Seasonal and diel patchiness of a Daphnia population: An acoustic analysis

Detailed information about the location and extent of zooplankton patches is fundamental to understand how abiotic and biotic forces interact to structure the spatial distribution of zooplankton. We mapped zooplankton patchiness in a Minnesota lake during spring, summer, and autumn with high-frequency (192-kHz) single-beam sonar. Conventional plankton samples of aggregations detected acoustically revealed that Daphnia pulicaria (mean body length 1.6 mm, mean target strength 2120 dB) scattered most (;63%) of the sound. Other taxa were smaller (,1⁄2 the length of D. pulicaria) and were usually less abundant and therefore scattered much less sound than D. pulicaria. Our acoustic estimates of Daphnia concentrations illustrate extreme patchiness, with concentrations varying by as much as four orders of magnitude over vertical distances of less than 1 m. Seasonal patterns of patchiness were related to predation by rainbow trout and to abiotic factors associated with stratification. Daphnia concentrations were highest from June to October in a deep-water ‘‘refuge zone’’ where oxygen concentrations were between 3 and 5 mg L21. These oxygen levels are suitable for Daphnia but are lower than those required by rainbow trout. Heterogeneity in Daphnia concentration along the lake’s long axis was highest in May and June, when the population resided primarily in the oxic hypolimnion during the daytime. From July to October, as oxygen concentrations declined in the hypolimnion, the population became more metalimnetic and more uniformly distributed in the horizontal dimension. A diel study of the population in October indicated that the patchiness of population also changed dramatically between day and night. During the day the population aggregated densely in a thin layer (;2 m thick) in the thermocline. After sunset the population dispersed into the epilimnion, where concentrations were ;100,000 m23 less than they were during the day in the thermocline. Patchy spatial distributions and the low resolution of conventional sampling methods have impeded analyses of zooplankton populations. Zooplankton concentrations have been shown to vary by a factor of 1,000 within distances of meters horizontally or vertically, and the sampling resolution of conventional plankton nets and hoses is usually too coarse to identify the spatial limits of aggregations precisely (Coyle 2000). Variation in population density estimates due to sampling often cannot be distinguished from real changes of population density, and the effects of biological processes are difficult to distinguish from those of advective transport (Megard et al. 1997). Acoustic and optical plankton samplers developed during recent decades are major advances. They have very high sampling rates and spatial resolution, comparable to modern instruments (e.g., CTD profilers) used to measure environmental variables. Large numbers of plank1 Present address: Biology Department, Hamline University, 1536 Hewitt Avenue, St. Paul, Minnesota 55104. Acknowledgments This research was financed by the Dayton and Wilkie Funds, the Minnesota Center for Community Genetics, the Itasca Biological Station Program, and the Minnesota Sea Grant College Program supported by the NOAA Office of Sea Grant, United States Department of Commerce, under grant NA46RG0101. The U.S. Government is authorized to reproduce and distribute reprints for government purposes, not withstanding any copyright notation that may appear hereon. This paper is journal reprint JR 491 of the Minnesota Sea Grant College Program. LKH was supported by a fellowship from the Minnesota Center for Limnology during the writing of this manuscript. We thank R. Sterner, R. Newman, D. Alstad, and two anonymous reviewers for constructive comments on earlier versions of this manuscript. ton samples can be obtained rapidly from large geographic areas. The new technology dramatically improves the ability to locate and describe aggregations and to identify how aggregations are related to physical variables (e.g., Ross et al. 1996; Megard et al. 1997; Zhou et al. 2001). High-frequency sonar (;200 kHz) can detect sound-scatterers in the size range of lake zooplankton. While sonar by itself cannot discern the biological identities of sound-scatterers in aggregations, it can be used in conjunction with plankton nets to identify scatterers and to gain other information that neither method can provide alone. A large number of acoustic samples can be acquired and displayed rapidly to delineate aggregations at high precision with respect to depth and geographic coordinates. Surveys along transects, therefore, can detect and record major features of spatial distribution that are invisible with conventional sampling. Because sonar data can be displayed instantaneously, they also can identify sites and depth increments for more intensive sampling with plankton nets and other devices. Zooplankton investigations are no longer constrained by limits imposed by conventional ‘‘blind sampling’’ with nets at localities and depths selected arbitrarily. Sonar may be used as a tool to assess the relative importance of abiotic versus biotic drivers of zooplankton patchiness (Zhou 1994; Pinel-Alloul 1995; Folt and Burns 1999). Abiotic factors that influence the spatial distribution of zooplankton include water movements (e.g., upwelling, Megard et al. 1997; Langmuir circulation, George and Edwards 1973) and thermal stratification (Pinel-Alloul et al. 1988; Pinel-Alloul and Pont 1991). Biological forces/activities that cause zooplankton to be distributed patchily include diel vertical migration (Young and Watt 1996), predation avoidance 2222 Hembre and Megard (Pijankowska and Kowalczewski 1997), searching for food (Tiselius 1992), and locating mates (Strickler 1998). Patchiness is ecologically significant because it influences the interactions among individuals and populations. For example, extremely dense patches of zooplankton exert intense grazing pressure on phytoplankton in localized areas (Hembre 2002), and competition for food among individuals in a densely aggregated population will be stronger than in a population that is more evenly distributed. Also, interactions between predators (e.g., fish) and their zooplankton prey will be greatly affected by the ability of the fish to locate patchily distributed zooplankton aggregations. Here we use single-beam, high-frequency sonar to investigate the properties of a population of a Daphnia pulicaria, a planktonic cladoceran (in the D. pulex group) that congregates below the mixed layer of a Minnesota lake (Long Lake) in the daytime during summer stratification. With a body length up to 3 mm, D. pulicaria is larger than most other zooplankton. The properties of D. pulicaria populations are of special interest because they and other largebodied cladocerans play a central role in lacustrine food webs. Large-bodied Daphnia feed efficiently on phytoplankton and can cause a ‘‘clear-water phase’’ in many lakes during late spring and early summer (e.g., Lampert et al. 1986; Luecke et al. 1990; Wright and Shapiro 1990; Hembre 2002). They are also important for fisheries management because they are often the preferred prey for size-selective planktivorous fish. The dominant planktivore in Long Lake is rainbow trout (Oncorhynchus mykiss), which is stocked annually by the Minnesota Department of Natural Resources (MDNR). Rainbow trout require cold water (,218C), and abundant oxygen (.5 mg L21; Wang et al. 1996), which largely restricts them to foraging in the metalimnion during summer stratification. Because D. pulicaria migrate below the mixed layer during the daytime, there is substantial overlap in their habitat space with trout. Because of their spatial proximity, D. pulicaria comprise most of the trout diet during the summer (Hembre 2002), as they do in other lakes stocked with rainbow trout (e.g., Bevelhimer and Adams 1993; Geist et al. 1993; Wang et al. 1996). Previous studies (Hembre 1996; Ross et al. 1996) of D. pulicaria in Long Lake used acoustic methods qualitatively to locate aggregations for sampling and to provide information about the overall spatial distribution of the population. To date, few acoustic studies have analyzed freshwater zooplankton populations quantitatively (exceptions: Rudstam et al. 1992; Melnik et al. 1993; Megard et al. 1997; Gal et al. 1999), and none have analyzed the target strength of Daphnia. The research presented here (1) shows the spatial variation in backscattered sound over a season, (2) determines the relative contributions by D. pulicaria and smaller zooplankton to the total strength of backscattered sound as well as the target strength D. pulicaria, and (3) identifies how abiotic and biotic factors interact to shape the spatial distribution of the population. Methods Study site—Long Lake is a dimictic, oligoto mesotrophic lake located in northwestern Minnesota (latitude 478179N, longitude 958179W). The lake has a single basin, is 2.4 km long, and has a surface area of 66.5 ha and a volume of 7.63 3 106 m3. The basin is symmetrical, relatively deep (maximum depth 5 24 m, mean depth 5 13 m), and has a small littoral zone (;15% of lake surface area) (Schmid 1965). Its depth and simple morphometry make it ideal for acoustic analysis, and because interactions in the pelagic zone dominate the ecology of the lake, an understanding of the spatial distribution and abundance of Daphnia is especially relevant. Fisheries management—Long Lake has been stocked with rainbow trout by the MDNR since 1961. Rainbow trout require streams with current-washed gravel to spawn. Such streams are not available to the trout in Long Lake, so trout abundance is not affected by reproduction. Instead, their abundance depends on the number stocked, natural mortality, and fishing mortality. Acoustic estimates of trout abundance indicate that fewer than 10% of the stocked trout remain in the lake after 12 months (Hembre 2002). In 1998, the MDNR switched from autumn to spring stocking of trout in Long Lake. So, on the first sampling date in 1998 (22 April), the lake had not been stocked with trout for 17 months (since November of 1996), and the acoustic estimate of trout density was very low (2.4 ha21). On 23 April 1998, 14,500 yearling rainbow trout were stocked (density 5 218 ha21). Thereafter, the density of trout decreased linearly from 191 ha21 on 25 May to 30 ha21 on 24 October (Hembre 2002); representing 80 to 12 times the density of trout present on 22 April. Instrument design and capabilities—We used a sonar system described earlier (Megard et al. 1997) to sample zooplankton in Long Lake at high resolution. The system consists of a Lowrance X-16 high-frequency (192-kHz) single-beam echosounder and a Loran-C navigation receiver connected to a portable computer. A narrow-beam transducer (48 half angle), directed vertically and suspended from the bow of the boat ;1⁄2 m below the lake surface emitted 100 ms acoustic pulses (often called ‘‘pings’’) at a rate of approximately one pulse per second. An analog-digital converter in the computer digitizes voltage variation due to sound back-scattered by zooplankton, fish, and other particles in 2,000 50-ms increments (bins), which correspond to depth increments of ;4 cm (28 acoustic samples m21). Backscattered sound from a near-field ‘‘dead zone’’ less than 1.2 m from the transducer is ignored. The software calculates the strength of backscattered sound from the digitized signal strength in terms of volume scattering strength, using the sonar equations (Urick 1983) to compensate for transmission losses due to beam-spreading as sound travels from the transducer to scatterers and returns to the transducer. Volume scattering strengths, calculated with reference to a standard tungsten–carbide sphere (Foote and MacLennan 1984), are displayed in the format of an echogram on the computer monitor and saved on the hard 2223 Acoustic analysis of Daphnia Fig. 1. Layers of sound scatterers detected while anchored at a sampling station (19 August 1999). Colors on echogram correspond to strength of backscattered sound (decibels, dB). Blue indicates weak backscattering, and orange–red indicates strong backscattering, as indicated by decibel values on the color scale. To calibrate the sonar information, sound scattered from depth increments (indicated by white rectangles) was compared with zooplankton densities calculated from plankton net samples. disk of the computer. Depths to be sampled with nets or other devices can be selected efficiently with respect to the locations of sound-scatterers, because echograms are displayed instantaneously. Volume backscattering strength, expressed in terms of decibels, is