The structure and function of the microbial and planktonic communities of the Bay of Quinte, Lake Ontario were studied for 8 years from 2000 to 2007. The bay has a long history of eutrophication and has undergone remediation efforts which included reductions in phosphorus loadings and the implementation of a long term research and monitoring program (1972–present) conducted by Fisheries and Oceans Canada along with other federal and provincial agencies. Microbial loop research was added in 2000 to the ongoing monitoring program which included nutrients, phytoplankton, zooplankton, benthos and fish so that a comprehensive picture of food web linkages would emerge. The structure of the microbial-planktonic food web was determined based on microscopic analysis of bacteria, autotrophic picoplankton (APP), heterotrophic nanoflagellates (HNF), ciliates, phytoplankton and zooplankton and was compared to a traditional grazing food chain consisting of phytoplankton and zooplankton. On a seasonal weighted mean basis, HNF biomass (fresh weight) of 1.7–8.4 g m−3 at Belleville and 1.1–6.3 C m−3 at Conway exceeded that of zooplankton in virtually all observations and was often greater than the combination of phytoplankton and zooplankton. Furthermore, the results showed that HNF contributed upwards to 85% of the organic carbon pool on a seasonal weighted mean basis. Various parameters in the upper bay relating trophic status to autotrophic communities were measured including: point source phosphorus loadings (>10 kg d−1); primary production (>300 g C m−2 y−1); chlorophyll a (>12 μg l−1) and phytoplankton biomass (>3 g m−3) which indicated that the upper bay remained eutrophic. This was also confirmed by the “battery of tests” strategy of ecological indicators developed in our laboratory to assess trophic status, health, and potential Beneficial Use Impairments. Based on our observations spanning 8 years, it was concluded that the microbial food web was dominated by heterotrophic communities which are still widely ignored in Great Lakes research and monitoring efforts. Our data clearly demonstrates that future studies and management strategies should include the “microbial loop” to obtain a holistic picture of the structure, function and dynamics of the lower food web.
Eutrophication has been one of the major stressors affecting the North American Great Lakes and the need to control eutrophication has driven research and management objectives since the 1970s. External phosphorus loadings, primarily from agricultural wastes and detergents, were identified as the probable cause of noxious algal blooms related to eutrophication (Vollenweider, 1971; Vallentyne, 1974; Schindler, 1978). Within the Great Lakes basin, the tightly coupled relationship between phosphorous loadings, primary production and algal standing crop (chlorophyll a) was modelled by Vollenweider et al. (1974) and resulted in a phosphorus abatement program being implemented by the governments of Canada and the United States in 1978 under the terms of the Great Lakes Water Quality Agreement (GLWQA) (International Joint Commission (IJC)), 1988.
In later revisions to the GLWQA, 42 Areas of Concern (AoC) were identified as having experienced severe environmental degradation (IJC, 1989) in order to provide a focal point for remediation efforts. These areas were identified as having at least 1 of 14 Beneficial Use Impairments (BUIs) (Hartig and Zarull, 1992). The Bay of Quinte, located on the north-eastern shores of Lake Ontario, has a long history of eutrophication and was designated an AoC because it was deemed to have 10 BUIs including “Eutrophication or undesirable algae” and “Degradation of phytoplankton and zooplankton communities” (BQ RAP, 1993). There is currently some controversy regarding the application of BUIs in the Great Lakes basin since objective, quantitative and defensible ecological parameters were never fully defined (Krantzberg, 2004; George and Boyd, 2007).
Fisheries and Oceans Canada began a long term research and monitoring effort in the Bay of Quinte in the early 1970s in response to eutrophication which included phytoplankton, zooplankton, nutrients and chlorophyll a, and these findings were essential to its designation as an AoC. Remediation efforts in the Bay of Quinte have included phosphorus load reductions, the establishment of targets for total phosphorus concentrations (30 μg l−1), phytoplankton biomass (4–5 g m−3) and chlorophyll a (12–15 μg l−1). The purpose of these targets was to provide some objective measures against which improvements in trophic state and the health of planktonic communities can be measured (BQ RAP, 1993) and, for the most part, focus on structural measurements of autotrophic organisms. Although microbial food web research was initiated in the Great Lakes in the late 1980s (Munawar and Weisse, 1989; Weisse and Munawar, 1989), the research and monitoring program of the Bay of Quinte was expanded in 2000 to include the microbial loop–bacteria, autotrophic picoplankton (APP), heterotrophic nanoflagellates (HNF) and ciliates–in order to assess the role of both autotrophs and heterotrophs on trophic state and ecosystem health.
A large body of research exists relating the structure and dynamics of autotrophic communities in the Great Lakes to trophic state (e.g. Munawar and Munawar, 1982; Munawar and Munawar, 1996; Millard et al., 1999; Fitzpatrick et al., 2007). Trophic state, however, is not simply a consequence of autotrophic processes. Recent work has shown that heterotrophic components of the microbial food web have a critical role in recycling both organic and inorganic carbon (Dodds and Cole, 2007) so a thorough examination of the microbial–planktonic food web is required for a holistic assessment of ecosystem health (Munawar et al., 2010). The microbial–planktonic food web includes bacteria, APP, phytoplankton, HNF, ciliates and zooplankton and inextricably links the primary producers with the nutrient recyclers and secondary consumers providing the energy that supports higher trophic levels (Pomeroy, 1974; Sherr et al., 1986; Sherr et al., 1988; Munawar et al., 1999). Heterotrophs, particularly bacteria, have been shown to have an increasingly important role in energy flux in oligotrophic versus eutrophic systems (del Giorgio et al., 1997; Cole et al., 2000; Biddanda et al., 2001). However, the impacts of cultural eutrophication (i.e. inputs of phosphorus and allochthonous matter, frequent algal blooms) have been shown to enhance heterotrophic microbial processes (Pace, 1993; Porter, 1996).
The aims of this paper are twofold. The first is to characterize the composition of the organic carbon pool of the Bay of Quinte into the major autotrophic and heterotrophic assemblages in order to uncover potential energy pathways and understand the linkages between organisms. The second is to comment on the health of the bay with respect to the remediation targets and beneficial use impairments. The implications for energy transfer from lower to higher trophic levels in the Bay of Quinte are discussed within the broader context of ecosystem health and potential impairments.
Materials and Methods
The Bay of Quinte is a large z–shaped embayment (254 km2) that is also relatively shallow (4–8 m), but is much deeper at the interface with Lake Ontario. One shallow station in the upper bay (Belleville–5 m) and one deep station in the lower bay (Conway–30m) (Figure 1) were sampled bi-weekly from May 1st to October 31st for each of the years from 2000 to 2007. Integrated epilimnetic water samples were collected from each site and subsamples were drawn for total phosphorus, chlorophyll a, microbial loop, ciliates and phytoplankton. Zooplankton was sampled from discrete depths. The Area of Concern is represented by the upper bay site while the lower bay site serves as a control.
Total phosphorus (TP) concentrations were determined with the automated ascorbic acid technique after persulfate digestion (Strickland and Parsons, 1968). Point source phosphorous loadings were compiled by P. Kintsler, Ontario Ministry of the Environment, Kingston, Ontario (pers. com.).
Microbial loop samples, including bacteria, autotrophic picoplankton and heterotrophic nanoflagellates, were fixed with 1.6% formaldehyde and enumerated using DAPI staining (Porter and Feig, 1980) under epi-fluorescence microscopy (Munawar and Weisse, 1989). Organic carbon was estimated as 200 fg C cell−1 for autotrophic picoplankton (APP), 10 fg C cell−1 for bacteria and 14 fg C cell−1 for heterotrophic nanoflagellates (HNF) (Sprules et al., 1999). Ciliate samples were preserved with Lugol's iodine (Lynn and Munawar, 1999) and enumerated following the Quantitative Protargol Staining technique (Montagnes and Lynn, 1993). Ciliate organic carbon was calculated as 11% of the fresh weight biomass (Turley et al., 1986).
Chlorophyll a concentrations were determined by filtering up to 1 L of water through Whatman GF/C filters followed by cold acetone pigment extraction and spectrophotometric analysis (Strickland and Parsons, 1968). Phytoplankton samples were fixed immediately with Lugol's iodine. Identification and enumeration followed the Utermöhl (1958) inverted microscope technique as described in Nicholls et al. (2002) which are broadly compatible with the established techniques of Munawar et al. (1987). Organic carbon was estimated for the various groups of phytoplankton using equations from the literature namely: Diatomeae (Strathmann, 1967), Chlorophyta, Chrysophyceae, Cryptophyceae and Dinophyceae (Verity et al., 1992), and Cyanophyta (Lee and Furman, 1987).
Zooplankton samples were collected using a 41 litre Schindler–Patalas trap fitted with a 64 μm mesh. Three depths were sampled at Belleville (1 m, 2 m, 3 m) and eight depths were sampled at Conway (1 m, 3 m, 5 m, 8 m, 10 m, 20 m and 25 m). Samples were preserved immediately with 4% sugared formalin. A single composite sample was constructed for each station-date by combining 50% of the sample from each depth. Biomass of each taxon was determined in each sample from length–weight regressions (Bowen and Johannsson, 2011). Organic carbon was estimated as 48% of the dry weight (Anderson and Hesson, 1991).
Annual primary production was estimated by 14Carbon uptake. Samples from each cruise were spiked with Na14CO3 and incubated for 2–4 hours at irradiance levels ranging from 10–120 μE s−1 m−2. Following incubation, acidification and bubbling was used to remove excess 14C and radioactivity was determined by liquid scintillation counting (Millard et al., 1999). Seasonal areal primary production was then estimated according to the model of Fee (1990). These estimates were increased by 10% in order to account for winter primary production (Fitzpatrick et al., 2007; Vollenweider et al., 1974).
Upper Bay (Belleville)
The biomass of a traditional planktonic food web (phytoplankton and zooplankton) and an integrated microbial-planktonic food web (phytoplankton, zooplankton and microbial loop) for Belleville is shown in Figure 2 for 2000–2007. The seasonal weighted mean fresh weight biomass was 2.4–5.5 g m−3 for phytoplankton and 0.5–3.2 g m−3 for zooplankton (Figure 2a). Phytoplankton contributed 57.5–85.7% of the biomass, on average, of a traditional planktonic food web. In comparison, an integrated microbial planktonic food web is shown in Figure 2b. The microbial loop contributes an additional 2.2–8.7 g m−3 of biomass and accounts for 22.6–63.5% of the combined biomass compared to 26.1–55.5% for phytoplankton and 8.1–28.6% for zooplankton. The largest single component of the microbial loop was heterotrophic nanoflagellates which ranged from 1.7–8.4 g m−3 followed by bacteria (0.1–0.5 g m−3), autotrophic picoplankton (0.1–0.3 g m−3) and ciliates (0.01–0.2 g m−3).
The size of the organic pool at Belleville ranged from 546.4–1837.8 mg C m−3 and there was considerable inter annual variability in the relative proportions of autotrophs (APP, phytoplankton) and heterotrophs (bacteria, HNF, ciliates, zooplankton) as shown in Figure 3. During 2000 and 2001, phytoplankton contributed 50% of the organic carbon of the combined microbial and planktonic food web. In 2002, HNF contributed slightly more organic carbon (48%) than phytoplankton (43%), and HNF dominated the organic carbon pool (60%) in 2003 and 2004. However, during 2005–2007, phytoplankton contributed upwards to 75% of the organic carbon.
Annual estimates of primary production, (mean) chlorophyll a, (mean) point source phosphorus loadings and (mean) total phosphorus concentrations are given in Table 1 for Belleville from 2000–2007. Primary production ranged from ≈300–425 g C m−2, chlorophyll a from 12–26 μg l−1, phosphorus loadings from 8.6–22.3 kg d−1 and total phosphorus from 28–42 μg l−1.
Lower Bay (Conway)
At Conway, the combined phytoplankton and zooplankton biomass of the traditional planktonic food web ranged from 0.8–1.3 g m−3 (on a seasonal weighted mean basis) with phytoplankton accounting for 50–81% of the total (Figure 4a). The combined microbial-planktonic food web for Conway is depicted in Figure 4b. The microbial loop contributes an additional 1.3–6.7 g m−3 of biomass at Conway which accounts for 56–86% of the combined food web compared to 7–36% for phytoplankton and 7–17% for zooplankton. Similar to Belleville, HNF were the largest component of the microbial loop (1.1–6.3 g m−3), followed by bacteria (0.09–0.4 g m−3), APP (0.04–0.1 g m−3) and ciliates (0.02–0.06 g m−3).
The size of the organic carbon pool at Conway ranged from 303.9–900.4 mg C m−3 and HNF were typically the largest component accounting for 39–77% of the organic carbon followed by phytoplankton ranging from 14–47% (Figure 5). Bacteria, APP and zooplankton accounted for less than 5% of the organic carbon and ciliates were typically <1%.
Traditionally, studies of the lower trophic levels of aquatic ecosystems have been restricted to analyses of the phytoplankton and zooplankton communities and this was the case at the inception of Project Quinte in 1972. However, microbial loop assessments were added beginning in 2000 and continue to the present. In our analysis from 2000–2007 (Figure 2a), we observed that phytoplankton biomass at Belleville was generally greater than 3 g m−3 (2.4–6.0) and typical of eutrophic conditions (Munawar and Munawar, 1982). Phytoplankton biomass was 2–5× greater than zooplankton biomass which is not unexpected in eutrophic environments since it reflects a lack of top down control (e.g. Bays and Crisman, 1983; Pace, 1986; Elser and Goldman, 1991). Phytoplankton biomass at Conway fell between 0.5–1.0 g m−3 (Figure 4a) on an annual mean basis which is typical of oligotrophic waters. Zooplankton biomass was about half of the phytoplankton biomass, although there was some variability. In 2000 for example, zooplankton biomass was only about 25% of the phytoplankton biomass, whereas in 2002 phytoplankton and zooplankton biomass were not significantly different and in 2003 zooplankton biomass exceeded phytoplankton biomass. On the whole, the relationship between phytoplankton and zooplankton biomass appears to be more closely related at Conway than at Belleville and broadly indicative of oligotrophic conditions.
However, our results demonstrate that the traditional food web analyses based mainly on phytoplankton and zooplankton, give only a partial picture of the community and do not account for the significant amount of biomass which the microbial loop (bacteria, APP, HNF, ciliates) may contribute to the pelagic food web. Recent and historic work conducted in Lake Ontario by our lab and by others have shown that heterotrophic protists can contribute significantly to plankton abundance and biomass (Munawar and Nauwerck, 1971; Pick and Caron, 1987; Munawar et al., 2010). At Belleville, microbial biomass was equal to zooplankton biomass at its lowest point and upwards to 7× greater than zooplankton biomass (Figure 2b). In fact microbial biomass was equal to or exceeded the combined phytoplankton and zooplankton biomass in 3 of the years studied (2002, 2003, 2004). HNF dominate the microbial loop and contributed 17.4–60.9% of the total microbial and planktonic biomass.
Similarly, at Conway the microbial biomass ranged from 1.3–6.7 g m−3 during the study period and was greater than the combined phytoplankton and zooplankton biomass in all of years studied (Figure 4b). The largest component of the microbial loop and by extension the microbial and planktonic food web was HNF which contributed 45.9–79.5% of the combined microbial and planktonic biomass. The high biomass of HNF and other microbes raises questions as to the nature of the organic carbon pool that supports the food web and whether or not energy is being diverted away from a traditional grazing food chain.
Organic carbon has long been considered to be a sound currency of energy exchange (Gosselain et al., 2000) so the structure of the organic carbon pool is an important consideration when determining ecosystem health because it provides an estimate of the amount of energy in the system available for transfer to higher trophic levels. Over the course of 8 years, the organic carbon pool at Belleville and the potential energy that it provides was evenly split between autotrophs and heterotrophs during the first 3 years; and then heterotrophs dominated for the next two followed by an overwhelming dominance by autotrophs once again (Figure 3). The largest contributors to the organic carbon pool in all of the years studied were HNF and phytoplankton. These two factors appear to be controlling the shift of the organic carbon pool from heterotrophic to autotrophic. Such dramatic swings in the structure of the food web were not observed between autotrophs and heterotrophs at Conway (Figure 5). However heterotrophs did supply 50–85% of the organic carbon (predominantly HNF). These observations are similar to our earlier studies in the offshore oligotrophic waters of Lake Ontario (Munawar et al., 2010). It is apparent that the potential dominance of HNF is not limited by the trophic state of the ecosystem.
The observed changes in the structure of the microbial–planktonic food web at Belleville, from being primarily heterotrophic to being primarily autotrophic, appear to be due to late summer blooms of filamentous diatoms and blue green algae such as Melosira spp., Anabaena spp. and Lyngbya spp. (Nicholls and Carney, 2011) which results in an overall increase in the supply of organic carbon from autotrophs. Importantly, there is no corresponding decline in the amount of organic carbon supplied by heterotrophs so algal blooms appear to increase the size of the organic carbon pool. However, the appearance of late summer algal blooms should not be interpreted as an increase in autochthonous production. This is supported by estimates of annual primary production in the upper bay which ranged from 300–425 g C m−2 y−1 (Table 1). Despite the inter-annual variability in primary production, such changes are unlikely to be statistically significant (Porta et al., 2005; Fitzpatrick et al., 2007) indicating that autochthonous production is stable over longer time frames.
The question of whether the microbial loop is a source or sink for organic carbon has been debated in the literature for some time (Wylie and Currie, 1991; Porter, 1996). Zooplankton, particularly Daphnia, play an important role in regulating HNF biomass (Pace, 1993; Porter, 1996). While Daphnia are the most prevalent zooplankter in the Bay of Quinte (Bowen and Johannsson, 2011), the amount of zooplankton at both Belleville and Conway stations is less than expected given its trophic state (Pace, 1986). Research conducted in coastal areas of the Great Lakes has shown that both HNF and Daphnia are important bacterivores (Hwang and Heath, 1997; Hwang and Heath, 1999), so the high proportion of HNF to zooplankton would suggest that HNF are out-competing zooplankton for food resources in the Bay of Quinte. This hypothesis assumes that bacteria are the sole food resource and there is sufficient bacterial production to support such a large standing crop of HNF. However, results from stable isotope (13C) experiments suggest that zooplankton preferentially graze autochthonous matter, particularly in eutrophic environments (Pace et al., 2004; Brett et al., 2009). As we noted in a previous work, very little is known about the dynamics of HNF in the Great Lakes and production and grazing experiments are badly needed (Munawar et al., 2010). With respect to food web structure in the Bay of Quinte, it appears that autochthonous production can be diverted away from a traditional grazing food chain towards a microbial food web and support a large biomass of heterotrophs sequestering the autochthonous energy (Munawar et al., 2010).
In the first part of our discussion on microbial–planktonic webs, we treated the Bay of Quinte as a geographic entity (i.e. a shallow embayment of Lake Ontario) which includes both eutrophic and oligotrophic conditions. In the second part of the discussion, we consider the Bay of Quinte as a management entity defined by its status as an Area of Concern. The AoC includes the eutrophic waters of the bay which are represented in this study by observations at the Belleville site. As noted, the Bay of Quinte has a long history of eutrophication and Project Quinte was developed as a multi-agency effort to control eutrophication under the terms of the GLWQA. Phosphorus load reductions were implemented beginning in 1978 and point source loads were immediately reduced from more than 200 kg d−1 to less than 70 kg d−1 (Minns et al., 1986). As a consequence, from 1977 to 1978, total phosphorus concentrations declined from 70–48 μg l−1, chlorophyll a declined from 40 to 21 μg l−1 and phytoplankton biomass from 12.5 to 4.3 g m−3 (K.H. Nicholls, Ontario Ministry of the Environment, Rexdale, ON, submitted; Nicholls and Carney, 2011). Despite this success, “Eutrophication or undesirable algae” and “Degradation of phytoplankton and zooplankton communities” in the upper bay (i.e. Belleville) were identified as 2 of the 10 Beneficial Use Impairments when the Bay of Quinte was designated as an AoC (BQ RAP, 1993) based on the data gathered from Project Quinte.
With respect to alleviating these Beneficial Use Impairments (BUIs), phosphorus abatement was continued and remediation targets for total phosphorus (30–35 μg l−1), chlorophyll a (12–15 μg l−1) and phytoplankton biomass (4–5 g m−3) were established to help measure the success of management efforts. A cap of 15 kg d−1 was also recommended for phosphorous loadings (Minns and Moore, 2004). All of these parameters were at or near target levels from 2000–2007 (Table 1), but it remains unclear whether or not these targets demonstrate enough of an improvement in the trophic status and health to conclude that the stated BUIs no longer apply. According to George and Boyd (2007), the concept of BUIs is vague and poorly defined from a policy perspective. Since precise criteria for the 42 AoCs in the Great Lakes basin were never established, it is difficult to evaluate the status of individual impairments (ibid). However, considerable progress has been made from an ecological perspective in developing integrated, sensitive and rapid indicators of ecosystem health. Consequently, a variety of multi-trophic tests and models are available which can be deployed to assess the status and recovery of BUIs.
In order to establish a scientifically sound basis for BUIs and assess progress towards recovery, our lab adopted a “battery of tests” strategy using multi-trophic indicators (Table 2). One such indicator is phytoplankton biomass and its species composition. The phytoplankton biomass at Bellville exceeded an annual mean concentration of 3 g m−3 establishing the upper bay as eutrophic. Species composition requires a detailed treatment of its own and hence is not included here due to lack of space. Another indicator chosen was Vollenweider's eutrophication model which classifies trophic state as a function of phosphorous loadings, primary production and chlorophyll a as a proxy for algal standing crop (Vollenweider et al, 1974; Munawar et al., 2010B). Table 1 shows the values of these parameters observed during our study (2000–2007) which is summarized as follows:
Point source phosphorus loadings: >10 kg d−1
Primary production: >300 g C m−2 y−1
Chlorophyll a concentration: >12 μg l−1
Vollenweider's model defines eutrophic conditions as those where algal standing crop (chlorophyll a) exceeds 9.0 μg l−1 and annual primary production exceeds 295.0 g C m−2 y−1, which confirms that the upper bay is eutrophic as indicated by the above observations. Furthermore, the model predicts that existing remediation targets will result in the continuation of eutrophic conditions. But according to the model, the predicted loadings of phosphorous required to sustain this level of primary production is greater than our current estimates of point source loadings and suggest that non-point sources of phosphorous have a critical role in determining the Bay of Quinte's overall trophic state. These non point sources would include: tributary loadings (estimated by Minns et al., 2004 to be >600 kg d−1), storm sewer overflows and recycling by dreissenid mussels. Furthermore, Vollenweider's model was developed for contained lake systems and does not fully incorporate the effect of flow in riverine systems. However, the model does appear to be applicable to the Bay of Quinte although it is a flow through system.
Summary and Conclusions
In conclusion, our long term study of 8 years (2000–2007) indicates that the upper Bay of Quinte (Belleville) continues to be eutrophic, while the lower bay (Conway) is oligotrophic. Given that the lower bay was intended as a control site, results from the upper bay show the Beneficial Use Impairments of “Eutrophication or undesirable algae” and “Degradation of phytoplankton and zooplankton communities” still persist. These findings are based on a “battery of tests” of widely accepted indicators and models of trophic status which rely on measures of the autotrophic community.
Our study also shows that the microbial loop components contribute significantly to the organic carbon pool of the Bay of Quinte and that heterotrophic nanoflagellates (HNF) can contribute more organic carbon than phytoplankton and zooplankton combined. However, we observed significant shifts in the structure of the food web from being primarily autotrophic to being primarily heterotrophic due to the presence or absence of algal blooms. Science has developed considerably since the 1970s and 1980s and the new frontiers of knowledge include sensitive techniques and models which need to be included in ecosystem health assessments. Consequently, future research, monitoring and management strategies, including refined criteria for Beneficial Use Impairments, should not ignore the important role of the microbial loop for a holistic evaluation of the structure, function and dynamics of the lower trophic levels.
We thank Drs. C.K. Minns, J.H. Leach and M.A. Koops for their constructive reviews of the manuscript. Thanks are also due to S. Blunt and L. Elder (AEHMS) for their technical editing. Finally, we gratefully acknowledge the hard work and perseverance of the DFO staff whose dedication to Project Quinte have made this research possible, including: E.S. Millard, M. Burley, J. Gerlofsma, C. Timmins, R. Bonnell and A. Bedford.