10.0 Amphibians and Reptiles

Impacts on Quality of Inland Wetlands of the United States:
A Survey of Indicators, Techniques, and Applications of Community Level Biomonitoring Data
Excerpts from Report #EPA/600/3-90/073
(now out of print)

10.1 USE AS INDICATORS

This discussion addresses the monitoring of "herptiles"—turtles, frogs, salamanders, snakes, crocodilians, and lizards that occur in wetlands. The life histories and requirements of amphibians differ greatly from those of reptiles, and species within each group also differ significantly. Most amphibian species and many reptiles spend all or critical parts of their life in wetlands. However, with only a few exceptions (Brooks and Croonquist 1990, Corn and Bury 1989), their responses to anthropogenic stressors in wetlands have barely been studied in the United States at the community level. Most recent ecological research on herptiles can be characterized as assessments of the occurrence and abundance of particular species in specific micro-habitats. Advantages and disadvantages of use of herptiles as indicators are shown in Appendix A. A possible approach for using assemblages of anuran amphibian species (frogs and toads) as indicators of wetland condition is described by Beiswenger (1988).

Enrichment/Eutrophication. The effects of enrichment on overall community structure of herptiles apparently have not been documented in wetlands, and indicator assemblages of "most sensitive species" remain speculative for this stressor. In southern England, Beebee (1987) found that the bullfrog, Bufo calamita, consistently selected the more eutrophic wetlands.

Organic Loading/Reduced DO. The effects of severe organic loading, e.g., from wastewater outfalls, on overall community structure of herptiles apparently have not been documented in wetlands, and indicator assemblages of "most sensitive species" remain undefined for this stressor. Toxicological data were reviewed by Birge et al. (1980). Anderson (1965) noted that a moderate amount of sanitary sewage pollution seemingly increased the dominance of soft-shelled and snapping turtles in parts of the Missouri and Mississippi Rivers, but heavy industrial waste nearly eradicated turtles for miles downstream, especially the Ouachita map turtle, in part a mollusk-eater.

Contaminant Toxicity. The effects of heavy metals, pesticides, oil, and other contaminants on the overall community structure of herptiles apparently have seldom been documented in wetlands, and indicator assemblages of "most sensitive species" remain speculative for such stressors. Speculation about causes of regionwide or even global declines in several wetland amphibians (e.g., northern leopard frog, boreal toad, spotted frog, tiger salamander in the Rocky Mountains) has often focused on either (a) effects of airborne contaminants on growth and development of tadpoles (Phillips 1990), or (b) effects of increased ultraviolet-B radiation as a result of trophospheric ozone depletion, since such declines have been noted in otherwise seemingly pristine wetlands.

Some laboratory based toxicological data for individual species may be found in USEPA (1986), EPA's "AQUIRE" database, and the U.S. Fish and Wildlife Service's "Contaminant Hazard Reviews" series that summarizes data on arsenic, cadmium, chromium, lead, mercury, selenium, mirex, carbofuran, toxaphene, PCBs, and chlorpyrifos. However, relatively few field data are available for judging which wetland species are most sensitive.

Acidification. Larval stages of amphibians have been suspected of being highly sensitive to acidification effects. Although impacts at the species level have most often been reported (e.g., Clark 1986a,b, Corn and Fogelman 1984), acidification impacts on the overall community structure of herptiles have been documented in wetlands only recently (e.g., Corn et al. 1989, Leuven et al. 1986). Turner and Fowler (1981) found significantly fewer species in wetlands with pH of less than 5.5.

A few species, e.g., wood frog ( Rana sylvatica ), are known to be particularly tolerant of acidic pH in bogs (Karns 1984). However, most amphibians require a pH of higher than 4.5 to 5.0 for embryo survival and metamorphosis (Corn et al. 1989, Freda 1986, Freda and Dunson 1986, Gosner and Black 1957, Karns 1984, Pierce 1985).

Acidic conditions in surface-mine (constructed) wetlands were implicated as a reason for reduced amphibian use by Hepp (1987). Based on a single pH measurement from each surface-mine pond, the mean pH at which various species occurred was given by Turner and Fowler (1981) as follows:

	pH	# of ponds
Northern Spring Peeper	5.2	16
Pickerel Frog	5.42	11
Red-spotted Newt	5.80	8
Gray Tree Frog	5.96	9
Bullfrog	5.91	6
American/Fowler's Toad	5.97	7
Northern Cricket Frog	6.00	1
Wood Frog	6.25	2
Green Frog	6.26	8
Spotted Salamander	6.32	8
Upland Chorus Frog	6.33	8

Similar types of data are presented by Clark (1986b) for Ontario wetlands.

Most of the true frogs are thought to be especially sensitive to acidic precipitation because they respire through their skin. During foggy periods such respiration may occur while they are out of the water. At such times, they may be directly exposed to airborne contaminants.

Salinization. The effects of salinization, e.g., from irrigation return water and oil drilling wastes, on overall community structure of herptiles apparently have not been documented in wetlands, and indicator assemblages of "most sensitive species" remain undefined for monitoring salinization. In softwater lakes and streams, moderate increases in water hardness and alkalinity can result in increased amphibian densities (Hepp 1987).

Sedimentation/Burial. Moderate increases in soft bottom sediments can increase habitat for overwintering turtles. However, excessive sedimentation can smother eggs of many amphibians and alter food sources. The North American dusky salamander ( Desmognathus fuscus ) and the spring salamander ( Gyrinophilus porphyriticus ) are reportedly very sensitive to effects of bank erosion, sedimentation, and turbidity (Campbell 1974, Orser and Shure 1972). However, the effects of sedimentation/ burial (e.g., of amphibian eggs) on overall community structure of herptiles apparently have not been documented in wetlands, and indicator assemblages of "most sensitive species" remain speculative for this stressor.

Turbidity/Shade, Vegetation Removal. Many herptiles are sensitive to the presence and type of vegetation and its juxtapositioning with open water, particularly in arid regions. In Colorado River riparian zones, lizards were most abundant in shoreline habitats, moderately dense in riparian habitats, and least dense in non-riparian or upland habitats; densities depended on insects inhabiting herbaceous debris heaps and litter piles washed up by the river (Jones and Glinski 1985, Warren and Schwalbe 1985). In more humid Oregon watersheds, amphibian richness, density, and biomass were less in logged watersheds than in unlogged watersheds, particularly when vegetation removal occurred primarily in headwater areas (Corn and Bury 1989). Herptile richness was also less in Pennsyvania watersheds with disturbed stream corridors than in those with intact riparian vegetation (Croonquist 1990). Use of riverine wetlands by herpetofaunas has been positively related to number of cover types, sinuosity, circumneutral pH, and gradual shoreline slopes (Hill 1986). Richness of breeding frogs may be related also to the variety of herbaceous plant forms in a wetland (e.g., Diaz-Paniagua 1987).

The community composition of Minnesota amphibians was found to be correlated with wetland vegetation form. The leopard frog ( Rana pipiens ) was found most frequently in sedge mat and less commonly in the very wet tamarack zone. The wood frog ( Rana sylvatica) was found primarily in the fir-ash zone with lesser numbers in the spruce and tamarack. Spring peeper ( Hyla crucifer ) and swamp frog ( Pseudacris nigrita ) were found in the two zones most distant from the pond, spruce and fir-ash (Marshall and Buell 1955).

Despite these initial efforts, indicator assemblages of "most sensitive species" of herptiles remain speculative in most of the U.S. for monitoring effects of vegetation removal, and the effects of vegetation removal on overall community structure of herptiles apparently have not been documented in wetlands.

Thermal alteration. Herptiles, as ectotherms, are particularly sensitive to thermal alteration of wetlands. Although a vast literature exists describing thermal preferenda of individual species, the effects of thermal alteration on overall community structure of herptiles apparently have seldom been documented in wetlands. Lack of comparative studies has resulted in a lack of information on most-sensitive indicator assemblages.

Dehydration/Inundation. Changes in wetland water level alter the quantity and quality of herptile habitat, and may trigger immigration, emmigration, and breeding of particular species and their predators (Pechmann et al. 1988). The effects of dehydration may be particularly severe if dehydration occurs during herptile hibernation, due to the effects of exposure and increased predation of eggs (Campbell 1974).

Impoundment has been reported to increase the regional populations of toads and turtles (Anderson 1965), or at least causes a shift in spatial distribution of habitat. However, inundation can reduce and alter the seasonal timing of flooding of downstream habitats. The resultant changes in vegetation and floodplain leaf litter accumulation can reduce both abundance and diversity of reptiles, as reported by Jones (1988) for Arizona riparian systems. Also, if inundation causes temporarily flooded wetlands to become connected to permanent waters, predatory fishes can gain access to the temporary wetlands, perhaps resulting in reductions in some amphibians (e.g., Dodd and Charest 1988). Temporary dehydration of wetlands may have the opposite effect, benefitting amphibians by reducing fish predation. The ratio of non-predatory to predatory salamanders can increase in wetlands following dry springtime conditions (Cortwright 1987).

Herptile taxa that characterize seasonally flooded wetlands or have terrestrial phases appear to resist effects of urbanization more than those that characterize permanently flooded wetlands or which spend their entire life cycle in wetlands (Minton 1968). In San Francisco, Banta and Morafka (1966) attributed the decline of the native California red-legged frog ( Rana aurora draytoni ) and the introduced leopard frog ( Rana pipiens ) to dehydration and filling of wetlands. Leopard frogs also declined in Colorado as a probable result of drying up of breeding ponds during a drought (Corn and Fogleman 1984). Vickers et al. (1985), studying aquatic and semi-aquatic amphibians in northern Florida cypress wetlands, found no change in mean numbers, numbers of species, or species diversity in ditched versus unditched wetlands. However, species richness declined and terrestrial species became more abundant with ditching.

Fragmentation of Habitat. We found no explicit documentation of herptile community response to fragmentation of regional wetland resources, although the presence of some individual species, e.g., spotted salamander, is known to sometimes depend on proximity to source ponds (Cortwright 1987). One can surmise that as the distance between wetlands containing herptiles becomes greater, and/or hydrologic connections become severed by dehydrated channels, dams, or (particularly) roads, species most dependent on wetlands and/or which do not disperse easily might be most affected (Campbell 1974, Croonquist 1990). In Oregon, Corn and Bury (1989) found that logging upstream from unlogged habitats had no effect on the presence, density, or biomass of any species inhabiting the unlogged habitat.

Other Human Presence. The introduction by humans of non-indigenous aggressive predators (e.g., bullfrog, snapping turtle, and several predatory fishes) into particular water systems has sometimes led to a decline in richness of indigenous frog communities (e.g., Hammerson 1982, Hayes and Jennings 1986, Moyle 1973).

10.2 SAMPLING METHODS AND EQUIPMENT

Some factors that could be important to measure and (if possible) standardize among wetlands when monitoring anthropogenic effects on community structure of herptiles include:

water depth, temperature (site elevation, aspect), conductivity and baseline chemistry of waters and sediments (especially pH, DO, and suspended sediment), current velocity, stream order or ratio of discharge to watershed size (riverine wetlands), shade, amount and distribution of cover (logs, crevices, etc.), ratio of open water to vegetated wetland, extent of plant litter and rotting logs, vegetation type, and the duration, frequency, and seasonal timing of regular inundation, as well as time elapsed since the last severe inundation or drought.

Sampling methods for wetland herptile communities are described in Bury and Raphael 1988, Corn and Bury 1990, Halvorson 1984, Jones 1986, Scott 1982, Vogt and Hine 1982, and others.

Because amphibian distribution and abundance has strong ties to seasonal hydrologic phenomena and the capability of particular species for dispersal, the temporal and spatial variability in amphibian community structure strongly reflects these factors. As is the case with sampling macroinvertebrates whose communities are similarly dependent on ephemeral hydrologic events, sampling amphibian communities can require several repeated visits to a wetland to fully describe community composition. Nonetheless, Corn and Bury (1989) assert that, at least for riparian communities of the Pacific Northwest, amphibian population densities are usually stable in undisturbed habitat and serve as better indicators of habitat quality than do similar densities of birds or mammals.

Herptiles can be sampled during the mid- to late growing season when maximum numbers of developing juveniles are present. However, many species are easy to find only after the first few days of rain following a drought, during late-summer thunderstorms, during the first spring thaw in northern areas, during mid-day basking hours, or at night (Kaplan 1981). Occasionally, traditional winter hibernation areas can be located and used to count individuals representing a larger (but undefinable) area. For Arizona, Jones (1986) noted trapping was most effective in riparian habitats between May and July.

Methods used in wetlands for herptile sampling potentially include pitfall traps (often with drift fences and baited), visual belt transects, direct capture methods, and vocalization recording.

Pitfall traps and funnels are perhaps the most widely used mechanism for capturing herptiles (Jones 1986). Animals enter and cannot find the opening to escape. They are subsequently identified, counted, measured, and released. To reduce loss of trapped animals to predation, traps and funnels can be checked regularly (at least every other day) and can be shaded, and/or filled with sufficient moist plant litter to minimize physiologic stress to animals. Pitfall traps are impractical in many wetlands where the water table is so close to the land surface that pits fill rapidly with water.

Pitfall traps and funnels often produce more species per sampling effort than direct capture methods (Jones 1986). The size of the trap, baits used, and trap placement can affect the species that are caught. Trap and funnel methods can provide relatively quantitative data, when arranged systematically and level-of-effort (e.g., "trap-hours") is standardized.

They involve emplanting a container in the soil, either on the periphery of the wetland or within it (if surface water is absent), with the lip of the container placed flush with the ground surface. Herptiles stumble in and cannot climb the steep sides to escape. Because some species can drown if the container fills with rainwater, Jones (1986) recommends placing floatable material (e.g., styrofoam) in the container to reduce mortality.

Funnel openings are usually oriented toward land for greatest effectiveness. Hoop funnel traps are generally used for turtles, and other funnel traps are used (particularly in deeper wetlands) for catching salamanders, frogs, and occasionally snakes. A special kind of floating pitfall trap can be used to sample basking turtles (see Jones 1986 for description). Aquatic turtles in a Missouri marsh were captured using hoop and net traps. Traps were baited with sardines, local fish, tadpoles, frog, crayfish, dragonfly larvae, snails, and clams (Kofron and Schreiber 1987).

The efficiency of traps and funnels can be increased by channeling movements of herptiles in the direction of the trap or funnel. This is commonly done with "drift fences" (Gibbons and Bennett 1974). These are fences constructed of wire screen or polyethylene plastic, with lengths upwards of 15 m. Traps are placed at both ends of the drift fence, along the fence at various points, or at the junction of several intersecting fences. The bottom edge of the fence is emplanted in the ground, or at least no space is provided for herptiles to crawl under the fence.

Drift fences and pit traps can be more effective and less biased than log-turning, walking transects, electroshocking in streams, or searching and digging through litter. However, they are expensive; time and cost estimates for drift fence trapping are provided by Gibbons and Semlitsch (1982). Jones (1986) comments that, for quantifying herptile communities, drift fence/pitfall trap methods are less effective for frogs, toads, large snakes, terrestrial turtles, and salamanders than for small snakes and lizards.

Sizes and shapes of containers and associated drift fences and their configurations vary greatly, depending partly on target species and wetland type. Vickers et al. (1985) sampled in and around cypress ponds using arrays of four 7.6 m lengths of 0.75 m high, 6 mm polyethylene drift fences arranged perpendicularly and attached at the center. The fences were held upright with wooden laths and buried to 10 cm depth to avoid animals passing underneath. Two aluminum screen wire funnel traps, 75 cm long with 20 cm entrance funnels were placed beside each drift fence. To insure that a trap would always fall on the ecotone regrardless of pond size, distance between arrays was standardized at one-half the pond radius. Working in peatland vegetation in Maine, Stockwell (1985) censused herptofauna in eight vegetation types -lagg, forested bog, wooded heath, shrub heath, moss lawns, pools, streamside meadow, and shrub thicket. Drift fences of free standing aluminum flashing were used as well as those of lath supported polyethylene. Pitfall traps were made of two #10 tin cans joined with tape and silicone. Funnels made from margarine tubs were used in the top of each trap to prevent escape and 2-3 cm of water placed in the bottom of each trap to prevent desiccation of captives. Similar traps were made by Jones and Glinski (1985) using double-deep 3 lb. coffee cans with a lid placed 15 cm over the top to prevent desiccation.

The above methods require multiple visits to a wetland, first to set up and later to check traps. Herptiles can also be monitored directly, that is, during a single visit, or without having to wait for traps to catch individuals. However, direct methods usually do not provide accurate quantitative data on abundance. Unless frequent visits are made and the correct microhabitats are searched at the proper times of year, direct methods are also unlikely to yield good estimates of species dominance or richness. However, they can provide a useful complement to trap methods, locating species that are not easily trapped.

The simplest type of direct search involves scanning a wetland with binoculars to observe the more obviously visible species such as basking turtles, frogs, and alligators. In some cases, floating egg masses of amphibians can also be detected visually and identified to species. Observational methods can be done formally, along defined transects. Searches on foot, perhaps employing many people shoulder-to-shoulder (e.g., Marshall and Buell 1955) have been used, but could be impractical and destructive of habitat in many wetlands. Long-handled nets can be used to surround logs and rocks as they are lifted to search for herptiles, so as to catch individual herptiles as they flee. In riverine wetlands, fine-mesh seines (see Fish section above) can be used for similar purposes.

To enhance opportunities for encountering herptiles during direct searches, electrofishing and identification of vocalizations and tracks can be used. Electrofishing methods (described in section 9, used in conjunction with sweep nets or seines, are particularly effective for retrieving larger salamanders and frogs. Because some species leave distinctive tracks, travel corridors can be searched periodically for tracks. Frogs can sometimes be located more easily at night, as their eyes reflect in the beam of a flashlight. Vocalizations of many frogs and toads are easily identified (commercially-available recordings are available to learn these) and can be used to augment observations. Frogs and toads can sometimes be induced to vocalize by introducing sharp, loud sounds or played-back tape recordings of vocalizations. Low-altitude overflights or aerial photography under favorable conditions can be used to identify alligator holes and paths.

10.3 SPATIAL AND TEMPORAL VARIABILITY, DATA GAPS

In general, quantitative data on structure of the entire herptile community of wetlands has not been uniformly collected from a sufficiently large, statistically-drawn sample of wetlands in any region of the country. Thus, it is currently impossible to state what are "normal" levels for parameters such as herptile density, species richness, or biomass, and their temporal and spatial variability, in any type of wetland.

We found only a few published studies that quantified the entire herptile community (or a large proportion of it) across a region and/or among a set of wetlands:

Brooks et al. 1985, 1987, 1989, Clark 1986b, Congdon et al. 1986, Corn and Bury 1989, Corn et al. 1989, Fowler et al. 1985, Gibbons and Semlitsch 1982, Jackle and Gatz 1985, Karns 1984, Hepp 1987, Stockwell 1985, Turner and Fowler 1981, Ward 1988.

We found no journal articles that quantified year-to-year variation in the entire community structure of herptiles in wetlands, but conceivably such unpublished data may be available from sites of the U.S. Department of Energy's National Environmental Research Park system, and sites of the National Science Foundation's Long Term Ecological Research (LTER) program. Some studies (e.g., Corn et al. 1989) have featured qualitative re-checking of wetlands known in previous decades to have particular species, but probably could not be termed "long-term monitoring."

Quantitative data on community composition of wetland herptiles is virtually lacking from all regions except the Southeast, Southwestern riparian areas, and parts of the Northeast and Pacific Northwest. Information on impacts is limited mostly to studies of hydrologic effects and vegetation removal; especially little is known of impacts from contaminants, salinization, sedimentation, and habitat fragmentation.

Qualitative lists of "expected" herptiles have been compiled by statewide herptile atlas projects in Illinois, Kansas, Massachusetts, Maine, and perhaps other states, as well as by less comprehensive surveys in various smaller areas. Species that show highest affinity for wetlands of various types might be identified by consulting with local herpetologists, Niering (1985), and the "Vertebrate Characterization Abstracts" database managed by The Nature Conservancy and various state Natural Heritage Programs. Limited qualitative information may be available by wetland type from some of the "community profile" publication series of the USFWS (Appendix C).