Survey Methods

There are multiple ways to sample amphibians. Techniques may be species-, habitat-, or season- specific. Every technique is accompanied by underlying biases, limitations, and assumptions that need to be considered carefully, in the context of the goals of each study, before the technique is implemented. Multiple techniques are often used in a single area because one technique is generally not adequate to document all species of interest. Heyer et al. (1994) provides an excellent description of many techniques and their assumptions (Heyer, W.R., M.A. Donnelly, R.W. McDiarmid, L.C. Hayek, and M.S. Foster. 1994. Measuring and monitoring biological diversity standard methods for amphibians. Smithsonian Institution Press, Washington, D.C.). The SE ARMI web site also provides good information on sampling design and technique limitations.

We have emphasized a metapopulation approach in our sampling design because amphibian populations are often linked among breeding sites by migration. The use of a metapopulation framework (in a large landscape, the sampling of groups of sites rather than individual sites) will improve the efficiency of field work and may allow us to examine the role of isolation in local population dynamics. In National Parks or other large tracts of land, we use drainages as the sampling unit, and use a selection scheme that provides relatively even spatial coverage of the study area.

We built a set of hierarchically nested drainage catchments from the USGS Elevation Derivatives for National Applications (EDNA) Project to determine our survey units. Drainage boundaries are overlaid onto National Wetlands Inventory (NWI) maps for the study area. This method allows us aggregate "drainages" to a hierarchic level as necessary. The criterion for describing a drainage is the inclusion of 10-50 identifiable NWI water bodies. This exercise results in a variable number of potential survey units depending on the size and topography of the study area.

We conduct surveys of all potential amphibian breeding sites within the chosen drainage during times when amphibian eggs, larvae, or metamorphs are likely to be present if the species occurs in the area.

Due to small size or lack of topographic relief, some survey areas do not lend themselves to using drainages as unit boundaries. In these cases such as our surveys on Bureau of Land Management (BLM) lands in Wyoming and at Great Sand Dunes National Park in Colorado, we use grid cells rather than drainages to define our study units.


Drainage delineation for Glacier National Park, Montana. 15 watersheds (green) were sampled in 2001, not including apex sites (white).

Drainage deliniation for Rocky Mountain National Park, Colorado. 8 drainages were sampled in 2003, not including apex sites.

We use Visual Encounter Surveys (VES) to document the presence of amphibians. Visual Encounter Surveys are effective in most habitats and for most species that breed in lentic (non-flowing) water. Surveys are conducted by a pair of workers walking slowly around the perimeter and shallows of a water body while searching for amphibians. Multiple transects are used for large wetland complexes such as marshes that cannot be surveyed completely. Most amphibians are identified by sight, but surveyors use dip nets in areas with vegetation or where water clarity is poor. We record all life history stages (eggs, larvae, juvenile, or adult) of each species observed; however, detection of evidence of breeding activity (presence of eggs or larvae) is emphasized so that future surveys can use changes in the number and distribution of breeding sites for each species to assess status and trends. Data collected yields information on the presence of a species but does not establish absence, nor does it provide reliable estimates of abundance.

Our goal is to survey a sub-sample of sites > 2 times within a short time period (e.g., 2 weeks) so that we can estimate species-specific detection probabilities. Collection of data in this manner allows us to estimate the proportion of the sampling area occupied by breeding populations of each species (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).

Proportion of Area Occupied (PAO) by breeding populations of each species is the primary response variable used for detecting trends of amphibians in the Rocky Mountain region. Proportion of Area Occupied is based on population models that incorporate detection probabilities along with the number of water bodies where each species is detected to estimate the area occupied by each species. For common species, PAO is based on the proportion of sites where breeding populations are detected. For rare species such as the boreal toad, PAO may be scaled up to the proportion of watersheds (or other large units of land) occupied rather than sites.

Survey methods, National Elk Refuge, Grand Teton National Park. Photographs courtesy of D. Patla

Data Collection

Collective images of team member using a PDA.

In general, our field data are housed in either an Excel* spreadsheet or an Access* database. We began our pilot year (2002) using Personal Data Assistants (PDAs) (e.g., palm pilot* or visor*) for all data collection in the field on all of our projects. Field technicians input data directly into a PDA using forms (the equivalent of datasheets) specifically designed for the task at hand. We are using Access* and Pendragon* to create dataforms (link to samples of our dataforms) that are loaded onto the PDA in the lab, and accessed and completed at the study site. Data is downloaded directly to the office PC through a "hot sync" from each PDA. We use waterproof cases (Aquapac*) to protect the PDA and a backup memory chip to prevent loss of data due to power or battery failure. Pendragon*, the application that enables data collection on the PDAs, was designed in Access. Therefore, all data that is collected with Pendragon creates an Access file.

These files can be linked together to match databases that have already been created locally in Access or exported to Excel. These files can be appended to existing Access data files or copied and pasted onto an existing Excel file. Because data can be downloaded directly from the PDA to a spreadsheet or a relational database, we find that using PDAs decreases transcription error, the number of illegible entries, and typographical errors. Using pre-designed forms on the PDA we can insure completeness of recorded field information (it is possible to disallow the movement to the next box on the form if the current box is not filled in). There is also a substantial reduction in the time from data collection to data analysis.

We estimate that we save a minimum of 60 hours per summer inputting the data and in the quality assurance phase of the project. This represents a substantial savings based on 1.5 weeks of technician time.

Our data will become part of the ARMI national database that is currently under construction (USGS-WERC and USGS-Patuxent).

Survey Methods

Drainage delineation for Glacier National Park, Montana. 15 watersheds (green) were sampled in 2001, not including apex sites (white).

Drainage deliniation for Rocky Mountain National Park, Colorado. 8 drainages were sampled in 2003, not including apex sites.

Survey methods, National Elk Refuge, Grand Teton National Park. Photographs courtesy of D. Patla

Data Collection

* Use of brand names is for informational purposes only and does not indicate endorsement by the U.S. Government.

Back To Top