Patuxent Wildlife Research Center

NAAMP III Archive - calling surveys
Home | Archive by Alphabetical Order | Archive by Category

Evaluation of Randomly Chosen Frog Call Routes in Illinois

Christopher A. Phillips, Jeanne M. Serb & John E. Petzing

Center For Biodiversity
Illinois Natural History Survey
Champaign, Illinois 61820

Recently, protocols have been proposed for long-term monitoring of amphibian populations using anuran calling surveys (Protocols and Strategies for Monitoring North American Amphibians, North American Amphibian Monitoring Program, various authors). The authors stressed that randomly chosen routes were vital to making the program statistically defensible. This statement has prompted vigorous discussion about the potential logistical difficulties associated with using randomly chosen call routes (probability sampling) over subjectively chosen routes (judgment sampling). One of the main concerns is that some of the randomly chosen routes will fail to include a sufficient quantity of potential amphibian breeding habitat (= 10 stops in the NAAMP protocols) in some areas of North America. The number or percentage of such failed routes in any geographic region is the level of nonresponse. It is important to identify the level of nonresponse before routes are assigned to volunteers for several reasons. First, routes with little or no potential amphibian breeding habitat will not keep the interest of the volunteer and participation will drop rapidly. Second, routes without enough stops do not add any information toward the goal of the survey. In a probability sample, nonresponse does not provide information relevant to the goals of the survey.

We field checked 75 random routes in Illinois according to the guidelines in the NAAMP protocols. [Click here to get more information on how the routes were chosen]. The protocol requires that each participant drive their routes in daylight, before the start of breeding season, and locate 10 stops along each route. These stops must be wetlands or potential amphibian breeding habitat and must be a minimum of 500 meters (0.3 miles) apart and a maximum of 200 meters (0.12 miles) from the road. Once ten stops are established, the route is terminated. This means that some routes will be longer than others, depending on the concentration of breeding habitat. A few other rules for choosing stops apply: stops should be on quiet roads (no traffic noise to interfere with hearing the breeding calls); stops should be safe (enough shoulder for parking, etc.); and stops should be on public, all-season roads.

In our study, 14 different field personnel evaluated from one to 20 call routes. The majority of the personnel were biologists from the Illinois Department of Natural Resources (IDNR) Division of Natural Heritage. Each field personnel was given copies of the maps of the routes in their area and asked to follow the instructions above. The recording of stops was standardized as much as possible by asking that the type of habitat at a stop be chosen from a list of wetland types. Stops were numbered consecutively and marked directly on the maps. That same number and the wetland type (from the list provided) that best described the stop were written on the back of the map. Descriptive comments for each stop were also encouraged. This procedure was followed until 10 stops were recorded or 30 km (20 miles) had been traveled.

Drainage Ditch in Iroquois County, Illinois
Seventy-five routes were evaluated. Two routes did not have any potential amphibian habitat because they went through big cities. Two routes had to be terminated before ten stops were reached because of road conditions. Four routes did not have ten qualifying stops within our pre-determined reasonable driving distance (30 km) because of current land use, mainly agricultural practices. This leaves 67 routes (89% of the routes evaluated) that had ten stops within 30 km (Fig 2). Driving distances (Fig 3) varied from 8 to 30 km.

The most frequent habitat type encountered was "stream" (49% of stops were streams, mean = 5 streams/route, range = 0-10 streams/route). The second most frequent habitat type was "drainage ditch" (24% of stops were streams, mean = 2 ditches/route, range = 0-9 ditches/route).

The level of nonresponse, 7%, was lower than expected based on our experience with amphibian breeding habitat throughout most of Illinois. One question that any level of nonresponse raises is how to deal with nonresponse routes. In our study there were three categories of nonresponse routes; big cities, road conditions, and land use. The road condition problem can be solved by slightly altering the route to go around the closed road. This solution was given in the modified protocols that were distributed to individuals conducting or contemplating anuran calling surveys. For the land use problem, one possibility is to extend the length of the routes until ten stops are accumulated. How long is too long is the obvious question in this case. Another solution for these routes and the big city routes is to replace them with new randomly chosen routes but with the restriction that they must come from the general vicinity of the failed route. A final solution for all failed routes is to not replace them and hope their elimination does not introduce bias. This would be a potential problem if all failed routes were clumped in a small geographical area within a state or province.

The Problem of Drainage Ditches. It could be argued that across large parts of central Illinois, drainage ditches provide marginal amphibian breeding habitat at best. Our initial field observations suggest that, in the highly agricultural Grand Prairie Natural Division of central Illinois, drainage ditches provide breeding habitat only for bullfrogs (Rana catesbeiana). This hypothesis will be rigorously field tested next spring. It should be noted that, in other parts of Illinois and Missouri, drainage ditches can provide breeding habitat for up to nine species of amphibians (Tom Johnson, pers comm).

If it can be demonstrated that drainage ditches provide poor amphibian breeding habitat (only one or two common species sporadically breed there), should they be eliminated from the anuran calling surveys? Under strict interpretation of the NAAMP protocols, they should be included because they have the potential to provide breeding habitat. If the natural condition of the landscape around a ditch improves then breeding opportunities and amphibian populations will increase. If ditches have been excluded from the surveys, then these increases will go undocumented. The results of the survey will be biased. For large portions of central Illinois however, it is unlikely that natural conditions will improve in the landscape surrounding ditches. This presents a dilemma in which we must decide how much we want to safeguard against possible bias versus how wisely we want to spend our limited resources.

One point of view is that including ditches on the slim possibility that conditions will improve is not worth the resulting loss of coverage in other areas. This is certainly our position in central Illinois where monitoring ditches is unlikely to contribute much to the overall goal of the NAAMP. The other issue is the loss of enthusiasm and morale that might accompany running call routes that include 70-80% ditches.

If ditches are excluded from the protocols (at least in some geographic regions), what effect would this have on levels of nonresponse? In our study, 14 additional routes failed to make ten stops in 30 km. This increased the nonresponse level to 31%. Eleven of the 14 new nonresponse routes were in the Grand Prairie Natural Division (central Illinois).

Response level was better than expected.

How do we replace failed routes?

Should ditches be included as stops? Only in certain geographical regions?

Route selection protocol
The actual driving routes are chosen by dividing the State into degree blocks (1 degree of latitude and one degree of longitude). Five random points are chosen in each degree block. From each of those points that fall within Illinois' boundaries (degree blocks at the edge of Illinois also include parts of neighboring states), the nearest rural road is chosen and the "route" begins there. It proceeds in a randomly chosen direction staying as close to that direction as available rural roads allow. This procedure results in a set of routes at random locations throughout Illinois. The NBS developed a computer program to generate these routes and print out detailed maps of each.


Table 1. Field personnel involved in anuran call route evaluation
Maggie Cole Amy Ragusa
Chris Dinesen Brad Semel
Terry Esker Jeanne Serb
Carla Murray Michelle Simone
John Petzing Todd Strolle
Chris Phillips Bob Szafoni
Mark Phipps Diana Tecic


Table 2. List of amphibian breeding habitat types used in anuran call route evaluation
Floodplain Forest Pond
Swamp Lake
Wet Shrubland Stream
Marsh Drainage Ditch
Sedge Meadow Other
Wet Meadow


U.S. Department of the Interior
U.S. Geological Survey
Patuxent Wildlife Research Center
Laurel, MD, USA 20708-4038
Contact: Sam Droege, email:
Last Modified: June 2002