Patuxent Wildlife Research Center

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

Preliminary Analysis
of the Marsh Monitoring Program's
1995 Calling Amphibian Surveys

Amy A. Chabot

Long Point Bird Observatory.
P.O. Box 160.
Port Rowan, Ontario. N0E 1M0.

[ Abstract ]


Wetlands across North America, and the plants and animals inhabiting them, face increasing threats from development and pollution. In developed areas of the Great Lakes Basin, two-thirds of the wetlands have been destroyed. Scientists around the world are concerned about the loss of marsh habitat and declining populations of amphibians and marsh birds. However, the extent of marsh bird and amphibian declines is very poorly documented.

The International Joint Commission has identified 43 Areas of Concern (AOCs) around the Great Lakes as being particularly stressed by pollutants and in urgent need of rehabilitation. In each of these areas, the Great Lakes Water Quality Agreement's guidelines have been exceeded and remedial measures are necessary to restore the 14 beneficial uses listed in the Agreement. Remedial Action Plans (RAPs) and Public Advisory Committees (PACs) have been established to rehabilitate the ecosystems in each Area of Concern.

Actions to rehabilitate wildlife habitats in Great Lakes AOCs are in their early stages and have emphasized primarily fisheries concerns, although efforts are being made to address broader wildlife and habitat concerns. Several wetland restoration and rehabilitation projects have been initiated in both Canada and the United States and more are planned. However, wetland rehabilitation efforts are often hindered by a lack of knowledge about present and historical marsh bird and amphibian population levels.

The Marsh Monitoring Program (MMP) was established to aid the conservation and rehabilitation of marshes in the United States and Canada by studying population changes and habitat requirements of marsh birds and amphibians. The MMP is a cooperative project of the Long Point Bird Observatory (LPBO) and the Canadian Wildlife Service. The program is supported by Environment Canada's Great Lakes 2000 Cleanup Fund and the U.S. Great Lakes Protection Fund.

Naturalists with an interest in birds or amphibians are being asked to participate in the MMP by undertaking evening surveys for calling amphibians or marsh birds during the spring and early summer. Volunteers have the option of adopting a marsh in one of the 43 AOCs, or they can set up their own route. Special emphasis is placed on the coastal Great Lakes AOC marshes. However, marshes throughout Ontario and the eight Great Lakes states can be monitored.

Initially, the MMP will identify a baseline survey and analysis of marsh bird and amphibian populations at sites where rehabilitation and restoration efforts have taken place or are planned, at wetlands within AOC watersheds and at other wetlands within the Great Lakes Basin. A network of volunteers will be established and an implementation strategy to ensure long-term viability of the program will be developed. The ultimate goal of the MMP is to monitor marsh bird and amphibian populations on a long-term basis by volunteer surveyors. The information gathered by the program will be used, in part, to assess the progress and success of marsh rehabilitation efforts, especially those in the Areas of Concern.

This paper presents the summary results of the exploratory analyses of the 1995 data. Much more thorough and advanced analyses will be completed later, based on field work conducted in 1995 and 1996.


Marsh Monitoring Program Survey Protocols

General Survey Protocol

In order to adequately instruct and train participants in the MMP, a training kit was produced and distributed to all participants prior to the 1995 surveys. The kit was comprised of the following: instructions for surveying marsh birds and amphibians and their habitats, a training tape, marsh bird broadcast tape, aluminum tags, data forms, and a window decal. The training tape was provided to help volunteers distinguish the various categories of amphibian choruses and aid in amphibian species identification.

MMP survey routes consist of up to 8 semi-circular sample stations. Amphibian surveys are conducted from a central point located on the edge of the sample station. Routes are generally accessed by walking along the edge of marshes. However, volunteers also have the option of conducting routes along roads, or using a boat. In very large marshes, several routes can be established by one or more volunteers. In smaller, or less accessible marshes, it may be feasible to establish only 1 or 2 stations. Marshes must be at least 2 ha in size to accommodate a sample station. Information on smaller marshes is very much needed to help determine the effects of marsh size on species diversity and abundance. Survey points at each station are permanently marked (usually with a metal pole and aluminum identification tag) for future relocation.

It should be noted that some components of the 1995 protocol were still being assessed on a trial basis. Based upon participants' feedback and preliminary data analyses, simplified versions of the protocols were designed for 1996.

Amphibian Monitoring Protocol

Amphibian surveys use an "unlimited distance" semi-circular sampling area. This is because it is nearly impossible to accurately estimate distance in the dark and to judge whether amphibians are calling from inside or outside a defined sample area. Amphibian stations are separated by at least 500 metres (550 yards). On routes established through the middle of a marsh, the semi-circular sample stations can be arranged back to back, so that they face in opposite directions. Routes established in this way maximize the number of stations which can fit in a marsh.

Each route is surveyed three times in the spring and early summer. Following a one minute settling period, each station is surveyed for 3 minutes, starting one-half hour after sunset and ending before midnight. Participants survey routes in their entirety, in the same station sequence, starting at about the same time on all visits.

Accurate counts of calling amphibians are often impossible, so frog and toad populations are assessed within four Call Level Codes: 0 = species not heard, 1 = no overlapping calls, 2 = some overlapping calls and 3 = full chorus. For Call Level Codes 1 and 2, an Abundance Count of the number of calling amphibians heard is also made.

All surveys are to be completed in conditions conducive to monitoring amphibians (warm, wet weather with little or no wind). The first visit should coincide with minimum night-time air temperatures of at least 8-12C and the first or second warm spring shower. Night-time air temperatures should be between 13-20C for the second survey, and 21-28C for the third survey.

In 1995, surveys in southern regions (below the 43rd parallel) were be conducted in the following three survey windows: 1 April to 15 April, 1 May to 15 May and 1 June to 15 June. Surveys in central regions (between 43rd and 47th parallels) were to be conducted between 15 April to 30 April, 15 May to 30 May and 15 June to 30 June. Surveys in northern regions (above the 47th parallel) were to be conducted between 1 May to 15 May, 1 June to 15 June and 1 July to 15 July.

Habitat Description

Due to their dynamic nature, marsh habitats are being monitored annually. This information will be used as a basis for statistical comparisons, and will be used to determine species habitat associations. In addition, this information can be used to track and relate changes in habitat and species occurrence over time.

Participants make a simple sketch map of the station and answer a series of multiple-choice questions. Submergent vegetation is recorded as dense, moderate, slight, none or unknown. Marsh size is categorized as huge (greater than 50 hectares), large (between 25 and 50 hectares), medium (between 5 and 25 hectares), small (between 2.5 and 5 hectares) or tiny (between 2 and 2.5 hectares). Marsh permanency is recorded as permanent, semi-permanent or seasonal. Participants assess the percent cover of emergent vegetation, open water, exposed mud, and exposed rock in the sample area, and categorize the number of live trees and snags.

In 1995, participants were asked to distinguish between six important groups of plants: narrow-leaved emergents, broad-leaved emergents, tall robust emergents, floating plants, shrubs and trees. The percent vegetation cover of each group was assessed using a Braun-Blanquet code (1 = < 5%, 2 = 5-25%, 3 = 25-50%, 4 = 50-75%, 5 = > 75%). The percent cover of the two dominant species was also assessed using Braun-Blanquet codes and the presence of other species was checked off on the data sheet.

Route Selection

Volunteers monitoring marshes in AOCs are usually assigned a route, while volunteers monitoring non-AOC marshes can simply choose an accessible route close to home that they are familiar with (i.e. the route is subjectively chosen). Volunteers are given the option of completing marsh bird surveys, amphibian surveys, or both on their route.

Statistical Analysis

Comparative analyses of marsh bird and amphibian data from AOC versus non-AOC marshes need to take habitat differences into account. Therefore, data from similar habitats only were used for comparison presented in this paper. For example, data from sites dominated by cattail marshes sampled in AOCs and non-AOCs were treated separately from grass/sedge marshes. To reduce the habitat effects in analysing both bird and amphibian data, comparisons for AOC and non-AOC marshes were conducted only for stations with Braun-Blanquet codes of 3 or greater for cattail and grass/sedge. Sample sizes of marshes dominated by other vegetation were too small for analysis purposes. Statistics used for each analysis are described briefly in the relevant sections of the results. All statistics were calculated using SAS, either with standard procedures or with custom-written programs prepared by C.M. Francis, LPBO's senior scientist, as required.


Volunteer Participation

Over 500 volunteers registered to participate in the MMP in 1995. Thirty-three percent of these volunteers submitted either bird or amphibian data, or both, and planned to participate in 1996. The volunteers who did not submit data were queried to determine their intentions to participate in the program in 1996. Forty percent of the registered volunteers established routes, but did not monitor them in 1995. These volunteers intended to conduct surveys in 1996. A small number of volunteers indicated that they were no longer interested in participating in the program.

Sample Sizes

Amphibian data were received for a total of 476 stations on 127 routes, with an average of 4 stations per route. Eighty-seven routes were established in Ontario and 40 were established in the United States. Twenty-two routes were established in AOCs in the United States and 32 routes were established in AOCs in Ontario. Surveys were conducted on 18 routes in non-AOC marshes in the U.S. and 55 routes were established in non-AOC areas in Ontario.

Survey Conditions

Forty-one percent of the first amphibian surveys in the southern region, 46% of the surveys in the central region and 50% of the surveys in the northern region were conducted during the recommended survey dates. Sixty-six percent of the second surveys in the southern region, 67% of the surveys in the central region and 75% of the surveys in the northern region were conducted during the recommended survey dates. Sixty-six percent of the third surveys in the southern region, 64% of the surveys in the central region and 100% of the surveys n the northern region were conducted during the recommended survey dates.

Volunteers began their first amphibian survey between 18:30 and 23:55, their second survey between 19:30 and 23:56 and their third survey between 18:10 and 23:55. Eighty-three percent of volunteers began their first survey after 18:00, which is approximately dusk in early spring. Ninety-one percent began their second survey after 18:00 and 94% began their third survey after 21:00 which more closely approximates dusk later in the summer.

The minimum requirement for air temperature was met on 75% of the first amphibian surveys, 79% of the second surveys and 58% of the third surveys. Ninety-five percent of the amphibian surveys were conducted when the Beaufort wind scale strength was less than level 3 as recommended. Eighty-seven percent of amphibian surveys were conducted in the recommended damp or hazy weather.

The above results demonstrated that most volunteers adhered to the survey guidelines when conducting their surveys. However, many volunteers felt that they had trouble meeting all of the survey conditions outlined in the protocol. These concerns have been addressed in the revised guidelines developed for 1996.

Survey Results


Not surprisingly, the marshes sampled in AOCs were more frequently close to one of the major Great Lakes water bodies than the non-AOC marshes (Wilcoxon 2-Sample Test, z = -5.19867, p = .001) (Table 1). Indeed, the majority of coastal Great Lakes marshes are found within the boundaries of the AOCs. While less than 5% of the AOC routes were located in marshes over 50 km distance from the Great Lakes, 37% of the non-AOC routes were located this far inland.

We will be more closely examining the effects of marsh location relative to the Great Lakes. While there is no scientific evidence to support the contention that Great Lakes coastal marshes differ from "inland" marshes, there is a general feeling that differences may occur.

Table 1. Distance of Routes Sampled in AOC and non-AOC Marshes to Nearest Great Lake.
Distance from Route to Great Lakes (km) AOC

% of Routes


% of Routes


% of Routes

0-1 33 19 25
2-5 22 14 17
6-10 14 5 9
11-20 14 2 8
21-50 12 23 18
50-100 5 26 17
>100 0 11 6

Both the AOC and non-AOC marshes sampled were most frequently permanent marshes (Table 2). The proportions of AOC and non-AOC stations in each category were the same.

Table 2. Wetland Permanency of AOC and non-AOC Marshes.
Wetland Permanency AOC

% of Stations


% of Stations


% of Stations

Permanent 86 86 86
Semi-permanent 4 4 4
Seasonal 10 10 10

Both the AOC and non-AOC marshes sampled tended to be greater than 50 ha in size (i.e. huge), although there was good representation in all size classes (Table 3).

Table 3. Size of Marshes Sampled in AOCs and non-AOCs.
Size of Marsh AOC

% of Stations


% of Stations


% of Stations

Tiny (2 - 2.5 ha) 23 25 24
Small (2.5 - 5 ha) 12 9 10
Medium (5 - 25 ha) 16 18 18
Large (25 - 50 ha) 19 10 14
Huge (>50 ha) 30 38 34

Marshes sampled in AOCs tended to have less submergent vegetation than those in non-AOCs (2 = 34.31, df = 4, p = .02) (Table 4). A lesser proportion of non-AOC marshes had no submergent vegetation and a slightly greater proportion of non-AOC marshes had dense submergent vegetation. The percentage of stations with slight, moderate or unknown submergent vegetation was approximately equal in the AOC and non-AOC marshes sampled.

Table 4. Submergent Vegetation at Stations Sampled in AOC and non-AOC Marshes.
Submergent Vegetation AOC

% of Stations


% of Stations


% of Stations

None 20 12 15
Slight 24 26 25
Moderate 22 24 24
Dense 11 16 14
Unknown 23 22 22

The percent emergent vegetation cover averaged 72%; there was no significant difference between AOC and non-AOC marshes (Table 5). Proportions of vegetation cover and open water were understandably negatively correlated, with most stations having an average open water cover of 25%. The average proportion of the station covered by exposed mud or rock was small.

Table 5. Average Percent Cover of Habitat Types on Routes Sampled in AOC and non-AOC Marshes.
Major Habitat Types AOC

% of Cover


% of Cover


% of Cover

Emergent Vegetation 72 73 72
Open Water 25 24 25
Exposed Mud 2 2 2
Exposed Rock 1 1 1

Most sites had few trees or snags (Tables 6 and 7). However, the greater proportion of the stations sampled in AOC marshes had more trees and fewer snags then did those in non-AOC marshes (2 = 9.90, df = 4, p = .042 and 2 = 12.25, df = 4, p = .016 respectively).

Table 6. Number of Live Trees at Stations Sampled in AOC and non-AOC Marshes.
Number of Live Trees AOC

% of Stations


% of Stations


% of Stations

0 35 38 37
1-5 14 21 18
6-10 14 11 12
11-50 24 22 23
>50 13 8 10

Table 7. Number of Snags at Stations Sampled in AOC and non-AOC Marshes.
Number of Snags AOC

% of Stations


% of Stations


% of Stations

0 55 46 50
1-5 27 27 27
6-10 6 10 8
11-50 8 14 12
>50 4 3 3

Tall robust emergents dominated the emergent vegetation cover in the majority of marshes sampled (Braun-Blanquet codes of 4 or 5) (Table 8). Both the AOC and non-AOC marshes sampled were frequently associated with a Braun-Blanquet code of 0 or 1 for floating plants (2 = 13.30, df = 5, p = .021). Sample stations in AOCs usually had little or no narrow-leaved emergents (Braun-Blanquet code 0 or 1), while non-AOC stations had slightly more narrow-leaved emergent cover (Braun-Blaunquet code 1 or 2) (2 = 73.84, df = 5, p < .001). Both AOC and non-AOC marshes usually had little or no broad-leaved emergent, shrub or tree cover (Braun-Blaunquet codes of 0, 1 or 2).

Cattail was monodominant at 33% of the AOC and 23% of the non-AOC stations sampled, and was the dominant vegetation at 27% of the total stations sampled (Table 9). Grass/sedge and grass in combination with other narrow-leaved emergents dominated the emergent vegetation cover at 13% of all marshes sampled, and was dominant at more non-AOC stations than AOC stations. Thirty percent of both AOC and non-AOC stations were dominated by other forms of emergent vegetation (including shrubs and trees) or no emergent vegetation. It is possible that some of these sites may not actually be suitable marsh sites and as such, may be unsuitable for analyses.

Table 8. Braun-Blanquet Codes of Vegetation on Routes Sampled in AOC and non-AOC Marshes.

Vegetation Class
% of AOC Routes % of Non-AOC Routes
Braun-Blanquet Codes Braun-Blanquet Codes
0 1 2 3 4 5 0 1 2 3 4 5
Broad Leaved Emergent 40 34 16 7 2 1 40 38 16 5 1 1
Narrow Leaved Emergent 30 38 14 7 6 5 14 23 26 16 12 10
Tall Robust Emergent 6 12 10 12 22 39 9 13 17 14 18 29
Floating Plant 41 36 13 5 4 1 33 36 18 9 2 2
Shrubs 36 39 18 5 2 1 27 40 25 6 2 1
Trees 34 43 14 4 2 2 37 47 12 3 1 1

Table 9. Dominant Vegetation at Stations Sampled in AOC and non-AOC Marshes.
Dominant Vegetation AOC

% of Stations


% of Stations


% of Stations

Cattail Monodominant 33 23 27
Cattail/Grass 5 12 9
Cattail/Shrub 8 9 8
Cattail/Other Species 16 10 13
Grass-Sedge Monodominant 4 7 6
Grass-Sedge/Other Species 4 9 7
Other 30 30 30

On the whole, the above results indicate that the habitat parameters sampled were more or less equally represented in the AOC and non-AOC marshes. Much more work needs to be done in this regard, but these preliminary results tend to support the contention that comparisons between AOC and non-AOC are valid, at least from a broad habitat perspective.


Twelve of the 13 species of frogs and toads found in the Great Lakes Basin were detected on the routes sampled in 1995 (Tables 10 and 11). Blanchard's Cricket Frog was not detected on any route. Eleven species were detected in cattail marshes and 10 species were detected in grass/sedge marshes. Fowler's Toad was not detected on any route in either type of marsh and Cope's Gray Treefrog was not detected on any route in grass/sedge marshes. American Toad, Spring Peeper and Green Frog were recorded at more than 50% of the routes monitored in both cattail and grass/sedge marshes. Bullfrog, Northern Leopard Frog, Gray Treefrog and Chorus Frog were recorded less frequently. The other species were only detected occasionally. With the exception of American Toad, all species tended to be heard less frequently in the AOC marshes sampled than in similar non-AOC marshes. However, only two species, Gray Treefrog and Wood Frog, were detected significantly less frequently on routes sampled in cattail marshes in AOCs than on those in similar non-AOC marshes (Fisher's Exact Test, p < .05). There were no significant differences in species presence on routes sampled in grass/sedge marshes. In the future, our studies will likely focus on a suite of "indicator" species, to be identified by a group of herpetological experts.

Table 10. Frog and Toad Species Presence on Routes Sampled in AOC and non-AOC Cattail Marshes.
Species % of AOC Routes % of Non-AOC Routes
Bullfrog 22 37
Green Frog 56 77
Wood Frog* 9 35
Mink Frog 0 5
Northern Leopard Frog 50 58
Pickerel Frog 3 2
Gray Treefrog* 25 51
Cope's Gray Treefrog 0 2
Spring Peeper 59 74
Chorus Frog 25 35
American Toad 72 63

* Significant difference in frequency of occurrence in marshes sampled in AOCs and non-AOCs (p < .05).

Table 11. Frog and Toad Species Presence on Routes Sampled in AOC and non-AOC Grass/Sedge Marshes.
Species % of AOC Routes % of Non-AOC Routes
Bullfrog 0 34
Green Frog 57 69
Wood Frog 0 27
Mink Frog 0 8
Northern Leopard Frog 29 35
Pickerel Frog 0 4
Gray Treefrog 43 62
Spring Peeper 57 81
Chorus Frog 14 15
American Toad 43 31

In a preliminary study, the habitat selection of the amphibians sampled was examined using a step-wise multiple logistic regression. The variables tested included location of the site in relation to the nearest Great Lake, permanency and size of the marsh, percent of emergent vegetation, and the amount of narrow-leaved emergent, broad-leaved emergent, tall robust emergents, floating plant, shrub, tree, purple loosestrife and cattail. Eight species (American Toad, Bullfrog, Chorus Frog, Green Frog, Gray Treefrog, Northern Leopard Frog, Spring Peeper and Wood Frog) showed highly significant correlations with habitat features. Bullfrog and Wood Frog were found less often in marshes sampled in AOCs. Wood Frog tended to be associated with small marshes that had shrubs, while Bullfrog and Spring Peeper were more often found in larger marshes. Spring Peeper also tended to be associated with marshes close to the Great Lakes that had high percent emergent vegetation cover, but less tall robust emergent cover. Bullfrog and Northern Leopard Frog were associated with marshes that had a lot of broad-leaved emergents, while Green Frog and Wood Frog were associated with those that did not. Gray Treefrog was usually found in marshes that had a high percentage cover of narrow-leaved emergents. Green Frog preferred marshes with floating plants.


Preliminary Statistical Analyses

Few differences were detected in the habitat parameters sampled in the AOC and non-AOC marshes. While most of the amphibian species monitored were found less frequently in the marshes sampled in AOCs than in the non-AOCs, significant differences were found for only two species. However, comparative analyses of AOC and non-AOC marshes must be considered preliminary since the non-AOC marshes sampled are not necessarily representative of all potential non-AOC marshes. The analyses presented were preliminary and were undertaken as a lead-in to more definitive analyses of the 1995 and 1996 data.

Selection of Survey Routes

A statistically sound procedure for selecting and allocating routes needs to be developed, bearing in mind constraints of volunteer availability and marsh access. The representativeness of existing non-AOC route allocation must then be examined.

From an analytical standpoint, all routes would ideally be randomly selected. This is difficult to achieve as routes must be selected within reasonable distance of volunteers who can survey them and that the marshes must be easily accessed (e.g. on foot or roadside). Nevertheless, it is very important that the routes be representative of marsh habitat within the study area. If routes run through atypical areas, then the trends observed may not be representative of those in the populations of marsh birds and amphibians within the study area. Because some areas have higher densities of marshes than others, it may be appropriate to stratify the sampling scheme, with more intensive sampling in high density areas. More information and further planning is required before implementing such changes.

Survey Protocols

Due to the difficulty volunteers experienced in conducting surveys in the proposed survey conditions, the amphibian survey guidelines were refined for 1996. Volunteers now are encouraged to focus primarily on ensuring that suitable temperature and weather conditions are met during the surveys. Dates are now given only as guidelines for surveys in the protocol.

The habitat description methodology has also been greatly simplified, based upon volunteer feedback and the preliminary analyses. While volunteers are still required to draw a simple sketch map of the station, the number of multiple choice questions has been greatly reduced. Braun-Blanquet codes are no longer used and the presence of additional, non-dominant species will no longer be noted. Instead, volunteers are now asked to select up to four kinds of herbaceous emergent vegetation that appears to dominate the sample area and estimate what percentage of the emergent vegetation is made up by each of these species.

The protocols will be reassessed upon completion of the 1996 surveys. They will be revised in 1997, if necessary, based upon volunteer feedback, staff and partnership input and the work being conducted by the North American Amphibian Monitoring Program. Final protocols, suitable for long-term monitoring, will be produced and distributed to all RAP teams and PACs in 1997.


We thank the following staff of Environment Canada for their help with the overall delivery of the Marsh Monitoring Program: Donna Stewart, Nancy Patterson, Lesley Dunn, Christine Bishop, Mike Cadman, Tammara Boughen and Brian McHattie. The direction of Michael Bradstreet, Jon McCracken and Charles Francis (LPBO) was also very much appreciated. Natalie Helferty provided valuable input and helped to coordinate the program in 1995. Kathy Jones has assisted with, and greatly furthered, the coordination of the program in 1996. We are also grateful to our other project partners, Federation of Ontario Naturalists, Great Lakes United, National Audubon Society, for promoting and disseminating the results of the program. We are especially indebted to the members of the program's Advisory Committee for their advice on the direction and delivery of the program.

We owe our deep gratitude to the funders who have made this project possible (Great Lakes Protection Fund, Great Lakes 2000 Cleanup Fund and Citizenship-Eco Action) and to the many volunteers who took part in the surveys. We are also grateful to the RAP coordinators and PAC chairpersons for their support and assistance in establishing surveys in the AOCs. Thanks are also extended to private landowners who allowed access to their marshes.


Bishop, C.A. 1994. Methods for evaluating the occurrence and relative abundance of amphibians and reptiles in natural and constructed habitats in Areas of Concern in the Great Lakes Basin. Canadian Wildlife Service, Canada Centre for Inland Waters, Burlington, Ontario. 19 pp. (+appendices).

Freedman, B. and N.L. Shackell. 1992. Amphibians in the context of a National Environmental Monitoring Program. In: Bishop, C.A. and K.E. Pettit (ed.). Declines in Canadian amphibian populations: designing a national monitoring strategy. Occasional Paper 76. Canadian Wildlife Service, Environment Canada, Ottawa, Ontario. 120 pp.

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