Metalimnetic Oxygen Depletion: Organic Carbon Flux and Crustacean Zooplankton Distribution in a Quarry Embayment

Particulate organic carbon (POC) flux and the distribution and abundance of crustacean zooplankton and bacteria associated with formation of a metalimnetic oxygen minimum were examined in a deep embayment of Kentucky Lake, Kentucky. POC measurements from sediment traps placed above and below the metalimnion yielded an estimate of the organic material that was metabolized in the metalimnion. This estimate was the molar equivalent of the oxygen that was depleted from the metalimnion. Calculated zooplankton respiration accounted for 26-31% of the observed oxygen loss, except in midsummer when it accounted for 15%. Estimated bacterial respiration accounted for >44% of the observed oxygen loss. The comparison of calculated oxygen demand with observed oxygen loss emphasizes the importance of in situ processes as the cause of the minimum and suggests that metalimnetic deficits may be useful to estimate productivity. The vertical distribution of three species of Daphnia changed as the oxygen minimum formed. Daphnia pulex became entirely hypolimnetic. Thus, changes in chemical structure influence spatial distribution of zooplankton species. Disappearance of oxygen from deep, dark layers of productive thermally stratified lakes is one of the classical dogmata of limnological knowledge (Birge & Juday, 1911). Under homothermal conditions, wind mixing keeps all depths oxygenated through photosynthetic oxygen production in the euphotic zone and atmospheric invasion at the surface. Organic matter, synthesized in the upper lighted layers, is decomposed by bacteria as it sinks, using dissolved oxygen (Henrici, 1939). When mixing is prevented by the thermal/density This study was supported by the Center for Reservoir Research and conducted at the Hancock Biological Station, Murray State University, Murray, Kentucky, U.S.A. We gratefully acknowledge the efforts of Gary Rice for field assistance and Jennifer Burch for zooplankton enumeration. Reviews of the manuscript by Drs. Alan W. Groeger, Michael L. Mathis, and David S. White are appreciated. Contribution no. 18 of the Center for Reservoir Research. TRANS. AM. MICROSC. Soc., 113(2): 105-116. 1994. ? Copyright, 1994, by the American Microscopical Society, Inc. This content downloaded from 207.46.13.193 on Thu, 08 Sep 2016 04:38:24 UTC All use subject to http://about.jstor.org/terms TRANS. AM. MICROSC. SOC. barrier that defines stratification, deep waters are no longer oxygenated, and the net respiratory losses result in oxygen depletion. Disappearance of oxygen from only the metalimnion is one of several variants of this phenomenon. The metalimnetic oxygen minimum, or negative heterograde oxygen profile (Hutchinson, 1957), is characteristic of productive lakes with steep-walled basins and voluminous hypolimnia. These conditions seem to be met often in river impoundments (Cole & Hannan, 1990). In the situation described here, dense metalimnetic populations of crustacean zooplankton were observed, suggesting that animal respiration might contribute significantly to the metalimnetic oxygen loss (Baker et al., 1977; Mindler, 1923; Patalas, 1963; Shapiro, 1960). We felt that if the sinking rate of particulate organic carbon slowed as it reached the density barrier of the upper metalimnion, then either POC would accumulate, or the particulate organic resources for bacteria and for crustacean filter feeders (consuming both POC and bacteria) would be enriched, defining a metabolically active layer that might favor this spatially dramatic oxygen depletion. Thus, our objectives were (1) to measure the flux of particulate organic carbon through the metalimnion in order to compare the POC loss with oxygen loss, (2) to estimate the relative contribution of bacteria and zooplankton to the metalimnetic oxygen depletion, and (3) to document the movements of crustacean zooplankton before and after the formation of the metalimnetic oxygen minimum. DESCRIPTION OF STUDY AREA Pisgah Quarry is a rectangular (ca. 4.3 ha), 33-m deep embayment of Kentucky Lake located approximately 21 km upstream from Kentucky Dam. A metalimnetic oxygen minimum has been observed here annually since 1977 (J. Sickel, personal communication). The basin was quarried during the construction of Kentucky Dam and inundated in 1944 when the reservoir filled. The embayment is isolated from the main portion of Kentucky Lake by a narrow, shallow (2 m) inlet. The quarry walls are vertical on three sides and steep on the fourth. MATERIALS AND METHODS The sampling schedule was designed to extend from before the onset of thermal stratification in the spring (mid-March) until after the fall mixing (late October) in 1989. Sampling intervals were approximately three weeks during this period. The intervals used for calculation of oxygen depletion and POC flux were 7 April-i May (I), 1 May-23 May (II), 23 May-15 June (III), and 15 June-12 July (IV). All samples were taken from a site near the center of the quarry. Temperature and oxygen profiles were measured electrochemically at 1-m intervals (Hydrolab Surveyor II). Light was measured at 1-m intervals (Li-Cor, model LI-185B). Sediment traps, a cluster of four PVC pipes (70 x 7.5 cm) closed at the bottom and open at the top (Hakanson & Jansson, 1983), were suspended with open ends at the top (6 m) and bottom (12 m) of the metalimnetic layer to 106 This content downloaded from 207.46.13.193 on Thu, 08 Sep 2016 04:38:24 UTC All use subject to http://about.jstor.org/terms VOL. 113, NO. 2, APRIL 1994 collect POC settling into and out of the metalimnion. A dense layer of 100 ml of 10% formalin in 50% NaCl was added to the bottom of each trap to arrest decomposition. When the traps were retrieved after each interval, the water in the traps was decanted and the sediment in the preserving layer removed. The sediment sample volume was brought to 1 L with distilled water. The POC content (carbohydrate) was measured by filtering 50-ml subsamples on glass-fiber filters (Whatman GF/F) and oxidizing the POC with dichromate (Strickland & Parsons, 1968). The difference between POC in upper and lower traps represents the amount of POC lost to decomposition in the 6-m layer between the traps. Results are expressed as ALtg POC/cm2/d. The molar equivalent of oxygen represented by the POC then was compared with the change in oxygen concentration during the same period. Bacterial enumeration was performed on water collected at 3-m depth intervals in a 2.2 L van Dorn-style water bottle. A single 25-ml subsample was taken from each sample depth and preserved in the field with 4% filtered, CaCO3 buffered formalin. One one-ml subsample was filtered (0.2 ,m), stained with 4'6-diamidino-2-phenylindole (DAPI), and counted with UV epifluorescence microscopy (Porter & Feig, 1980). One 500-ml subsample was filtered for chlorophyll-A analysis by extraction in 90% acetone (Clesceri et al., 1989). Three replicate zooplankton samples were collected at 3-m intervals on each sampling day with a 15-L Schindler plankton trap fitted with a 63-,im Nitex? sieve bucket (Schindler, 1969). Vertical series were collected at 4-h intervals from noon to 0800 h on 1-2 May and 12-13 July to determine diurnal distribution patterns of zooplankton in relation to the metalimnion before and after the formation of the oxygen minimum. Zooplankters were stored in 70-ml polystyrene tissue-culture flasks and preserved in 3% formalin. Crustacean zooplankters were enumerated without subsampling at magnifications of 50100 x. Respiratory oxygen consumption during each interval was estimated by calculations from published respiration rates of various zooplankton groups (Chaston, 1969; Comita, 1968; Ivanova, 1970; Kibby, 1971; Moshiri et al., 1969; Richman, 1958). The bacterial respiration rate was estimated from data of Kusnetzow & Karsinkin (1931). Respiration rates expressed as carbon were converted to oxygen consumed by applying the mean oxy-caloric coefficient of Winberg et al. (1934) (1 ml O2/mg carbon) and the energy to carbon relation for aquatic invertebrates (10.98 cal/mg carbon) derived by Salonen et al. (1976). Oxygen consumption estimates were converted to mg of oxygen per sampling interval for each of the four intervals. When specific respiration rates were not available, calculations were made using values for the closest allied taxon. RESULTS Analyses presented here are based on the period from 7 April, after stratification was well established, to 12 July, when dissolved oxygen in the metalimnion reached a minimal concentration (Fig. 1). Temperature and oxygen. Thermal stratification was first observed between depths of 7 and 10 m on 24 March. The vertical dimension of the metalimnion 107 This content downloaded from 207.46.13.193 on Thu, 08 Sep 2016 04:38:24 UTC All use subject to http://about.jstor.org/terms TRANS. AM. MICROSC. SOC.