epilimnetic zooplankton samples (following depth protocol for water sampling; see above) were collected using a 64-μ m mesh, 50-cm diameter, flow-metered (Rigosha) net raised at a minimum rate of 0.7 m second−1. Zooplankton were preserved in the field in 70% ethyl alcohol after anesthetization with antacid tablets. Zooplankton (other than Cercopagis) were counted using the stratified regime of Cooley et al., (1986) and biomass determined by length weight regressions (Ora Johannsson, unpublished data). To separate Cercopagis, the sample was first passed through a 400-μ m mesh sieve. The 400-um fraction was transferred to a Petri dish and examined under a microscope and all Cercopagiswere removed. Up to 60 animals were measured, and weights were determined using the regression equation of Makarewicz et al. (2001). If more than 60 Cercopagis were present, a subsample of 200-250 individuals was counted, dried, weighed, and used to estimate abundance and biomass in the entire sample. Cryovial bullet tubes were numbered and precisely weighed using a Mettler balance. Cercopagis from both the subsample and remainder of the sample were rinsed to remove trace ethanol and other zooplankton, transferred to the cryovials, freeze dried, and weighed. The weight of the counted fraction was then used to determine the weight of the uncounted fraction.
T-tests were used to compare means for each parameter to identify spatial differences during each season. Significantly different means were identified by a p-value adjusted for multiple comparisons using the Bonferroni correction. Spatial categories include east and west (separated by 77.8° W longitude line), referred to hereafter as regional comparisons, and nearshore and offshore (separated by the 30-m bathymetric contour), referred to hereafter as habitat comparisons. Analysis of variance (ANOVA) was used to compare means across seasons. Significant results, identified by a p-value < 0.05, were examined using the Tukey-Kramer HSD test (JMP IN 5.1.2; SAS Institute) to identify which pairs of means were different. Water quality parameters (Secchi disc, TP, SRS, and chlorophyll a) and zooplankton mean length (ZML) data were normally distributed and therefore were not transformed. In the case of SRP, transformation did not improve normality; analyses were performed on untransformed data. Zooplankton density and biomass data were log10 transformed after adding one. Zooplankton mean length was calculated from species average lengths after weighting for species density. Chlorophyll a concentrations were unadjusted for phaeophytin.
Programs: Intensive surveys conducted on Lake Ontario during the 1970s found more temporal than spatial variability in chlorophyll a and zooplankton (El-Shaarawi and Kwiatkowski, 1977; Minns, 1984), and a research program was designed accordingly. The Department of Fisheries and Oceans Bioindex Program (1981–1995) sampled two fixed stations (midlake station 41 and Kingston Basin station 81) on a weekly basis from April to October (Johannsson et al., 1998; Johannsson, 2003; Millard et al., 2003). Recognizing that periodic, large-scale, spatial research was also essential in the overall evaluation of ecosystem change, the Department of Fisheries and Oceans participated in a study assessing whole lake, multi-trophic level productivity in Lake Ontario in 1990 (the Lake Ontario Trophic Transfer or “LOTT” Project). The LOTT Project sampled five to eight depth-defined sites on six north-south transects spaced evenly from west to east across the lake in spring, summer, and fall (Millard et al., 2003). In addition, Environment Canada (Surveillance Program, 1969-2003) conducted spatially intensive, whole-lake surveys for nutrients in spring (April) and summer (late July – early August) each year: a total of 98 stations per cruise (Millard et al., 2003). The USEPA (GLNPO Limnology Program, 1986–2003) sampled eight offshore sites in the spring (April) and eight offshore and five nearshore sites in late summer (August) of each year (Makarewicz et al., 1995; Barbiero and Tuchman, 2001; Barbiero et al., 2001). We obtained historical data collected by these sampling programs for comparison with data collected in 2003. All four sampling programs were instrumental in tracking the long-term decrease of phosphorus and the early impacts of dreissenid mussels on the Lake Ontario ecosystem.
Methods: Surveillance Program nutrient data were collected in spring from surface waters (1 m sampling depth) of open lake stations (sounding depth >100 m). Bioindex Program nutrient concentrations were obtained from integrated samples (0 to 20 m) collected weekly from stations 41 and 81 (spring 1981–1995). Spring averages for each year were calculated using values from April 1 to the onset of stratification. LOTT nutrient (May 22–31, 1990) and chlorophyll a (August 1, 1990 and July 14–15, 1995) data were from integrated samples collected from a depth of 0 to 20 m or 0 to bottom minus 1 m. EPA nutrient data (1986–2002) from stations 41 and 81 were collected in April and were integrated from a depth of 0 to 20 m.
Nutrient analyses for Surveillance, Bioindex, and LOTT were performed by the National Laboratory for Environmental Testing, Environment Canada, Burlington, Ontario (Environment Canada 1997). Soluble nutrients were analyzed onboard ship during Surveillance cruises. Bioindex and LOTT samples were filtered onboard, stored at 4°C and returned to the lab for analysis. Total phosphorus was analyzed by acid persulfate digestion followed by automated colorimetric molybdate stannous chloride method (Philbert and Traversy, 1973). For silica, water was filtered through a 0.45-micron membrane filter and then analyzed by the autoanalyzer heteropoly-blue method (Philbert and Traversy, 1973). EPA TP data were composites of water samples taken at discrete depths with Niskin bottles (spring: surface, 5 m, 10 m, and 20 m) mounted on a SeaBird Carousel. Sample processing techniques for those data are described in detail by Barbiero and Tuchman (2001).
Zooplankton samples were collected weekly at stations 41 and 81 during the Bioindex Program from 1981–1995. A 64-μ m mesh, 50-cm diameter, metered net was towed from 20 m depth (or 1 m above thermocline) to the surface. We compared LOLA densities from May, August and September to Bioindex densities (April 15– May 15; Aug 1 – Aug 31; Sep 1 – Sep 30). Zooplankton were not collected during the LOTT study.
In 2003, lake-wide mean spring TP concentration was 7.5 μ g L−1, and it increased over the course of the sampling season, reaching 11.3 μ g L−1 in the fall (Table 1). Lake-wide fall TP was significantly higher than spring TP (ANOVA F(2, 66) = 3.99, p = 0.02; Tukey-Kramer HSD). There were no significant differences in TP concentrations between nearshore and offshore habitats for any season, but TP was significantly higher in the western region in summer (Table 1 and 3b). Mean seasonal SRP concentrations were below 1.0 ug L−1 in 2003 (Table 1). For SRP, there were no significant seasonal changes lake-wide, no significant differences in concentrations between nearshore and offshore habitats for any season, and no significant differences between eastern and western regions for any season.
Mean seasonal SRS concentrations were highest in spring (734 μ g L−1) and lowest in summer (265 μ g L−1) (Table 1). Lake-wide SRS levels were significantly higher in spring compared with summer and fall (ANOVA F(2, 66) = 60.9, p < 0.0001; Tukey-Kramer HSD). A comparison of habitat by season found SRS to be significantly higher in the nearshore in the fall (t-test; p < 0.0001) (Table 1 and Table 3a), but a comparison of region by season found no differences.
Chlorophyll a and water clarity
Mean chlorophyll a concentrations did not exceed 2.7 μ g L−1 in 2003 (Table 1). Lake-wide chlorophyll a was significantly higher in fall compared with summer, and significantly higher in summer compared with spring (ANOVA F(2, 66) = 13.9, p < 0.0001; Tukey-Kramer HSD). A comparison of habitat by season found chlorophyll a to be significantly higher in the offshore in the fall (t-test; p = 0.0006) (Table 1 and Table 3a), but there were no significant differences between eastern and western regions for any season. Mean lake-wide Secchi disc depth was highest in spring (9.8 m) and declined through the summer reaching a low of 6.7 m in the fall (Table 1). Lake-wide Secchi depth was significantly higher in spring compared with summer and fall (ANOVA F(2, 66) = 8.26, p = 0.0009; Tukey-Kramer HSD). Secchi depth was significantly higher in the eastern region in spring (t-test; p = 0.0003) (Table 1 and Table 3b).
Epilimnetic zooplankton density and biomass increased significantly from spring to fall (ANOVA; F(2,57) = 72.6, p < 0.0001and F(2,57) = 47.5, p < 0.0001, respectively) (Table 2). Regional comparisons of epilimnetic zooplankton biomass, density, and length by season showed a significant difference only for length; zooplankton length was higher in the east in spring (t-test; p = 0.003) (Table 2 and Table 3b). Habitat comparisons of epilimnetic zooplankton biomass, density, and length by season showed no significant differences in any season. Epilimnetic zooplankton mean length significantly decreased from 675 μm in spring to 524 μm in summer and 470 μm in fall (ANOVA F(2,57) = 15.1; p < 0.001) (Table 2). This size decrease is consistent with the composition shift from copepods (particularly cyclopoids) in spring to generally smaller cladocerans in summer and fall.
We assessed change in zooplankton community composition by dividing epilimnetic zooplankton biomass into six groups: bosminids, daphnids, invasive predatory cladocerans, other cladocerans, calanoid copepods, and cyclopoid copepods. Bosminids and daphnids were dominant in the fall and summer seasons, respectively, each accounting for 46% of the biomass at those times (Figure 7). Cyclopoid copepods were dominant (85% of the total epilimnetic biomass) in spring 2003 (Figure 7). Regional comparisons of zooplankton mean length for dominant (> 10% of the biomass) community groups for each season found a difference only in spring; cyclopoids were significantly higher in the east (t-test; p = 0.008) (Table 2 and Table 3b). Invasive predatory cladocerans (B. longimanus and C. pengoi) accounted for 7% of the total epilimnetic biomass in summer and 3% in fall. An examination of zooplankton species composition by biomass showed that Diacyclops thomasi and Limnocalanus macrurus were the dominant spring species, Daphnia retrocurva and Bosmina longirostris dominated the summer period, and Eubosmina coregoni and D. retrocurva were dominant in the fall (Table 4).
Neither Cercopagis pengoi nor Bythotrephes longimanus were observed in spring. Epilimnetic density (and biomass) of Cercopagis averaged 18.6 m−3 (0.9 mg m−3) in summer and 131.2 m−3 (1.9 mg m−3) in fall (Table 2). Cercopagis density and biomass were higher in western Lake Ontario during summer and higher in eastern Lake Ontario in fall (Table 2 and Table 3b). The species made up a large proportion (17%) of the total zooplankton biomass in western Lake Ontario in summer. Bythotrephes was only observed in fall in the Kingston Basin, with a lake-wide average density of 0.6 m−3.
Open-lake, long-term, spring TP has been steadily decreasing in Lake Ontario since 1977 (Figure 2). Offshore TP concentrations as measured by the Canadian Bioindex Program approached the target level of 10 μ g L−1 in 1985 (Millard et al., 2003) and, with the exception of 1991, remained close to the target until 1995. Since that time, spring TP levels have continued to decline, reaching a low of 4.32 μ g L−1 (USEPA; station 41) in 2001. However, EPA samples were processed by a different method and it is unknown if the continued downward trend is real or an artifact of differing methods. Surveillance data appear to be leveling off during the same time period, and comparable values between LOTT (1996) and LOLA data support this trend.
Phosphorus enrichment stimulates silica uptake by diatoms, and Schelske et al. (1986) contended that TP concentrations in the Great Lakes in the 8.0 – 25.0 μ g L−1 range were required to deplete silica. Open-lake spring TP concentrations have been below 8.0 μ g L−1 since 1996; while silica concentrations have been gradually increasing (Figure 3). Millard et al. (2003) found no significant increase in spring silica at either station 41 or 81 for the 1981 to 1995 time period, and Surveillance data for the same time period support this finding (Figure 3). However, Surveillance data plotted from 1969–2003 show an increasing trend, and addition of data from the LOLA project for stations 41 and 81 to the long-term Bioindex data shows a trend that mirrors this increase. With silica concentrations well above growth limiting levels (300 ug L−1; Schelske et al., 1986), it appears that diatoms may be experiencing P-limitation. Millard et al. (2003) questioned whether offshore spring diatom biomass could be controlled by P supply, but noted the possibility of co-limitation by both light and P supply (Healy, 1985). Spring diatom biomass was lower in 2003 (0.1 g m−3) (Munawar et al., this volume) compared to historic levels (1970, 0.9 g m−3; 1978, 0.4 g m−3; 1990, 4.2 g m−3; Munawar and Munawar, 2003) while spring Secchi depth has increased (Figure 4), indicating that P-limitation may be more important now than in the past.
Summer chlorophyll a concentrations at station 41 in Lake Ontario, 1981–1995 did not decline in tandem with spring TP (Millard et al., 2003) (Figure 5). However, the chlorophyll a decline was significant at station 81 during the same time period. In 2003, mean chlorophyll a concentrations at stations 41 and 81 were lower than any recorded during 1981–1995.
Offshore epilimnetic zooplankton species observed in 2003 did not differ greatly from those observed in 1992-1995 by Johannsson (2003). Cercopagis pengoi, Alona rectangula, and Polyphemus pediculus were the only species present in 2003 that were absent in 1992–1995. Daphnia pulicaria and Ceriodaphnia quadrangula, were present in 1992-1995 but not in 2003. Epilimnetic zooplankton density and biomass were high in the early 1980s and declined to relatively stable levels by the late 1980s which continued until 1995 (Johannsson, 2003). Summer and fall densities are typically an order of magnitude higher than spring densities. In 1987 and 1991, spring densities approached those of fall (Figure 6). In summer of 2003, zooplankton density was as low as 17,000 m−3 and biomass had decreased to 25 mg m−3. Spring and summer zooplankton densities in 2003 were the lowest recorded for the 1981 to 2003 time period (Figure 6).
Lake Ontario's offshore waters remain oligotrophic, with TP and chlorophyll a concentrations at or near record low levels in 2003. Nearshore and offshore habitats in Lake Ontario are similar with respect to nutrient levels and chlorophyll a, an indication that the lake's nearshore waters are oligotrophic as well. These findings appear to contradict recent reports describing Lake Ontario's ‘nearshore’ zone as water-quality impaired—high phosphorus levels, algal blooms, and increased abundance of the benthic filamentous alga Cladophora (Makarewicz et al., 2006). However, the similarity between nearshore and offshore habitats in our study is a reflection of our operational definition of nearshore. We classified sites as nearshore based on bottom depths of less than 30 m, but actual depths of nearshore sites in our study ranged from 10 to 28 m. Other studies examining differences between nearshore and offshore waters in Lake Ontario have used a wide range of definitions (20 m – 100 m) for the boundary between the two habitats (e.g. Munawar et al. 2003; Millard et al., 2003), but none have used depths less than 10m (e.g. Hall et al., 2003). Our study did not include the coastal zone – areas with bottom depths in the range of 1–5 m (Makarewicz et al., 2006) – a habitat that requires further study.
Regional (east vs. west) comparisons in 2003 showed spring Secchi depth higher in the east, summer TP concentrations higher in the west, and spring epilimnetic zooplankton mean length higher in the east. The deeper eastern Secchi depth in spring was not explained by other variables sampled; there were no significant differences in either TP or chlorophyll a between eastern and western regions at that time. It is possible that inputs from the Niagara River increased turbidity in the western region during spring. The higher summer TP concentration in the west may be attributed to localized phytoplankton development in the western end of the lake. Although chlorophyll a concentrations were not significantly higher in the west at that time, satellite imagery (http://oceancolor.gsfc.nasa.gov; August 18, 2003) shows higher surface chlorophyll a concentrations, particularly at northernmost sites on the western transect. These sites had higher TP concentrations than the two sites closest to the Niagara River.
Implications of a changing ecosystem
Millard et al. (1996) reported a decline of as much as 30% in primary productivity between 1972 and 1992 and commented that should that decline cascade up the food chain, that fish managers would be charged with the difficult task of preparing stakeholders to expect lower fish production. Should they persist, the record low nutrient and zooplankton levels observed in 2003 will have several food web implications. Low zooplankton biomass may translate into a decrease in populations of top predators. Changes in nutrient cycling may become more evident as well. Hecky et al. (2004) hypothesize the existence of a ‘nearshore phosphorus shunt’, a consequence of dreissenid mussel establishment whereby nutrients are sequestered in nearshore/coastal areas and along the bottom slope. If true, nutrient levels will remain very low in offshore and nearshore waters while the coastal zone will continue to experience water quality impairment and eutrophic conditions. In any case, most of the surface area of Lake Ontario is now oligotrophic and the significance of the microbial food web is critical to the functioning of this ecosystem (Mills et al., 2003). Heath et al. (2003) noted that the significance of the microbial food web in transporting carbon and phosphorus to higher trophic levels increased as waters became more oligotrophic, and that bacterial biomass could actually exceed phytoplankton biomass under such circumstances, a scenario likely reflective of the Lake Ontario ecosystem.
Lower phosphorus concentrations and high water clarity in Lake Ontario in 2003 may impact food quality of zooplankton by altering C:N:P ratios. Research has indicated that the ratio of carbon to nutrients in algal food can be an important indicator of food quality for some invertebrates (Hessen, 1992, Sterner, 1993, Sterner et al., 1993, Sterner and Hessen, 1994, Schulz and Sterner, 1999, Schulz and Sterner, 2000, Elser et al., 2000). High algal C:P ratios are indicative of poor food quality for several important zooplankton consumers, such as large daphnids, because elemental phosphorus is required for rapid growth and reproduction. Low P content in algal food has been demonstrated to reduce growth rate and productivity of daphnids and other fast-growing species (Sterner and Schulz, 1998; Sterner and Elser, 2002). In addition, high C:N or low N:P ratios can also be indicative of protein limitation for zooplankton such as copepods (Russell-Hunter, 1970; Hatcher, 1994). In 2003, spring and summer mean phytoplankton biovolumes were the lowest ever reported for Lake Ontario while the proportion of cyanobacteria, a poor quality food source, was high (Munawar et al., this volume). Summer biomass of Cryptophyta, a high quality algal food resource for zooplankton, declined to 0.05 g m−3 (Munawar et al., this volume) a level less than one-third of that reported in 1995 (Johannsson et al., 1998). These changes associated with the increased oligotrophication of Lake Ontario will disrupt the food supply to zooplankton, and the microbial food web may become a more important pathway of energy transfer to zooplankton.
Nutrients, food quality, invasive species, and predation are all factors that can affect zooplankton abundance, biomass, and community structure. Johannsson (2003) reported that the Lake Ontario zooplankton community responded to declines in TP with changes in both species composition and biomass. Chydorus sphaericus and Ceriodaphnia lacustris, species preferring nutrient-rich conditions, declined or disappeared in the early 1990s. These species were present at very low densities (< 300 m−3) in summer and fall of 2003. Summer biomass and density of both cladocerans and cyclopoids fell by approximately 50% from the early 1980s to the mid 1990s (Johannsson, 2003). By 1995, mean summer cladoceran density and biomass at station 41 were 51,000 m−3 and 81 g dry wt m−3, respectively, and spring TP concentration was 9.6 ug L−1. The spring TP concentration for the offshore of Lake Ontario in 2003 was 7.5 ug L−1, and we saw a marked decline in zooplankton abundance and biomass, suggesting decreased productivity. Lake wide mean summer epilimnetic zooplankton density and biomass in 2003 were 15,000 m−3 and 24 g dry wt m−3, respectively; these means were lower than any summer data observed at station 41 in 1995. Decreases in biomass could also have resulted from increased predation by planktivorous fish, mysids or predatory zooplankton, in particular the recent invaders, Cercopagis pengoi and Bythotrephes longimanus.
Invasive predatory cladocerans have been shown to affect Great Lakes' zooplankton populations. Pangle et al. (2007) demonstrated that Bythrotrephes longimanus decreased the productivity of Daphnia retrocurva and Bosmina spp. However, B. longimanus was absent in our samples collected in the summer of 2003. Several recent studies have also assessed the predatory impacts of Cercopagis in Lake Ontario. Warner et al. (2006) observed significant declines in abundance of bosminids, Diacyclops thomasi, and copepod nauplii during peak Cercopagis abundance in nearshore Lake Ontario. Similar declines were not observed during times when Cercopagis was absent, suggesting predation by Cercopagis was responsible for declines in small zooplankton. Benoit et al. (2002) found that invasion by Cercopagis was correlated with a decline in juvenile epilimnetic cyclopoids, and that production of juvenile copepods decreased both through direct predation and through a shift in copepod vertical distribution to colder waters. Decreases in Bosmina longirostris abundance in the presence of Cercopagis were noted as well (Warner et al., 2006; Benoit et al., 2002). Laxson et al. (2003) found that Cercopagis fed on small-bodied zooplankton (D. retrocurva and B. longirostris) in laboratory experiments, and noted a decline in the abundance of D. retrocurva, B. longirostris, and Diacyclops thomasi between 1999 and 2001 coinciding with an increase in abundance of Cercopagis. Cercopagis densities in 2003 were lower than those reported by Laxson et al. (2003) and Warner et al. (2006), and lake wide mean density reached 131.2 m−3 in fall, a time when Cercopagis was dominant in eastern Lake Ontario. The temporal scale of our study is coarser compared to those mentioned and so it is difficult to assess predatory impacts based on our data. Fall Cercopagis, cyclopoid, and bosminid densities were all higher in eastern waters, but the differences were not significant.
Managers will continue to struggle with the issue of sustaining fishery resources given the nature of changes in Lake Ontario's lower food web. Persistent low levels of total phosphorus, soluble reactive phosphorus, chlorophyll a, and epilimnetic zooplankton biomass will continue to stress alewife populations both through reduced food availability and reduced food quality for zooplankton. Alewife may be forced to seek alternative food such as mysids, a shift that may have both positive and negative energetic consequences. Because the success of efforts to maintain recreational fisheries and to restore self-sustaining populations of native species depends on the condition of the lower food web, long-term assessment is critical to measure the effectiveness of remedial actions, to better understand how stressors manifest themselves across habitats and impact fish communities, and to make recommendations for future actions.
We thank the crews of the USEPA R/V Lake Guardian and CCGS Limnos. We acknowledge the leadership of Fred Luckey and Jack Kelly of the USEPA and Vi Richardson of Environment Canada in the LOLA program. This work was funded by EPA Grant CR-83209001 to Cornell University and by a Canada Ontario Agreement (COA) Grant from Ontario Ministry of Natural Resources. This is contribution # 253 of the Cornell Biological Field Station.