Ecological Applications, 14(1), 2004, pp. 113–131 q 2004 by the Ecological Society of America
ESTABLISHMENT OF A METALIMNETIC OXYGEN REFUGE FOR ZOOPLANKTON IN A PRODUCTIVE LAKE ONTARIO EMBAYMENT ROBERT A. KLUMB,1,3 KIETHA L. BUNCH,1 EDWARD L. MILLS,1 LARS G. RUDSTAM,1 GARY BROWN,2 CHARLES KNAUF,2 RICHARD BURTON,2 AND FREDRIK ARRHENIUS1,4 1Department
of Natural Resources, Cornell Biological Field Station, 900 Shackelton Point Road, Bridgeport, New York 13030 USA 2Monroe County Health Department, Environmental Health Laboratory, 740 East Henrietta Road, Rochester, New York 14623 USA
Abstract. Hypolimnetic oxygen injection is a management tool used to improve water quality by preventing anoxia and associated phosphorus release from the sediments. An additional benefit would be formation of a low-oxygen refuge for large-bodied herbivorous zooplankton in the metalimnion. The magnitude and timing of hypolimnetic oxygenation was deliberately manipulated during summer 1997 in Irondequoit Bay, a eutrophic Lake Ontario embayment (New York, USA), to maintain metalimnetic dissolved oxygen (DO) concentrations in the range of 1–2 mg/L. We assessed the vertical distributions of fish and zooplankton in 1997, a year with controlled oxygenation, and 1996, a year with oxygenation but without a deliberate attempt to create amenable refuge conditions. After initiation of oxygenation, June through early August metalimnetic DO in 1997 was 1.8 6 0.5 mg/L (means 6 2 SE), whereas metalimnetic DO in 1996 was 3.8 6 1.0 mg/L. Hydroacoustic surveys in 1997 and gillnet catches in both years indicated that .96% of planktivorous fish, alewives (Alosa pseudoharengus) and emerald shiners (Notropis atherinoides), were restricted to depths with DO . 4.0 mg/L. A refuge effect in 1996 was slight and evidenced only by significantly larger zooplankton in the metalimnion while densities of Daphnia galeata medotae and D. retrocurva were similar or significantly greater in the epilimnion. In 1997, the refuge effect was strong despite an almost twofold increase in planktivorous fish abundance compared to 1996. Overall zooplankton densities were similar in the epilimnion and metalimnion while zooplankton size and densities of both daphnids were significantly greater in the metalimnion. Densities of an invertebrate predator, Mesocyclops edax, were also significantly higher in the metalimnion compared to the epilimnion. Despite high fish abundance in 1997, mean summer (late May through August) chlorophyll a decreased 29% compared to 1996, consistent with expectations from grazing by large Daphnia protected in the refuge. Presence of large-bodied zooplankton in 1997 despite high planktivore abundance demonstrated that successful creation of a low-oxygen refuge from fish predation was possible in an open Great Lakes embayment where traditional biomanipulation via fish removal was not practical. Key words: alewife; Alosa pseudoharengus; biomanipulation; Daphnia galeata mendotae; Daphnia retrocurva; dissolved oxygen; embayments; Lake Ontario; Mesocyclops edax; predation; refuge; zooplankton.
INTRODUCTION Vertical oxygen structure provides a framework shaping the composition and distribution of freshwater fish and zooplankton communities (Fast 1971, Herberger and Reynolds 1977). Selective feeding by many fish can eliminate large-bodied zooplankton like Daphnia spp. (see Plate 1). (Brooks and Dodson 1965). In thermally stratified and productive lakes, high biological oxygen demand (BOD) leads to low oxygen conManuscript received 5 February 2002; revised 1 March 2003; accepted 15 April 2003. Corresponding Editor: J. M. Melack. 3 Present address: U.S. Fish and Wildlife Service, Great Plains Fish and Wildlife Management Assistance Office, 420 South Garfield Avenue, Suite #400, Pierre, South Dakota 57501 USA. E-mail: robert
[email protected] 4 Present address: Institute of Marine Research, National Board of Fisheries, P.O. Box 4, SE-453 21 Lysekil, Sweden.
centrations or anoxia below the metalimnion. Because most fish are not tolerant of low oxygen (Doudoroff and Shumway 1970, Davis 1975), large zooplankton can escape predation by inhabiting and seeking refuge in low-oxygen areas (Sih 1987, Hanazato et al. 1989, Wright and Shapiro 1990). Daphnia spp. are relatively large-bodied herbivorous crustaceans in lentic ecosystems and a preferred food source for many fishes. If a low-oxygen refuge becomes available in thermally stratified lakes, Daphnia frequently concentrate within low-oxygen depths during the day to evade fish predators (Lampert 1987, Stirling et al. 1990, Luecke and Teuscher 1994). Daphnia have been experimentally induced to migrate to less favorable oxygen conditions in response to increased concentrations of fish kairomones (Lass et al. 2000). Other studies have documented use by Daphnia of naturally
113
114
Ecological Applications Vol. 14, No. 1
ROBERT A. KLUMB ET AL.
PLATE 1. (Left) Photograph of the herbivorous zooplankter Daphnia galeata mendotae. Photo credit: Cornell Biological Field Station. (Right) Aerial view of Irondequoit Bay, New York, USA, an urbanized Great Lakes embayment attached to Lake Ontario (lower right corner). The bay is located within the metropolitan area of Rochester, New York (upper right). Photo credit: Rochester Institute of Technology.
formed low-oxygen refuges in productive lakes prone to prodigious algal growth and summer oxygen depletion (Hanazato et al. 1989, Tessier and Welser 1991). Because planktivorous fishes can eliminate Daphnia from lakes (Wells 1970, O’Gorman et al. 1991), a lowoxygen refuge should improve survival of these herbivores, potentially improving water clarity from increased grazing on phytoplankton (Carpenter et al. 1985). Oxygenation of subsurface waters has reproduced low-oxygen refuge-like conditions in field experiments (Field and Prepas 1997, Gemza 1997) but few researchers have specifically attempted to create and maintain a low-oxygen refuge for zooplankton to escape predation in field studies. The objective of this study was to assess whether a metalimnetic refuge for zooplankton could be developed through hypolimnetic oxygen injection in a productive, urban embayment. The goal to maintain oxygen concentrations between 1 and 2 mg/L in the metalimnion was based on avoidance and tolerance levels established from laboratory experiments and field observations on fish (Doudoroff and Shumway 1970, Davis 1975, Rudstam and Magnuson 1985) and cladocerans (Herbert 1954, Nebeker et al. 1992). Our definition of a ‘‘low-oxygen’’ zooplankton refuge is within the range of previously cited values (Wright and Shapiro 1990). For example, the refuge in Hart Lake, British Columbia, Canada, was maintained near 2.0 mg/L (Gemza 1997). Specifically, we hypothesized that successful refuge creation would be evidenced by higher densities of large-bodied zooplankton (e.g., Daphnia spp.) and larger mean sizes of zooplankton in the low-
oxygen metalimnion in contrast to the well-oxygenated epilimnion. METHODS
Study site Irondequoit Bay (see Plate 1) is a 679-ha productive embayment located within an urbanized watershed on the south shore of Lake Ontario in New York, USA (Fig. 1) and is composed of two longitudinally aligned basins (Bannister and Bubeck 1978). The north basin has a maximum depth of 24 m while the south basin has a maximum depth of 12 m. The far northern and southern areas of the bay are shallow and heavily vegetated with macrophytes in the summer. Wetlands in the south connect the bay with its largest tributary, Irondequoit Creek. A shallow, narrow channel permits boat traffic and fish movements to and from Lake Ontario. The main planktivorous fish in Irondequoit Bay are alewives (Alosa pseudoharengus), emerald shiners (Notropis atherinoides), and young-of-the-year (YOY) gizzard shad (Dorosoma cepedianum). As a result of past management actions, Irondequoit Bay’s trophic status has shifted from a highly eutrophic to a near-mesotrophic state (Bannister and Bubeck 1978, Wetzel 1983). An approximately 10 m thick layer of organic detritus exists on the bottom of Irondequoit Bay (Monroe County Health Department [MCHD], unpublished data). Therefore, an aluminum sulfate seal was applied in 1986 to immobilize phosphorus within the sediments, inhibit internal phosphorus cycling, and reduce nuisance algal blooms (Welch and Cooke 1999). However, continued oxygen depletion in summer led
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
115
to the adoption of hypolimnetic oxygen injection as a method to maintain the aluminum sulfate seal and further improve water quality. A gravity-driven oxygen release system was installed whereby pure oxygen passes from an aboveground pressurized storage tank through bifurcated pipe to five diffuser grids suspended approximately 2 m above the lake bottom (Fig. 1). Starting in 1993, .100 000 kg (range: 109 000– 176 000 kg) of pure oxygen were injected at a rate of 1500–1900 kg/d into the hypolimnion each summer. Before oxygen supplementation in 1993, the summer (May–October) oxygen concentration from 1986 to 1992 in the metalimnion of Irondequoit Bay was 0.2 6 0.1 mg/L (means 6 2 SE; MCHD, unpublished data). Oxygen was injected into the north and south basins in 1996 without the goal of maintaining a metalimnetic zooplankton refuge; the targeted DO concentration in the metalimnion was 4 mg/L. In 1997, oxygen injection occurred only in the north basin. Unlike 1996, oxygen diffusion was regulated in 1997 to maintain dissolved oxygen (DO) concentrations in the metalimnion within 1.0–2.0 mg/L. Daily amounts of oxygen injected into the north and south basins from 10 June to 14 September 1996 averaged 1600 kg/d (total 158 000 kg). In 1997, 1700 kg/d (total 136 000 kg) of oxygen was injected from 13 July to 29 September through the three diffuser grids in the north basin.
Water chemistry Temperature and dissolved oxygen (DO) profiles were determined during the day (0700–1400) in the north basin (Fig. 1) from late May to August in 1996 (N 5 26) and from late May to mid-October in 1997 (N 5 19). Water clarity, as determined from Secchi disk depth, was recorded to the nearest 0.5 m. Vertical profiles of temperature (in degrees Celsius) and DO (in milligrams per liter) were measured with a Hydrolab Surveyor II model SVR2-SU (Hydrolab, Austin, Texas, USA) at 0.5- or 1.0-m intervals from surface to the lake bottom. Temperature probes were calibrated weekly with a mercury thermometer and DO probes were calibrated with lake water samples using the azide modification of the Winkler titration (APHA 1982). The location of the metalimnion was determined based on temperature changes .0.78C/m. To assess potential grazing effects by zooplankton, we assessed seasonal changes in algae concentrations indexed by chlorophyll a. Integrated water samples from the epilimnion, metalimnion, and hypolimnion were collected in the north basin in 1996 (N 5 19) and 1997 (N 5 20) with a nalgene tube (1.9 cm inner diameter), fixed with MgCO3, and vacuum-filtered within 4 h of collection through a Whatman 934-AH glass fiber filter that retains particles .1.5 mm. Chlorophyll a was determined after acetone extraction with a spectrophotometer (Strickland and Parsons 1972).
FIG. 1. Bathymetric map of Irondequoit Bay, New York, USA, illustrating the locations of five oxygen diffusers, the sampling site for water chemistry and zooplankton in the north basin, and the gillnet sampling location in the south basin studied in 1996 and 1997.
Fish Community composition and depth distribution of pelagic fishes were sampled every two weeks with monofilament vertical gillnets. In 1996, fish were collected on five dates from 30 May to 29 July. Fishes were collected on 10 dates from 28 May to 15 September during 1997. Seven vertical gillnet panels that consisted of an individual bar mesh of 6.25, 8.0, 10.0, 12.5, 15.0, 18.75, or 33.0 mm were set at a depth of 10–12 m. Gillnet panel size was 3 3 12 m, and the nets fished the entire water column. Nets were set for 1–2 h approximately 1 h before sunset (1730 to 2130) in the northeast corner of the south basin (Fig. 1). These mesh sizes were used because fish of 50–200 mm total length (TL) are captured with approximately equal efficiency (Warner et al. 2002). Therefore, catch-per-uniteffort (CPUE) was reported as the total catch in all mesh sizes per hour for each 1-m depth increment.
116
Ecological Applications Vol. 14, No. 1
ROBERT A. KLUMB ET AL.
On 10 dates following gillnet sampling in 1997, we conducted nighttime (2200–0300) hydroacoustic surveys with a 70-kHz split beam Simrad EY500 echo sounder (half-power beam width 11.3 8, pulse length 0.2 or 0.6 m; Simrad, Horten, Norway). The transducer was suspended 0.5 m under the water surface and towed at 1.5–2.0 m/s along a transect traversing the longitudinal axis of the bay. Acoustic data were collected for all surveys with a minimum target strength (TS) threshold of 270 dB and a minimum threshold for volume back-scattering of 280 dB. The hydroacoustics unit was calibrated every two weeks with a standard copper sphere of known TS (239.1 dB). Mean densities of pelagic fish were calculated using echo integration with Simrad EP500 software version 5.2 (Simrad 1996). Only targets larger than 261 dB were included in the analysis because these targets represented fish larger than approximately 20 mm TL (Warner et al. 2002). Reported densities of targets larger than 246 dB represent juvenile and adult alewives greater than 90 mm TL (Warner et al. 2002). Hydroacoustic estimation of fish density can be confounded by bubbles (Rudstam and Johnson 1992) formed as byproducts during methanogensis, denitrification, or sulfate reduction in anoxic sediments (Wetzel 1983). Distinctively ‘‘non-fish’’ targets appeared on the echograms in the hypolimnion beginning on 2 July 1997. Bubbles appeared as inverted teardrops, in contrast to the half-crescent moons of individual fish targets typically seen in hydroacoustic echograms. These bubbles were not present before 2 July. Therefore, we collected stationary acoustic data on the night of each subsequent fish survey for 15–20 min at two sites in the north basin to evaluate release period, terminal height of ascent, and reflective properties (i.e., target strengths) of rising bubbles in the water column. Also on 15 September 1997, seven gillnet panels of 3 m length and 22 m depth were set in the north basin. These 22-m gillnets fished the entire water column in the north basin and were of the same bar mesh sizes as the 3 3 12 m nets previously described. No fish were caught below 11 m, a depth corresponding to ,3.0 mg/L DO. Target strengths of bubbles and fish overlap, which required that the contribution of bubbles be removed from the overall TS distribution. Bubble traces were observed to rise to shallow depths on the echograms. However, of all the traces identified during the stationary surveys from 16 July through 15 September in water with DO , 2.0 mg/L, only 1.4% (eight traces) could have been fish, and all but one of these traces were from small targets (259 to 261 dB). Therefore, we felt it was reasonable to assume that all acoustic targets in water with DO , 2.0 mg/L were bubbles. With this assumption, the contribution of bubbles to the overall TS distribution (i.e., fish abundance) was removed by subtracting the density of targets found in
water with DO , 2.0 mg/L in proportion to their TS from fish densities at other depths.
Zooplankton Zooplankton were collected at approximately weekly intervals from the epilimnion, metalimnion, and hypolimnion in the north basin (Fig. 1). In 1996, zooplankton were collected on 14 dates from 5 June to 5 August while samples were collected on 19 dates from 29 May to 1 October 1997. In 1996, vertical integrated samples of zooplankton were collected in the epilimnion, metalimnion, and hypolimnion with a 1-m diameter, 153-mm mesh closing net. In 1997, integrated samples were collected from the surface to 4 m and at discrete depths from 5 to 13 m (not all depths sampled on all dates) and at 17 and 22 m using a peristaltic pump. Fifty liters of water were filtered (flow rate 5 4.5–5.0 L/min) at individual depths and zooplankton were concentrated by sieving through a 153-mm mesh. Zooplankton were anesthetized with antacid tablets prior to preservation in 4% sugar formalin (in 1996) or 90% ethanol (in 1997); samples were visually inspected (in 1997) for red coloration (i.e., indication of hemoglobin synthesis in daphnids). To assess variation of zooplankton density from samples collected with the peristaltic pump, replicate samples were collected at random depths throughout the season in 1997. Counts of dominant cladoceran and copepod species varied no more than 18%. No mechanical damage to zooplankton from the peristaltic pump was observed; intact organisms were retrieved from 20 m to the surface. The crustacean zooplankton assemblage and size structure were determined from a minimum of two 1mL subsamples. Adult zooplankton were identified to species when possible (Torke 1974, Balcer et al. 1984) using a compound microscope (magnification 403). Neonnates, nauplii, and immature copepodites were not identified to species. Densities of each species captured were calculated based on total volume of water filtered (net efficiency assumed 100%) and pooled in the following taxons: bosminids (Bosmina and Eubosmina spp.), daphnids (Daphnia spp.), copepods (calanoid and cyclopoid), other cladocerans (Ceriodaphnia, Chydorus, and Diaphanosoma spp.), and nauplii. Reported zooplankton densities (number of individuals per liter) are means calculated from counts for individual depths within the epilimnion, metalimnion, and hypolimnion. Seasonal metalimnetic densities of two important largebodied herbivores (Daphnia galeata medotae, D. retrocurva) and one predacious copepod (Mesocyclops edax) are reported separately. M. edax was investigated to assess potential use of the refuge by an invertebrate predator, which may reduce refuge effectiveness for protection of Daphnia (Horppila et al. 2000). Zooplankton body length was measured to the nearest 0.01 mm using a light microscope and an ocular micrometer at 4, 10, or 203 magnification. Body length
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
was measured as the distance from the anterior region of the helmet or head to the base of either the tail spine of cladocerans or caudal rami of copepods. At least 100 zooplankton from a single or multiple 1-mL subsamples were measured; if densities were exceptionally low the entire sample was counted and measured.
Statistical analyses Effectiveness of the DO refuge during 1996 and 1997 was assessed by analyzing seasonal mean zooplankton density and size in the epilimnion, metalimnion, and hypolimnion with ANOVA models (Sokal and Rohlf 1981). Repeated-measures ANOVA was used with sampling dates considered as replicates for each depth stratum; however, multiple samples for a given week were averaged to align data in the figures. Because zooplankton sampling methods differed between years our statistical analyses of zooplankton densities and sizes were restricted to within-year comparisons among the three depth strata. Tukey-Kramer multiple comparison tests were used to determine if significant differences existed between depth strata by calculating 95% comparison limits around the mean difference of density or size (Sokal and Rohlf 1981). To stabilize variances, all zooplankton density data were log10(x 1 1) transformed. Mean zooplankton size was analyzed without transformation. All statistical analyses were performed with NCSS 2000 (Hintze 1998), and significance was set at a 5 0.05 for the ANOVA models and for calculating experiment-wise confidence intervals for the Tukey-Kramer multiple comparison tests. We assumed any differences observed at our single sampling station were representative of conditions within Irondequoit Bay. Comparisons of zooplankton refuge establishment between years were based on the ecological interpretation of the trends in zooplankton density and size among the depth strata for each year in conjunction with the relative abundance of planktivorous fish and metalimnetic DO concentrations. Oxygenation was manipulated to create refuge conditions in one year (1997) and a single year (1996) with oxygen injection but without a deliberate attempt to create a refuge was considered a ‘‘control.’’ Of course, differences between the two years other than metalimnetic DO concentrations could cause changes in zooplankton sizes or densities, and these possibilities are discussed. RESULTS
Vertical structure of temperature, dissolved oxygen, and chlorophyll a Thermal structure was not affected by hypolimnetic injection of oxygen and a well-defined metalimnion formed by late May or early June in the north basin of Irondequoit Bay during 1996 and 1997. Late May– August epilimnetic temperatures in 1996 and 1997 were 22.9 6 0.78C and 21.7 6 1.88C, respectively (means 6 2 SE). Epilimnetic temperatures approached
117
highs of 24–258C throughout August 1996 and the second half of July 1997. Metalimnetic temperatures averaged 17.6 6 0.88C in 1996 and 15.1 6 1.08C in 1997. Hypolimnetic temperatures increased from 6–7 8C in late May to 11–128C by August of both years. Mean depth and thickness of the metalimnion differed slightly between years; starting depth was 5.3 m in 1996 and 6.1 m in 1997. Width of the metalimnion from May through August averaged 6.1 m in both years. Despite continuous oxygen supplementation in the hypolimnion of ;1700 kg/d starting on 10 June in 1996 and 13 July 1997, DO concentrations in the hypolimnion dropped below 0.5 mg/L (Fig. 2). Dissolved oxygen concentrations in the hypolimnion declined at similar rates (0.13 mg·L21·d21) from late May to early July in both years despite oxygenation starting approximately one month earlier in 1996. Slight metalimnetic increases of DO were evident in July, August, and September of 1997. Metalimnetic DO concentrations after hypolimnetic oxygenation differed between years (Fig. 2). In 1996, DO concentrations in the metalimnion were generally above 4.0 mg/L from May through the third week of July and only became ,2 mg/L during the first week of August. In contrast, 1997 metalimnetic DO concentrations declined below 4.0 mg/L by the first week of July and remained within the targeted refuge concentration of 1–2 mg/L for four straight weeks. In 1996, metalimnetic DO was 2.8 6 0.7 mg/L from late May through August compared to 1.5 6 0.7 mg/L in 1997. However, in 1996 during the time coinciding with our zooplankton collections (late May to early August) metalimnetic DO averaged 3.8 6 1.0 mg/L. During 1997, the metalimnion became essentially anoxic (DO , 0.2 mg/L) in late September as fall mixing began. Chlorophyll a concentrations in the epilimnion were variable in 1996 while concentrations remained relatively uniform throughout the summer of 1997 (Fig. 3). Epilimnetic chlorophyll concentrations in 1996 from late May to early August were often $15 mg/L while those in 1997 at this time were ,15 mg/L. Mean summer (late May–August) chlorophyll a concentrations were 5.2 mg/L (29%) greater in 1996 compared to 1997 but Secchi depths were similar (Table 1). In 1996, epilimnetic chlorophyll a concentrations and Secchi depths declined throughout August, to levels lower than 1997; however, concurrent zooplankton data was lacking. Attributing changes in chlorophyll to grazing by zooplankton protected by the refuge was confounded by high seasonal variability. For both years, coefficients of variation were .33% for chlorophyll a and .20% for Secchi depth; considerable overlap existed among 1996 and 1997 values (Fig. 3).
Fish Relative abundance (CPUE) of planktivores was higher in 1997 compared to 1996. Alewives were the only planktivore captured in 1996, whereas alewives
118
ROBERT A. KLUMB ET AL.
Ecological Applications Vol. 14, No. 1
FIG. 2. Dissolved oxygen concentrations in the epilimnion (closed circles), metalimnion (dashed line), and hypolimnion (open circles) of the north basin of Irondequoit Bay, New York, during 1996 and 1997. Arrows indicate initiation of hypolimnetic oxygenation.
and emerald shiners were the principal species caught in 1997 (Figs. 4 and 5). Overall CPUE in 1996 was 128 6 86 fish/h while relative abundance nearly doubled in 1997 to 234 6 125 fish/h. Inshore movement of spawning alewives into Irondequoit Bay was intense from late May to mid-June of both years (CPUE . 100 fish/h), and abundance then declined in late June through July. Emerald shiners were first seen on 2 July 1997 in very high abundance (CPUE . 650 fish/h) and remained in the bay throughout the summer. During 1997, the first appearance of YOY alewife (.50 mm TL) was on 6 August. Gizzard shad YOY (75–105 mm TL) were only caught in September of 1997 and in low relative abundance (CPUE , 20 fish/h). Abundance of hydroacoustic targets generally corresponded with gillnet CPUE (Fig. 5). There was a concomitant decline in the relative frequency of targets greater than 246 dB on 2 July, when yearling and older alewife gillnet CPUE declined 82%. Also on 2 July, a
pronounced increase in targets less than 246 dB was evident, which corresponded with the high influx of emerald shiners in the bay. Vertical distributions of fish illustrated the shrinking of fish habitat in association with summer oxygen depletion (Figs. 4 and 5). In May, alewives were predominantly captured at mid-depth, 5–7 m in 1996 and 8–11 m in 1997. By mid-July, DO concentrations began to fall below 4 mg/L in the metalimnion, and alewives shifted to shallow depths (surface to 6 m). Acoustic fish targets also filled the entire water column in May and early June and generally disappeared below 6 m in mid-July. Depth of capture in vertical gillnets in 1996 and 1997 in relation to DO concentrations coincided with the depth distributions of acoustic targets recorded in 1997; fish were generally concentrated in water with DO . 4.0 mg/L (Fig. 5). However, .4.0% of fish and acoustic targets were observed at depths with DO # 4.0 mg/L.
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
119
FIG. 3. Chlorophyll a concentrations in the epilimnion (closed circles), metalimnion (dashed line), and hypolimnion (open circles) with corresponding Secchi depths (open triangles) for the north basin of Irondequoit Bay, New York, in 1996 and 1997. Arrows indicate initiation of hypolimnetic oxygenation.
Mean densities of 261 to 246 dB targets (small fish) and targets greater than 246 dB (mainly adult alewives) were almost always greatest in the epilimnion (Fig. 5). For all fish targets, seasonal densities in the epilimnion, metalimnion, and hypolimnion were 5502 6 1764, 1699 6 1646, and 267 6 358 fish/ha, respectively. Densities of alewife-sized targets (greater than 246 dB) in the epilimnion were greatest (range 1218– 2056 fish/ha) during the spawning season from late May to mid-June. Small fish targets were observed in the hypolimnion on 2 July, the date of peak emerald shiner abundance. After corrections for bubble densities, few acoustic fish targets during August and no fish targets during September remained in the hypolimnion.
Zooplankton During 1996 and 1997, the zooplankton assemblage in Irondequoit Bay consisted mainly of three cladocerans, Bosmina longirostris, D. galeata mendotae, and D. retrocurva, and two cyclopoid copepods, Tropocyclops prasinus and M. edax. Eubosmina coregoni, Diaphanasoma spp., Acanthocyclops vernalis, and
Diacyclops thomasi biscupidatus were common but collectively amounted to ,2% of all zooplankton enumerated. Calanoid copepods, Daphnia ambigua, Daphnia pulicaria, and Ceriodaphnia quadrangulata, were rare. No red coloration was evident in any zooplankton samples during 1997, indicating no hemoglobin production by daphnids. Seasonal changes in assemblage composition were generally similar within the epilimnion, metalimnion, and hypolimnion in both years (Fig. 6). Bosminids (essentially B. longirostris) peaked in early June and comprised 60–95% of the zooplankton assemblage in the epilimnion from late May to early July in both years. However, in late May of 1997, the zooplankton assemblage was principally copepods (mainly D. bicuspidatus) before shifting a week later to B. longirostris. In both years, the assemblage shifted again in mid-July to daphnids (D. retrocurva in 1996 and a mix of D. retrocurva, D. galeatea, and D. longiremis in 1997) and copepods (M. edax in both years). At this time, B. longirostris essentially disappeared from the water column. The longer time series available in 1997 encom-
Ecological Applications Vol. 14, No. 1
ROBERT A. KLUMB ET AL.
120
TABLE 1. Secchi depth (in meters), chlorophyll a (in micrograms per liter), crustacean zooplankton density (number of individuals per liter), and size (in millimeters) for the epilimnion, metalimnion, and hypolimnion of the north basin of Irondequoit Bay, New York, USA, during 1996 and 1997 (means 6 2 SE).
Variable
N
Epilimnion
1996 Secchi depth Chlorophyll a
40 19
2.0 6 0.3 17.8 6 3.7
··· 6.3 6 3.1
6 6 6 6 6
141 6 43b 1.5 6 1.1a,b 8.5 6 4.0b 6.6 6 3.9 0.48 6 0.09b
Density All species D. galeata mendotae D. retrocurva M. edax Zooplankton size 1997 Secchi depth Chlorophyll a Density All species D. galeata mendotae D. retrocurva M. edax Zooplankton size
14 14 14 14 14 14 14 19 19 19 19 19
227.9 3.1 38.1 14.2 0.40
32.6a 2.1a 17.2a 9.7 0.06a
1.9 6 0.3 12.6 6 2.0 102.9 1.8 1.5 2.1 0.44
6 6 6 6 6
66.0a 1.8a 1.6a 1.6a 0.03a
Metalimnion
··· 4.6 6 1.0 104.5 10.4 8.1 13.4 0.59
6 6 6 6 6
64.3a 7.8b 6.2b 7.6b 0.04b
Hypolimnion ··· 2.3 6 0.8 76.9 0.2 2.5 6.9 0.46
6 6 6 6 6
24.9c 0.2b 1.5c 4.0 0.07a,b
··· 2.1 6 0.9 39.9 4.0 1.1 7.5 0.57
6 6 6 6 6
18.4b 6.8a,c 1.3a,c 5.2b,c 0.03b,c
Notes: Repeated-measures ANOVA models used the factors date and depth to assess the effectiveness of a low-oxygen metalimnetic refuge for zooplankton. Mean zooplankton densities and sizes with different superscripts are significantly different based on Tukey-Kramer 95% experiment-wise confidence intervals. Mean chlorophyll and Secchi depth were calculated for time periods where 1996 and 1997 data coincided (late May through August). Additional data for 1997 are presented in Fig. 3.
passed a second peak of B. longirostris in late August and September, although this peak comprised at most 65% of the zooplankton assemblage. Seasonal mean densities (numbers per liter) of all zooplankton differed between 1996 and 1997 (Fig. 7) and among some depth strata. Densities of zooplankton in the epilimnion rarely declined below 150 zooplankton/L in 1996, whereas, after a peak of B. longirostris on 11 June, densities were rarely above 100 zooplankton/L in the epilimnion during July and August of 1997. In 1996, mean densities declined with depth from the epilimnion to the hypolimnion (Table 1). Depth stratum significantly described variance in zooplankton density in both years. In 1996, densities were significantly higher in the epilimnion compared to the metalimnion and significantly higher in the metalimnion vs. the hypolimnion. During 1997, mean zooplankton densities were similar in the epilimnion and metalimnion with densities significantly lower in the hypolimnion. Mean densities of D. galeata mendotae and D. retrocurva were highest in the epilimnion in 1996 and in the metalimnion in 1997 (Table 1; Fig. 8) providing indication of successful refuge formation in 1997. Depth stratum significantly described variation in the densities of both daphnids in both years. In 1996, densities of D. galeata mendotae were similar in the epilimnion and metalimnion with significantly lower densities in the hypolimnion (Table 1). D. retrocurva densities in 1996 were highest in the epilimnion and sig-
nificantly declined with depth. In 1997, densities of both daphnids were significantly greater in the metalimnion than in the epilimnion or hypolimnion (Table 1; Fig. 8). Mean densities of M. edax were highest in the epilimnion in 1996 and in the metalimnion in 1997 (Table 1) indicating the refuge in 1997 also protected this invertebrate predator. Densities of M. edax did significantly differ with depth in 1997 but not in 1996 (Table 1). Significantly greater densities in 1997 were seen in the metalimnion than the epilimnion while densities in the metalimnion and hypolimnion were similar. Metalimnetic peaks of M. edax generally corresponded with low densities of D. galeata mendotae and D. retrocurva during 1997 (Fig. 8). In both years, largest mean zooplankton size was observed in the metalimnion or hypolimnion (Table 1, Fig. 9). In 1996, zooplankton size remained small (,0.4 mm) during May and June at all depths when alewife relative abundance was highest (Fig. 4) and gradually increased in July and August. In 1997, mean zooplankton size declined below 0.5 mm at all depths on 25 June, the first sample taken after the peak abundance of alewives was observed on 18 June (Fig. 5). Sample date and depth stratum significantly described variance in zooplankton size in both years (Table 1). In both years, significantly larger zooplankton were found in the metalimnion compared to the epilimnion. Zooplankton in the hypolimnion were significantly
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
February 2004 TABLE 1.
121
Extended.
ANOVA factor Date
Depth
df
F
P
df
F
P
··· ···
··· ···
··· ···
··· ···
··· ···
··· ···
3.50 1.98 5.07 4.48 10.54
0.003 0.067 ,0.001 ,0.001 ,0.001
35.64 6.15 33.42 1.29 5.38
,0.001 0.007 ,0.001 0.294 0.011
··· ···
··· ···
··· ···
··· ···
3.39 4.02 5.09 5.64 4.97
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
8.44 9.37 13.70 15.90 52.60
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
13, 13, 13, 13, 13,
26 26 26 26 26
··· ··· 18, 18, 18, 18, 18,
36 36 36 36 36
larger than the epilimnion in both years and significantly larger than the metalimnion in 1997. DISCUSSION
Refuge establishment In both study years, metalimnetic DO concentrations after oxygenation were higher than historical levels. In 1997, we were able to maintain metalimnetic oxygen levels between 1 and 2 mg/L whereas metalimnetic DO in 1996 was higher and refuge concentrations were sustained for a shorter time despite similar hypolimnetic oxygen injection rates. Only by the first week of August in 1996 did DO levels fall within the targeted refuge concentrations, while in 1997 refuge concentrations generally were met from the second week of July through the third week of August (Fig. 2). High biochemical oxygen demand (BOD) can impede formation and maintenance of a low-oxygen refuge in eutrophic systems. Shapiro and Wright (1984) were unable to establish a DO refuge since quantities of oxygen injected into the metalimnion could not compensate for the high BOD in Round Lake, Minnesota, USA. Despite hypolimnetic aeration throughout the summer of a small kettle lake in Ontario, Canada, high BOD resulted in anoxia after mid-July or August in the hypolimnion; hydrogen sulfide accumulated in the metalimnion and restricted zooplankton to the epilimnion (Taggart 1984). Successful establishment of a DO refuge will likely only occur when internal cycling of phosphorus and external nutrient loading are reduced. Epilimnetic summer levels of soluble reactive phosphorous in Irondequoit Bay ranged from 1200 to 6000
2, 2, 2, 2, 2,
26 26 26 26 26
··· ··· 2, 2, 2, 2, 2,
36 36 36 36 36
mg/L in 1970 and 1971 (Bannister and Bubeck 1978). In part due to the accumulating effects of the detergent ban on phosphates in the 1970s, elimination in the 1970s and 1980s of raw sewage inputs during storm events, and the alum treatment in 1986, epilimnetic levels of total phosphorus in 1996 and 1997 were generally below 75 mg/L (MCHD, unpublished data). In addition to physically creating suitable metalimnetic DO levels for zooplankton, exclusion of zooplankton predators is also required to establish a refuge. In Irondequoit Bay, gillnet catches in 1996 and 1997 combined with hydroacoustic data from 1997 indicated that fishes were prevalent at depths where DO concentrations exceeded 4.0 mg/L with few fish present (,4%) at lower concentrations (Figs. 4 and 5). Prolonged exposure to DO concentrations below 4.0 mg/L may be harmful to temperate freshwater fish (Doudoroff and Shumway 1970, Davis 1975), and few fish are typically found at depths with DO , 2.0 mg/L (Rudstam and Magnuson 1985, Aku and Tonn 1999). In two small, inland New York lakes, alewives were abundant in the water column where DO was 3–4 mg/L but avoided concentrations below 2 mg/L (L. G. Rudstam, unpublished data). However, fish have been documented to forage temporarily in waters with , 2.0 mg/L (Talbot and Kramer 1986, Luecke and Teuscher 1994, Aku and Tonn 1999). Our hydroacoustic data also indicated that fish may have temporarily entered the refuge to forage as seen by echoes counted at depths greater than 14 m in late June and early July of 1997 (Fig. 5). High numbers of large daphnids in the metalimnion (Figs. 8 and 9) indicated that short foraging bouts by fish into low-
122
ROBERT A. KLUMB ET AL.
Ecological Applications Vol. 14, No. 1
FIG. 4. Vertical distributions of fish in relation to temperature and dissolved oxygen (DO) concentrations in Irondequoit Bay, New York, based on the catch-per-unit-effort (CPUE) from vertical gillnets during 1996. Depths are in 1-m increments (i.e., 1 5 depth interval 0–1 m). Gillnets only fished from the surface to 10–12 m. Alewives were the only planktivore species captured in 1996.
oxygen waters apparently were not sufficient to erode the efficacy of the refuge in Irondequoit Bay during 1997. Luecke and Teuscher (1994) also found that high densities of daphnids persisted in a low-oxygen refuge despite periodic foraging by trout. Higher metalimnetic DO concentrations in 1996 provided more available foraging habitat in the water column for planktivorous fish compared to 1997, which likely reduced the ability of a low-oxygen refuge to protect large daphnids in 1996. Our study is the first attempt to successfully form and maintain a zooplankton refuge in a deep Great Lakes embayment with a pelagic fish community dominated by alewives. Predation by alewives is well known to structure zooplankton communities (Brooks and Dodson 1965, O’Gorman et al. 1991). Alewives feed on zooplankton using particulate, gulping, and filter-feeding behaviors (Janssen 1978) and can feed at night (Janssen et al. 1995). Larval and juvenile life stages are especially effective predators due to their high numbers and growth rates, consuming an estimated 50% of the total annual production of zooplank-
ton eaten by all alewives (Hewett and Stewart 1989). Whether in enclosures or whole-lake experiments, most studies of zooplankton refuges have been done in small eutrophic lakes dominated by centrarchids and percids (Shapiro et al. 1975, Wright and Shapiro 1990, Tessier and Welser 1991, Mittelbach and Osenberg 1993). Exceptions to these studies include ciscos ( Coregonus artedi) in Canada (Aku et al. 1997), a mixed salmonidsmelt community in Japan (Hanazato et al. 1989), and smelt (Osmerus eperlanus) in Finland (Horppila et al. 2000). In this study, we observed an exceptional test of refuge effectiveness in 1997 due to the high densities of emerald shiners, a species characterized by sporadic huge year classes (Smith 1985). Emerald shiners were abundant in Irondequoit Bay from 2 July to 15 September based on gillnet CPUE (28–671 fish/h) and hydroacoustic densities of emerald shiner-sized targets (4113–11 702 fish/ha). In contrast, we observed no emerald shiners in 1996. Alewife numbers have declined during most of the 1990s in Lake Ontario (O’Gorman and Stewart 1999), lessening their predation on the
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
pelagic eggs and larvae of emerald shiners (Crowder 1980). Emerald shiners are omnivorous but their diets consist mainly of large cladocerans (Hartman et al. 1992) with peak feeding observed during daylight hours (Lippold 1998). Therefore, in regards to zooplankton predation, emerald shiners are functionally equivalent to alewives. Densities of planktivores in Irondequoit Bay were likely among the highest documented in any study of a zooplankton refuge and indicated potential predation on zooplankton was high throughout the summer of 1997. Densities of alewife-sized hydroacoustic targets (TS greater than 246 dB) were 1896–2416 fish/ha in early summer while the mean density of all fish targets including YOY was 7468 fish/ha (range: 2936–12 406) during 1997 (Fig. 5). These densities are conservative because the top 2 m of the water column was not sampled by hydroacoustics gear. After oxygenation in Amisk Lake, Alberta, Canada, densities of adult ciscoes were 1900 fish/ha with a peak of 7400 fish/ha when including YOY (Aku and Tonn 1997), densities comparable to those we observed in Irondequoit Bay. In the lakes studied by Mittelbach and Osenberg (1993), densities of juvenile and adult bluegills ( Lepomis macrochirus) were less than 1600 fish/ha and 400 fish/ha, respectively, based on SCUBA counts. Young-of-year alewives (20–50 mm TL) would be included in our acoustic estimates and are probably present throughout the summer even though we did not catch these fish in our gill nets until August. By August, the largest YOY alewives were susceptible to capture in our gill nets; mean TL in August was 61 mm. We did not include densities of larval fish targets less than 261 dB (i.e., fish larvae ,20 mm TL) in this study because we could not reliably distinguish small fish from invertebrate targets. Consequently our acoustics-based estimates of fish density were conservative since larval fish (alewives and Lepomis spp.) are present in Irondequoit Bay during July and August (Klumb et al. 2003), exerting additional predation pressure on the zooplankton. Hydroacoustics has been used previously in eutrophic lakes to show restriction of fishes to shallow depths where DO concentrations generally are greater than 4 mg/L (Shapiro et al. 1975, Wright and Shapiro 1990, Luecke and Teuscher 1994, Aku et al. 1997, Horppila et al. 2000). Although no observations concerning bubble release from sediments are mentioned in these studies, bubble formation does occur and can bias fish density estimates in eutrophic, thermally stratified lakes suffering periodic anoxia.
Zooplankton community dynamics Refuge size and density of predators differed between years (1996 vs. 1997) and likely affected zooplankton dynamics and effectiveness of the metalimnetic DO refuge in Irondequoit Bay. In 1996, DO concentrations remained too high in the metalimnion ( .2.6 mg/L) during July and likely failed to protect large
123
zooplankton. Alewives were still captured at middepths (4–8 m) in July, and mean zooplankton size in 1996 was small (,0.6 mm) in the epilimnion and metalimnion (Fig. 9), indicating high planktivory rates (Mills et al. 1987). Despite high fish abundance in 1997, the DO refuge protected large zooplankton and kept the mean size in the metalimnion 0.14 mm larger than the epilimnion. Compared to 1996, zooplankton in the metalimnion and hypolimnion were 23% and 24% larger in 1997. Summer mean zooplankton densities in the epilimnion also suggest high planktivory in 1997. Zooplankton densities were generally less than 100 zooplankton/L from July to August 1997 compared to densities greater than 150 zooplankton/L in 1996 (Fig. 6). Sustaining metalimnetic DO concentrations in 1997 at or below 2.0 mg/L for four consecutive weeks was likely sufficient time to influence zooplankton densities and size structure. Turnover rates of Lake Ontario cladocerans generally occur within 5–18 d at the temperatures observed in Irondequoit Bay (Borgman et al. 1984, Cooley et al. 1986). Our findings demonstrate that in 1997 large herbivorous zooplankton were rare in the oxygen-rich epilimnion compared to the lowoxygen metalimnion of Irondequoit Bay. Daphnia galeata mendota and D. retrocurva became increasingly abundant in the metalimnion during hypolimnetic oxygen injection in 1997 (Fig. 8), while only juvenile and small-bodied cladocerans and copepods ( ,0.60 mm) existed in the epilimnion where vertebrate predators were plentiful. Other studies have found that zooplankton communities responded to an available refuge from fish predation. In Amisk Lake, Alberta, Canada, Field and Prepas (1997) found that zooplankton migrated deeper and abundance of two large-bodied Daphnia spp. was greater in a basin treated by hypolimnetic oxygen injection compared to an untreated basin. High densities of D. longiremis and D. pulex in Amisk Lake corresponded with low hydroacoustic fish densities at oxygen levels avoided by cisco, the dominant planktivore (Aku et al. 1997, Field and Prepas 1997). In Harp Lake, hypolimnetic oxygen injection successfully produced an oxygen refuge and D. galeata mendota dominated the summer zooplankton community, contrary to the traditional shift to small (non-daphnid) cladocerans in the absence of a refuge (Gemza 1997). Refuge DO levels observed in Irondequoit Bay could induce negative physiological responses in Daphnia. Dissolved oxygen concentrations below 2.0 mg/L negatively affect respiration and elicit hemoglobin production in some species (Fox and Phear 1953, Heisey and Porter 1977). In Irondequoit Bay during 1997, lack of hemoglobin production indicated stress from low DO in the refuge was not impairing the health of daphnids. Colder temperatures in the refuge may have slowed metabolic rates (Orcutt and Porter 1983), offsetting some of the energetic costs of reduced respi-
124
ROBERT A. KLUMB ET AL.
Ecological Applications Vol. 14, No. 1
FIG. 5. Vertical distributions of fish in relation to temperature and dissolved oxygen (DO) concentrations in Irondequoit Bay, New York, based on the catch-per-unit-effort (CPUE) from vertical gillnets and densities estimated with hydroacoustics during 1997. Hydroacoustic targets were divided into two groups representing small (20–90 mm total length) fish (261 to 246 dB) and adult (.90 mm total length) fish targets (.46 dB). Depths are in 1-m increments (i.e., 1 5 depth interval 0–1 m). Gillnets only fished from the surface to 10–12 m.
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
FIG. 5
Continued.
125
126
ROBERT A. KLUMB ET AL.
Ecological Applications Vol. 14, No. 1
FIG. 6. Crustacean zooplankton assemblages in the (A) epilimnion, (B) metalimnion, and (C) hypolimnion of the north basin of Irondequoit Bay, New York, in 1996 and 1997. No zooplankton were collected after the first week of August in 1996.
ration efficiency associated with inhabiting the lowoxygen region. Diel vertical migrations may have allowed D. galeata mendotae access to food and oxygen in the epilimnion, enabling survival within a low-oxygen environment. However, we sampled zooplankton only in the daytime and are uncertain whether largebodied species vertically migrated into the epilimnion at night. Diel migration is a common behavior of Daphnia to maximize predator avoidance by staying at depths avoided by fish during the day and maximize food intake by migrating upward into the chlorophyllrich epilimnion at night (Zaret and Suffern 1976, Lampert 1987). Although the DO refuge protected zooplankton from fish predation, concurrent protection of invertebrate predators can reduce herbivorous zooplankton popu-
lations within the refuge (Hanazato 1992, Horppila et al. 2000). Invertebrate predators can be the main source of mortality for freshwater zooplankton, often exceeding predation by fishes (Lane 1979). High densities of M. edax were observed in the DO refuge of Irondequoit Bay in 1997. Cyclopoid copepods can tolerate hypoxia and withstand several hours of anoxic conditions (Tinson and Laybourn-Parry 1985, Williamson 1991). Mesocyclops edax is an omnivorous cyclopoid copepod but invertebrate prey can increase its reproductive success (Adrian and Frost 1993). Prey of M. edax includes copepod nauplii and rotifers (Brandl and Fernando 1979), although zooplankters larger than themselves can be successfully ingested (Williamson 1983), and predation rates on cladocerans, as the percentage of total numbers in the population eaten per day, may be
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
127
FIG. 7. Mean densities of all zooplankton species in the epilimnion (closed circles), metalimnion (dashed line), and hypolimnion (open circles) of the north basin of Irondequoit Bay, New York, in 1996 and 1997. Arrows indicate initiation of hypolimnetic oxygenation. No zooplankton were collected after the first week of August in 1996.
as high as 4–16% (Brandl and Fernando 1979). Despite the predation risk from M. edax, mean metalimnetic Daphnia density in 1997 (D. galeata and D. retrocurva) was 21 Daphnia/L from July through September. However, predation by M. edax on small-bodied zooplankton within the refuge could have contributed to the increased overall mean zooplankton size within the metalimnion in 1997.
Chlorophyll and water clarity Epilimnetic summer mean chlorophyll a in Irondequoit Bay was approximately 29% lower in 1997 than in 1996, and the seasonal patterns were different (Fig. 3). A large chlorophyll peak was observed in mid-July of 1996, whereas chlorophyll concentrations were relatively uniform during the summer of 1997. Corresponding Secchi depth from late May to early August was 13% deeper in 1997 compared to 1996 and increased 2 m from 30 July to 20 August 1997 (Fig. 3) when DO levels in the metalimnion were ,2 mg/L. In August of 1996, when metalimnetic DO concentrations were 1–2 mg/L, chlorophyll decreased over 75% and Secchi depth increased 2 m. Similar epilimnetic summer (late May–August) total phosphorous concentrations in 1996 (32 6 6 mg/L) and 1997 (31 6 4 mg/L) indicated differences in chlorophyll among years was likely not due to enrichment (MCHD, unpublished data). Therefore, we suggest that lowered chlorophyll
levels in both years could have resulted from grazing by daphnids migrating to the epilimnion from the refuge. However, decreased chlorophyll and increased Secchi depth from 1996 to 1997 were also within the annual fluctuations for the years 1993 to 1996 when oxygenation took place without the intent to create a metalimnetic DO refuge. From 1993 to 1995, summer (late May–August) chlorophyll a was 10.3 6 0.8 mg/L and Secchi depths averaged 2.0 6 0.2 m (MCHD, unpublished data). Oxygenation with the aim to create a metalimnetic DO refuge for zooplankton has continued in Irondequoit Bay from 1998 to 2002. Available zooplankton data during that period further suggest that metalimnetic DO refuge protected large-bodied zooplankton. Mean zooplankton size in Irondequoit Bay during July from 1998 to 2000 measured from samples collected with a 153-mm meshed net vertically hauled across 10 m of the water column was large (0.70 mm) and indicative of low predation rates (Mills et al. 1987, O’Gorman et al. 1991). In contrast, zooplankton sizes during this period ranged from 0.44 to 0.64 mm at three other Lake Ontario embayments (E. L. Mills, unpublished data). Natural or human-induced changes in trophic levels can cascade through the food web (Carpenter et al. 1985), forming the basis for biomanipulation as a management strategy (Shapiro et al. 1975). However, ef-
128
ROBERT A. KLUMB ET AL.
Ecological Applications Vol. 14, No. 1
FIG. 8. Mean densities of Daphnia galeata mendotae, D. retrocurva, and Mesocyclops edax in the epilimnion (closed circles), metalimnion (dashed line), and hypolimnion (open circles) of the north basin of Irondequoit Bay, New York, in 1996 and 1997. Arrows indicate initiation of hypolimnetic oxygenation. No zooplankton were collected after the first week of August in 1996.
fects of planktivorous fish removal or piscivorous fish additions on phytoplankton abundance are often dampened in eutrophic lakes (McQueen et al. 1989). By creating a DO refuge for zooplankton in Irondequoit Bay, the ‘‘biomanipulation’’ in our study was one trophic level lower compared to traditional projects that physically removed fish biomass (reviewed by Hansson et al. [1998]). Increased densities of large Daphnia in the refuge formed in Irondequoit Bay during 1997 after physical manipulation of the environment were associated with a 29% reduction in chlorophyll compared to 1996. Formation of refuges for large zooplankton contributed to successful biomanipulation in shallow lakes. Macrophyte beds are used by large Daphnia during the day as a refuge from fish predation (Stansfield et al. 1997, Moss et al. 1998), and these refuges contribute
to the stability of the macrophyte-dominated clear-water state of shallow lakes (Scheffer et al. 1993). A metalimnetic oxygen refuge can have the same function in deep lakes but it is important to maintain oxygen levels high enough to support Daphnia but low enough to exclude fish (1–2 mg/L). Such specific oxygen levels are difficult to maintain without deliberate manipulation as in this study. Therefore effective low-oxygen refuges for Daphnia are likely more transient phenomena in natural lakes.
Management implications Our study suggests that water clarity improvement in eutrophic ecosystems through manipulation of metalimnetic oxygen levels to create a refuge for largebodied zooplankton grazers has promise as a management tool. Water quality in an open system such as
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
129
FIG. 9. Mean total lengths of all crustacean zooplankton species in the epilimnion (closed circles), metalimnion (dashed line), and hypolimnion (open circles) in the north basin of Irondequoit Bay, New York, in 1996 and 1997. Arrows indicate initiation of hypolimnetic oxygenation. No zooplankton were collected after the first week of August in 1996.
Irondequoit Bay is influenced directly by an urban watershed and indirectly by intense planktivory associated with the seasonal migration of alewives originating from Lake Ontario (O’Gorman et al. 1991). Consequently, fish removal to encourage increased grazing by large-bodied zooplankton was not a practical solution to water quality management in this large embayment. Diligent monitoring of metalimnetic oxygen concentrations is necessary to ensure adequate DO levels exist to establish a zooplankton refuge. This monitoring should not be considered a drawback since current technology allows DO to be remotely sensed in situ and electronically accessed in real time. Where hypolimnetic oxygen injection is needed to prevent mobilization of phosphorus from sediments, creation of a low-oxygen zooplankton refuge could provide additional benefits toward the goal of improved water clarity. ACKNOWLEDGMENTS We thank James Brien, Bas Benneji, and Mark Van de Weerd for their assistance in the field. This project was funded by a grant from the Monroe County Health Department Environmental Health Laboratory to E. L. Mills and L. G. Rudstam. F. Arrhenius was supported by a postdoctoral grant from the Swedish Council for Forestry and Agricultural Research. Early versions of this manuscript were greatly improved by comments from David Warner, and Patrick Sullivan provided statistical advice. This is contribution No. 214 of the Cornell Biological Field Station.
LITERATURE CITED Adrian, R., and T. M. Frost. 1993. Omnivory in cyclopoid copepods: comparisons of algae and invertebrates as food for three, differently sized species. Journal of Plankton Research 15:643–658. Aku, P. M. K., L. G. Rudstam, and W. M. Tonn. 1997. Impact of hypolimnetic oxygenation on the vertical distribution of cisco (Coregonus artedi) in Amisk Lake, Alberta. Canadian Journal of Fisheries and Aquatic Sciences 54:2182–2195. Aku, P. M. K., and W. M. Tonn. 1997. Changes in population structure, growth, and biomass of cisco (Coregonus artedi) during hypolimnetic oxygenation of a deep, eutrophic lake, Amisk Lake, Alberta. Canadian Journal of Fisheries and Aquatic Sciences 54:2196–2206. Aku, P. M. K., and W. M. Tonn. 1999. Effects of hypolimnetic oxygenation on the food resources and feeding ecology of cisco in Amisk Lake, Alberta. Transactions of the American Fisheries Society 128:17–30. APHA [American Public Health Association]. 1982. Standard methods for the examinations of water and waste water. Fourteenth edition. Washington, D.C., USA. Balcer, M. D., N. L. Korda, and S. I. Dodson. 1984. Zooplankton of the Great Lakes. University of Wisconsin Press, Madison, Wisconsin, USA. Bannister, T. T., and R. C. Bubeck. 1978. Limnology of Irondequoit Bay, Monroe County. Pages 106–221 in J. A. Bloomfield, editor. Lakes of New York State. Volume II. Ecology of the lakes of western New York. Academic Press, New York, New York, USA. Borgmann, U., H. Shear, and J. Moore. 1984. Zooplankton and potential fish production in Lake Ontario. Canadian Journal of Fisheries and Aquatic Sciences 41:1303–1309.
130
ROBERT A. KLUMB ET AL.
Brandl, Z., and C. H. Fernando. 1979. The impacts of predation by the copepod Mesocyclops edax (Forbes) on zooplankton in three lakes in Ontario, Canada. Canadian Journal of Zoology 57:940–942. Brooks, J. L., and S. I. Dodson. 1965. Predation, body size, and composition of plankton. Science 150:28–35. Carpenter, S. R., J. F. Kitchell, and J. R. Hodgson. 1985. Cascading trophic interactions and lake productivity. Bioscience 35:634–639. Cooley, J. M., J. E. Moore, and W. T. Geiling. 1986. Population dynamics, biomass, and production of the macrozooplankton in the Bay of Quinte during changes in phosphorus loadings. Pages 166–176 in C. K. Minns, D. A. Hurley, and K. H. Nicholls, editors. Project Quinte: pointsource phosphorus control and ecosystem response in the Bay of Quinte, Lake Ontario. Canadian Journal of Fisheries and Aquatic Sciences Special Publication. National Research Council of Canada, Ottawa, Ontario, Canada. Crowder, L. B. 1980. Alewife, rainbow smelt and native fishes in Lake Michigan: competition or predation? Environmental Biology of Fishes 5:225–233. Davis, J. C. 1975. Minimal dissolved oxygen requirements of aquatic life with emphasis on Canadian species: a review. Journal of the Fisheries Research Board of Canada 32: 2295–2332. Doudoroff, P., and D. L. Shumway. 1970. Dissolved oxygen requirements of freshwater fishes. Food and Agricultural Organization of the United Nations. FAO Fisheries Technical Paper 86. Fast, A. W. 1971. Effects of artificial destratification on zooplankton depth distribution. Transactions of the American Fisheries Society 100:355–358. Field, K. M., and E. E. Prepas. 1997. Increased abundance and depth distribution of pelagic crustacean zooplankton during hypolimnetic oxygenation in a deep, eutrophic Alberta lake. Canadian Journal of Fisheries and Aquatic Sciences 54:2146–2156. Fox, H. M., and E. A. Phear. 1953. Factors influencing hemoglobin synthesis by Daphnia. Proceedings of the Royal Society London B, Biological Sciences 141:179–189. Gemza, A. F. 1997. Water quality improvements during hypolimnetic oxygenation in two Ontario lakes. Water Quality Research Journal of Canada 32:365–390. Hanazato, T. 1992. Direct and indirect effects of low-oxygen layers on lake zooplankton communities. Archiv fur Hydrobiologie Ergebninsse der Limnologie 35:87–98. Hanazato, T., M. Yasuno, and M. Hosomi. 1989. Significance of a low oxygen layer for a Daphnia population in Lake Yunoko, Japan. Hydrobiologia 185:19–27. Hansson, L.-A., H. Annadotter, E. Bergman, S. F. Hamrin, E. Jeppesen, T. Kairesalo, E. Luokkanen, P.-A. Nilsson, M. Sondergaard, and J. Strand. 1998. Biomanipulation as an application of food-chain theory: constraints, synthesis, and recommendations for temperate lakes. Ecosystems 1: 558–574. Hartman, K. J., B. Vondracek, D. L. Parrish, and K. M. Muth. 1992. Diets of emerald and spottail shiners and potential interactions with other western Lake Erie planktivorous fishes. Journal of Great Lakes Research 18:43–50. Heisey, D., and K. G. Porter. 1977. The effect of ambient oxygen concentration on filtering and respiration rates of Daphnia galeata mendotae and Daphnia magna. Limnology and Oceanography 22:839–845. Herberger, F., and J. B. Reynolds. 1977. Abundance, composition, and distribution of crustacean zooplankton in relation to hypolimnetic oxygen depletion in west-central Lake Erie. U.S. Fish and Wildlife Service Technical Paper Number 93. Herbert, M. R. 1954. The tolerance of oxygen deficiency in the water by certain cladocera. Memorie dell’Istituto Italiano di Idrobiologia Dott Marco de Marchi 8:97–107.
Ecological Applications Vol. 14, No. 1
Hewett, S. W., and D. J. Stewart. 1989. Zooplanktivory by alewives in Lake Michigan: ontogenetic, seasonal, and historical patterns. Transactions of the American Fisheries Society 118:581–596. Hintze, J. L. 1998. Quick start and help manual NCSS 2000 statistical system for Windows. NCSS, Kaysville, Utah, USA. Horppila, J., T. Malinen, L. Nurminen, P. Tallberg, and M. Vinni. 2000. A metalimnetic oxygen minimum indirectly contributing to the low biomass of cladocerans in Lake Hiidenvesi—a diurnal study on the refuge effect. Hydrobiologia 436:81–90. Janssen, J. 1978. Feeding-behavior repertoire of the alewife, Alosa pseudoharengus, and the ciscoes Coregonus hoyi and C. artedii. Journal of the Fisheries Research Board of Canada 35:249–253. Janssen, J., W. R. Jones, A. Whang, and P. E. Oshel. 1995. Use of the lateral line in particulate feeding in the dark by juvenile alewife (Alosa pseudoharengus). Canadian Journal of Fisheries and Aquatic Sciences 52:358–363. Klumb, R. A., L. G. Rudstam, E. L. Mills, C. P. Schneider, and P. M. Sawyko. 2003. Importance of Lake Ontario embayments and nearshore habitats as nurseries for larval fishes with emphasis on alewife (Alosa pseudoharengus). Journal of Great Lakes Research 29:181–198. Lampert, W. 1987. Vertical migration of freshwater zooplankton: indirect effects of vertebrate predators on algal communities. Pages 291–299 in C. Kerfoot and A. Sih, editors. Predation: direct and indirect impacts on aquatic communities. University Press of New England, Hanover, New Hampshire, USA. Lane, P. 1979. Vertebrate and invertebrate predation intensity on freshwater zooplankton communities. Nature 280:391– 392. Lass, S., M. Boersma, and P. Spaak. 2000. How do migrating daphnids cope with fish predation risk in the epilimnion under anoxic conditions in the hypolimnion? Journal of Plankton Research 22:1411–1418. Lippold, M. A. 1998. Daily food ration and gastric evacuation rates of emerald shiners. Thesis. State University of New York, Buffalo, New York, USA. Luecke, C., and D. Teuscher. 1994. Habitat selection by lacustrine rainbow trout within gradients of temperature, oxygen, and food availability. Pages 133–149 in D. J. Stouder, K. L. Fresh, and R. J. Feller, editors. Theory and application in fish feeding ecology. Belle W. Baruch Library in Marine Sciences Number 18. University of South Carolina Press, Columbia, South Carolina, USA. McQueen, D. J., M. R. S. Johannes, J. R. Post, T. J. Stewart, and D. R. S. Lean. 1989. Bottom-up and top-down impacts on freshwater pelagic community structure. Ecological Monographs 59:289–309. Mills, E. L., D. M. Green, and A. Schiavone, Jr. 1987. Use of zooplankton size to assess the community structure of fish populations in freshwater lakes. North American Journal of Fisheries Management 7:369–378. Mittelbach, G. G., and C. W. Osenberg. 1993. Stage-structured interactions in bluegill: consequences of adult resource variation. Ecology 74:2381–2394. Moss, B., R. Kornijow, and G. J. Measey. 1998. The effects of nyphaeid (Nuphar lutea) density and predation by perch (Perca fluviatilis) on the zooplankton communities in a shallow lake. Freshwater Biology 39:689–697. Nebeker, A. V., S. T. Onjukka, D. G. Stevens, G. A. Chapman, and S. E. Dominguez. 1992. Effects of low dissolved oxygen on survival, growth and reproduction of Daphnia, Hyalella and Gammarus. Environmental Toxicicology and Chemistry 11:373–379. O’Gorman, R., E. L. Mills, and J. S. DeGisi. 1991. Use of zooplankton to assess the movement and distribution of
February 2004
METALIMNETIC ZOOPLANKTON OXYGEN REFUGE
alewife (Alosa pseudoharengus) in south-central Lake Ontario. Canadian Journal of Fisheries and Aquatic Sciences 48:2250–2257. O’Gorman, R., and T. J. Stewart. 1999. Ascent, dominance, and decline of the alewife in the Great Lakes: food web interactions and management strategies. Pages 489–513 in W. W. Taylor and C. P. Ferreri, editors. Great Lakes fisheries policy and management: a binational perspective. Michigan State University Press, East Lansing, Michigan, USA. Orcutt, J. D., Jr., and K. G. Porter. 1983. Diel vertical migration by zooplankton: constant and fluctuating temperature effects on life history parameters of Daphnia. Limnology and Oceanography 28:720–730. Rudstam, L. G., and B. M. Johnson. 1992. Development, evaluation, and transfer of new technology. Pages 507–523 in J. F. Kitchell, editor. Food web management: a case study of Lake Mendota. Springer-Verlag, New York, New York, USA. Rudstam, L. G., and J. J. Magnuson. 1985. Predicting the vertical distribution of fish populations: analysis of cisco, Coregonus artedii, and yellow perch, Perca flavescens. Canadian Journal of Fisheries and Aquatic Sciences 42:1178– 1188. Scheffer, M., S. H. Hosper, M.-L. Meijer, B. Moss, and E. Jeppesen. 1993. Alternative equilibria in shallow lakes. Trends in Ecology and Evolution 8:275–279. Shapiro, J., V. Lamarra, and M. Lynch. 1975. Biomanipulation: an ecosystem approach to lake restoration. Pages 85–96 in P. L. Brezonik and J. L. Fox, editors. Proceedings of a symposium on water quality management through biological control. University of Florida, Gainesville, Florida, USA. Shapiro, J., and D. I. Wright. 1984. Lake restoration by biomanipulation: Round Lake, Minnesota the first two years. Freshwater Biology 31:1–12. Sih, A. 1987. Prey refuges and predator–prey stability. Theoretical Population Biology 31:1–12. Simrad. 1996. Simrad EY500 portable scientific echosounder. Version 5.2. Instruction manual. Simrad, Horten, Norway. Smith, C. L. 1985. Inland fishes of New York. New York State Department of Environmental Conservation, Albany, New York, USA. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. Second edition. W. H. Freeman, New York, New York, USA. Stansfield, J. H., M. R. Perrow, L. D. Tench, A. J. D. Jowitt, and A. A. L. Taylor. 1997. Submerged macrophytes as refuges for grazing Cladocera against fish predation: observations on seasonal changes in relation to macrophyte cover and predation pressure. Hydrobiologia 342:229–240.
131
Stirling, D. G., D. J. McQueen, and M. R. S. Johannes. 1990. Vertical migration in Daphnia galeata mendotae (Brooks): demographic responses to changes in planktivore abundance. Canadian Journal Fisheries and Aquatic Sciences 47:395–400. Strickland, J. D. H., and T. R. Parsons. 1972. A practical handbook of seawater analysis. Bulletin of the Fisheries Research Board of Canada Number 167. Taggart, C. T. 1984. Hypolimnetic aeration and zooplankton distribution: a possible limitation to the restoration of coldwater fish populations. Canadian Journal of Fisheries and Aquatic Sciences 41:191–198. Talbot, A. J., and D. L. Kramer. 1986. Effects of food and oxygen availability on habitat selection by guppies in a laboratory environment. Canadian Journal of Zoology 64: 88–93. Tessier, A. J., and J. Welser. 1991. Cladoceran assemblages, seasonal succession and the importance of a hypolimnetic refuge. Freshwater Biology 25:85–93. Tinson, S., and J. Laybourn-Parry. 1985. The behavioural responses and tolerance of freshwater benthic cyclopoid copepods to hypoxia and anoxia. Hydrobiologia 127:257– 263. Torke, B. G. 1974. An illustrated guide to the identification of the planktonic crustacea of Lake Michigan, with notes on their ecology. Special Report Number 17. Center for Great Lakes Studies, University of Wisconsin, Milwaukee, Wisconsin, USA. Warner, D. W., L. G. Rudstam, and R. A. Klumb. 2002. In situ target strength of alewives in freshwater. Transactions of the American Fisheries Society 131:212–223. Welch, E. B., and G. D. Cooke. 1999. Effectiveness and longevity of phosphorus inactivation with alum. Journal of Lake and Reservoir Management 15:5–27. Wells, L. 1970. Effects of alewife predation on zooplankton populations in Lake Michigan. Limnology and Oceanography 15:559–565. Wetzel, R. G. 1983. Limnology. Second edition. Saunders College Publishing, Philadelphia, Pennsylvania, USA. Williamson, C. E. 1983. Behavioral interactions between a cyclopoid copepod predator and its prey. Journal of Plankton Research 5:701–711. Williamson, C. E. 1991. Copepoda. Pages 787–822 in J. H. Thorp and A. P. Covich, editors. Ecology and classification of North American freshwater invertebrates. Academic Press, New York, New York, USA. Wright, D., and J. Shapiro. 1990. Refuge availability: a key to understanding the summer disappearance of Daphnia. Freshwater Biology 24:43–62. Zaret, T. M., and J. S. Suffern. 1976. Vertical migration in zooplankton as a predator avoidance mechanism. Limnology and Oceanography 21:805–813.
