Calling-Amphibian Surveys


Group Leader: Christine Bishop. Reviewed by AMP participants and edited by Leslie Mertz.

In the eastern sections of North America, most anurans call during their breeding season. The timing and extent of those calls vary by species, but by establishing sampling periods for groups of species that call at the same time of year, it is possible to detect most species along roadside surveys. These surveys attract volunteers, and training tapes make it possible to involve even inexperienced observers (Kline, 1995 Shirose et al., 1995).

Current calling-amphibian programs are concentrated in the northern and the eastern sections of the continent. By adjusting the timing of surveys and perhaps the choices of species to be sampled, these recommendations should also be applicable for southeastern anuran populations. Because most of the information available for evaluation came from the Northeast, the recommendations developed here reflect that geographic bias. As discussed in the Western Amphibians section of this document, relatively few western anurans would likely fit this model. However, three species, the western, boreal, and Pacific chorus frogs (Pseudacris triseriata, P. maculata, and P. regilla) are suitable for this type of survey and have been incorporated into the project.

Calling-amphibian surveys are inexpensive to manage, and variations in index numbers are low enough that a moderate number of survey routes in a state or province will have the power to detect major population and distribution changes in most frog and toad species (Kline, 1995 Mossman et al., 1995). Our goal is to design a network of calling-amphibian surveys in North America, creating a statistically defensible means of detecting population changes and/or change in geographic distributions in calling anurans.

Automated surveys still demand more testing and standardization before they can be used as full replacements for the calling-amphibian surveys delineated here. In addition , most project leaders find that automated surveys are often prohibitively expensive.

The following principles guided the development of the Critical Elements.

Calling-Count Surveys Should:

1. Provide early warning of declines in population size or occurrence of anurans.

2. Promote public involvement in protection of amphibians and educate the public about amphibians.

3. Be adaptable for studying the fluctuations and trends in local, regional and continental anuran populations.

4. Contribute toward separating chronic changes (especially declines) from natural fluctuations.

5. Complement intensive monitoring and research on amphibians.

6. Contribute toward defining the environmental stressors that affect amphibian populations.

Call Counts - Critical Elements

1. Program Overlap. Calling counts need not supplant other ongoing monitoring programs for amphibians. More than one method of monitoring amphibians is acceptable and valuable. Trends can be compared among methods. Parallel trends among several programs would increase the credibility of conclusions drawn from monitoring efforts (see Appendix 1 for further discussions regarding sampling).

2. Standardization of Listening Time. The number of detections of new species rapidly falls off after the first minute of listening time (Shirose et al. 1995). Time for an individual survey stop along a route has been set at three minutes. Observers can listen for up to two minutes longer than the initial three minutes if noise from traffic, etc., is interfering with the ability to hear calls. We also advise volunteers to wait one minute after arriving at the point and then begin the three-minute survey. This will reduce the effect any disturbance made by the volunteers' approach may have had on anuran calling.

Because individuals cannot be discerned in most calling amphibian choruses, the Wisconsin system of calling chorus categories should be used (Mossman, 1994). The training tape should contain examples of these categories along with tests to make sure the observers are all calibrated in the sense that they can all correctly identify anuran calls.

3. Training. Identification training tapes need to developed for each state and province. Tapes should include the range of calls for each species as well as a test section at the end to verify the observer's abilities. Training seminars for volunteers are also useful.

4. Validation. Validation studies for call counts should be conducted across each new biogeographic area where a monitoring program is established. This validation will determine those species that occur within the region and that will be detected by the calling survey. Use of an automated monitoring system is of great value in determining the species that actually are calling each evening versus those heard by volunteers on the three-minute survey (Dorcas et al. 1995).

5. Sampling Frame. We recommend that a stratified, randomized method be employed to determine sampling locations. We differentiated between a large "area" of similar habitat (e.g., a large bog several hectares in size) and a large area with several discrete but similar habitat types or "basins" (e.g., many farm dugouts - water-filled pits used as watering holes - across a large agricultural landscape) that was to be surveyed. See Appendix 1 for further discussions on the important topic of sampling frame and its history and problems in developing extensive monitoring programs for wildlife.

To better sample rare or highly specialized anurans in some regions, further stratification beyond a general approach to placing points at wetlands may be needed.

Investigations from the Wisconsin system indicate that 40 survey routes in a state/province of that size were adequate to get good estimates of almost all of the species (Mossman et al., 1995). These results need to be confirmed for other regions of the continent.

The suggested approaches are:

Approach #1

a) A random starting point and direction would be assigned (for practical purposes, these could be randomly selected road intersections) b) Wetlands along the roadside segment would be identified without prior knowledge of their capacity as amphibian breeding sites. c) A volunteer or technician would drive or walk the route listening only at the wetlands that were identified in (b). d) Wetlands must be a minimum of 500m apart on both driving and walking surveys. The 500m distance corresponds to the minimum distance at which an observer is no longer able to hear calling frogs (Helferty, 1995).

Approach #2

a) A grid would be placed on a map of the area to be surveyed. b) For each grid wetlands would be identified (see #6 below) without prior knowledge of their capacity as amphibian breeding sites. c) Wetlands would be randomly selected from the group described in (b) and a route would be established among them. d) Wetlands must be at least 800m apart on a driving survey and no more than 500m apart on a walking survey.

6. Determining Potential Wetlands. Wetlands can either be determined by using existing wetland maps (e.g., National Wetland Inventory in the U.S.) or by giving volunteers a clear set of guidelines on how to determine whether a tract is a potential wetland site. Because many wetland maps do not delineate some types of vernal pools, there are advantages to using observers or technicians to do so.

It is best to scout the area during or just prior to the breeding season to determine wetland suitability. Many wetland areas can dry up during the summer. Identification of potential wetland sites should be done during the day and should not take into account the presence or absence of amphibians during that time. The best approach would be to have an separate group establishing the points from those running the surveys.

7. Index of Change. What is our "Index of Change" in population or species occurrence and population size? We recommend that the data collected in this fashion could be statistically analyzed at the route, physiographic region, state/province and continental levels. Data could also be compared among points and routes with similar habitats, e.g., data at all vernal pools among routes across a particular biophysical region.

8. Sampling Period. Based on the experience of the participants at the Toronto (1995) meeting, volunteers can become bored or they will not participate in a survey that requires too much time in the evening, especially after sunset. We recommend that a route should require no more than one to two hours of observer time to complete. The preferred period would be one hour to complete the route itself, because of commuting time from the volunteer's home to the route. The observer should begin the counts one-half hour after sunset, (approximately when birds stop calling) and complete the entire survey within two hours.

9. Sampling Site Destruction. Wetland breeding sites for amphibians come and go. It is to be expected that some sites will be better over time (for example, beaver activity may form a new pond), and others will disappear (for example, under a new shopping center parking lot). The stops should not be changed to incorporate new sites or eliminate sites that are no longer available. This is important because even if the main wetland is destroyed, other water sources, such as water control pits and drainage ponds, may still provide suitable frog habitat.

Under unusual circumstances it may be possible that new forms of wetland may be developed where none formerly existed. These new wetlands would likely to be sampled unequally from existing sampling sites. If these new wetlands occur in places that were not initially determined to be within the realm of possible wetland locations (for example, the creation of new cattle tanks in upland areas that previously did not support anuran populations), then the survey will have to be statistically adjusted to account for this new habitat. A review of wetland trends every 5 years should catch any such changes.

10. Historical Data. Some regions will be confronted with the problem of how to respond to the protocols outlined in this document when they already have calling surveys of different types in place. This requires careful consideration. See Appendix 1 for a discussion of the problems of interpreting biased samples and transitions to a new system. A transition from old surveys to new can be made using a period of years where both surveys are conducted.

If historical routes were chosen and developed in an unbiased manner (e.g., they were randomly or systematically placed and not simply chosen by the observer), then it should be possible to incorporate these surveys into a new system. However, if the current system is biased (for example, the Wisconsin method samples locations only where amphibians are currently breeding), then it is probably best to redesign the system and begin again. Note that the example of the Mourning Dove call counts given in Appendix 1 did just that. Even after going through an extended transition period, the 10 years or so of data were eventually completely dropped from all analyses in deference to the replacement system.

At this point most groups have spent relatively few years sampling amphibians. Compared to the desired longevity of such systems, it makes sense to consider transitions to a better system as soon as possible.

11. Special Habitat or Program Surveys. Some surveys sample only selected marshes or areas readily accessible from backyard porches. The data collected for those specific locations can only be compared among themselves and not to other habitats. For example, surveys from back porches will tell you about trends in calling amphibians heard from the porches of people inclined to participate in such surveys. The conclusions drawn from such a survey would have to be limited to the world of amphibian-loving people's backyards and associated natural areas.

Many of these specialized surveys perform very important educational, research or conservation purposes. However, in almost all cases they cannot be statistically meshed with a regional approach. 12. Road Conditions. Driving surveys should be conducted on public, all-weather, all-season roads only.

13. Survey Species and Survey Timing. The coordinators of a monitoring program must tailor the frequency of surveys to capture the seasonal differences in calling rates of anurans in their region. Additionally, not all anurans in a region would be expected to occur frequently enough on this type of survey to warrant consideration.

For example, we know that rare species, species that call very late in the evening, species that call underwater, or species that don't call at all would not be expected to be adequately monitored in an extensive survey conducted within two hours after sunset. Automated surveys, in some cases, may be able to fill in this gap.

14. Time/ Date/ Location Information. The monitoring points must be marked and described by the observer and maps kept on permanent file at a central office. Ideally, the points should be recorded using a global positioning system, however, this is not always feasible. Therefore, we recommend that volunteers record the latitude and longitude coordinates including topographic map number for each point, provide a verbal description of the location using semi-permanent landmarks such as bridges, curves in the road, geographic features, barns, houses, etc. (so that someone can get to within 30-50 meters of the site). These points must be a) recognizable by volunteers; b) as permanent as possible; c) known to the observer/ program coordinator; and d) identifiable on a topographic map.

On each survey, the observer must record: a) observer name, address, phone number with area code, county and closest major population center; b) time of day with start and finish time, and exact time the observations were made at each point; and date of survey with the month spelled out and the year indicated.

15. Habitat Data. It was agreed by those at the Toronto (1995) meeting that some habitat data should be recorded at each point. The following list is the minimum required habitat data necessary: a. A general description of the wetland /habitat type, e.g., vernal pool, pond in wood lot, marsh, bog, fen, creek, slough, farm dugout.

b. A general evaluation of water level in the wetland at the time of the survey (if possible).

c. Any major changes to the habitat type since the previous survey.

d. Any major changes to the habitats adjacent to the monitoring point habitat since the previous survey.

16. Weather Conditions. The following data should be recorded during each survey:

a. wind speed (Beaufort scale or exact speed if available)

b. air and water temperature to be taken at the start and end of the survey (we recognize that on some types of surveys water temperature is not easily or consistently sampled and several procedures are under consideration).

c. percent cloud cover (per stop).

d. precipitation conditions (e.g., drizzle, fog, rain, downpour) (per stop).

e. barometric pressure (at start).

f. any noise which interfered with hearing calls (per stop).

17. Quality Control

An adequate training tape is necessary to teach observers to differentiate species' calls, to know those species expected to call at the same time of year, to use correct methods for counting the numbers of anurans calling, and to evaluate calling intensity. This tape should also include bird calls and songs that are likely to be heard at the time of the survey (see 18).

A strategy to ensure that observers collect data under the correct weather conditions be implemented. Correct weather conditions include periods following a rain or periods of high humidity. This could involve observers calling a regional coordinator for advice or implementing a toll-free phone number that observers could dial to ask for advice from a more distant coordinator. We must stress the importance of collecting the data under suitable weather conditions to ensure high quality and comparable data among sites.

Data screening and computer editing procedures should be developed that can detect observer problems and incorrectly entered information.

18. Night Birds. Owls, caprimulgids (e.g., nighthawks, whippoorwills) and several species of marsh birds call frequently during the same periods in which amphibians are being monitored.

Similar to some amphibian species, many of these birds are suspected to be declining. This group of birds is also largely undersampled by the traditional bird monitoring programs such as the North American Breeding Bird Survey. Counting these species is a natural addition to calling-amphibian species. This group's calls are easily identified and can readily be included on any calling-amphibian training tape (see 17). Counting these species adds diversity to the experience of the observer without losing any of the amphibian information. Calling-amphibian surveys should not be modified to increase the number of detections of birds, but all observers should be counting night birds as well as amphibians.

19. Data Standardization. To create a truly compatible network of monitoring programs, means must be found to standardize survey techniques and to share the resulting data. The points made above are intended to standardize methodology. Additional efforts need to be directed towards standardization of the entry, editing, reporting and storage of the resulting data. Such efforts are a prerequisite to the occurrence of large regional analyses. We would like to begin discussions of database standards and data code books for calling amphibian data.

Research Needs

1. Involve volunteers that normally would not conduct surveys at night (e.g., school groups). Validation is needed of the results of surveys during the day with and without using broadcast tapes. An approach using school groups may also enable more data to be collected on urban natural areas such as ravines, creeks etc.

2. Confirm the distance required between observation points to avoid overlap which may lead to detection of the same calling species. At present, some preliminary testing along shoreline marsh wetland areas at Long Point, in Ontario, has indicated that a minimum distance of 500m is necessary to ensure that calling heard at one point is not detected at the next point surveyed. It is probably necessary to test that distance with other species groups.

3. Continue to evaluate whether the data collected is valuable only as presence/ absence data or if the data is indicative of population fluctuations within a site. This would be possible by continuing to assess the relationships between call counts, call intensity, and chorus and population size as evaluated by mark-recapture or other intensive population size estimates. This needs to be assessed among several populations of the same species over several years.

4. Continue to evaluate automated call-count systems vs. call-count accuracy and comparability of data collected under various weather conditions by human observers.

5. Continue to determine the cost of collecting data and the quality of the data under combinations of different survey lengths and timings to determine if this method of collecting data produces high-quality information and is cost-effective.

6. Compare trends in presence/absence and population fluctuations among different call-count approaches, e.g., roadside call counts; single-pond surveys; larval surveys; and automated systems.

7. Develop analytic procedures that will make the best possible use of the calling and environmental data gathered, ideally generating regional statistical models of the probability that not hearing a species indicates that the species is absent at a site, route or area. Analytic procedures could also study the degree of statistical independence of individuals stations along a route.

8. Use portable anemometers to recelebrate the Beaufort Scale to the kind of wind disturbance observers are likely to notice (movements of dead spring herbage, particular species of trees), rather than the British marine/urban phenomena of the standard scale.

9. Look into designing ancillary programs to investigate colonization rates of new ponds and other created environments. Locations in North America where extensive monitoring using call-counts is currently under way (or has been conducted in the past* but was discontinued). The monitoring program may not apply to the entire state or province, depending on observer participation or limitations in coordinator's time and funding for the program.


Dorcas, M. E. 1995. Monitoring anuran populations on the Savannah River Site using automated recording systems. Abstract of paper presented at the Second NAAMP conference, Toronto.

Helferty, N. 1995. Amphibian Protocol for Marsh Monitoring Program. Abstract of paper presented at the Second NAAMP conference, Toronto.

Kline, J. 1995. Assessment of precision and accuracy of calling amphibian surveys conducted by volunteers in Southeast Wisconsin USA. Abstract of paper presented at the Second NAAMP conference, Toronto.

Mossman, M. 1994. Wisconsin Frog and Toad Survey Instructions. Endangered Species Branch Wisconsin DNR, Madison. 17p.

Mossman, M., P. Rasmussen, J. Sauer, and S. Droege. 1995. Sample size estimation for amphibian calling surveys an some surprising trends from an 11-year analysis of Wisconsin Frog and toad Survey data. Abstract of paper presented at the Second NAAMP conference, Toronto.

Shirose, L. J., C. A. Bishop, D. M. Green, C. J. MacDonald, R. J. Brooks, N. J. Helferty. 1995. Validation study of a calling amphibian survey in Ontario. Abstract of paper presented at the Second NAAMP conference, Toronto.

Back to Amphibian Home Page