Using Artificial Refugia to Sample Hylid Frogs in Okefenokee National Wildlife Refuge

Jennifer S. Staiger¹, Lora L. Smith², William J. Barichivich¹, and C. Kenneth Dodd, Jr.<sup>1

1</sup>U.S.G.S. Florida Integrated Science Center, 7920 N.W. 71^st St., Gainesville, FL 32653
²Joseph W. Jones Ecological Research Center, Rt. 2, Box 2324, Newton GA 31770

Presented at the Joint Meeting of the American Elasmobranch Society, American Society of Ichthyologists and Herpetologists, Herpetologists' League, and Society for the Study of Amphibians and Reptiles, Kansas City, Missouri.  July 3-8, 2002

Abstract: As part of the Department of Interior's initiative to survey amphibian populations on federal lands, a long-term research and monitoring project is being conducted at the Okefenokee National Wildlife Refuge, Georgia.  Many amphibian sampling techniques are being used, including artificial refugia (PVC pipes) targeting hylid frogs.  Sixteen intensive survey sites were visited an average of 13 times during quarterly and special sampling periods from January 2001 to January 2002. Each intensive site has an array of 30 pipes in trees and/or in the ground, and represents one of six habitat types. Hyla femoralis, H. cinerea, and H. squirella used the refugia (692, 80, and 16 total captures, respectively), with the number of captures generally increasing over time. Recaptured individuals accounted for 46.4% of all H. femoralis caught, 38.7% of all H. cinerea, and 81.2% of H. squirella captures. The majority of H. femoralis captures occurred in forested wetland and pine flatwood sites (36.6% and 16.9%, respectively) whereas most H. cinerea captures were in lake sites (78.7%).  All H. squirella captures were in pine flatwood and ephemeral pond sites (81.2% and 18.8%, respectively).  Hyla femoralis was the only species found at least once at each site and during each sampling period.

Introduction

To address concerns about amphibian population declines, the United States Department of the Interior (DOI) instituted the national Amphibian Research and Monitoring Initiative (ARMI). The goal of the initiative is to promote the inventory and long-term monitoring of the status and trends of amphibian populations on DOI lands.  The Okefenokee National Wildlife Refuge (ONWR) is one of several primary index sites in the southeastern United States where research and monitoring efforts are focused. The research initiative at ONWR includes a baseline inventory to determine presence and distributions of amphibians, as well as sampling to determine species richness, habitat associations, and proportion of area occupied (PAO).

A variety of amphibian sampling techniques are being used at intensive monitoring sites in ONWR, including artificial refugia (PVC pipes) targeting hylid frogs.  Because traditional herpetofaunal sampling methods tend to underrepresent arboreal anurans (Dodd 1991, Greenberg et al. 1994), the use of artificial refugia can be an effective tool when conducting amphibian surveys or long term monitoring of treefrogs (Boughton et al. 2000, Moulton et al. 1996). The data obtained with this method in ONWR are used to evaluate its utility for mark-recapture studies and sampling across different habitats.

Description of the study site: The Okefenokee Swamp is a 200,000 ha freshwater wetland located in southeast Georgia and extreme northeast Florida.  The ONWR encompasses approximately 158,000 ha of the swamp and 98% of the refuge is a designated National Wilderness Area.  The refuge is known for its cypress (Taxodium spp.) dominated wetlands, but also contains dense shrub thickets, extensive wet prairies, lakes, and surrounding pine flatwoods. The area contains significant amphibian biodiversity (21 anurans and 17 salamanders, including 5 species of concern) and the region represents the southern limit of the range of several species.

Methods 

Sixteen intensive monitoring sites, stratified across the major vegetation communities within ONWR, were sampled an average of 13 times from January 2001 to January 2002.  Surveys included quarterly intensive sampling at all sites, one 24 hr pipe survey at four sites, and a preliminary sample at three sites. The monitoring sites represent six habitat types: forested wetland (FW1-4), lake (L1-2), pine flatwoods (PFW1-2), ephemeral pond (P1-2), shrub wetland (ShW1-2), and wet prairie (Pr1-4).

At each site an array of 30 PVC pipes was placed to detect and sample hylid frogs (Figs. 1, 2).  Pipes were placed between late October and December 2000, at least two months before the first sample. Fifteen pipes (5 cm diameter x 60 cm long, bottom capped) were mounted vertically approximately 2 m high on tree trunks (Fig. 3), and 15 pipes (5 cm diameter x 91 cm long) were placed upright in the ground with approximately 60 cm of pipe exposed (Fig. 4).  All pipes were placed in the ground at the wet prairie sites.

Frogs found in the pipes during sampling periods were identified to species, measured (snout-vent length, wet body mass), marked, and released at the point of capture.  Individual frogs were marked with a unique series of toe clips (Hero, 1989).

Results and Discussion 

Three hylid species, pine woods treefrog (Hyla femoralis), green treefrog (H. cinerea), and squirrel treefrog (H. squirella), were found in pipe retreats at ONWR (Figs. 5, 6, 7).  A total of 788 captures of the three species was recorded during quarterly monitoring sessions and other surveys, of which 365 (46.3%) were recaptures.  Hyla femoralis were found most often, with 692 total captures, followed by H. cinerea with 80, and H. squirella with 16 captures. Recaptured individuals accounted for 46.4% of all H. femoralis caught, 38.7% of all H. cinerea, and 81.2% of H. squirella captures. Recaptures at individual sites ranged from zero to 54.0% and averaged 31.8%.  No other hylid species was detected by the other sampling methods (e.g. automated remote call recorders, litter searches) used in the refuge.

The majority of H. femoralis captures occurred in forested wetland and pine flatwood sites (36.6% and 16.9%, respectively) while most H. cinerea captures were in lake sites (78.7%, Fig. 8).  All H. squirella captures were in pine flatwoods and ephemeral pond sites (81.2% and 18.8%, respectively, Fig. 8).  Hyla femoralis was the only species found at least once at each site and during each sampling period.  The number of captures of each species varied over the year (Fig. 9).

More captures came from ground pipes than tree pipes (630 total for ground, 157 total for tree). The effect of pipe type on capture success was significant (Table 1), though there was no significant interaction effect between species and pipe type.  The four prairie sites were excluded from this analysis because these sites had no tree pipes.

The results presented here point to the utility of artificial refugia as a hylid frog sampling method. The method can provide data on species richness and diversity, on population demographics, and for estimating population sizes, survival rates, and capture probabilities with mark-recapture programs such as MARK. The data can also be used in the PAO analysis advocated by ARMI (MacKenzie et al. in press).

Table 1. General linear model table of likelihood-ratio statistics, corrected for overdispersion, for species and pipe type effects on number of captures.  

Source     Num DF    Den DF     F Value     Pr > F    Chi-Square     Pr > ChiSq
Species          2             20           46.59       <.0001        93.18            <.0001
Pipe type        1             20           22.96       <.0001        22.96            <.0001

Literature Cited

Boughton, R.G., J. S. Staiger, and R. Franz. 2000. The use of PVC pipe refugia as a sampling technique for hylid treefrogs. Am. Midl. Nat.144:168-177.

Dodd, C.K. Jr. 1991.  Drift fence associated sampling bias of amphibians at a Florida sandhill temporary pond.  J. Herpetol. 25:296-301.

Greenberg, C.H., D.G. Neary, and L.D. Harris.  1994.  A comparison of herpetofaunal sampling effectiveness of pitfall, single-ended, and double-ended funnel traps used with drift fences.  J. Herpetol. 28:319-324.

Hero, J.-M.  1989. A simple code for toe clipping anurans.  Herpetol. Rev. 20:66-67.

MacKenzie, D.I., J.D. Nichols, G.B. Lachman, S. Droege, J.A. Royle, and C.A. Langtimm. Estimating site occupancy rates when detection probabilities are less than one. Ecology (in press).

Moulton, C.A., W.J. Fleming, and B.R. Nerney.  1996.  The use of PVC pipes to capture hylid frogs.  Herpetol. Rev. 27:186-187.

Acknowledgments

We thank the rest of the Okefenokee field crew: Kevin G. Smith, Kristina Sorensen, Maya J. Zacharow, Audrey K. Owens, Gary L. Hill, and Robert J. Lewis.  A special thanks to Robert E. Bennetts and Catherine A. Langtimm for statistical guidance. We also thank Amy D. Hester and Kelly L. Keefe for fieldwork and the refuge personnel for permission to work in the swamp.

For more information: jennifer_staiger@usgs.gov.