Behind the berms: comparing water quality and zooplankton communities between coves of varying connection to Harlan County Reservoir, Nebraska

Abstract Mason BE, Koupal KD, Wuellner MR, Kreitman JW. 2023. Behind the berms: comparing water quality and zooplankton communities between coves of varying connection to Harlan Country Reservoir, Nebraska. Lake Reserv Manage. 39:327–339. Coves are an important feature of reservoirs throughout the United States. Over time, sediment deposition and lateral drift can form a berm in the cove mouth that can restrict or eliminate exchange of surface water between the main reservoir and the coves. Little attention has been given to water quality and the zooplankton communities within these coves that provide unique habitats. This study compared several water quality parameters and the zooplankton communities between disconnected coves, connected coves, and the main reservoir of Harlan County Reservoir, Nebraska. Temperature, dissolved oxygen, pH, dissolved nitrates, and dissolved phosphates were similar between both cove types and the main reservoir. Parameters indicating water clarity and productivity were significantly different among the 3 habitats; disconnected coves had the highest turbidity and chlorophyll a (Chl-a) readings, while the main reservoir had the lowest. Similarly, Secchi depth was also lowest within disconnected coves and highest within the main reservoir. Zooplankton densities were highest in disconnected coves compared to connected coves or the main reservoir. Zooplankton communities in disconnected coves were largely dominated by rotifers, Ostracoda, Cyclopoida, Bosmina, Leptodora, and Ceriodaphnia, while the main reservoir had higher densities of Harpacticoida, Calanoida, and Daphnia. Interestingly, zooplankton communities were largely similar between connected coves and the main reservoir, indicating an influence of surface water exchange in these coves. As reservoirs in the United States continue to age, cove isolation is likely to become more common. Results from this study can help predict the potential changes that may occur as coves become disconnected.

Coves are common features within reservoirs and are typically formed by the flooding of former tributaries following dam construction (Miranda et al. 2014).These habitats may support more abundant and diverse fish communities (Gido and Matthews 2000), possibly owing to differences in water quality and primary and lower trophic productivity compared to the parent reservoir.For instance, previous research has found that reservoir cove habitats have more variable water temperatures (Marsh and Langhorst 1988, Slipke and Maceina 2005, Slipke et al. 2005); higher nitrogen, phosphorous, and chlorophyll a (Chl-a) concentrations (MDNR 2017); and higher densities of aquatic vegetation (Ferrer-Montaño and Dibble 2002), woody debris, and detritus (Matthews 1998).Such differences may subsequently lead to more diverse and abundant zooplankton communities compared to the corresponding main reservoir (O'Brien and de Noyelles 1974, Blancher 1984, Canfield and Jones 1996).While coves have previously been recognized for their ecological uniqueness, they remain understudied compared to mainstem reservoirs (Matthews 1998).
As reservoirs age, coves can become disconnected from the parent reservoir due to lateral drift of shoreline sediments (Marsh andLanghorst 1988, Mueller 1995) or by backfill of sediments from the main lake (Slipke et al. 2005, Slipke andMaceina 2007).These processes form a sediment berm at the cove mouth that may restrict water exchange with the main reservoir to high water events or groundwater infiltration (Slipke and Maceina 2005).Consequently, water quality within a cove may degrade over time, similar to that observed for disconnected oxbows (Miranda 2005) and floodplain lakes (Winemiller et al. 2000).Further, lateral movement of aquatic biota between coves and the main reservoir may also be limited (Slipke and Maceina 2005, Slipke et al. 2005, Allen 2007, Frisch et al. 2012).Eventually, aquatic community dynamics within coves and the main reservoir may be affected if isolation persists (Mason et al. 2022, Ruoss et al. 2023).Although the importance of water connection between coves and the main reservoir has been recognized, little research has documented differences in abiotic factors and biotic communities of coves with varying connection history to the mainstem reservoir (Miranda 2017).
Harlan County Reservoir, Nebraska, has numerous coves, several of which have developed sediment berms of various heights over time.Water quality parameters and zooplankton assemblages within the main waterbody of Harlan County Reservoir are well documented (Peterson et al. 2005, Maline et al. 2011, Olds et al. 2011, Olds et al. 2014); however, little attention has been given to cove habitats.The objective of this study was to compare water quality parameters and zooplankton assemblages between disconnected and connected coves and the mainstem of Harlan County Reservoir by season (spring, summer, and fall) for 2017 and 2018.We hypothesized that water quality and zooplankton communities in the main reservoir would be more similar to connected coves than disconnected coves.Further, we expected that water quality would be poorer and the zooplankton community would be dominated by more tolerant taxa in disconnected coves compared to connected coves and the main reservoir.

Study area
Harlan County Reservoir (surface area = 5400 ha at full conservation pool; storage capacity ~1 billion m 3 ) is located on the Republican River near the Nebraska and Kansas border (Figure 1).Dam construction was completed in 1952, and the dam is operated by the US Army Corps of Engineers primarily for flood control and irrigation purposes (USACE 2011).Mean depth of the reservoir at full conservation pool is 4 m and maximum depth is 18 m (Uphoff et al. 2013); however, water elevation can vary up to 3 m on an annual basis (Diffendal et al. 2002).
Seven coves located throughout Harlan County Reservoir were selected for this study based on their connection status (i.e., connected vs. disconnected; Figure 2).Cove connection was based on measurements of the sediment berms for each cove, taken by Flatwater Group, Incorporated, in 2017, compared to the main reservoir water elevation recorded on the dam spillway (USBOR 2020; Figure 3).Four coves (Bone, Gremlin, Patterson, and Prairie Dog) were classified as "connected" to the main reservoir and 3 coves (Indian Hill, Methodist, and Tipover) were classified as "disconnected" in this study.Disconnected coves were defined as those having <1 m of water connection at the berm saddle point (i.e., the lowest point atop the berm crest; Hanslow et al. 2000) at any time throughout the study.All disconnected coves had been isolated from the main reservoir for 4 yr or longer at the beginning of this study.Indian Hill, Methodist, and Tipover coves have been disconnected from the main reservoir since September 1993, February 2012, and June 2012, respectively, assuming sediment berm heights have remained consistent since their last connection event (Figure 3).All other coves were classified as connected and retained connection to the main reservoir via routine dredging or natural protection from erosion and sediment berm development.None of the coves included in this study changed in classification from their original connection designation over the course of this study.

Methods and materials
Sampling took place during spring (May), summer (Jul), and fall (Sep/Oct) in 2017 and 2018.Sampling locations within the mainstem reservoir were standardized and consistent with sites established as part of a long-term effort to capture the spatial and temporal variability of water quality and zooplankton (see Peterson et al. 2005, Olds et al. 2011; Figure 1).In short, 4 locations were randomly selected within 3 zones (Riverine, Transitional and Lacustrine upper, middle, and lower) and were sampled in both years.In addition, 4 locations within each of the 7 coves were randomly selected and standardized across both years to provide spatial representation of available conditions (Figure 2).
A suite of water quality metrics was measured at all sampling sites following protocols established by Olds et al. (2011).Water temperature (C) and dissolved oxygen (DO; mg/L) were measured at 1 m increments throughout the entire water column using a YSI Pro20 probe.Because Harlan County Reservoir does not maintain thermal stratification (Olds et al. 2011), temperature and DO readings were averaged among all depths at a given site for each sampling event.
Water transparency was indexed using a Secchi disk (cm).All remaining water quality parameters were measured from an integrated water column sample collected from a Van Dorn bottle sampler.Water samples were collected at 3 m increments, starting at 1 m below the surface.Relative Chl-a (relative florescence units; RFU) was measured using a Turner Design AquaFluor model 8000-010 fluorometer, and pH was measured using an Oakton series 11 pH meter.Turbidity (FAU), dissolved nitrates (mg/L), and dissolved phosphates (mg/L) were measured using a Hach model DR/870 colorimeter.Subsamples of water were filtered using a 1 µm syringe filter to eliminate suspended particles and improve accuracy in 2018.Thus, only data for 2018 were included in the analyses for nitrates and phosphates, as the 2017 samples reflected concentrations with suspended particles.
Zooplankton were collected in tandem with water quality at all main reservoir and cove sites during both years following previously established protocols for this reservoir (Peterson et al. 2005, Olds et al. 2014).A circular framed (0.5 m diameter) simple plankton net (80 μm mesh) was lowered to the bottom of the water column and pulled vertically to the surface.Captured zooplankton were stored in 95% ethanol and transported to the laboratory at the University of Nebraska at Kearney.Samples were diluted with tap water to a known volume and mixed to suspend zooplankton.Four subsamples of 1 mL were drawn using a Hensen-Stempel pipette and placed on a Ward (1955) counting wheel for identification and enumeration by taxa group; taxa groupings were consistent with previously published work for Harlan County Reservoir (Maline et al. 2011, Olds et al. 2014).All counts were averaged among the 4 subsamples.Tow depth and net circumference were used to calculate the volume of water filtered (L) for each sample.Densities of all taxa combined and for each taxonomic group were calculated by dividing the average count per volume of water sampled for each site (number/L).
Differences in water quality and total zooplankton density between connected and disconnected coves and the main reservoir by season were evaluated using general linear mixed models in SAS, Version 9.4.Models accounted for our study design by considering individual coves or zones of the reservoir as random variables nested within their classified type (i.e., connected or disconnected cove or main reservoir).In addition, we accounted for repeated measures within an individual cove or reservoir zone (n = 4) across seasons (n = 3) and years (n = 2) for all variables except for dissolved nitrates and phosphates.If differences were noted between connected and disconnected coves and the mainstem reservoir, a follow-up Tukey's test was used to determine which habitat type or types (i.e., connected coves, disconnected coves, or the main reservoir) differed.Significance was determined at α = 0.10.
Nonmetric multidimensional scaling (NMDS) using Bray-Curtis distance metrics was used to visualize differences in zooplankton taxa assemblage between connected and disconnected coves and the mainstem reservoir.A stress of <0.2 is considered suitable for interpreting ecological patterns (Clarke 1993), so the lowest number of axes with the stress of <0.2 was chosen for the final plot.A one-way analysis of similarity (ANOSIM, 999 permutations) was used to determine whether zooplankton communities differed between the main reservoir, connected coves, and disconnected coves (α = 0.05; Clarke 1993).Ordination and ANOSIM calculations were generated using the "vegan" and "MASS" packages in R (R Core Team 2020) version 4.0.0.
Total zooplankton density varied between 14 and 12,639 individuals/L (mean = 640 ± 108 individuals/L).Densities of zooplankton only differed between habitat types (F = df 2, P = 0.02; Figure 7) and not by season (F = 2.27, df = 2, P = 0.11).Total zooplankton densities were substantially higher in disconnected coves but were similar between connected coves and the main reservoir (Figure 7).Zooplankton community composition also appeared to differ between habitat types (Figures 8 and 9).Daphnia lumholtzi, Leptodora, and Harpacticoida were absent from disconnected coves and no Chydoridae or Leptodora were found in the main reservoir throughout this study.All other taxa were found in all 3 habitat types.All taxa collected were found in connected coves.Percent composition of taxa were more similar between the main reservoir and connected coves than between the main reservoir and disconnected coves (Figure 8).Calanoida and nauplii composed a smaller portion of the zooplankton community in disconnected coves compared to the main reservoir or connected coves, and Bosmina composed a larger proportion of the community in disconnected coves compared to the other 2 habitat types (Figure 8).Daphnia composed a smaller proportion of the community in connected and disconnected coves compared to the main reservoir (Figure 8).Rotifers composed a higher proportion of the community in both cove types compared to the main reservoir and encompassed more than half of all zooplankton sampled within disconnected coves (Figure 8).
Nonmetric multidimensional scaling ordination showed that zooplankton communities were distinct between the 3 habitat types (stress = 0.16; ANOSIM R = 0.39, P < 0.01; Figure 9).Sites within the main reservoir were grouped relatively closely in ordinal space, indicating consistency in zooplankton assemblage between locations (Figure 9).Main reservoir sites showed higher associations with Harpacticoida, Calanoida, and Diaphansoma, D. retrocurva, and D. lumholtzi (Figure 9).Disconnected coves, conversely, had larger distribution in ordinal space compared to the main reservoir and connected coves, indicating more variable zooplankton assemblages.Dominant taxa in disconnected coves nauplii, Ostracoda, Cyclopoida, Bosmina, Leptodora, and D. pulicaria (Figure 9).No overlap of convex hulls in ordinal space occurred between the main reservoir and disconnected coves, indicating distinct assemblages within each habitat (Figure 9).Connected coves also had greater variability in assemblages compared to the main reservoir, but not as much as disconnected coves (Figure 9).The convex hull of connected coves overlapped entirely with that of the main reservoir and  partially disconnected coves, suggesting some degree of similarity 9).

Discussion
Several water quality parameters important for biological activity were found not to differ between habitat types within our study.Interestingly, these findings are in contradiction to other studies regarding temperature in isolated reservoir coves.Water temperatures within disconnected coves in Lake Mohave, Arizona, and Demopolis Reservoir, Alabama, function somewhat independently from the main reservoir (i.e., disconnected coves were warmer during the spring and summer and cooler in the fall and winter compared to main reservoir; Marsh andLanghorst 1988, Slipke et al. 2005).Thermal differences noted in other studies could be attributed to coves being shallower and smaller and requiring less energy to decrease or increase temperature compared to their parent reservoir (Wetzel 2001a).Why these patterns for water temperature did not occur in this study is unclear.One potential explanation is that Harlan County Reservoir is relatively shallow (mean depth of 4 m  at conservation pool) compared to other reservoirs, high wind events that limit of a thermocline (Olds et al. 2011).The synchronization of temperatures between habitats could also be driven by unique dynamics of groundwater exchange between disconnected coves and the main reservoir in Harlan County Reservoir.Ground and surface water exchange may also explain the relative consistency of pH, dissolved nitrates, and dissolved phosphate between disconnected coves, connected coves, and the main reservoir found in this study.Alternatively, these variables may also simply be regulated similarly by other processes not unique to any particular habitat type, such as nutrient uptake by plants and algae (Kennedy andWalker 1990, Mosley 2015) or minerals of the parental substrate buffering the pH of the water (Langmuir 1997).
All habitats in this study also had similar DO levels across all seasons and years sampled in this study, and most measurements of DO were within values acceptable for most warm-and coolwater fish species (Doudoroff and Shumway 1970).No sampling, however, was completed during the winter, and ice cover could lead to lower concentrations and, subsequently, localized winterkill events (Cole andHannah 1990, Stefan andFang 1997).Adequate DO concentrations are critical for nearly all aquatic organisms, and reductions of DO can stress fish, aquatic amphibians, and aquatic invertebrates, potentially leading to mortality (Rottmann et al. 1992).Furthermore, because disconnected coves are isolated waterbodies, aquatic organisms may have limited opportunities to seek refugia from low DO concentrations if they occur.This limiting factor could impact the biotic communities of disconnected coves, even if occurring for only a short timeframe, promoting species assemblages that are tolerant of low DO concentrations (Mason et al. 2022).
Unlike water temperature, pH, dissolved nutrients, and DO, water parameters indicating water clarity and productivity were different between habitat types.SDT was shallowest within disconnected coves, while the main reservoir had the deepest recorded depth, which likely influenced the higher and lower turbidity readings, respectively, observed in our study (Bachmann et al. 2017).Differences in water clarity between the 3 habitats could be related to water depth.Shallow habitats are often more vulnerable to wind-driven sediment resuspension, potentially contributing to increased turbidity and lower SDT readings, even with little wind and mild turbulence (Miranda 2005, Knight et al. 2008).Within Harlan County Reservoir, disconnected coves varied little in depth (all ~1 m deep).In contrast, depths of connected coves and main reservoir sites varied between 1 and 5 m, and between 3 and 12 m, respectively.
Primary productivity within each habitat may also influence water clarity, as suspended algae in the water column can result in increased turbidity (Wetzel 2001b).As available Chl-a levels within habitats showed trends similar to the turbidity (high in disconnected coves, and low in the main reservoir), perhaps the observed turbidity in Harlan County Reservoir results from high biogenic turbidity.The exchange of surface water between connected coves and the main reservoir could help explain the intermediate water clarity and productivity values observed within connected coves, compared to disconnected coves and the main reservoir.Similar to turbidity, differences in Chl-a between habitats in our study could also be related to water depth, but due to different processes.Because Chl-a samples were taken via an integrated water sample of the water column, samples from sites with deeper depths (i.e., main reservoir and some connected cove sites) could be diluted, as most of the primary production occurs in the first few meters (Kimmel et al. 1990).
Similar to some water quality parameters, densities of total zooplankton differed between the disconnected coves and the other 2 habitat types in this study.Marsh and Langhorst (1988) also observed higher densities of zooplankton in disconnected coves compared to the main reservoir within Lake Mohave, Arizona.Differences in abundance of zooplankton could be related to primary production within each habitat (Canfield and Watkins 1984, Dodson 1992, Canfield and Jones 1996, Shuter and Ing 1997).Large differences in zooplankton densities between habitats could also be related to water depth.Maline et al. (2011) found that zooplankton densities within Harlan County Reservoir in May were higher at 1 m depth, compared to those taken deeper within the water column.If the majority of zooplankton are the surface grazing where most photoactivity (Adams et al. 1974, Maline et al. 2011) and habitats differ in water column depth, then zooplankton densities from the main reservoir and connected coves could be diluted compared to shallower disconnected coves.
The different connection regimes of coves could also be driving differences in zooplankton communities.The community similarities between connected coves and the main reservoir likely indicate that the assemblages in these habitats are affected by the interchange of water and organisms.Disconnected coves, conversely, are secluded and have little direct influence from the main reservoir, likely producing the unique assemblages within.In addition, differences in water quality parameters could contribute to differences in zooplankton assemblages within disconnected coves.For example, Daphnia spp.have limited tolerance to suspended solids and high turbidity (Arruda 1983, McCabe andO'Brien 1983), possibly explaining the lower abundances of Daphnia spp.found in our study within connected and disconnected coves.Rotifers, however, are more tolerant of turbid waters (Bernot et al. 2004) and were found in higher abundance within both cove habitats of Harlan County Reservoir.High rotifer abundance was also observed by Marsh and Langhorst (1988) within Lake Mohave, Arizona, where rotifers were 50% more abundant within disconnected coves compared to the main reservoir.Low DO levels have also been shown to influence zooplankton communities.Rotifers are relatively tolerant to low DO concentrations compared to crustacean zooplankton; thus, rotifers are often more abundant in oxygen-stressed systems (Karpowicz et al. 2020).Although DO levels were relatively similar between habitat types in this study, the high abundance of rotifers within disconnected coves may indicate these habitats undergo occasional periods with limited oxygen availability, allowing the more tolerant taxa to become dominant.Lower abundances of some taxa such as Daphnia spp.can also influence densities of other taxa such as rotifers and Bosmina due to decreases in competition (DeMott andKerfoot 1982, Wolfinbarger 1999).These changes and shifts in community structure within isolated habitats would likely continue over time, particularly as isolation is maintained longer.
As more reservoirs reach the end of their original life expectancy and increase in functional age, changes in reservoir features related to senescence will likely become more common (Miranda and Krogman 2015).As part of the senescence process, cove disconnection within reservoirs will be a reoccurring issue for water and fisheries managers.Little research to date on the ecological change that may occur within reservoir coves due to disconnection over time has been completed, but research on other disconnected aquatic habitats may provide some insights.For example, oxbow lakes have been found to be warmer, more turbid, and have lower DO concentrations as they are disconnected for longer times from rivers (Miranda 2005).If reductions in water quality are persistent, intolerant taxa in aquatic communities within disconnected coves may become rare or locally extirpated (Karpowicz et al. 2020, Mason et al. 2022, Ruoss et al. 2023).In order to improve water quality, reconnection via cove renovation or other means may be necessary.Cove reconnection could also benefit the reservoir fish community by increasing available spawning habitat and providing unique zooplankton assemblages as a food resource.Alternatively, keeping coves disconnected from the main reservoir could have positive outcomes such as by creating a buffer between nutrient runoff and the main waterbody and containing algal blooms (Izydorczyk et al. 2008) or through supporting greater system diversity by supporting unique fish (Mason et al. 2022) and zooplankton assemblages.Reservoir managers planning cove renovations will need to weigh the costs and benefits of reconnecting isolated coves to improve access to the main reservoir and cove water quality, versus maintaining disconnection to preserving the unique ecological features.

Figure 1 .
Figure 1.Map of Harlan County Reservoir, nebraska (adapted from aerial imagery taken by the Usda-nRCs, 13 July 2016; surface water elevation ~591 m.a.s.l.) and surrounding towns.Reservoir zones were previously established by Peterson et al. (2005).Black triangles represent sampling stations within each zone.

Figure 2 .
Figure 2. Map of coves within Harlan County Reservoir, nebraska (adapted from aerial imagery taken by the Usda-nRCs, 13 July 2016; surface water elevation ~591 m.a.s.l.).Black triangles represent sampling stations within each cove.

Figure 3 .
Figure 3. elevation of the sediment berms disconnecting coves from the main reservoir, compared to end-of-month water level elevation of Harlan County Reservoir since January 1990.the gray line indicates the water elevation of the main reservoir, recorded at the dam spillway (UsBoR 2020).Horizontal black lines indicate minimum water level required for connection (heights of sediment berm plus 1 m) for each corresponding cove.vertical dotted lines indicate 1 January of the year labeled below.initial berm height was calculated by Flatwater group, inc., 2017.

Figure 4 .
Figure 4. Mean (a) water temperature, (B) do, and (C) pH across different habitat types in Harlan County Reservoir, nebraska.error bars denote 1 standard error.

Figure 5 .
Figure 5. Mean (a) sdt, (B) turbidity, and (C) relative Chl-a across different habitat types in Harlan County Reservoir, nebraska.the numerical axis for sdt has been inverted to better represent depth within the water column from the water surface.error bars on all graphs denote one standard error.Lowercase letters denote significant differences based on a tukey test (α = 0.10).

Figure 6 .
Figure 6.Mean (a) dissolved nitrates and (B) dissolved phosphates across different habitat types in Harlan County Reservoir, nebraska.error bars on all graphs denote one standard error.

Figure 7 .
Figure 7. Mean total zooplankton density across different habitat types in Harlan County Reservoir, nebraska.error bars on all graphs denote one standard error.Lowercase letters denote significant differences based on a tukey test (α = 0.10).

Figure 8 .
Figure 8.Comparison of the percent composition of zooplankton taxa groups within different habitat types (main reservoir, connected coves, and disconnected coves) in Harlan County Reservoir, nebraska.Percent composition was calculated based on the total density of zooplankton across locations of those habitat types and season.Zooplankton taxa groups included in this figure consist of those with ≥1% of the total density of zooplankton for at least one of the 3 habitat types.