USGS - science for a changing world

South Florida Information Access (SOFIA)

projects > use of amphibian communities as indicators of restoration success in the greater everglades > work plan

Project Work Plan

Department of Interior USGS GE PES
Fiscal Year 2012 Study Work Plan

Project Title: Use of Amphibian Communities as Indicators of Restoration Success in the Greater Everglades
Investigator(s): Susan C. Walls (SESC) and J. Hardin Waddle (NWRC)
Principal Investigator Contact Information: Susan C. Walls, U.S. Geological Survey, Southeast Ecological Science Center, 7920 NW 71st Street, Gainesville, FL 32653. Office: (352) 264-3507; FAX: (352) 395-6608; E-mail:

Statement of Problem: Hydrology is a major driving factor in Everglades habitats and strongly influences the occurrence of amphibians across the landscape. Previously, a model of amphibian occupancy in relation to hydrology and habitat (The Stressor Response Model, funded through the USGS GE PES) was created for use as a restoration evaluation tool. This model uses estimates of occupancy for assemblages of amphibians. In turn, this community estimate serves as an index, a target for restoration assessment and, in a spatial framework, as a tool for evaluating alternatives. This index is designed to be used to predict the effects of various water management scenarios on an amphibian community. The goal of the current phase of this work is two-fold: (1) to extend the existing Stressor Response Model to include the central Everglades, and (2) to collect additional parameter estimates that can be used to validate this newly-developed model with respect to predicted responses of an amphibian community to hydrologic and habitat restoration in the Picayune Strand Restoration Project watershed. This model also serves as a promising tool for assessing the potential impact of climate change on amphibians occupying a hydrologically modified/restored landscape.


  1. Use occupancy estimates generated from previous field work to extend the Stressor Response Model to include the Greater Evergades (previously funded as restoration science through USGS GE PES).
  2. Estimate amphibian site occupancy on lands included in the Picayune Strand Hydrological Restoration Project (adjacent to the Greater Everglades), versus occupancy on adjacent conservation lands with relatively intact hydrology. These parameter estimates may be used to validate the newly-developed expansion of the Stressor Response Model.
  3. Generate products on estimating site occurrence of amphibians, along with predicting responses of amphibians to hydrological restoration and future climate change across this landscape.
  4. Generate additional products that will include reports, data, presentations, summary fact sheet for SFNRC/CESI website, and peer-reviewed publications. Forecast potential responses of the amphibian community to the planned hydrological restoration of Picayune Strand.

Long-term goals:

  1. Provide insight into management options with respect to the effects of various water management scenarios on the amphibian community in the Picayune Strand.
  2. Refine the Stressor Response Model to address how climatic variation may impact the hydrology of sites and, thus, their occupancy by amphibians.

Development of a model of amphibian occurrence and species richness for the greater Everglades (J. H. Waddle). - Amphibians in the Everglades, even the more terrestrial ones like toads and treefrogs, require water in which to lay eggs and develop as larvae. The hydrology of a site interacts with the habitat of a site to determine the suitability for the occurrence of each species. Some amphibians require shallow, short-hydroperiod wetlands for breeding where fish and other aquatic predators are absent, present but in low density, or mediated by extensive vegetative cover to offer protection to vulnerable eggs and tadpoles. Other species require long hydroperiods or even permanent water to accommodate long larval periods. Therefore the interaction of habitat and hydrology can be an important determinant to which species of amphibians will be present at a site. By understanding the effect of hydroperiod and habitat on the occurrence of amphibian species we can build a model to predict individual species presence and the overall species richness of amphibians at a site. This type of model can be used as a tool to help us evaluate alternative restoration plans.

Taking data from a number of USGS and University of Florida datasets including Rice et al. (2004) and Rice et al. (2005), we performed a multi-species (community) occupancy analysis (Kery and Royle 2008) using habitat and hydroperiod as covariates. The resulting output from this model was used to build linear relationships with habitat and hydroperiod for 12 species of anuran amphibians in the Greater Everglades (modeled below as psi (ψ); interpreted for this model as a Habitat Suitability Index (HSI)). This produces a linear model that can be used to derive the estimate of psi for any cell as long as habitat and hydroperiod inputs are available. To create a model-based tool to project amphibian HSI we have created a base map of habitat for the Greater Everglades. The hydroperiod of a site can be determined from various datasets including the EDEN data library or the RSM model. The HSI of a cell is determined by first obtaining the habitat category of the cell and the hydroperiod for the year. Those numbers are then inserted into the below equation:

logit(ψi) = b0i + b1iHammock + b2iPineland + b3iPrairie + b4iSlough + b5iSwamp + b6iHydroperiod

The model should run a separate equation for each species (i). The habitat in each cell can take one of the five values, so the other habitats are removed from the equation so that only one of the betas 1-5 are used in any cell. Hydroperiod is calculated by appropriate methods depending on the potential use of the model, but then must be back-transformed using the following equation to input into the above linear model:

  (Hydroperiod - 195.5048)
Hydroperiod =  

Finally to convert the logit of the psi value to a number on the real scale we use the following formula:


1 + ex

Where x is the value from the linear model.

This result will be a value between 0 and 1 that represents the probability of occurrence or the habitat suitability of the cell for the species.

The resulting HSI can then be used to draw from a binomial distribution with probability of success (i.e. probability of value = 1) = HSI. Thus occurrence of the species at the site is stochastic but based on real estimates of occurrence probability derived from our original sample data. Species richness at a cell is simply the sum of all species present in the cell.

This model was presented at a workshop on 13 December 2011 and after peer review will be made available for use by managers and decision makers in the Greater Everglades. At least one peer reviewed publication and a presentation at the INTECOL conference in June 2012 are planned.

Estimation of amphibian site occupancy in the Picayune Strand: Methodology. – Since 2008, we have conducted surveys for anuran amphibians within the area targeted in the Picayune Strand Restoration Project (former Southern Golden Gates Estates), as well as the adjacent Belle Meade Conservation and Recreation Lands (CARL) area; together, these two units comprise Picayune Strand State Forest (PSSF). We also conducted surveys in the adjacent Fakahatchee Strand Preserve State Park (FSPSP). Belle Meade and FSPSP are reservoirs of natural habitat that may serve as vital refugia of native flora and fauna for recolonizing Picayune Strand following its restoration. With prior funding from the USGS GE PES Initiative, we expanded our amphibian monitoring in Southwest Florida to include the Picayune Strand and adjacent land management units. In 2008, a total of 27 sites were selected for monitoring in PSSF and FSPSP, and both manual vocalization and visual encounter surveys were conducted to assess amphibian occupancy. In FY 2009, we modified our protocol to use automated recorders (ARU; Song Meter SM1, Wildlife Acoustics) instead of manual surveys. We deployed an ARU at each site and programmed them to record for 5 min at the beginning of every hour, from 1800-0600 h each day. Loggers were installed on trees approximately 2 m above the ground at each site. With FY 2010 funding, we will continue our monitoring at these same sites. Song Scope Bioacoustics Monitoring Software (from Wildlife Acoustics) will be used to build species-specific "call recognizers" and to otherwise aid in identifying frog calls at each site. Recordings will be viewed manually as well in the Spectrogram view of Song Scope to locate unique visual signatures of each species. All identified potential calls will be verified by listening to each. A decibel level > 40 dB will be considered to be "at the site" and not a distant call.

Data are being used to generate unbiased estimates of site occupancy for all amphibian species detected. These components, in turn, influence site occupancy and the overall community composition at a site. We will use a dynamic occupancy model (MacKenzie et al. 2003; Royle and Kéry 2007) to derive estimates of species-specific probabilities of occurrence (ψ), extinction (ε), and colonization (γ), summarized by land management unit (e.g., Belle Meade vs, PSSF, and FSPSP). This model has recently been extended for the joint analysis of communities of species (Walls et al. 2011). In these models replicated observations of apparent detection/non-detection data are used to estimate the ecologically relevant model parameters (ψ, ε, and γ) while accounting for errors in detection of each species. These parameters can be formulated as a function of changes in hydrology (based on both restoration activities and yearly cumulative rainfall) and any restoration-related habitat shifts; thus, they will change from one year to the next. However, the expectation is that sites will become persistently wetter over time as hydrology is restored in these areas. In relatively dry years (i.e., before hydrology is restored and/or in drought years), wetlands with currently long hydroperiods may shift to short or intermediate in hydroperiod, increasing local extinctions of species adapted to permanent or semi-permanent wetlands. Also, in dry years local colonizations may occur as individuals disperse from drying wetlands to inundated ones. In relatively wet years (i.e., as hydrology becomes progressively restored and/or in years with above average rainfall), currently short-hydroperiod wetlands may disappear as all sites become persistently inundated, thus increasing local extinctions of amphibians adapted to ephemeral habitats. The dynamic occupancy model will enable us to predict how large-scale shifts in hydroperiod may affect estimates of species' occurrence and attribute whether altered estimates of occupancy in assemblages of amphibians are associated with changes in extinction, colonization, or both.

Literature Cited:

Kéry, M. and J. A. Royle. 2008. Hierarchical Bayes estimation of species richness and occupancy in spatially replicated surveys. Journal of Applied Ecology 45:589-598.

MacKenzie, D.I., J.D. Nichols, J.E. Hines, M.G. Knutson, and A.B. Franklin. 2003. Estimating site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology 84:2200-2207.

Rice, K. G., J. H. Waddle, M. E. Crockett, B. M. Jeffery, and H. F. Percival. 2004. Herpetofaunal Inventories of the National Parks of South Florida and the Caribbean: Volume I. Everglades National Park. U.S. Geological Survey, Open-File Report 2004-1065, Fort Lauderdale, FL, USA.

Rice, K. G., J. H. Waddle, B. M. Jeffery, A. N. Rice, and H. F. Percival. 2005. Herpetofaunal Inventories of the National Parks of South Florida and the Caribbean: Volume III. Big Cypress National Preserve. U. S. Geological Survey, Open-File Report 2005-1300, Fort Lauderdale, FL, USA.

Royle, J.A. and M. Kéry. 2007. A Bayesian state-space formulation of dynamic occupancy models. Ecology 88:1813-1823.

Walls, S.C., J.H. Waddle, and R.M. Dorazio. 2011. Estimating Occupancy Dynamics in an Anuran Assemblage from Louisiana, USA. The Journal of Wildlife Management 75:751-761.


This project is anticipated to end in FY 12. The amphibian model is in the final stages of completion, which is expected to occur by the end of January, 2012. Following peer review this model will be made available for use by managers and decision makers in the Greater Everglades. At least one peer reviewed publication and a presentation at the INTECOL conference in June 2012 are planned.

Field activities (monitoring of amphibians) ended in September, 2011. We are now in the process of reviewing recordings to identify frog calls. Since September, 2009 we have acquired approximately 390 hours of recordings for each of 27 sites. Once these data are transcribed from the recordings and entered into a database, we will proceed with our planned occupancy analysis, the preparation of manuscripts for peer reviewed journals and presentations at meetings of professional societies. Ultimately, these data may be used to validate the predictive model developed for amphibians in the Greater Everglades.


We will publish peer-reviewed journal articles on the refinement of the amphibian stressor response model and other results from previous and on-going funding. We will present results of our study at national and international meetings, and produce data/metadata, summary reports and a summary fact sheet for the SFNRC/CESI website and GEER conference.

Product Milestone Date
*Fact Sheet for GEER Conference

*Reports/progress summaries for Picayune Strand Forest, Fakahatchee Strand Preserve State Park, and coordinator of the Greater Everglades Priority Ecosystems Science program

12/2010, 2011, 2012
*Oral and poster presentations at professional meetings

07/2010, 2011, 2012
*Additional journal articles as data become available (e.g., "Use of amphibians to establish a baseline for measuring restoration success at a hydrologically degraded site in Southwest Florida, USA". Goal for submission: the journal Restoration Ecology). FY 2012, FY 2013

Relevance and Benefits:

As with our previously funded work, this project continues to address several science objectives in the USGS Science Plan in Support of Everglades Restoration. Primarily, this work is concentrated under the second main goal: "Activities to restore, protect, and manage natural resources on DOI lands in South Florida." The tasks directly address four science objectives:

The need for monitoring and modeling amphibian populations during CERP is specifically mentioned in the DOI Science Plan in Support of Everglades Restoration under several projects such as the Picayune Strand (Southern Golden Gate Estates) Hydrologic Restoration Project. In the DOI Science Plan, the need for monitoring and simulation projects for indicator species is listed in at least the Ten Mile Creek Reservoir Assisted Stormwater Treatment Area, Henderson Creek/ Belle Meade Restoration, Southwest Florida Feasibility Study, Florida Bay and Florida Keys Feasibility Study, Landscape-Scale Modeling, and Everglades National Park Fire Ecology Science Action Plan projects. The need to develop models simulating response of species sensitive to change in hydrology, especially those of threatened and endangered species; and determine response of key indicators to changes in water management is described as a research area/restoration goal of the South Florida Ecosystem Restoration Task Force (SFERTF).

Communication Plan, Technology and Information Transfer:

The results of our work will aid in understanding how amphibians respond to wetland restoration, modification, and creation efforts, as well as hydrological variation over time and space. This information will be useful if restoration plans for the Picayune Strand are refined with respect to the hydrological and habitat needs of amphibians. This information will also be useful for natural resource managers and conservation biologists affiliated with federal (U.S. Fish & Wildlife Service, U.S. Forest Service, National Park Service, and the USGS ARMI program), state (Fakahatchee Strand Preserve State Park, Picayune Strand State Forest, Florida Fish and Wildlife Conservation Commission) and non-governmental conservation organizations such as The Nature Conservancy and Partners for Amphibian and Reptile Conservation. Academicians at universities will benefit as well from the intensive data collection proposed in our study. Our publications will be made available by posting them on the USGS ARMI website. Publications will also be distributed to the agencies listed above for their consideration in hydrological restoration and wetland management.

This work is being conducted in partnership with USGS's Amphibian Research and Monitoring Initiative (ARMI). Reports and publications will be made available through the USGS ARMI database. ARMI has a single, multifaceted amphib­ian database that links field data with statistical parameter estimates for species being studied, health and disease clinical analyses, and geospatial information on potential species ranges. Metadata summarize the goals, locations, and target spe­cies of all field data. A complement to ARMI's amphibian database components is the collection of environmental geo­spatial layers that compose ARMI's geospatial database. The ARMI database already meets NBII metadata standards and is linked to existing NBII data searches.


Cooperators and Partners:

Facilities, Equipment, and Study Area(s): All of our work will be conducted in the field; thus, no laboratory equipment or facilities are needed.

Our focal study site, the Picayune Strand State Forest (PSSF), is targeted for hydrological restoration through the Picayune Strand Restoration Project (PSRP), one of 60+ planned projects to be implemented under CERP. The goal of the PSRP is to counteract the overdrainage that resulted from a failed real estate development project (the Southern Golden Gates Estates) of the 1960's. The development of this area (by the Gulf American Corporation) involved dredging 48 miles of canals and building 290 miles of shell-rock roads. Restoration plans for this area include the installation of a combination of spreader channels, canal plugs, road removal and pump stations in the Western Basin and Big Cypress of Collier County. Thusfar, by 2006 the northern 7 miles of the Prairie Canal (the eastern-most of four large canals originally constructed to provide drainage and flood protection for the planned residential development) were plugged. Most roads adjacent to the canal have also been removed and exotic plant species were removed from the canal banks. The effort completed thusfar is expected to be especially beneficial in terms of reducing drainage of the adjacent Fakahatchee Strand State Preserve.

The area targeted in the Picayune Strand Restoration Project (the former Southern Golden Gates Estates) is bordered to the east by Fakahatchee Strand Preserve State Park (FSPSP) and, to the west, by the Belle Meade Conservation and Recreation Lands (CARL) area. Belle Meade, together with the former Southern Golden Gates Estates, now comprise the Picayune Strand State Forest (PSSF). Both Belle Meade and FSPSP are reservoirs of natural habitat that, in Belle Meade, is largely hydrologically intact. The FSPSP is threatened by unnatural patterns of water flow and unrestricted use in the private ownerships of the area. Belle Meade includes some of the most extensive examples of remaining old-growth wet flatwoods in southwest Florida, along with high quality, undisturbed subtropical dwarf cypress savanna communities. Fakahatchee Strand is the best example of a strand swamp in the United States, and contains the largest concentration and the greatest diversity of native orchids in North America. These two areas are therefore vital refugia of native flora and fauna that can recolonize the Picayune Strand following its restoration. The results of our field surveys in 2008 revealed that both Belle Meade and FSPSP had significantly higher species richness of amphibians than did Picayune Strand. Moreover, 80% of the sites at which Pinewoods Treefrogs (a pine habitat specialist) were detected were within the boundaries of the Bell Meade area. Because of their proximity to Picayune Strand, along with their high species richness of amphibians, Belle Meade and FSPSP will likely play a crucial role in the recovery of anuran amphibians in this ecosystem.

Animal Welfare: Our study does not require handling of any animals as it is based on recording calling anurans.

Legal and Policy-Sensitive Aspects: The PI already possesses a Scientific Collecting Permit from the Florida Fish and Wildlife Conservation Commission that authorizes any necessary handling of amphibians in PSSF and FSPSP (permit # WX08477; expiration date 12/31/2011).