amphibian, mark-recapture, metapopulation dynamics, site fidelity, survival, VIE tags, artificial treefrog shelters, genetic variation, environmental variation
As urban areas expand, agricultural lands become increasingly important habitat for many species. Compared to some types of agricultural land-use, ranchlands provide vast expanses of minimally modified habitat that support many threatened and endangered species. Conservation biologists can promote ecologically sound management approaches by quantifying the effects of agricultural practices on resident species. I examined the effects of pasture management, cattle grazing, and landscape characteristics on both adult and larval treefrogs in a ranchland in south-central Florida. I experimentally determined optimal deployment of artificial treefrog shelters constructed of polyvinylchloride (PVC) pipe to efficiently sample adult treefrogs (Chapter 1). Seventy-two shelters were hung on oak trees (Quercus virginiana) and cabbage palm trees (Sabal palmetto) with smooth trunks or boots (residual palm fronds), at all possible combinations of three heights (2, 3, and 4 m), four compass directions (N, S, E, and W) and two water levels (with or without 10 cm). Shelter residence was completely dominated by the Squirrel Treefrog, Hyla squirella (N = 65). Significantly fewer H. squirella were found in shelters on palms with boots than on smooth palms or oak trees (0.29 ± 0.21 [mean ± 1 SE hereinafter] versus 1.3 ± 0.21 and 1.1 ± 0.21, respectively), and shelters with water had slightly more H. squirella than those without (1.5 ± 0.19 versus 0.88 ± 0.19, respectively). Orientation and height did not affect the number of treefrogs encountered; thus, the optimal protocol is to deploy shelters on either smooth palms or oak trees, with water, at 2 m height for easy sampling, and in random compass orientations. I used this protocol to sample H. squirella in woodlots surrounding twelve wetlands and examined how time, frog stage and sex, and landscape features influenced treefrog survival, recapture and site fidelity (Chapter 2). I deployed 15 shelters/ha of woodlot within a 100 m buffer around each wetland. I sampled shelters three times during the fall breeding season, removed all shelters to force frogs to overwinter in natural refugia, and replaced shelters for the final spring sampling. During sampling periods, I sexed, measured, and individually marked each frog using visual implant elastomer (VIE) tags. I used Program MARK to build linear models that included either gender group (female, male or juvenile) or life history stage (adult, juvenile) and either time (sampling interval 1, 2, or 3) or season (fall, spring). I used the most informative model as a null model to assess effects of landscape covariates on survival and recapture. Females had higher survival than either males or juveniles, for which estimates were similar (0.867 vs 0.741 and 0.783, respectively). Survival did not vary over time, although there was some support for an effect of season, with lower survival during the final over-wintering period than in the fall intervals (relative variable importance: group = 0.730; stage = 0.134; time = 0.200; season = 0.310). Adults had higher recapture rates than juveniles (average recapture 0.214 vs 0.102), and recapture for both stages varied over time, with highest recapture in sampling interval two (relative variable importance: group = 0.262; stage = 0.514; time = 0.513; season = 0.229). Hyla squirella was extremely site loyal; no individuals moved between sampling sites, and 95% of recaptured individuals were in their original shelter. Strong terrestrial site fidelity calls into question the traditional "ponds as patches" metapopulation view of treefrog population dynamics. Area of woodlot within 250 m was the most important landscape variable in explaining additional variation in both survival and recapture. Frogs had higher survival and lower recapture in larger woodlots, indicating that intact, contiguous woodlots are higher quality habitat than more fragmented woodlots. Neither survival nor recapture varied with wetland grazing treatments or between pasture types. Finally, I experimentally assessed the effects of cattle grazing and pasture management on larval H. squirella. I selected four wetlands: two in semi-natural pastures (SN) and two in intensively managed pastures (IM). One wetland in each pasture type was fenced so that it was released from cattle grazing (R). I collected three clutches of H. squirella eggs (Clutches A, B, and C) and reared tadpoles in the laboratory until Gosner stage 25. In each wetland, I deployed a total of 50 tadpoles from each clutch into 105 L pens constructed of plastic laundry baskets and mesh window screening. Clutch significantly affected tadpole survival, with Clutch A having the highest percent survival, followed by Clutch B and finally Clutch C (41.66, 32.11 - 53.95 [mean, 95% confidence limits hereinafter]; 9.00, 6.76 - 11.88; 2.89, 2.02 - 4.01, respectively). Wetland type also affected survival, with SN wetlands supporting significantly higher survival than IM wetlands (SN-R: 53.95, 32.88 - 88.13; SN-G: 18.95, 11.30 - 31.36 vs IM-R: 7.32, 4.13 - 12.49; IM-G: 1.09, 0.29 - 2.39). Genetic variation in survival confirms the potential for H. squirella to adapt to rangeland management, but extremely low survival of some clutches indicates that few clutches may be able to survive in low quality wetlands, such as IM-G wetlands. Higher survival in SN pasture wetlands suggest this is a superior habitat and future management objectives should conserve semi-natural pastures and limit further modification of intensively managed pastures, including removing woodlots near wetlands.
Master of Science (M.S.)
College of Sciences
Length of Campus-only Access
Masters Thesis (Open Access)
Windes, Kathryn, "Treefrog (hyla Squirella) Responses To Rangeland And Management In Semi-tropical Florida, Usa" (2010). Electronic Theses and Dissertations. 4412.