Pond-breeding amphibian community composition in Missouri.
Understanding the mechanisms that influence patterns of community structure is a fundamental goal in ecology. Knowledge of these mechanisms can be critical in evaluating the strength of species interactions, how food webs function, and how both local and regional biodiversity is maintained (Morin, 2011). Yet baseline information on community structure in natural systems is often lacking or poorly developed for comparisons of communities across large geographic areas (e.g. entire states), which is needed to understand if and how spatiotemporal changes occur in community composition.
Information on the mechanisms that influence amphibian community structure is increasingly necessary and important as many communities face threats such as invasive species, habitat alteration, or climate change (Parmesan, 2006; Semlitsch et al., 2009) and are experiencing widespread declines (Stuart et al., 2004). Management decisions hinge upon correct assessments of how these threats influence communities or ecosystems, resulting in a need for contemporary studies on community structure as a reference point for tracking potential changes.
Pond-breeding amphibians are model taxa for examining patterns of community structure. Populations are centered around breeding ponds, with larval stages occupying ponds for a few weeks to several months prior to completing metamorphosis (Wilbur, 1980). Aquatic stages face a variety of threats including contamination and invasive species (Blaustein et al., 2011), making population monitoring critical to determine both current patterns of community structure and potential factors already present that influence species composition. We examined pond amphibian community structure at breeding sites across the state of Missouri and examined whether the presence of American Bullfrogs, Lithobates catesbeianus, and fish influenced these patterns of diversity.
We sampled for amphibians in 210 ponds and wetlands between 2002 and 2012 in Missouri, U.S.A. (see electronic appendix 1 https://mospace.umsystem.edu/xmlui/handle/10355/40553). Sampling at these sites was conducted with drift fences around ponds or a combination of dipnetting and funnel trapping within ponds. Drift fences and pitfall buckets were installed and maintained around two ponds at the University of Missouri's Thomas S. Baskett Wildlife Research and Education Area (38.74847N, -92.20211) Boone County in 2002, and around five ponds at the Missouri Department of Conservation's (MDC) Daniel Boone Conservation Area (DBCA; within 1km of 38.771952N, -91.388054W), Warren County in 2005 (Hocking et al., 2008; Semlitsch et al., 2009).
Dipnet and funnel trap sampling was conducted at 49 constructed sites in 26 northern Missouri counties in 2006 (methods and sites in Shulse et al., 2010). Sampling at these sites was conducted in three rounds, March/April, May/June, and July/August of 2006, with one set of dipnets and one night of trap sampling per site visit. We also sampled at 154 sites at Fort Leonard Wood (FLW; within 4km of 38.74847N, -92.20211W), Pulaski County, in 2012. Sampling the FLW sites was conducted in February/ March and April/May of 2012, with three consecutive days of dipnetting and nights of trapping conducted concurrently in each round (Peterman et al., 2014). Dipnet sweeps and number of funnel traps were scaled to the surface area of the ponds (one per 25 [m.sup.2] of pond surface area; Shulse et al., 2010), and the location of dipnet sweeps and placement of traps within each pond were proportional to the aquatic habitat types present
Amphibians and fish were identified in the field to species when possible and released at site of capture; species of leopard and Pickerel Frog tadpoles were combined into one category as they are visually indiscernible. Fish species included Mosquitofish (Gambusia sp), sunfish (Lepomis sp), bass (Micropterus sp), catfish (Ictaluridae), and minnows (Cyprinidae).
Statistical Analyses: We characterized amphibian communities in Missouri using data collected by all survey methods but used data from sites surveyed with dipnets and funnel traps for statistical analysis because they represented a greater range of sites in Missouri. The reduced dataset was further truncated to include only those taxa that were identified to the level of species, and life stages were merged into a single binary response variable equivalent to 'detected' or 'not detected' for each locality. The binary dummy variable was used to calculate diversity at three taxonomic levels: (1) 'Total', for all amphibian taxa excluding American Bullfrogs; (2) 'Salamanders', for all seven caudate taxa; (3) 'Frogs', for 12 anuran taxa excluding American Bullfrogs. We calculated diversity as the number of species detected within the taxonomic group at a site.
To determine whether American Bullfrogs, fish, or their interaction predicted total amphibian species diversity, salamander species diversity, or frog species diversity in Missouri, we fitted generalized linear mixed-effect regression (GLMER) models with a binomial error term in the package 'lme4' (Bates et aL, 2014). In these models American Bullfrog presence, fish presence, and their interaction were fixed effects; pond was nested in ecoregion as random effects to capture regional variation. We included ecoregion as a random effect to account for the potential for species assemblages to differ based on ecoregion-specific attributes. The study sites were located in either the Ozark Plateau (n=174) in the southern portion of the state or the Central Dissected Till Plains (n = 36) in the North (Etheridge, 2009). Model fit was evaluated based on likelihood ratio tests and Akaike's Information Criterion (AIC) following maximum likelihood estimation and the Laplace approximation.
We detected 20 of the possible 27 pond-breeding amphibian species in the study areas (Daniel and Edmond, 2013; Fig. 1). The most common species encountered were anurans: Green Frogs (Lithobates clamitans, 56% of sites), American Bullfrogs (L. catesbeianus, 55%), and members of the leopard frog complex (50%). Frequently encountered caudates were Central Newts (Notophthalmus viridescens louisianensvr, at 49% of sites), Ringed Salamanders (Ambystoma annulatum, 42%), and Spotted Salamanders (A. maculatum, 40%). Pond-breeding species that were within our sampling range but were not encountered included the Northern Crawfish Frog (L. areolatus), Plains Spadefoot Toad (Spea bombifrons), Eastern Narrow-mouthed Toad (Gastrophryne carolinensis), and Four-toed Salamander (Hemidactylium scutatum). Tiger Salamanders (A. tigrinum) and Western Narrow-mouthed Toads (G. olivacea) were encountered at one site each. Fish were encountered at 22% (n=46) of the ponds sampled.
Overall, American Bullfrogs were positively associated with total amphibian diversity while fish were negatively associated with diversity (Fig. 2). Both single factor GLMER models (using fish or American Bullfrogs as fixed effects) performed better than the null model containing only random effects (Table 1A). The model with both fixed effects performed better than either single factor model ([X.sup.2] = 35.65, df=1, P=2[E.sup.-9]), and adding the interaction term (fish*American Bullfrogs) provided no benefit (Table 1A). For the two parameter model, random effect variances were small (Pond=0.08, Ecoregion=0.03), and the change in the slope parameter estimated for the fixed effects were both significant (American Bullfrogs: Z=6.454, P=1[E.sup.-10]; Fish: Z=-5.988, P=2[E.sup.-9]), indicating American Bullfrogs are positively associated with increased diversity and fish are negatively associated (Table 2A).
Salamander diversity was slightly higher in the presence of American Bullfrogs, but there was a strong negative association between fish presence and salamander diversity (Fig. 2). The single factor model including American Bullfrogs as a fixed effect showed no improvement over the null model, but the model containing fish presence as a fixed effect was informative ([X.sup.2]=45.17, df=1, P=2[E.sup.-11]; Table 1B).
The model with both fixed effects performed better than the fish-only single factor model ([X.sup.2]=6.94, df=1, P=0.008), but adding the interaction term (fish*American Bullfrogs) provided no benefit (Table IB). For the two parameter model, random effect variances were again small (Pond=0.00, Ecoregion=0.85), and the change in the slope parameter estimated for the fixed effects were both significant (American Bullfrogs: Z=2.626, P=0.009; Fish: Z=-5.856, P=5[E.sup.-9]), meaning American Bullfrogs are slighdy positively associated with increased salamander diversity and fish are negatively associated with salamander diversity (Fig. 2B).
For frog diversity the single factor model with fish as a fixed effect showed no improvement over the null model, but the model containing American Bullfrog presence demonstrated a significant improvement ([X.sup.2]=37.44, df=1, P=9[E.sup.-10]; Table 1C). The model with both fixed effects performed better than the American Bullfrog single factor model ([X.sup.2]=8.50, df=1, P=0.004), and frog species diversity was best predicted by the model including the interaction of American Bullfrog and fish presence ([X.sup.2] = 4.56, df = 1, P=0.033; Table 2C). For the interaction model, random effect variances remained small (Pond=0.13, Ecoregion=0.00) and significant estimates were found for American Bullfrogs (Z=7.108, P=1[E.sup.-12]), and the interaction term (Z=-2.24, P=0.025), but not for fish (Table 2C). This means American Bullfrog presence was positively associated with increased diversity of other frogs, but when fish and American Bullfrogs were both present, frog diversity was slightly reduced (Fig. 1C).
Habitat generalists such as Lithobates catesbeianus, L. clamitans, L. sphenocephalus, Anaxyrus americanus, Notophlhalmus viridescens louisianensis, Hyla chrysoscelis/versicolor, Pseudacris crucifer and P. maculata were encountered across our sites, as expected from distribution records (Daniel and Edmond, 2013). Two species that had historically wide distributions but were rare in our surveys were L. palustris and Ambystoma tigrinum. Low detection of /.. palustris may be explained by true absence in the areas or that they were present in the larval stage where they would have been classified in the "Leopard Frog complex" or "Lithobates sp." categories, given the difficulty in differentiating their tadpoles from those of other leopard Frog species. Historically, A. tigrinum had a wide distribution across the state but there are few current site records (Daniel and Edmond, 2013), suggesting our low encounter frequency may be a result of low site occupancy by A. tigrinumor could be an indication of population decline. Lack of detection of Gastrophryne may be due to the timing of our sampling combined with their rapid larval development. Other species, such as L. areolata, have a more restricted range outside of the areas we sampled. More targeted surveys and data are needed for these species of concern, which will help create a clearer understanding of rare species and overall community structure.
The results of our analyses indicate the presence of American Bullfrogs, fish, and their interaction influenced the community composition of amphibians but in opposite patterns. American Bullfrogs generally had a positive relationship with all groupings of species, whereas fish presence was negatively associated with species diversity. Although negative impacts of American Bullfrogs on all life stages of several other anuran species, particularly those where American Bullfrogs have been introduced, have been documented (Keisecker and Blaustein, 1998; Lawler et at, 1998; Boone et al., 2004; Pittman et al., 2013), we found an overall positive association of American Bullfrogs with amphibian species diversity across Missouri. Our results should not be interpreted as American Bullfrogs causing an increase in amphibian species diversity. Rather, we feel that other factors that were not included in our study, such as diversity of terrestrial landscape around the pond, diversity of aquatic microhabitats within the pond, and hydroperiod contribute to the increased amphibian species diversity at the sites where amphibian diversity was high, and these were sites that were also associated with American Bullfrogs. The fish-free American Bullfrog ponds in our study were typically large, permanent ponds set in a diverse terrestrial landscape that included field, forest and edge habitats that would be inhabited by a wide variety of pond-breeding amphibians that would use the pond. These permanent ponds provided deep waters for species with larvae that overwinter (L. catesbeianus and L. clamitans) and peripheral shallow waters for species with larvae that develop more rapidly such as toads (Anaxyrus sp,), as well as accommodating species with intermediate larval development times such as H. chrysoscelis/ versicolor. These sites also had diverse aquatic microhabitats including open water, leaf litter, submerged, floating and emergent vegetation, offering an assortment of substrate options for opposition and larval habitat. Fish were encountered in fewer of our study sites and had a negative overall association with all species groupings. The negative impacts of fish, native and introduced, on both salamander and frog species has been widely documented (Hecnar and M'Closkey, 1997; Smith et al., 1999; Shulse et al., 2010; Drake et al., 2014). Semlitsch et al. (2015) reported amphibian abundance, species richness, and diversity decreased with an increase in pond size due primarily to the presence of fish.
Our summary and analysis of pond-breeding amphibian community structure on a state-wide scale provides a baseline of information and a beginning assessment of potential factors that impact community structure on a large scale. It also affords an opportunity to subsequently examine how and if spatiotemporal changes occur in amphibian community composition.
Acknowledgments.--We thank J. Heemeyer for her assistance in the field. This research was approved by the University of Missouri Animal Care and Use Committee (7403) and supported by the DoD Strategic Environmental Research Development Program (RC2155). One year of research at University of Missouri's Thomas S. Baskett Wildlife Research and Education Area was funded by USGS Amphibian Research and Monitoring Initiative.
Bates, D., M. Maechler, B. Bolker, and S. Walker. 2014. Ime4: Linear mixed-effects models using Eigen and S4. R package version 1.1-7, http://CRAN.R-project.org/package=Tme4.
Blaustein, A. R., B. A., Han, R. A. Relyea, P. T. Johnson, J. C. Buck, S. S. Gervasi, and L. B. Kats. 2011. The complexity of amphibian population declines: understanding the role of cofactors in driving amphibian losses. Annals of the New York Academy of Sciences, 1223:108-119.
Boone, M. D., E. E. Little, R. D. Semlitsch, and S. F. Fox. 2004. Overwintered bullfrog tadpoles negatively affect salamanders and anurans in native amphibian communities. Copeia, 2004:683-690.
Daniel, R. and B. Edmond. 2013. Atlas of Missouri amphibians and reptiles for 2012. http:// atlas.moherp.org/pubs/atlas12.pdf. Accessed 19 Aug 2013.
Dodd, C. K. 2013. Frogs of the United States and Canada. Johns Hopkins University Press: Baltimore. 982p. Drake, D. L., T. L. Anderson, L. M. Smith, K. M. Lohraff, and R. D. Semlitsch. 2014. Predation on eggs and recently-hatched larvae of endemic Ringed Salamanders (Ambystoma annulatum) by native and introduced aquatic predators. Herpetologica, 70:378-387.
Ethridge, M. 2009. The Ozark Highlands: U.S. Geological Survey Fact Sheet 2009-3065, 2 p.
Hecnar, S. J. and R. T. M'Closkey. 1997. The effects of predatory fish on amphibian species richness and distribution. Biol. Conseru., 79:123-131.
Hocking, D. J., T. A. G. Rittenhouse, B. Rothermel, J. Johnson, C. A. Conner, E. B. Harper, and R. D. Semlitsch. 2008. Breeding phenology of pond-breeding amphibians in Missouri oak-hickory forests. Am. Midi. Nat., 160:41-60.
Kiesecker, J. M. and A. R. Blaustein. 1998. Effects of introduced Bullfrogs and smallmouth bass on microhabitat use, growth and survival of native Red-legged Frogs. Conseru. Biol., 12:776-787.
Lawler, S. P., D. Dritz, T. Strange, and H. Holyoak. 1998. Effects of introduced mosquitofish and bullfrogs on the threatened California Red-legged Frog. Consent. Biol., 13:613-622.
Morin, P. J. 2011. Community ecology. John Wiley & Sons. 424p.
Parmesan, C. 2006. Range-restricted species distribution shrinks with current climate change patterns. Annu. Rev. Ecol., Evol., and Syst., 37:637-669.
Peterman, W. E., T. L. Anderson, D. L. Drake, B. H. Ousterhout, and R. D. Semlitsch. 2014. Maximizing pond biodiversity across the landscape: a case study of larval ambystomatid salamanders. Animal Consent., 17:275-285.
Pittman, S., M. S. Osbourn, D. L. Drake, and R. D. Semlitsch. 2013. Predation of juvenile pond-breeding salamanders during initial movement out of ponds. Herpetol. Consent. Biol., 8:681-687. Semlitsch, R. D., T. L. Anderson, D. L. Drake, B. H. Ousterhout, and W. E. Peterman. 2015. Intermediate pond sizes contain the highest density, richness, and diversity of pond-breading amphibians. PLOS One, 10:e0123055, doi:10:1371/journal.pone. 0123055.
__B. D. Todd, S. M. Blomquist, A. J. K. Calhoun, J. W. Gibbons, J. P. Gibbs, G. J. Graeter, E. B.
Harper, D. J. Hocking, M. L. Hunter, Jr., D. A. Patrick, T. A. G. Rittenhouse, and B. B. Rothermel. 2009. Effects of timber harvest on amphibian populations: understanding mechanisms from forest experiments. BioScience, 59:853-862.
Shulse, C. D, R D. Semlitsch, K M. Trauth, and A. D. Williams. 2010. Influences of design and landscape placement parameters on amphibian abundance in constructed wetlands. Wetlands, 30:915-928. Smith, G. R , J. E. Rettig, G. G. Mittelbach, J. L. Valiulis, and S. R. Schaack. 1999. The effects of fish on assemblages of amphibians in ponds: a field experiment. Freshwater Biol., 41:829-837.
Stuart, S. N., J. S. Chanson, N. A. Cox, B. E. Young, A. S. L. Rodrigues, D. L. Fischman, and R. W. Wallace. 2004. Status trends of amphibian declines and extinction worldwide. Science, 306:1783-1786.
Wilbur, H. M. 1980. Complex life cycles. Annu. Rev. Ecol. Syst., 11:67-93.
DANA L. DRAKE and BRITTANY H. OUSTERHOUT, Division of Biological Sciences, University of Missouri, Columbia, 65211; JARRETT R. JOHNSON, Department of Biology, Western Kentucky University, 1906 College Heights Blvd., #11080, Bowling Green, 42101; THOMAS L. ANDERSON and WILLIAM E. PETERMAN, Division of Biological Sciences, University of Missouri, Columbia, 65211; CHRISTOPHER D. SHULSE, Missouri Department of Transportation, P.O. Box 270, Jefferson City, 65102; DANIEL J. HOCKING, Department of Natural Resources and the Environment, 114James Hall, University of New Hampshire, Durham, 03824; KENTON L. LOHRAFF, US Army IMCOM & FLW, DPW Natural Resources Branch, 1334 First St., Bldg 2222, Fort Leonard Wood, Missouri 65473; ELIZABETH B. HARPER, Division of Natural Resource Management and Ecology, Paul Smith's College, Paul Smiths, New York 12970; TRACY A. G. RITTENHOUSE, Department of Natural Resources and the Environment, University of Connecticut, 1376 Storrs Road, Unit 4087, Storrs, 06269; BETSIE B. ROTHERMEL, Archbold Biological Station, 123 Main Dr., Venus, Florida 33960; and LORI S. EGGERT and RAYMOND D. SEMLITSCH, Division of Biological Sciences, University of Missouri, Columbia, 65211. Submitted 16 January 2014; Accepted 28 February 2015.
(1) Corresponding author: e-mail: Drake.Dana.L@gmail.com
TABLE 1.--Comparison of generalized linear mixed models for the effect of American Bullfrogs and fish on species diversity. All models include the random effect variables 'Pond' nested within Ecoregion'. The models with the lowest AIC values are in bold. Based on the deltaAIC values there is some support for the interaction models even though the "best" models seem to be the additive models for 'Total' and 'Salamanders' [DELTA] Response Model Df AIC AIC logLik A. Total Diversity Null (intercept) 3 905.63 53.67 -449.82 Bullfrog 4 885.60 33.64 -438.80 Fish 4 887.49 35.53 -439.74 Bullfrog+Fish 5 851.96 0.00 -420.98 Bullfrog*Fish 6 852.07 0.11 -420.03 B. Salamanders Null (intercept) 3 615.94 48.11 -304.97 Bullfrog 4 616.56 48.73 -304.28 Fish 4 572.77 4.94 -282.38 Bullfrog+Fish 5 567.83 0.00 -278.92 Bullfrog*Fish 6 569.83 2.00 -278.92 C. Frogs Null (intercept) 3 760.26 44.49 -377.13 Bullfrog 4 724.82 9.05 -358.41 Fish 4 760.73 44.96 -376.37 Bullfrog+Fish 5 718.32 2.55 -354.16 Bullfrog* Fish 6 715.77 0.00 -351.88 TABLE 2.--Regression coefficients for fixed and random factors of the best supported models (See Table 1). American Bullfrog effects are positive and larger for frogs than salamanders, fish effects are negative and larger for salamanders than frogs. Random effect variances are small, and for salamanders all random variation is explained by Ecoregion, with the opposite being true for frogs. The frog interaction term shows the amount that the increase in diversity associated with American Bullfrog presence is reduced when fish are also present Pr(> [absolute Fixed value of Response Model Effect Est. SE Z z]) A. Total Bullfrog+ (Intercept) 0.940 0.152 6.166 7.01E-10 Species Fish Bullfrog 0.574 0.089 6.454 1.09E-10 Fish -0.736 0.123 -5.988 2.12E-09 B. Bullfrog+ (Intercept) -0.253 0.674 -0.376 7.07E-01 Salamanders Fish Bullfrog 0.297 0.113 2.626 8.64E-03 Fish -1.398 0.239 -5.856 4.73E-09 C. Frogs Bullfrog+ (Intercept) 0.153 0.113 1.357 1.75E-01 Fish Bullfrog 0.945 0.133 7.108 1.18E-12 Fish 0.223 0.315 0.709 4.78E-01 Interaction -0.791 0.353 -2.240 2.51E-02 Random Response Model Effect Var SD A. Total Bullfrog+ Pond 0.080 0.283 Species Fish Ecoregion 0.031 0.176 B. Bullfrog+ Pond 0.000 0.000 Salamanders Fish Ecoregion 0.849 0.921 C. Frogs Bullfrog+ Pond 0.133 0.365 Fish Ecoregion 0.000 0.000
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Notes and Discussion|
|Author:||Drake, Dana L.; Ousterhout, Brittany H.; Johnson, Jarrett R.; Anderson, Thomas L.; Peterman, william|
|Publication:||The American Midland Naturalist|
|Date:||Jul 1, 2015|
|Previous Article:||Characterization and classification of vernal pool vegetation, soil, and amphibians of pictured Rocks National Lakeshore.|
|Next Article:||Amphibians and Reptiles in Minnesota.|