Final Report: Environmental Factors That Influence Amphibian Community Structure and Health as Indicators of Ecosystems

EPA Grant Number: R825867
Title: Environmental Factors That Influence Amphibian Community Structure and Health as Indicators of Ecosystems
Investigators: Beasley, Val , Richards, Carl , Schotthoefer, Anna , Lieske, Camilla , Johnson, Catherine , Murphy, Joe , Johnson, Lucinda , Piwoni, Marvin , Schoff, Pat , Cole, Rebecca
Institution: University of Illinois at Urbana , Illinois Waste Management and Research Center , University of Minnesota - Duluth , United States Geological Survey [USGS]
Current Institution: University of Illinois at Urbana , Illinois Waste Management and Research Center , United States Geological Survey [USGS] , University of Minnesota - Duluth
EPA Project Officer: Packard, Benjamin H
Project Period: June 1, 1998 through May 31, 2001 (Extended to September 30, 2002)
Project Amount: $1,299,991
RFA: Ecosystem Indicators (1997) RFA Text |  Recipients Lists
Research Category: Ecosystems , Ecological Indicators/Assessment/Restoration


Multiple factors in agricultural regions are likely to be involved in disease outbreaks and population declines in amphibians; however, the relative importance of such factors is unclear. Studies are needed to examine amphibian responses to an array of environmental stressors at multiple spatial scales in largely agricultural ecosystems, because the lack of knowledge hinders ecosystem planning, management, and monitoring. The overall goal of this research project was to assess the influence of landscape patterns, biotic interactions, water quality, and selected contaminants on the community structure and health of anuran amphibians (frogs, toads) in the upper Midwestern United States. The specific objectives of this research project were to: (1) compare the relative influences of wetland- and watershed-scale factors (landscape and ecological data) on indicators of amphibian community structure and health; and (2) assess whether amphibian community structure and health are indicative of ecological conditions. Studies were conducted in two tiers:

· Tier 1. Satellite-derived landscape, ecological survey, amphibian community structure, and amphibian general abundance data from 64 sites in 13 regional watersheds in Illinois, Wisconsin, and Minnesota.

· Tier 2. Aerial photography, water quality, sediment quality, elemental, nutrient, organic contaminant, local habitat, ecological, and comprehensive amphibian morphologic, pathologic, parasitologic, and residue data from 36 wetland catchments in Minnesota, many of which also were examined in Tier 1.

We explored the relative influences of site-specific, local landscape, and broad landscape phenomena on the distributions of amphibian species as well as the health status of a target species, the northern leopard frog Rana pipiens (termed leopard frog below). Other anurans were examined when few R. pipiens were available. The findings of this study should assist landowners, managers, and regulators in managing and setting priorities for further research on environmental factors that specifically influence amphibian community structure and health.

Summary/Accomplishments (Outputs/Outcomes):

Analyses are ongoing, and thus some findings should be considered preliminary.

Agriculture altered physical and chemical properties of wetlands and local biological communities. Anuran species richness was highest in less fragmented landscapes with less agriculture. Multiple small habitat patches that included forest cover generally supported more amphibian species than larger, more isolated patches with little forest. Detected impacts varied, depending on the scale at which landscapes were analyzed. To utilize species diversity as an indicator of wetland quality, landscape structure, and composition should be examined at spatial scales relevant to taxa of interest.

We identified 11 species of anurans in 80 study wetlands. The numbers of species identified in each wetland ranged from one to nine, with most sites supporting four to six species. Species compositions at sites varied from year to year. At spatial scales less than or equal to 1 km, there were significant negative relationships between the number of sensitive anuran species present and the proportion of agriculture in the landscape. Anurans were considered sensitive to wetland and landscape conditions if they were habitat specialists, or if they appeared to be declining in the region, based on the literature and species-habitat models developed in this research project. The eight most commonly encountered frog species examined in our analyses have a range of associations with habitat variables at the spatial scales examined (site-specific [1 km], local landscape [2 km], and broad landscape [10 km]. No single habitat feature or set of features had a consistent overriding influence on occupancy of all or even most frog species at study wetlands (Johnson et al. 2003; unpublished data). The number of frog species was strongly associated with landscape metrics only at the broadest scales examined. The nine study wetlands with the highest species richness (less than six species) were bordered at least partially by woodland, which reflects the critical importance of forest cover to certain anurans. Patch edge density, proximity to similar wetland types, and spatial distribution and interspersion of cover types also influenced species occupancies. Habitat fragmentation is very important in amphibian declines, but measures of fragmentation should be carefully weighed. In our study area, wetlands, forests, and other habitat components exist mainly as numerous small patches in a region dominated by agricultural fields. Unlike a few isolated "large" habitat patches, the many small habitat patches in our study area seemed to support amphibian persistence, likely because of the small distances across fields to other suitable habitats. The strong relationship between landscape patch interspersion and the occurrence of species, such as spring peepers and gray tree frogs, is consistent with the natural history of these species. Both species exhibited a strong tie to forest cover at varying spatial scales. Although these anurans favored highly fragmented landscapes in this region, their presence was associated with low edge contrast; thus, when they were present, much or all of the land immediately surrounding the wetland was neither agricultural nor urban.

Aquatic macroinvertebrate species diversity was reduced as agricultural land increased. Species richness of macroinvertebrates was significantly lower for wetlands in landscapes dominated by agriculture at spatial scales ranging from 500 m to 1 km. This relationship was even stronger for those taxa that are relatively intolerant of pollution (e.g., Odonata). Patch edge density, proximity to similar wetlands, and spatial distribution and interspersion of cover types were significantly correlated with invertebrate species metrics.

Higher dissolved oxygen (DO) concentrations and plants that increase DO in the water were associated with greater catch of leopard frog metamorphs. Aquatic DO and coontail (Ceratophyllum demersum), which oxygenate the water, but not duckweed (Lemna spp.), which rests on the surface and blocks sunlight, were associated with catch of leopard frogs. Optimal DO was around 14 mg/L. Low day-time DO in ponds (below 5 mg/L) may have harmed tadpoles. The lower limit for leopard frog tadpoles to rest on the bottom instead of staying at the surface is about 3 mg/L. Wetlands at six sites with pasture, or that were greater than 30 percent row crop agriculture within 50 m of the wetland edge, had DO levels below 5 mg/L in May of 1999 and 2000. The presence of pasture within 100 m of the wetland was associated with significant reductions in summer DO concentrations. Plants in the water column also may benefit anurans by providing refuge from predators and parasites, surfaces on which periphyton food of tadpoles grows, removal of toxicants such as nitrite and nitrate, and competition with toxigenic blue-green algae.

Wetlands of the Tier 2 study area had relatively low contamination with nutrients and pesticides. Many farms in the area rely substantially on crop rotation for weed control and/or are involved in organic agriculture. Such methods, in combination with buffers, may have limited nutrient and pesticide concentrations in Tier 2 wetlands, compared to much of the Midwest. Phosphate, which was associated with low DO and can contribute to eutrophication, increased with proximity to crop fields and pastures. At most sites, aqueous nitrate and nitrite levels were below known toxic concentrations for anurans. Each year, mean nitrate was 1.2-1.5 mg/L, but maximums were about 24 mg/L; thus, nitrate at some wetlands approached thresholds for toxicity to frogs. In 2000, nitrite was detected at two sites. Although the higher concentration exceeded a published LC50 for 15-day continuous exposures of Rana pretiosa tadpoles, it was below previously reported 15-day LC50s for other anuran species. Less than 70 pesticides and pesticide degradates were assayed; 23 were never detected in wetlands. Herbicides detected were usually below known toxic levels. Frog tissues did not contain many target insecticide contaminants that were above detection limits. Also, tissue concentrations of the dichlorodiphenyltrichloroethane (DDT) breakdown product dichlorodiphenyldichloroethylene (DDE), were below published thresholds for toxicity, and there was no association between intersex and the infrequent detection of DDE.

Malformation rates in Minnesota frogs were almost four times higher than historic rates, and most malformations were missing or attenuated limbs. To assess prevalences, forms, and distributions of malformations, 5,975 metamorphs from nine species were examined in 1998-2000, and especially in 2001. Respective malformation rates were 2.3 percent, 1.6 percent, 1.4 percent, and 1.8 percent. Our sites in northern Illinois, southern Wisconsin, and east-central Minnesota had similar malformation rates. We examined more of the widespread leopard frogs than any other anurans, but many grey tree frogs (Hyla spp.), green frogs (Rana clamitans), and wood frogs were examined in 2001. Of the species examined in the highest numbers, the malformation rate was similar, with the lowest rate being 1.4 percent in leopard frogs and the highest rate being 2.9 percent in wood frogs. The overall malformation rate of approximately 2 percent was almost 4-fold greater than the reported background rate of 0.5 percent in frogs from the mid-1960s. The vast majority of malformations in the current study were shortened limbs or limb segments. The near absence of taumely (bony triangles) and polymely (multiple limbs), or polydactyly (multiple toes), and the few skin lesions, all of which have been much more common in "hot spot" collections, was surprising. We observed each of these malformation types, but they represented far lower numbers than in recent studies, even though species distributions were similar. The qualitative differences may relate to differences between a "background" malformation rate, as revealed in this study, and "acute" rates at hot spots that have attracted recent studies (Schoff, et al., 2003).

Intersex was associated with grasslands/pastures, but not with atrazine exposure. Research is needed on exposures to estrogenic compounds from plants and herbivores, as well as triazine herbicides. Intersex gonads (ovotestes) were noted in 19 frogs from nine sites, and 9 frogs from five sites in 1999 and 2000, respectively. Affected individuals at impacted sites included 8-36 percent of males in 1999 and 3-27 percent of males in 2000. Sites with maximal percentages of intersex individuals each year yielded no intersex individuals in the alternate year. Atrazine, the widely used and relatively stable herbicide, was detected in water at 39 percent of wetlands in April and May 1999 (greater than or equal to 1 µg/L), and 89 percent of wetlands in June and July 1999 (0.04-0.44 µg/L). In June and July 2000, atrazine was found in water at four wetlands (1.5-11 µg/L). A higher detection limit in 2000 (1 µg/L) likely precluded more detections. Measured concentrations were below those reported to be lethal to amphibians, but intersex gonads have been reported in frogs exposed at 0.1-20 µg/L. We found no association between atrazine and intersex, and we repeatedly captured our target number of frogs (25) at contaminated sites. In Minnesota, the summer season is short and chemical application and runoff are likely to occur later than in much of the corn belt to the South. Further studies on atrazine and intersex should focus on thresholds of toxicity and the functional significance of intersex states. Higher proportions of intersex were significantly associated with larger areas of grasslands and pastures. Whether hormones from grazing animals or phytoestrogens from legumes, such as alfalfa-affected amphibians, should be studied.

Heavy metals in Minnesota frogs often appeared high and were associated with low body condition, but drought in 2000 likely reduced body condition, limiting the ability to discern possible associations with toxic stress. We examined correlations between exposure to metals and amphibian health and residues. Many reports on toxicity to amphibians describe only lethality to tadpoles and, for most elements, tissue residues associated with amphibian toxicoses are unknown. Aluminum (Al), lead (Pb), and zinc (Zn) are known to delay metamorphosis and alter predator avoidance. Also, tadpoles reduced growth and were more susceptible to predation after Al exposure. Although all of our study sites had high aqueous Al, and several of our sites had acidic water, which increases Al toxicity, impacts on frog health were not associated with Al levels. Aquatic coppper (Cu) was high at some sites, and Cu in frog liver was high compared to residues in birds and mammals, but levels were below those of control frogs from studies of Cu toxicity. Although aqueous iron (Fe) was high compared to published toxicity thresholds for tadpoles, and liver Fe was high compared to other species groups, liver Fe was not in a range associated with consistent toxic effects in other species, and no lesions of Fe toxicosis were noted. Although aqueous Pb was high in some wetlands, levels were below those previously associated with reduced activity in tadpoles, tissue Pb was not elevated, and high Pb exposure was not accompanied by lesions of Pb toxicosis. High aqueous Zn was noted in some wetlands. At two sites, concentrations exceeded a reported LC50 for tadpoles. No correlation was detected between liver Zn and water or sediment Zn. In a report on the study of tadpoles, sublethal Zn exposure decreased growth, altered behavior, and produced high-liver residue concentrations. In our frogs, liver Zn was often less than or equal to those concentrations. Although lesions were not correlated with Zn residues, body condition declined with increasing tissue Zn. One pooled liver specimen contained arsenic (As) at a level consistent with toxic effects in birds and mammals. Liver lesions were noted in frogs with high liver As, but similar lesions were found in frogs from other sites with low As, and no intestinal lesions of As toxicosis were noted. More than one-half of 30 frogs, however, from the site with the highest As residues, were in poor body condition. In 2000, body condition scores also were negatively correlated with vanadium, cobalt, selenium, and cadmium. Drought conditions in 2000 probably contributed to the concentration of elements in food, water, and frog tissues, as well as accelerated metamorphosis (Lieske, et al., 2002). Drought-induced acceleration of anuran metamorphosis has previously been associated with lower body condition. Research is needed to understand interactions among increased metal exposures and other aspects of drying aquatic habitats. Studies involving exposures to metals via food and water at various pH levels should monitor lethality, behavioral effects, immunosuppression, lesions, and tissue residues.

Potentially lethal fungal pathogens are present in Minnesota frogs. Although leopard frogs may be less susceptible to chytrid infections than mink frogs, they experienced an outbreak of mesomycetozoa infection. Batrachochytrium dendrobatidis (chytrid fungus; cause of widespread lethal epizootics in frog) was found in 26 mink frogs and wood frogs from five wetlands in 1999. No chytrid organisms were identified in leopard frogs, even at sites where the other frog species were affected. Prevalence of chytrid-infected wetlands in Minnesota may be more extensive than we identified, assuming leopard frogs are less susceptible than other frog species to infection. The site at which mink frogs were most heavily infected with chytrids was among those with the highest concentrations of polyaromatic hydrocarbons (PAHs), suggesting a need to explore immunotoxicity of PAHs to frogs. Organisms consistent with fungi of the Mesomycetozoa (Rhinosporidium, Dermocystidium, Ichthyophonus, and others) infected four of six frogs (mainly leopard frogs) at one site in 2000. They were disseminated throughout most tissues, especially the liver and spleen, where 50-95 percent of the tissue appeared to consist of organisms. Mesomycetozoa have been associated with tadpole die-offs in several states (David Green, personal communication). Considering the prevalence of infection of mink frogs with chytrids, the exposure to PAHs, and the high infection rates and severity of Mesomycetozoa at one of our study sites, further research on fungal pathogens and PAH exposures of Minnesota frogs is warranted.

The majority of parasites in Tier 2 leopard frogs are trematodes: echinostomes and fibricola sp. More than 128,000 parasites were counted from 492 frogs of 18 sites in 1999 and 12 sites in 2000. Parasite species included 14 trematodes, 5 nematodes, a larval cestode, larval mites, 5 protozoan blood parasites, and a leech. About 97 percent were larval trematodes, 1-2 percent were adult trematodes, and 1 percent were nematodes. Larval echinostome trematodes (Echinostoma and Echinoparyphium) that encyst in kidneys occurred at each site. Fibricola sp., a larval trematode that largely infects muscle, was the most abundant parasite. Ribeiroia ondatrae, the larval trematode that induces limb malformations in frogs, was found at low mean abundances at six sites in 1999 and two in 2000. Trematodes use snails in their life cycles, and asexual reproduction therein, greatly increases infective stages. The most widely distributed snail species found in study wetlands were Planorbella (equal to Helisoma) trivolvis, Physa gyrina, Stagnicola exilis, and Gyralus parvus. Infections with echinostomes were most frequently encountered in snails of P. trivolvis, P. gyrina, and S. exilis. Of malformed frogs examined for parasites, none were found to harbor R. ondatrae, and no frogs found to have R. ondatrae were malformed. Evidence suggests, however, that R. ondatrae may have been involved in observed malformations. In 1998, 36.4 percent of frogs captured at one site had malformations of types associated with R. ondotrae. Parasitology data were not generated that year, but all frogs collected at the site in 1999 were infected with R. ondatrae. Also, two malformed mink frogs were discovered at a site in 1999, where R. ondatrae infections were observed in one of three leopard frogs examined for parasites. Similarly, one leopard frog from another site was found to have polymelia in our pathology studies, and R. ondatrae was found in leopard frogs of our parasitology studies from that site. The limited association between malformations and Ribeiroia infections reflects a need for studies into whether the parasites are cleared from frogs that develop malformations, as well as research on trauma- and toxicant-induced teratogenesis.

Trematode prevalence in snails increased with cattails and conductivity. Conductivity, chloride (Cl) concentrations, and dominance of cattails (typha) in wetlands increased as agricultural land increased. High percentages of agriculture near wetlands were associated with higher conductivity, higher Cl concentrations, and domination of aquatic plant communities by cattails. Cl also was correlated with the density of roads within 50 m, and the two sites with the highest Cl were downslope and within 20 m of highways. By August 2000, 64 percent of study sites were dry, and Cl in water increased compared to 1999. Maximal aqueous Cl was 172 mg/L; however, no site exceeded a published toxic concentration for aquatic life (230 mg/L). High conductivity and high cover of cattails also were correlated with high prevalence of trematode infections in snails. In particular, the prevalence of larval echinostomes in P. trivolvis snails tended to increase in wetlands with a high cover of cattails.

Trematode infections tended to occur in a greater percentage of frogs and at higher intensities in areas where more forest cover was available to definitive hosts. The distributions and abundances of echinostomes, Fibricola sp., and R. ondatrae in frogs consistently increased in relation to availability of forest within 1 to 10 km of wetlands. The increased abundance of parasites in frogs of more forested/less agricultural areas probably reflects responses by bird and mammal definitive hosts to required habitats. Future long-term studies are needed to understand determinants of parasite prevalences and abundances in frogs, including factors that influence the use of wetlands by avian and mammalian definitive hosts, as well as the abundances of snails and other intermediate hosts.

Drought was associated with reduced parasite richness, prevalences, and abundances. Parasite species richness, prevalences, and abundances declined in 2000, compared to 1999, which may have been related to drought in the study area. Declines in parasitism because of drought conditions might occur because definitive hosts spend limited time at drying/dry sites. As a result, deposition of infective parasite ova is reduced, fewer snail intermediate hosts survive in drying sites, drought conditions accelerate metamorphosis (reducing the time of greatest susceptibility to infection), and high temperatures and drying reduce survival of free-living infective stages of parasites.

Impacts of echinostome and Ribeiroia infection depend on the stage at which exposure occurs. In laboratory work that accompanied our field studies, the ability of R. ondatrae to induce malformed limbs was stage-specific (study was funded entirely by the U.S. Geological Survey) (Schotthoefer, et al., 2002). Most malformations occurred in tadpoles infected during the limb bud phase of development. In research with both R. ondatrae and echinostomes (the latter supported by the U.S. Environmental Protection Agency), mortality in tadpoles infected during early life stages often exceeded 90 percent, which was much higher than tadpoles infected at later stages (Shotthoefer, et al., 2002; Schotthoefer, et al., 2003). It seems likely that trematode infections impact not only malformation rates, but also kidney function; however, additional research is needed to deduce population outcomes and optimize management options.

Amphibian, macroinvertebrate, and parasite communities reflect environmental conditions. The congruency among the occurrence of certain amphibian species, macroinvertebrate community structure, parasite species richness (the latter possibly reflecting vertebrate species richness or abundance), and landscapes with comparatively low impacts because of habitat fragmentation and loss suggests that amphibians and amphibian health have value as indicators of ecological conditions.

Journal Articles on this Report : 4 Displayed | Download in RIS Format

Other project views: All 47 publications 5 publications in selected types All 4 journal articles
Type Citation Project Document Sources
Journal Article Rohr JR, Civitello DJ, Crumrine PW, Halstead NT, Miller AD, Schotthoefer AM, Stenoien C, Johnson LB, Beasley VR. Predator diversity, intraguild predation, and indirect effects drive parasite transmission. Proceedings of the National Academy of Sciences of the United States of America 2015;112(10):3008-3013. R825867 (Final)
R833835 (Final)
R835188 (2014)
R835188 (Final)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: PNAS-Abstract & Full Text HTML
  • Other: PNAS-Full Text PDF
  • Journal Article Schoff PK, Johnson CM, Schotthoefer AM, Murphy JE, Lieske C, Cole RA, Johnson LB, Beasley VR. Prevalence of skeletal and eye malformations in frogs from north-central United States: estimations based on collections from randomly selected sites. Journal of Wildlife Diseases 2003;39(3):510-521. R825867 (Final)
  • Abstract from PubMed
  • Full-text: BioOne-Full Text PDF
  • Abstract: BioOne-Abstract
  • Other: JWD-Full Text PDF
  • Journal Article Schotthoefer AM, Koehler AV, Meteyer CU, Cole RA. Influence of Ribeiroia ondatrae (Trematoda: Digenea) infection on limb development and survival of northern leopard frogs (Rana pipiens ): effects of host stage and parasite-exposure level. Canadian Journal of Zoology 2003;81(7):1144-1153. R825867 (Final)
  • Full-text: Faxitron-Full Text PDF
  • Abstract: NRC-Abstract
  • Journal Article Schotthoefer AM, Cole RA, Beasley VR. Relationship of tadpole stage to location of echinostome cercariae encystment and the consequences for tadpole survival. Journal of Parasitology 2003;89(3):475-482. R825867 (Final)
  • Abstract from PubMed
  • Full-text: ResearchGate-Abstract and Full Text PDF
  • Abstract: BioOne-Abstract
  • Supplemental Keywords:

    watershed, sediments, risk assessment, health effects, ecological effects, integrity, biotic integrity, chemicals, toxics, metals, organics, ecosystem, indicators, aquatic, aquatic biota, habitat, integrated assessment, Midwest, Minnesota, MN, Illinois, IL, Wisconsin, WI, agriculture, amphibian, community structure, indicator, microbiology, herbicides, parasitic infection., RFA, Scientific Discipline, Ecosystem Protection/Environmental Exposure & Risk, Ecology, exploratory research environmental biology, Ecosystem/Assessment/Indicators, Microbiology, Ecological Effects - Environmental Exposure & Risk, Ecological Risk Assessment, Biology, Ecological Indicators, pesticide exposure, ecological exposure, watersheds, aquatic biota , landscape indicator, amphibians, ecosystem integrity, parasitic infection, multiple spatial scales, biotic integrity, contaminant impact, water quality

    Relevant Websites: Exit

    Progress and Final Reports:

    Original Abstract
  • 1998
  • 1999 Progress Report
  • 2000 Progress Report
  • 2001 Progress Report