Assessing the Impact of Algae Management Strategies on Anurans and Aquatic Communities

Residential areas are increasing on the landscape but their ability to provide suitable habitat is often based on management for recreational use and aesthetics. Amphibians rely on both aquatic and terrestrial habitat making them susceptible to changes in land-cover and land-use. As anthropogenic land-use change increases, it is imperative to assess how pond management practices impact aquatic communities. We assessed the impact of Aquashade (a common non-toxic pond dye) and copper sulfate (a toxic algaecide) on American toad (Anaxyrus americanus), northern leopard frog (Lithobates pipiens), and Cope’s gray treefrog (Hyla chrysoscelis) metamorphosis in outdoor mesocosm experiments. We also evaluated the relative impact of tadpole grazing and chemical treatment on phytoplankton and periphyton abundance. We found no signicant effects of pond management treatment on anuran metamorphosis, suggesting that addition of Aquashade and copper sulfate does not signicantly impact anurans under these experimental conditions. However, while we found no differences in phytoplankton and periphyton abundance due to pond management treatment, presence of tadpoles signicantly decreased phytoplankton and periphyton abundance over time. This result suggests that the creation of suitable pond habitat for anuran tadpoles may be an ecient and ecologically benecial form of pond management treatment to maintain water quality.


Introduction
Loss of wetlands poses a major threat to aquatic species; however, local creation of ponds and wetlands has been generally increasing in residential areas having the potential to restore habitat to maintain native biodiversity. Residential development has increased with human population growth further extending the impacts of anthropogenic change on the landscape (McKee et al. 2003;Theobald 2004;Ewers and Didham 2006). For instance, Smith et al. (2002) found that at least 2.6 million small, humanmade ponds have been created in the conterminous U.S., and these impoundments are a major sink for sediments (Renwick et al. 2005). Nevertheless, the management of new pond habitat can vary drastically based on individual preferences in aesthetics, recreational use, and cost (Blaine et al. 2012), positively or negatively altering habitat quality and biodiversity potential (Hansen 2005;Cushman 2006;Urban and Roehm 2018). Given that residential land-use is increasing, it is imperative to determine the impact of management practices on aquatic communities and local species that inhabit them (Locke et al. 2019).
Large-lot residential areas are desirable for landowners looking for residential space with increased exposure to wildlife and recreational area Blaine et al. 2012;Robinson 2012). While residential development increases habitat fragmentation, it can alternatively improve habitat suitability for native ora and fauna through pond creation and increased habitat heterogeneity (Urban and Roehm 2018). However, the habitat suitability of large lot residential lands will depend on the management practices of individual landowners. For instance, Goldberg and Waits (2009), found that over 50% of landowners surveyed stocked their ponds with sh for recreational use, which would alter aquatic food webs and eliminate or reduce diversity of taxa like amphibians that are sensitive to sh predation (Walston and Mullin 2007;Vonesh et al. 2009; Ade et al. 2010). Furthermore, landowners may choose to use chemical management to decrease aquatic vegetation and improve lawn aesthetics even though this may increase nitrogen and phosphorus loads on the landscape and negatively impact aquatic ecosystems (Loman and Lardner 2006;Peltzer et al. 2008;Carey et al. 2013;Martini et al. 2015). Blaine et al. (2012) found that 73% of surveyed landowners use chemical management and moreover the amount of chemicals applied is higher in residential areas compared to agriculture counterparts. Given the increase in residential areas it is imperative to determine the impact of common residential chemical management on aquatic communities.
Aquashade and copper sulfate are two common chemicals used to manage residential ponds suffering from aquatic plant overgrowth (Lynch 2009). Copper sulfate has historically been used to directly decrease algal growth, although it is known to be toxic to other aquatic organisms (Lynch 2009;Garcia Munoz et al. 2009;U.S. EPA. 2013). Two studies on northern leopard frog (Rana pipiens) tadpoles, found that when exposed to copper sulfate concentrations > 0.15 ppm, mortality signi cantly increased (Lande and Guttman 1973;Chen et al. 2007). Additionally, when exposed to > 0.20 ppm copper sulfate, Garcia Munoz et al. (2009) found Natterjack toad (Epidalea calamita) tadpoles had signi cantly decreased growth compared to control toads, but the timing of copper sulfate application mattered. For example, larval anurans exposed at Gosner stage 25 had increased time to metamorphosis and exhibited signi cant physiological impairments such as edema, resulting in high mortality (Carvalho and Fernandes 2006;Chen et al. 2007;Gurkan and Hayretdag 2012).
Aquashade is a non-toxic, pond dye that blocks photosynthetic light and indirectly controls aquatic plant growth (U.S. EPA. 2005;Suski et al. 2018). Though Aquashade was non-toxic when applied to an entire lake ecosystem at 1.5 ppm, Aquashade decreased aquatic photic depth by 50%, phytoplankton biomass by 60%, and zooplankton biomass by 55% (Batt et al. 2015). Contrastingly, Aquashade applied at the recommended dose (2.0 ppm) in a small-scale mesocosm experiment resulted in no change in phytoplankton biomass and an increase in zooplankton abundance; however, species composition became less diverse with Aquashade exposure (Suski et al. 2018). Ludwig et al. (2010) found Aquashade applied at twice the recommended dose (4.0 ppm) had no direct impact on several sh species' ngerling abundance, weight, or survival. While these results suggest that Aquashade could be an effective product for aquatic plant management, information about effects to other non-target aquatic organisms, including amphibians, is lacking and needs further investigation.
Phytoplankton and periphyton are essential in bottom-up aquatic ecosystem processes; therefore, it is important to understand how various management strategies impact aquatic resources and communities (Mallory and Richardson 2005;Rowland 2017). While herbicides are often chosen to decrease aquatic plant growth, they can indirectly and negatively impact larval anurans by limiting their primary food resource (Loman 2001;Boone and James 2003;Relyea 2005). Native biological control offers another option for controlling overproduction of algae. Because tadpoles graze on both phytoplankton and periphyton (Dickman 1968;Loman 2001;Pryor 2003), tadpoles could provide a natural form of algal bioremediation (Rowland 2017).
We conducted two experiments to determine the effects of algal management using Aquashade and copper sulfate on anuran metamorphosis. Additionally, to determine how tadpole presence compared with chemical management on the impact of phytoplankton and periphyton abundance, we exposed ponds to Aquashade, copper sulfate, and tadpoles. We hypothesized that addition of pond management treatments (chemical addition) would negatively impact anuran metamorphosis resulting in decreased survival, decreased mass at metamorphosis, and increased time to metamorphosis, and that chemical addition would reduce phytoplankton, periphyton, and zooplankton abundance. We predicted that Aquashade would indirectly impact anuran metamorphosis through reduced food resources (phytoplankton and periphyton abundance) due to a darkening effect from dye application; whereas, we predicted that the toxicity of copper sulfate would have a direct negative impact on anuran metamorphosis. Furthermore, we predicted that tadpole grazing and chemical management, via Aquashade and copper sulfate application, would similarly reduce algal resources.

Materials And Methods
We collected nine partial northern leopard frog (Rana pipiens) egg masses from a vernal pool at Talawanda High School in Oxford, OH on 21-22-March-2017, seven partial American toad (Anaxyrus americanus) egg strings from Rush Run Wildlife Area in Somerville, OH on 6-April-2017, and six partial Cope's gray treefrog (Hyla chrysoscelis) egg masses from six breeding pairs at Miami University's Ecological Research Center (ERC) in Oxford, OH on May 27-28, 2017. We placed eggs in containers with natal pond water and brought them back to the lab in an environmental chamber at 23° C where they were monitored daily for hatching. Hatched tadpoles were placed in new containers of dechlorinated water and monitored daily until they reached free swimming stage, Gosner stage 25 (Gosner 1960). We changed the water in tadpole containers daily and fed tadpoles TetraMin sh akes until they were added to outdoor mesocosms.
In 2017, we set up a total of 50 aquatic communities in polyethylene cattle tank mesocosms (1.83 m in diameter, 1,480 L volume, Figure  Tadpoles were added to mesocosms once they had reached the free-swimming stage, Gosner stage 25 (Gosner 1960). Speci cally, in the spring breeding mesocosms, we added 30 northern leopard frog tadpoles on 1-April-2017 and 30 American toad tadpoles on 12-April-2017 (experimental day 0). We added 30 Cope's gray treefrog tadpoles to the summer breeding mesocosms on 10-June-2017. All mesocosms were covered with a screen lid (2 mm x 2 mm) to prevent non-target species from entering the arti cial community.
In 2019, we conducted another study with four partial egg masses of Cope's gray treefrogs from four breeding pairs on 27, 28, and 29-May-2019 collected from the ERC. Based on the results from our initial experiment we chose to only include treefrogs to reduce experimental complexity and focus on additional aquatic community impacts. We set up a total of 24 aquatic communities on 27-May-2019 in mesocosms (Figure 1b) following the same methodology as in 2017. On 7-June-2019 (experimental day 0) we added 30 Cope's gray treefrog tadpoles to the 12 mesocosms randomly assigned to the tadpole management treatment to evaluate the impact of tadpole presence versus absence on the abundance of periphyton and phytoplankton separate from pond management treatments.
In 2017, we added Aquashade and copper sulfate ("pond management treatments") at two concentrations (high and low). The high concentration treatment was the recommended dose of each chemical, calculated to the volume of our cattle tank mesocosms (1000 L), and the low concentration treatment was half of the calculated recommended dose. We added 1.90 ppm (mg/L) and 0.95 ppm of Aquashade, and 0.0392 ppm and 0.0196 ppm of copper sulfate as the high and low treatment, respectively. We applied each treatment by adding the measured amount of each chemical to a watering can lled with city tap water and distributed each evenly across the mesocosm; the watering can was rinsed clean in between each treatment application. Pond management treatments were applied to mesocosms one week after tadpole addition (19-April-2017 and 19-June-2017 respectively) to allow for tadpole acclimation. One-hour post dosing, we collected a 1 L sample of water from each cardinal direction in each mesocosm, which was further combined by treatment. From this solution we collected a 50 mL composite sample for each recommended dose treatment and sent the sample to Mississippi State University for chemical analysis to con rm our treatment dose. Results from chemical analysis conducted by Mississippi State University from the initial experiment, showed no detection of copper sulfate or tartrazine (a yellow dye in Aquashade); however, erioglaucine (a blue dye in Aquashade) was detected at 4.82 ppm.
In 2019, we added Aquashade and copper sulfate on 11-Jun-2019, at the recommended concentration (1.90 ppm Aquashade and 0.0392 ppm copper sulfate), following methods used in 2017. We used the recommended concentration due to our previous results not showing any treatment differences due to chemical concentration. Our experimental design included four pond management treatments and two tadpole treatments for six total treatments; each treatment had four mesocosm replicates, totaling 24 mesocosms in our design (Figure 1b). Twenty-four hours after pond management treatment addition, we collected a 1 L composite sample (instead of 50 ml) of each treatment (following methods in 2017) and sent it to Mississippi State University for chemical analysis. We collected a larger water sample for chemical analysis in this experiment to have a higher likelihood of detecting our treatment chemicals since they are at a relatively low concentration. Results from this chemical analysis showed copper sulfate detection at 0.031 ppm, erioglaucine detected at 0.392 ppm, and tartrazine was not detected. We collected another composite sample one week after initial treatment addition to determine if any degradation had occurred. Results from those samples sent to Mississippi State University detected 0.020 ppm copper sulfate, 0.210 ppm of erioglaucine, indicating a 30% degradation over seven days; tartrazine was not detected.
We monitored mesocosms daily for metamorphosed individuals, presence of at least one front limb (Gosner stage 42;Gosner 1960). We brought all metamorphs to the lab and weighed them once their tails were fully resorbed (Gosner stage 46), to the nearest 0.001 g to determine mass at metamorphosis; additionally, survival and time to metamorphosis were determined as anuran data endpoints. All metamorphosed individuals were euthanized in 1% buffered MS-222. We terminated the mesocosm experiment on 25-July-2017 (experimental day 116) and 7-August-2017 (experimental day 57) after only one metamorph had been removed over a 7 day period. In 2019, the experiment was terminated on 1-August-2019 (experimental day 56), after only one metamorph had been removed over a 10 day period. We drained all mesocosms using a mesh lter and searched through the leaf litter for remaining tadpoles. In 2017, no tadpoles were found in mesocosms, and in 2019, only two tadpoles were found.
In addition to anuran responses, we measured light availability, zooplankton abundance, and phytoplankton and periphyton abundance, to determine the indirect effects of our pond management treatments on the aquatic communities. Throughout both experiments we measured light availability, using a LI-COR PAR (photosynthetically active radiation) sensor, LI-190R, at three depths (above the surface of the water, directly under the surface of the water [7 cm], and at the bottom of each mesocosm [37 cm], to determine the impact of darkening due to pond management treatment. Additionally, in 2019, we measured total zooplankton abundance twice during the experiment; 24 hours after chemical dosing and one week prior to experimental shut down. Each zooplankton sample was identi ed and grouped by cladocerans, copepods, and ostracods. We collected a 1 L composite sample of water from each mesocosm, using a polyvinyl chloride (PVC) water sampling device. Each sample was ltered using a zooplankton concentration net and preserved with 80% ethanol in a 40 mL glass scintillation vial.
We also measured periphyton abundance by deploying two pool noodles with glass microscope slides inserted vertically in each mesocosm; samples were collected weekly (experimental days 29, 36, 44, and 50). We collected two microscope slides and scraped each (15.24 each, 30.48 total area) using a razorblade to remove periphyton onto a Merck Millipore 0.7 mm pore diameter ber glass lter and preserved in a glass scintillation vial with 15 ml of 80% buffered acetone. We measured phytoplankton abundance weekly (experimental days 29, 36, 44, and 50) by collecting a 100 mL composite sample (1 L samples were taken from each of the four cardinal directions in each mesocosm, and a 100 mL sample was taken from that) from each mesocosm. We ltered all samples onto a berglass lter and preserved them in a 40 mL glass scintillation vial with 15 mL of 80% buffered acetone. All samples were stored in a 4° C refrigerator for 24 hours and analyzed for chlorophyll a using a Turner 10AU uorometer. Phytoplankton and periphyton samples out of range were diluted by adding 1 mL of sample to fteen milliliters of 80% buffered acetone; the resulting chlorophyll a value was multiplied by a factor of sixteen.
All statistical analyses were conducted in R version 3.5.1. Mesocosm was the experimental unit and all anuran species were analyzed separately for each year. We conducted an analysis of variance (ANOVA) to determine differences in mass at metamorphosis and time to metamorphosis for each anuran species among treatments. We analyzed differences in survival (metamorphs alive at end of the mesocosm experiment) across treatments using a generalized linear model (GLM) with a binomial distribution (R package lme4). We conducted a repeated-measures ANOVA to determine differences in light intensity over time across treatments, and phytoplankton and periphyton abundance over time by pond management treatment and tadpole presence. Chlorophyll a abundance was log transformed to meet the assumption of normality. Furthermore, we analyzed differences in total zooplankton abundance over time by pond management treatment with a repeated-measures ANOVA. Total zooplankton abundance was rank transformed to meet the assumption of normality. We conducted a Dunnett's multiple comparison test as a post hoc analysis on all ANOVAs regardless of whether the "pond management treatment" was signi cant from controls.

Results
In 2017 and 2019, light intensity was signi cantly decreased in Aquashade treated mesocosms, compared to control ponds at both low and high depths, regardless of Aquashade concentration (Table 1, Figure 2). Copper sulfate exposure, nor the presence/absence of tadpoles (in 2019) did not signi cantly alter light intensity. Overall, we found no impact of pond chemical management on anuran time to metamorphosis, mass at metamorphosis, or survival for any anuran species in 2017 or 2019 (Table 2, Figure 3). In the 2019 study, gray treefrog tadpole presence was associated with a signi cant decrease in the abundance of both periphyton and phytoplankton chlorophyll a (Table 3, Figure 4); in contrast, pond chemical management had no impact on periphyton or phytoplankton chlorophyll a abundance. In 2019, total zooplankton abundance signi cantly decreased over time; however, there were no signi cant differences due to chemical management or tadpole presence (Table 4, Figure 5).

Discussion
Creation of aquatic habitat can bene t local wildlife (Urban and Rohem 2018) and provide ecosystem services like sediment collection (Renwick et al. 2005). However, the quality of habitat depends on management practices which in uence the habitat's ecological value (Bastian et al. 2014;Nassauer et al. 2014;Kremen et al. 2018). It is imperative to understand the impact of common pond management strategies on aquatic communities because it can help inform local management recommendations and maintain valuable habitat. Overall, the results from these experiments indicate that common chemical pond management strategies, when applied at recommended doses, should have little impact on anuran metamorphosis, but also suggest that tadpoles could serve as a form of bioremediation and provide a natural ecosystem service via algal grazing.
Interestingly, it was tadpole presence that appeared to have the strongest impact on the aquatic environment. Tadpole presence resulted in a signi cant reduction in both periphyton and phytoplankton abundance, suggesting both are components of the gray treefrog tadpole's diet (Dickman 1968;Morin and Johnson 1988;Loman 2001), and that gray treefrog populations could control algal overgrowth.
Periphyton and phytoplankton abundance were observed to initially decrease when tadpoles were present and then increased after most individuals metamorphosed due to decreased grazing pressure. Morin and Johnson (1988) observed similar patterns of periphyton and phytoplankton abundance over time due to variation in tadpole density and grazing pressure, further supporting tadpoles as natural grazers in aquatic ecosystems. While tadpoles are generally not considered in aquatic algal management, creating quality habitats for amphibians could allow for a natural solution with amphibians maintaining aquatic algal growth via natural grazing as an ecosystem service (Holomuzki 1998;Connelly et al. 2008). Anurans are environmental health indicators in pond ecosystems (Mendelson et al. 2006) and promoting natural anuran colonization as a pond management strategy could bene t ecosystem health and maintain anuran abundance and diversity in human-dominated systems (Connelly et al. 2008).
While we initially predicted that Aquashade would indirectly and negatively impact arti cial pond communities, our results suggest Aquashade would be a safe pond management application for larval anurans. However, we did nd that Aquashade reduced light penetration in the ponds, similar to other studies including Madsen et al. (1999) who observed a 10% decrease in light intensity with Aquashade application at the recommended dose (measured at 33 cm with a peak absorbance at 630 nm) similar results were observed by Tucker and Mischke (2020). With decreased light availability, Aquashade can reduce phytoplankton and periphyton resources, thereby indirectly impacting aquatic grazers like zooplankton and tadpoles (Ludwig et al. 2010;Batt et al. 2015). While signi cant differences in light availability are observed from Aquashade treatment, differences in periphyton and phytoplankton abundance were not observed due to chemical management in 2019, suggesting that there was still su cient light penetration in our mesocosms to maintain algal growth to support larval anurans (Tucker and Mischke 2020). The results of our study suggest that if individual pond managers use Aquashade at recommended doses, aquatic community should not be negatively impacted.
While copper sulfate is a known to be toxic at certain concentrations (Lande and Guttman 1973;McKim et al. 1978; Gurkan and Hayretdag 2012), we did not observe signi cant impacts on anuran development or survival from copper sulfate exposure. However, this is not unexpected as we chose concentrations at the recommended or lower than the recommended application dose. Moreover, studies that found negative effects on amphibians exceeded the recommended copper sulfate application rate, typically using concentrations above 0.10 ppm. For example, Garcia Munoz (2009) used 0.20 ppm and Chen et al.
(2007) used 0.30 ppm and 0.40 ppm and found signi cantly increased mortality at these concentrations, which may occur when land occupants use of concentrations above recommended application rates when managing their ponds (Blaine et al. 2012).
In conclusion, the results from our experiments suggest that the use of pond management chemicals for aquatic vegetation management at recommended doses does not appear to signi cantly impact anuran metamorphosis, and that tadpoles can provide natural algal management in ponds. Landowners can create habitat that will attract anurans by leaving at least some natural habitat around the pond for terrestrial overwintering and foraging and by not stocking ponds with sh, which are predators of larval anurans (Porej and Hetherington 2005;Hamer and Parris 2011). These management practices could ultimately improve the local habitat matrix and increase anuran abundance and diversity across the landscape. As residential development increases it is important to further research on sustainable management practices that balance residential landowner preferences while improving the ecological value of the land.

Declarations
Funding: This study was funded by Miami University.
Con icts of Interest/Competing Interests: All authors certify that they have no a liations with or involvement in any organization or entity with any nancial interest or non-nancial interest in the subject matter or materials discussed in this manuscript.
Ethics Approval: The research conducted in this manuscript were covered by Miami University IACUC protocol #827: Interactive effects of contaminants, invasive species, pathogens, and habitat manipulation on amphibian populations and community structure: how multiple stressors may contribute to amphibian declines & in uence community dynamics. Additionally, collection of egg masses was covered by the Ohio Division of Wildlife Wild Animal Permit # 20-177 to M. D. Boone.
Consent to Participate/Publication: All authors contributed to the study conception and design. Material preparations, data collection and analysis were performed by Courtney Dvorsky, Kambrie Riddle, and Michelle Boone. The rst draft of the manuscript was written by Courtney Dvorsky, and all authors commented on previous versions of the manuscript. All authors read and approved the nal manuscript.
Availability of Data, Material, and Code: The datasets and code generated and/or analyzed during the current study are available from the corresponding author on reasonable request.    Represents phytoplankton (a) and periphyton abundance (b) with tadpole presence over time (experimental days). Tadpole presence signi cantly decreased periphyton and phytoplankton abundance over time