Proceedings of the Marsh Bird Monitoring Workshop

              Proceedings of the Marsh Bird
Monitoring Workshop
U.S. Fish & Wildlife Service
U.S. Geological Survey
Proceedings of the Marsh Bird
Monitoring Workshop
U. S. Fish and Wildlife Service
Patuxent Research Refuge
National Wildlife Visitor Center
Laurel, Maryland
26-28 April 1998
April 1999
Compiled by:
Christine A. Ribic, U. S. Geological Survey (USGS)–Biological Resources Division (BRD), Wisconsin
Cooperative Wildlife Research Unit, University of Wisconsin, 1630 Linden Drive, 226 Russell Labs,
University of Wisconsin, Madison, WI 53706-1598
Stephen J. Lewis, U. S. Fish and Wildlife Service (USFWS), Bishop Henry Whipple Federal Building, 1
Federal Drive, Fort Snelling, MN 55111-4056
Scott Melvin, Massachusetts Division of Fisheries and Wildlife, Westboro, MA 01581
Jon Bart, USGS–BRD, Snake River Field Station, 970 Lusk Street, Boise, ID 83706
Bruce Peterjohn, USGS–BRD, Patuxent Wildlife Research Center, Breeding Bird Survey, 12100 Beech
Forest Road, Laurel, MD 20708-4038
Workshop Steering Committee:
Christine Ribic (Chair), USGS–BRD, Wisconsin Coop. Wildlife Research Unit, Madison, WI
David Dolton, USFWS, Office of Migratory Bird Management, Denver, CO
Sam Droege, USGS–BRD, Patuxent Wildlife Research Center, Laurel, MD
Marshall Howe, USGS–BRD, Patuxent Wildlife Research Center, Laurel, MD
Stephen J. Lewis, USFWS, Fort Snelling, MN
Scott Melvin, Massachusetts Division of Fisheries and Wildlife, Westboro, MA
For additional copies of this report contact:
David Dolton
U.S. Fish and Wildlife Service/MBMO
PO Box 25486 DFC
Denver, CO 80225-0486
U.S. Fish & Wildlife Service
U.S. Geological Survey
Proceedings of the Marsh Bird Monitoring Workshop iii
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Need for a Workshop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Scope of the Monitoring Program. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Workshop Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Objective 1: Review Current Monitoring and Research Efforts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Objective 2: Develop Standardized Protocols for Monitoring Marsh Birds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Report from Group 1: Field protocols for marsh bird monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Preliminary recommendations for survey protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Objective 3: Develop Sampling Schemes at Multiple Scales. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Report from Group 2: Statistical design and sampling considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Report from Group 3: Implementing a Marsh Bird Monitoring Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Information Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Topics Needing Further Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Objective 4: Enhance Communication Among Biologists and Managers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Letters of Support (USGS-BRD and USFWS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Literature Cited. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Appendix 1. List of Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Appendix 2. Topics for the Discussion Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Appendix 3. Common and Latin Names of Bird Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Appendix 4. Abstracts of Invited Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Table of Contents
The Workshop Steering Committee would like to
thank all of the participants of the workshop for
their enthusiasm and willingness to help in
developing a Marsh Bird Monitoring Program. We
also thank Deb Ziebarth, Department of Wildlife
Ecology, University of Wisconsin-Madison, and
Linda Lee, U. S. Fish and Wildlife Service, Office of
Migratory Bird Management, for administrative
support. The U. S. Fish and Wildlife Service’s Office
of Migratory Bird Management and U. S. Geological
Survey’s Biological Resources Division provided
financial support for the workshop.
iv
Acknowledgments
A Marsh Bird Monitoring Workshop was held at the
National Wildlife Visitor Center located on the
Patuxent Research Refuge in Laurel, Maryland, in
April 1998, to discuss strategies for monitoring
populations of wetland-dependent bird species.
Inconspicuous marsh birds (e.g., rails, bitterns,
moorhens, gallinules, snipe and coots) include both
game and nongame species. These birds are sought
by hunters and birders alike. Some (black and
yellow rails and American and least bitterns) are U.
S. Fish and Wildlife Service species of management
concern because they are thought to be rare or
declining, and many are included on state lists of
rare or endangered species. However, inconspicuous
marsh birds are difficult to detect and inhabit areas
that are often not readily accessible. Therefore, they
are poorly surveyed by the Breeding Bird Survey
and other existing monitoring programs.
The workshop brought together experts actively
working on marsh bird monitoring issues. The
workshop reviewed current monitoring and research
efforts related to marsh bird population assessment,
initiated development of standardized protocols for
monitoring marsh birds, identified information
needs necessary for developing protocols and
sampling schemes for monitoring marsh birds at
national, regional, and local scales, and established a
Steering Committee to continue the development of
a Marsh Bird Monitoring Program.
Priority information needs identified by the
participants are to understand statistical issues and
potential biases of call-playback techniques and how
to choose habitats for sampling. Secondary
information needs are to evaluate: current wetland
databases, how to accommodate habitat change over
time, what habitat variables to collect, whether
counts should be made in the morning or evening,
how the index is linked to population size, and
geographical differences in the timing of the
breeding season. A tertiary information need relates
to data quality issues, such as the use of volunteers.
Several topics were identified as needing further
discussion. These include the geographic scale of the
program, what trend changes will be acceptable to
fulfill differing objectives, and how monitoring of
local management impacts might be evaluated by
the program.
A new Steering Committee composed of
representatives from U. S. Federal and state
agencies and Canadian organizations was
constituted to further develop a Marsh Bird
Monitoring Program. Activities of this committee
will be communicated at the following web site:
http://www.mp1-pwrc.usgs.gov/marshbird/
1
Executive Summary
Inconspicuous marsh birds (e.g., rails, bitterns,
moorhens, and gallinules) are important components
of the biodiversity of wetland ecosystems and
include both game and nongame species. These
birds are sought by hunters and birders alike, and
some (black and yellow rails and American and least
bitterns) are U. S. Fish and Wildlife Service
(Service) species of management concern because
they are thought to be rare or declining. These
species also appear on many state agency lists of
threatened and endangered species; they and their
habitats are afforded varying levels of state
regulatory protection. Many units of the National
Wildlife Refuge System provide important habitat
for these species, and the birds have undoubtedly
benefitted from wetland restoration activities
undertaken through the Service’s Partners for
Wildlife and the North American Waterfowl
Management Plan programs. However, little is
known about the abundance, population trends, or
management needs of these species.
Inconspicuous marsh birds are difficult to detect and
inhabit areas that are often not readily accessible.
Therefore, they are poorly surveyed by the North
American Breeding Bird Survey and other existing
monitoring programs. A number of efforts have
been made to standardize marsh bird surveys using
taped playback response. Notable among these is
the work by Gibbs and Melvin (1993), Bird Studies
Canada’s (formerly Long Point Bird Observatory)
Marsh Monitoring Program, and the USGS
Biological Resource Division’s (BRD) listserver
discussion group on marsh bird monitoring. The
Service’s Office of Migratory Bird Management
(MBMO), through its Webless Migratory Game Bird
Research Program, is currently funding a number of
studies across the country that are aimed at
developing marsh bird monitoring techniques and
determining the relationship between indices
derived from call-response surveys and other
measures of abundance. Finally, a number of
national wildlife refuges have monitored marsh
birds using a variety of approaches with a wide
range of scientific rigor.
There are several factors that limit the utility of
data obtained from current marsh bird monitoring
efforts. First, protocols are not standardized,
making it difficult to compare results from different
surveys. Second, a statistically based sampling
framework for conducting these surveys has not
been developed. A variety of sampling schemes is
needed because monitoring is desired at several
levels, specifically, large-scale monitoring to
determine a species’ range-wide or regional status
and to set hunting regulations on a flyway basis,
local monitoring on refuges and similar areas to
determine presence, habitat associations, and
responses to management, and nonrandom
monitoring of special sites like the Great Lakes
Areas of Concern that are the focus of Bird Study
Canada’s Marsh Monitoring Program. Finally, it is
not known what the relationship is between indices
obtained in call-response surveys and actual
population levels of the target species.
To address the aforementioned needs, a workshop
was held at the National Wildlife Visitor Center,
Laurel, Maryland, in April 1998, to bring together
biologists who are actively working on marsh bird
monitoring issues. Invited speakers presented
overviews of their work and then the group, through
a facilitated discussion, attempted to reach
consensus on standardized protocols and sampling
schemes for monitoring marsh birds, and to
prioritize research needs. Workshop participants are
listed in Appendix 1.
Objectives
The workshop objectives were to:
1) Review current monitoring and research efforts
related to marsh bird population assessment,
focusing primarily on rails, bitterns, pied-billed
grebes, common moorhens, and American coots.
2) Develop standardized protocols for monitoring
marsh birds.
3) Develop sampling schemes for monitoring marsh
birds at national, regional, and local scales.
4) Enhance communication among biologists and
managers interested in marsh bird monitoring
issues.
3
Need for a Workshop
Objective 1 was met by presentations of current
research (summarized below). To answer Objectives
2–4, the workshop participants divided themselves
up into three working groups: Field Protocols for
Marsh Bird Monitoring, Statistical Design and
Sampling Considerations, and Implementation of a
Marsh Bird Monitoring Program. The charges for
each group are in Appendix 2. Each group was free
to modify the questions posed and a group report
was made to the assembled scientists. These reports
are summarized below.
Scope of the Monitoring Program
The participants defined the primary species of
interest as those species that are marsh- dependent
and not well covered by existing monitoring
programs. The primary species are grebes, rails,
bitterns, moorhens, gallinules, snipe, and coots.
Secondary species are herons, black and Forster’s
terns, sedge and marsh wrens, red-winged and
yellow-headed blackbirds, Franklin’s gull, sharp-tailed
and LeConte’s sparrows, common
yellowthroat, willow and alder flycatchers, cranes,
and belted kingfisher. See Appendix 3 for Latin
names.
The participants further agreed to limit the program
to the breeding season. It was noted that migration
routes and wintering ground monitoring would be
important but, at this time, beyond the scope of the
workshop. The participants further agreed that
trend information at the continental or rangewide,
national, regional/provincial, and state levels should
be considered in the discussion groups.
4 Proceedings of the Marsh Bird Monitoring Workshop
Objective 1: Review current monitoring and research
efforts related to marsh bird population assessment,
focusing primarily on rails and bitterns
Researchers and agency biologists presented
current work related to marsh bird monitoring. The
presentations are summarized here by common
themes. Abstracts of the presentations can be found
in Appendix 4.
The first speaker, Sam Droege, set the development
of a Marsh Bird Monitoring Program in perspective
by reviewing general approaches to developing
monitoring programs. Pertinent points made
included the need to define what species are going to
be covered by a new program, what magnitude of
change is to be detected, and how often a program is
going to be wrong in finding a change when there
isn’t one. Mr. Droege argued that a monitoring
program, by its nature, should be willing to say
there is a change before the decline is obvious, if the
program is to serve as an early warning system.
Otherwise the monitoring program merely
documents that species have declined. He also
pointed out that any program is vulnerable if there
is no evidence that the index being used in the
program actually reflects the population levels of the
target species. Finally, a set of constituents that
need to know or care about marsh birds were
discussed, including hunting regulators, refuge
managers (due to the Refuge Organic Act),
conservationists, citizens, hunters, and anti-hunters.
The majority of speakers dealt with Field
Methodology issues (Downs and Anderson; Gibbs,
Melvin, and Crowley; Haramis and Kearns; Kirsch;
Legare, Eddleman, Buckley, and Kelly; Monahan
and Faulk; Paine; Skoloda and Ribic; Slack and
Mizell; Therres, Brinker, and Tango; Vogel,
Helmers, Fredrickson, and Humburg; Walther and
Hohman) (Table 1). Most of the speakers dealt with
rails; six of the eight presented information on sora
and Virginia rail. Most of the presentations included
information gleaned from point-count or transect
surveys coupled with call-playback surveys, leading
to information on multiple species. A few studies
used other techniques to try to understand the
relationship of call-playback surveys to species
numbers. Some information was presented on rails
during the non-breeding seasons.
Another set of speakers (Adamcik; Francis and
Weeber; Gibbs and Melvin; Sauer; Shieldcastle;
Weeber, McCracken, and Francis) dealt with the
development of Marsh Bird Monitoring Programs.
Current monitoring programs and their drawbacks
were discussed. Site selection was a critical issue.
Issues of the number of samples needed to detect
certain trends with specified Type I error and power
were presented. Typically, the number of routes may
look high enough to detect trends but the number of
routes on which specific species are actually
detected can be low. These problems are not atypical
when dealing with multiple species, leading to the
recommendation that surveys need to be planned
using the most variable species. Data needs for
developing a monitoring program were presented.
Problems with implementation were discussed. One
issue was the use of refuges in such a program and
that addressing refuge needs in the program would
be critical in eliciting the participation of refuges.
The last speaker, Paul Adamus, presented
information on the movement of the U.S.
Environmental Protection Agency to develop an
Index of Biotic Integrity for wetlands using marsh
birds. This program is currently in development but
gives another indication that marsh birds are of
interest to other Federal agencies not typically
considered to be concerned with this group of birds.
Objective 2: Develop standardized protocols for
monitoring marsh birds
Report from Group 1: Field protocols for marsh bird
monitoring
The group discussed issues related to standardizing
field protocols for use in a Marsh Bird Monitoring
Program.
1. Priority species
There was general agreement that priority species
in a Marsh Bird Monitoring Program should include
all rails, American and least bittern, common snipe,
and pied-billed grebe. American coot, common
5
Workshop Results
moorhen, and purple gallinule could also be included
in areas where they occur at low to moderate
densities. However, in some regions these three
species may be adequately monitored through
existing programs that track populations of breeding
waterfowl. Participants recognized that marsh bird
surveys will likely yield detection rates that are
significantly higher for many other species than
those obtained on Breeding Bird Survey routes.
2. Chronology of surveys
It was agreed that first priority should be to develop
monitoring programs for marsh birds on the
breeding grounds. However, participants recognized
that there may be interest in developing monitoring
programs on wintering grounds or in migration
areas for some species. Subsequent discussions
focused primarily on field protocols for monitoring
on the breeding grounds.
3. Use of playbacks
The issue of whether broadcasts of taped
vocalizations (playback surveys) should be used to
increase detection rates of some species of marsh
birds received considerable discussion and was one
of the major unresolved issues of the workshop.
Participants identified the following benefits of using
playbacks:
a. They significantly increase detection rates for
some species over passive listening. Table 2
summarizes participants’ species-specific
recommendations regarding improvements in
detection rates with playbacks.
b. They probably reduce variance in mean detections
per survey station or per mini-route (multiple
stations), by reducing numbers of “0” counts. This
may be especially important in regions where
target species occur at low to moderate densities,
for example, in northeastern North America, and
where marsh birds may be presented with fewer
intraspecific stimuli to vocalize than in other
regions (i.e., few calling individuals per wetland).
c. They probably increase proficiency of field staff.
By repeatedly playing tapes and eliciting more
responses from target species, field staff become
more familiar with marsh bird vocalizations.
d. They probably increase long-term interest of field
staff, especially volunteers. Playbacks
significantly increase detections of several target
species. It may be difficult to maintain the interest
of volunteers if they seldom detect target species
during surveys that consist only of passive
observations.
e. Cooperators will likely want to use playbacks for
intensive local surveys, to monitor patterns of
distribution or habitat use, determine presence or
absence of protected rare species, or monitor
effectiveness of habitat management. It may be
inefficient to conduct intensive local surveys with
playbacks and extensive regional surveys without
playbacks.
6 Proceedings of the Marsh Bird Monitoring Workshop
Table 1. Number of studies, by species, season studied, and techniques used, for presentations made in the
Field Methodology section of the Marsh Bird Monitoring Workshop.
(Abstracts are in Appendix 4. Primary and secondary species were defined by the participants. Latin names of species
mentioned are in Appendix 3.)
Primary Species Secondary Species
Rails Bitterns Moorhens Grebes Snipe Coots/
Gallinules
Season and Technique
Breeding
Playback or Passive
Surveys 8 4 3 4 1 2 4
Telemetry 2
Trapping 1
Nest Monitoring 1 1 1 1 1 1 1
Visual Observation 1 1
Other 1 1
Migration
Playback 2 1 1 1
Telemetry 1
Wintering
Telemetry 1
However, participants also recognized potential
disadvantages of using playbacks as part of
extensive regional or continental monitoring efforts:
a. Playbacks introduce an additional source of bias.
It would be necessary to standardize, as nearly as
possible, playback equipment, tapes, and how they
are used (for example, volume, distance from
observer).
b. It may not be feasible to expect that all
participants will be willing to purchase adequate
playback equipment. On the other hand,
volunteers that conduct Breeding Bird Surveys
are expected to use binoculars, which are often
2–3X more expensive than adequate playback
equipment.
Table 2. Summary of recommendations as to whether
playback surveys can increase detection rates of
marsh birds on the breeding grounds.
(Latin names of species are in Appendix 3.)
Are detection
rates improved
Species with playback surveys?
Virginia rail Yes
Sora Yes
King rail Yesa
Clapper rail Yesa
Black rail Yes
Yellow rail Probably no
American bittern Uncertain
Least bittern Probably
Common moorhen Yes
Purple gallinule Uncertain
American coot Probably yes at low
breeding densities
Pied-billed grebe Probably yes at low
breeding densities
Other grebes No
a King and clapper rails may respond well enough to each other’s
vocalizations that only one or the other need be included in a
playback survey protocol.
Participants agreed that a research priority should
be to determine detection rates for target species
both with and without use of playbacks. Power
analysis should then be used to estimate differences
in species-specific sample size requirements for
monitoring programs using playback versus passive
observations only. Data sets already exist for a few
species, for example, sora and Virginia rail, but for
most species, additional research will be required.
The prime determinant of whether or not to use
playbacks will be whether sample sizes adequate to
monitor population trends can be obtained for target
species in all or most regions using only passive
surveys.
If playbacks are to be used, additional research is
needed to support development of standardized
methods, including which calls to broadcast,
sequence of calls, sound levels, and types of
equipment to use.
Participants generally agreed upon interim
guidelines for playback survey methodology. At each
survey station, the following sequence of playback
and passive observation/listening was
recommended:
a. Begin with an initial passive listening period of 3-5
minutes.
b.Broadcast approximately 1 minute of vocalizations
for each species. For each species, broadcast the
type of call that is most likely to elicit a response
(e.g., the “whinny” call for soras, the “grunt” call
for Virginia rails).
c. Insert 30 seconds of silence after each minute of
vocalizations for each species.
d. Broadcast vocalizations of the least intrusive
species first, for example least bittern or black rail
before American bittern and pied-billed grebe.
e. At each station, end the series of broadcasts with
a “post-playback” listening period of
approximately 1minute.
Playback vocalizations should be broadcast in all
directions over suitable wetland habitats (i.e., the
playback arc should be 360 degrees from the
observer). However, playback arc may be as little as
180 degrees at some survey stations along the edge
of wetlands or along dikes.
4. Spacing of survey points
Spacing of survey points will be a compromise
between spacing large enough to minimize double-counting
of loudest species and smaller spacing that
reduces travel time between stations and maximizes
the number of points that will fit in a given wetland.
Participants recommended a minimum spacing of
200-300 m between survey stations.
5. Seasonal timing of monitoring
Participants recognized that peak vocalizations
occur in late spring and early summer. Regional
differences in breeding phenology may exist and
may be of sufficient magnitude to require regional
differences in timing of surveys.
Proceedings of the Marsh Bird Monitoring Workshop 7
Participants recommended that marsh bird
monitoring on the breeding grounds should take
place during a 6-8 week period that is consistent
over as large a region as possible (e.g., 15 May to 30
June). Sampling each survey station 2-3 times per
season will likely reduce effects of annual variation
in phenology. However, the relative benefits of
reducing bias associated with annual variation in
phenology should be compared to the cost of
reduced sample size that will occur if points are
sampled multiple times in a year.
6. Time of day
Most participants believed that detection rates of
most species are highest in early morning, from
approximately 1/2 h before sunrise to 4 h after
sunrise. The logistics of accessing and returning
from relatively large or remote wetlands are easier
in early morning than in early evening or at night.
Detection rates of secondary species, for example
wetland passerines, may be highest in early
morning, and wind speeds may be less in early
morning than in late afternoon or early evening.
Both of these assumptions need to be tested.
Bird Studies Canada already has in place a
relatively large volunteer program based on
monitoring between 6 and 9 p.m. This protocol
results in only 2-4 stations per day, versus > 10 that
can be done between 1/2 hour before to 4 hours after
sunrise. However, if a regional monitoring program
is based on volunteers, more volunteers may be
available and willing to participate in early evening
than in early morning.
7. Should surveys be done on foot or by canoe?
It was agreed that, when possible, surveys by canoe
are preferable because they are quieter, often faster,
and place the observer at the interface between
water and emergent vegetation where many target
species are most likely to be detected. However,
there was also agreement that surveys by canoe are
often not practical or necessary, for example, at
remote palustrine wetlands where canoe access is
not possible, at relatively small or narrow wetlands,
and where easy foot access is possible to points at
the edge of representative portions of the wetland
being sampled.
8. Weather
Participants agreed that the Breeding Bird Survey
weather standards should be observed for marsh
birds surveys (i.e., no surveys during periods of
sustained precipitation or when winds exceed 20
kph).
9. Habitat data collection
Participants agreed that collection of data on habitat
conditions at survey stations or for entire wetlands
where surveys are conducted is desirable, to be able
to relate trends in habitat conditions to patterns and
trends in marsh bird abundance and distribution.
Participants agreed this topic needs further
discussion.
10. Training
Participants agreed that some form of training is
desirable for individuals that will be conducting
marsh bird surveys. This is especially true when
monitoring will be done by large groups of
volunteers. This subject needs further discussion.
The group did recommend that training should focus
on primary target species.
Preliminary Recommendations for Survey Protocol
The recommendations of Group 1 were discussed
with the larger group and the following preliminary
protocol was agreed upon. We note that this protocol
will change as more research is conducted and a
Marsh Bird Monitoring Program is developed.
a. Surveys can be conducted in the morning (1/2 h
before sunrise to 4 h after sunrise). Note that
detection rate for snipe may be lower on morning
surveys. Evening surveys are recommended if
snipe is of primary concern. Field protocol for
common snipe can be found in the abstract by
Downs and Anderson (Appendix 4).
b. Surveys should be done when there is no steady
precipitation and low wind (less than 20 knots).
Use the Breeding Bird Survey recommendations
for further weather restrictions.
c. Survey points should be a minimum of 200 m
apart. Distances up to 800m may be appropriate
in certain wetland types such as tidal marshes
(due to the probability that birds will be heard at
much longer distances in these marshes
compared to other wetland types). Use some
type of marking or mapping system so that the
point can be found again for repeated surveys
(e.g., a sketch map, marking of the point, use of
Global Positioning System).
d. At a point, a 5-minute passive count should be
made first, followed by tape playback, and then 1
minute of passive listening. For tape playback, 5-
15 seconds of a species’ call should be repeated 3
times with 5 seconds of silence between repeats
for a total of 30-60 seconds per species. Thirty
seconds of silence should be used before
8 Proceedings of the Marsh Bird Monitoring Workshop
beginning another species’ call. Use the call most
likely to elicit a response (Table 3).
We suggest using the Cornell Laboratory of
Ornithology’s Library of Natural Sounds as a
source for recordings of vocalizations. The order
of calls on the tapes should be from the least
intrusive to the most intrusive calls. An example
of this ordering would be the calls of black rail,
followed by least bittern, other rails, American
bittern, and, lastly, pied-billed grebe. How loud to
play the tape was not discussed though it was
noted that quality of the tape and equipment
would impact the quality of the sound if the
volume was turned up to the maximum. Also,
having the tape played too loudly could drown
out some responses or inhibit nearby birds from
vocalizing.
e. The number of surveys to be done at a survey
point depends on the objectives of the survey. For
a monitoring program, 1 survey may be
adequate. For local information, we recommend
at least 2 surveys during a 6-week period during
the breeding season; surveys should be made at
least 1 week apart. Three surveys per season
may be necessary to accurately determine
presence/absence of some species. Local
information sources should be used to define the
breeding season for a specific location.
f. The number of survey points to be done depends
on the objectives of the study, size of wetlands,
and spacing between points. No general number
can be given. For example, the number of points
per route for a regional monitoring program is
still in need of research.
g. Use an arc that encompasses all wetland habitat
adjacent to the survey station.
h. The habitat focus of the survey should be on
emergent vegetation.
i. Points should be as close to or as far into the
wetland as possible/feasible. Canoeing or walking
into the wetland to place points are preferred.
Doing surveys along a dike or along a shoreline
are permissible where the preferred options are
not possible/feasible.
j. The survey point should consist of a 100 m fixed-radius
half-or whole-circle and then an area of
unlimited radius. Birds should be noted as being
within the 100 m radius or outside of the fixed
radius.
k. Collect information on the major habitat within
the fixed radius and beyond (where possible).
Some suggested habitat information includes:
area and percent coverage of sedge/grass or
cattail/bulrush, open water, aquatic bed (from
Cowardin et al. 1979), shrubs, scrub, flooded
timber; overall wetland area.
l. All participants should be trained in species
identification (in the lab as well as in the field),
estimation of distance, habitat assessment,
establishment of points, and the mechanics of
running a survey.
Table 3. Suggested calls to use in a playback survey.
Species Call
Sora Whinny
Virginia rail Grunt
King/clapper rail Mating call
Black rail Kic-kic-kerr call
Common moorhen/Gallinule Primary advertising call
Bitterns Advertising call
American coot Advertising call
Objective 3: Develop sampling schemes for
monitoring marsh birds at national, regional, and
local scales
Report from Group 2: Statistical design and
sampling considerations
The group discussed issues of scale, sampling frame,
sampling design, and field protocols, and reached
the following conclusions:
1. Scale of the survey.
A strong consensus existed that the marsh bird
survey should be multi-scale, responding to both
local and regional needs. National needs could then
be addressed by combining results from different
regions.
2. Sampling framework.
There was a consensus that the sampling program
should be based on sampling wetlands (i.e., habitat-based)
and that the development of a sampling
frame is a necessary prerequisite to developing a
sampling design. There was uncertainty about what
information is available to develop a list of habitats
(i.e., sampling frame). Information from which a list
of habitats could be constructed may already exist
for many local and regional (e.g., statewide) areas.
At the national level, no habitat-based frame
currently exists. An alternative approach would be
an “area-based” frame from which smaller areas
could be selected—for example squares 10 km on a
Proceedings of the Marsh Bird Monitoring Workshop 9
side—and sampling would then be conducted within
such areas. More information is needed on the
current and imminent availability of habitat-based
frames, especially at the national level, and on the
relative utility of habitat-based and area-based
frames, especially at the regional and national level.
A first step using either approach should be to
identify all areas that should definitely be sampled
(e.g., Federal refuges, state wildlife management
areas, waterfowl production areas), and these should
be designated as comprising a separate stratum
which would be sampled with 100% selection
probabilities (i.e., all units in this stratum would be
selected). This step is appropriate from a statistical
standpoint and will insure that high-interest areas
are included in the monitoring program.
3. Sampling plan to choose sites.
The group agreed that a sampling plan to develop
strategies for choosing sites and stations (including
geographical distribution) needs investigation. Such
factors as specific objectives, costs, and results from
simulations should be used to develop specific
sampling plans. We did not feel that blanket
statistical recommendations could usefully be made
at this early stage in the development of monitoring
programs. Some of the related issues that need
consideration are the relative need for, and difficulty
of, sampling edge vs. interior areas, the role of
volunteers in selecting survey locations, and
whether surveys can include both whole and half
circles around the observer.
4. Field protocols.
The group discussed statistical issues related to field
methods, identifying factors that affect detectability,
and factors that might cause biases in average
detectability through time or across other domains
of interest (e.g., regions, habitats). The possibility
that such biases might occur will be one of the most
serious problems in interpreting results from the
survey unless analysts can be sure that any such
biases are substantially smaller than observed
trends (otherwise observed trends might be due
entirely to changes in detectability). The group
agreed that this issue warrants careful study to
determine how factors such as density, habitat,
timing, and playbacks affect detectability.
Studying effects of playbacks will be important
considering the popularity of using this technique
and the fact that the use of playbacks in a
monitoring program was recommended by the Field
Methodology Discussion Group (see above). There
are many issues. Only a few are discussed here. One
important question concerns playback response and
population density. Specifically, is there a nonlinear
relationship between the number of calls from
playback and density (i.e., does the number of calls
level off when density is high?). Another issue is the
possible movement of birds towards the observer.
This means that one cannot census a plot of known
size (i.e., densities or indices cannot be calculated;
only presence/absence information is obtained) and
that this movement may depend on the habitat and
density of the species (it cannot be assumed that
there is a standard movement towards the observer
across all habitats and densities). Such information
would lead to guidelines for how field methods
should be standardized.
In conclusion, the group targeted investigating
sampling frames, sample selection designs based on
them, and studying detectability as issues of
importance from a statistical standpoint.
Report from Group 3: Implementing a Marsh Bird
Monitoring Program
This discussion group initially attempted to identify
the important needs that might be addressed by the
implementation of an operational Marsh Bird
Monitoring Program. These needs are summarized
in order of importance as evaluated by the group:
1. Estimates of annual population trends.
Many aspects of the biology of marsh birds remain
relatively poorly known as compared with other
groups of birds. Even fairly basic information such
as distribution during the breeding and winter
seasons, timing and status of migrants, and specific
habitat preferences throughout the year are poorly
documented in the literature in many geographic
areas. While obtaining these data may not be the
primary purpose of a marsh bird monitoring
program, some monitoring data may be helpful in
“filling the gaps” of our knowledge of these species.
The ability to obtain basic population trend
estimates for marsh birds is very important because
existing surveys do not adequately sample most
species. Trend estimates would be needed at various
geographic scales, particularly state/provincial,
regional, and continental estimates. While local
trends of marsh birds may be of interest to land
managers, these survey methods probably lack
sufficient precision to reliably estimate local
population trends. Ideally, marsh bird surveys
conducted at the state/provincial scale would
contribute to trend estimates at larger geographic
scales. The adoption of consistent survey
methodologies is critical to accomplish this goal.
However, sample sizes and survey designs necessary
to obtain trend estimates at the state/provincial
10 Proceedings of the Marsh Bird Monitoring Workshop
scale may be different from those of regional and
continental surveys. Greater sampling effort would
be needed to obtain reasonable trend estimates at
smaller geographic scales, and the distribution of
survey points may be different to meet the specific
needs of the states/provinces conducting the
surveys. However, the participants agreed that
these specific sample designs could probably be
accommodated within the framework of sample
designs for the regional/continental surveys.
In addition to the basic need to obtain population
trend estimates, this monitoring program would be
important for the following purposes:
  To obtain adequate population trend data for
establishing hunting regulations for hunted marsh
bird species.
  To better understand the population trends of rare
or declining marsh bird species at the Federal and
state/provincial levels, and to better ascertain the
status of species of management concern before
they require formal listing as threatened or
endangered.
  To help state agencies develop information on
population trends, distribution, and effects of
management as a basis for regulatory protection for
species of marsh birds already listed as endangered,
threatened, or of special concern under state laws.
  To meet various Federal legislative requirements
pertaining to the need for monitoring bird
populations, including the Migratory Bird Treaty
Act, Fish and Wildlife Conservation Act, and
National Wildlife Refuge System Improvement Act.
  To focus research and management activities for
the benefit of marsh bird species.
2. Habitat assessment.
To better understand the overall importance of
population trend estimates, trends must be placed
within the context of changes occurring within the
wetland habitats occupied by these species. Hence,
there is a need to:
  Obtain information on the habitat associations of
the various marsh birds. Since bird- habitat
associations can vary geographically, this
information is needed throughout the ranges of
these species.
  Obtain information on the trends of these habitats.
As was true for the population trend estimates,
habitat trend estimates will also be needed at
various geographic scales.
3. Monitor the effectiveness of management actions.
At the local scale of individual wildlife refuges or
similar management areas, there is a need to
understand how marsh birds respond to
management activities. Projects studying the effects
of management activities on marsh birds require
knowledge of population levels within the area of
interest, knowledge of the specific bird-habitat
associations of the region, and possibly some
demographic data to understand the factors
responsible for the observed changes in population
levels. Additionally, the survey design and allocation
of survey points would reflect the specific questions
being addressed at each area, and may be very
different from the survey design used to collect
large-scale population trend data.
The group participants recognized the need to
include a variety of stakeholders in monitoring
efforts for marsh birds, expanding beyond the
agencies and participants present at the workshop.
Various governmental and non-governmental
organizations, private industry, researchers, and the
general birding public will need to be involved in
various aspects of this program for these monitoring
efforts to be successful. Additionally, marsh bird
monitoring efforts have to focus beyond the goal of
simply obtaining population trend estimates; they
must also relate to wetland habitat management,
development of regional conservation plans for
birds, research, and actions devoted to the recovery
of declining species.
The group participants recognized that the three
needs listed above cannot be accomplished with a
single survey design. Hence, a tiered approach
should be considered that would allow managers to
collect the desired data according to their needs.
The first tier of data collection consists of the survey
design and methodology needed to estimate the
population trends of marsh birds. A consistent
methodology should be followed at all locations, and
the survey design should allow for the estimation of
population trends at several geographic scales.
The other tiers of data collection would be optional
and depend upon the needs of the local habitat
manager or the state/province coordinating the
survey efforts. The second tier of data collection
would emphasize the habitat characteristics at the
survey sites. While the complexity of these habitat
data depends upon the specific management or
research questions under study, in general, these
data should be fairly easy to obtain by the survey
participants and should not interfere with the bird
population surveys. A recommended set of habitat
parameters and survey methods should be
developed to promote the consistent collection of
Proceedings of the Marsh Bird Monitoring Workshop 11
habitat data across sites. The third tier of data
collection would relate to specific management
issues that require additional study. While the
nature of these surveys would vary from site to site
depending upon the management questions under
study, these surveys should try to adopt the survey
methodology used to estimate population trends
while allocating samples to meet specific needs, such
as to compare treatment with control sites. In this
way, some data could contribute to the first-tier
surveys that estimate population trends, while all of
the data would be available to address the specific
management questions.
Uncertainties
The group participants identified a number of
questions that need to be addressed prior to the
actual implementation of an operational Marsh Bird
Monitoring Program. The answers to these
questions will significantly influence the direction
and success of these monitoring efforts.
1. Observer-related issues.
Recruiting and retaining qualified observers to
conduct marsh bird surveys is fundamental to the
success of the program. Paying biologists to conduct
every survey is probably not a realistic option for a
continental monitoring program, because the costs
would greatly exceed the available funds, and
coordinating the logistics of many field crews
scattered across the continent would be a very time-consuming
task for the agency responsible for the
monitoring program. Hence, a volunteer-based
effort comparable to the Breeding Bird Survey
(BBS) is a more realistic approach for a continental
Marsh Bird Monitoring Program. However, there
also are several problems associated with the use of
volunteers to conduct these surveys:
  The expertise in bird identification required to
conduct these surveys will dictate the number of
potential volunteers in an area. If the marsh bird
surveys are restricted to a relatively small number
of species, then the potential pool of qualified
volunteers will be larger since it may be possible to
provide the necessary training to people with more
limited bird identification skills so they can
participate in these surveys. Conversely, marsh bird
surveys that require the identification by voice of all
species present in the wetland would require a much
greater level of bird identification skills, comparable
to those required for the BBS, and would reduce the
pool of qualified volunteers to a relatively small
proportion of the birders living in an area.
  Volunteers will not be evenly distributed across
the landscape, but will tend to be concentrated near
urban areas. Hence, rural areas may have few
potential volunteers so that some volunteers may
have to drive considerable distances to conduct
marsh bird surveys. A survey method that requires
multiple visits to each site may discourage regular
coverage of distant locations.
  The availability of volunteers to conduct marsh
bird surveys may be a limiting factor since many
qualified birders also participate in other bird
monitoring programs. This problem may be
compounded by survey methods for marsh birds
that require relatively greater effort on the part of
the volunteer, either to conduct multiple surveys or
to gain access to survey sites, for example.
While any Marsh Bird Monitoring Program will
require volunteers to collect data, this program may
also have to rely on paid biologists to participate in
order to achieve consistent coverage across the
continent. Hence, cooperation with Federal, state, or
other agency biologists will be important to obtain
coverage at some managed sites that may not be
easily accessed by volunteers. In more remote areas
where volunteers and agency biologists are scarce,
some paid field crews may be necessary to conduct
these surveys.
2. Access/logistic issues.
The methods and survey design implemented for
this marsh bird monitoring program will play an
important role in the long-term success of these
efforts. The methods should be kept as simple as
possible to collect accurate data but not impose
hardships on the volunteers collecting the data.
These methods should also minimize the logistical
considerations of the agency managing the
monitoring program. For example, the use of tape
playbacks may increase detection rates for some
species but would require the development of
region-specific tapes, and the necessity to routinely
provide the participants with new copies of the tapes
and broadcast equipment so that the same methods
are uniformly applied at each site. Considerable
resources would be necessary to provide the tapes
and broadcast equipment to observers scattered
across a continent.
The survey design should try to select sites that
have relatively easy access, either near roads or on
public lands, rather than a strictly random process.
Factors such as requirements for multiple visits
each year, the need to use boats to conduct surveys,
obtaining permission to visit sites on private
property, and crossing difficult terrain in darkness
would discourage participation by volunteers and
reduce the number of surveys conducted each year.
Incorporating off-road sites into the sampling
12 Proceedings of the Marsh Bird Monitoring Workshop
scheme may complicate the ability of some Federal
and state biologists to participate on surveys located
on private property. Such factors should be
implemented in the monitoring program only if
there are no acceptable alternatives.
3. Funding issues.
Questions regarding how to obtain the necessary
funding for a Marsh Bird Monitoring Program were
largely beyond the realm of this discussion group.
Some source of consistent funding will be required
to implement and operate this monitoring program,
but whether these funds are derived from additions
to base budgets or the redirection of existing funds
would have to be decided by others. However, the
coordinating agency(ies) should develop
partnerships with a broad spectrum of
governmental and non-governmental groups to
assist in the development and implementation of
these monitoring efforts. These partnerships may
reduce some of the funding requirements for both
the initiation and long-term operation of marsh bird
monitoring.
4. Coordination role.
Questions regarding who will coordinate the
development and operation of the marsh bird
monitoring program were also beyond the realm of
this discussion group, although the group
participants felt strongly that a Federal agency
should take the lead in this process. Since the trend
data for some species would be used in the
establishment of annual hunting regulations, the
U.S. Fish and Wildlife Service will want to be closely
involved with most aspects of this program;
coordination by a non-governmental agency may not
be appropriate. Since USGS Biological Resources
Division has a mandate to monitor the continental
wildlife resources and has considerable expertise in
monitoring non-game bird populations at large
geographic scales, this agency should also be
involved in various aspects of any Marsh Bird
Monitoring Program.
Regardless of which agency takes the lead
coordination role for marsh bird monitoring,
operating this monitoring program will require a
sizable commitment of resources. Unless large
amounts of new funding are available, one possibility
would be to have one agency coordinate the
operational aspects of the monitoring efforts, while
another agency would manage the data and
coordinate additional research studies.
Programmatic responsibilities could be shared
through other means, but various options should be
explored to reduce the amount of funding required
by a single agency to initiate marsh bird monitoring
efforts.
5. Habitat information.
Many uncertainties surround the collection of
habitat data. For example, which parameters to
measure, how to measure them, and to what
distance from the sample point should habitat data
be collected are some of the topics that remain
controversial. These issues need to be resolved
before a habitat component can be incorporated into
the marsh bird monitoring program and sampling
methodologies can be developed that would be
consistent across sites. Whatever methods are
developed, they should be as simple as possible since
adding a detailed habitat component to the
monitoring protocol would increase the cost of
training/supervision to prepare volunteers to collect
habitat data and increase difficulty of recruiting and
retaining qualified volunteers by adding additional
work.
Habitat data will also be needed at large geographic
scales in addition to the site-specific information.
Defining wetland physiographic strata at regional
scales may be important in any analysis of
population trends. Understanding regional trends in
habitats may be valuable in identifying factors
influencing marsh bird populations. Wetland
habitats at large geographic scales are currently
classified by the National Wetlands Inventory
(NWI), but additional information and ground
truthing may be needed to more precisely establish
the distribution and abundance of wetlands at the
landscape level.
6. Outreach/public education.
The group participants recognized an important
need to broaden the base of support for marsh bird
monitoring efforts and the conservation of these
species. The desirability of creating partnerships
with a variety of governmental and non-governmental
agencies to promote the marsh bird
monitoring program is described above.
Additionally, marsh birds need to receive a higher
level of recognition in ongoing bird conservation
efforts, such as the conservation planning processes
underway within Partners in Flight, the North
American Waterfowl Management Plan (NAWMP),
and the U. S.Shorebird Conservation Plan. Given
that marsh birds occupy the same habitats as
waterfowl, there is a natural linkage to include all
marsh-inhabiting birds within the NAWMP.
Increasing the visibility of marsh birds within these
planning processes will be necessary to obtain and
maintain the necessary funding to implement an
operational Marsh Bird Monitoring Program.
Proceedings of the Marsh Bird Monitoring Workshop 13
Recommendations
After discussing these topics, the group participants
developed the following list of recommendations
concerning the implementation of a Marsh Bird
Monitoring Program:
1. Overall coordination of the marsh bird monitoring
program should be at the level of the Federal
government in the United States and at a national
level within Canada, with assistance provided by
non-governmental organizations, states, and
provinces.
a. A central database should be created for this
program, which should be managed by the
Federal government.
b. The BBS should be used as a model for
establishing a coordination network of volunteers
willing to conduct these surveys.
2. Coordination between the U.S. Fish and Wildlife
Service and USGS Biological Resources Division
will be required to develop the budget initiatives and
necessary institutional support to implement marsh
bird monitoring across North America. These
agencies cannot be successful by themselves, but
will have to develop partnerships with a variety of
governmental and non-governmental agencies to
support these efforts. For example, the
International Association of Fish and Wildlife
Agencies and Waterfowl Flyway Councils should be
enlisted to elevate the priority of marsh bird
monitoring efforts within the states and provinces.
These activities also should be coordinated with the
Partners in Flight, NAWMP and shorebird
conservation planning activities.
3. The survey methods should be as simple as
possible but still collect adequate data for estimating
population trends at various geographic scales. The
survey design should consider the accessibility of
sites during the site-selection process. Feasibility of
the program may mean the use of boats and use of
sites far removed from roads may have to be
minimized. Volunteers will be required to collect
much of the survey data, and their long-term
participation in marsh bird monitoring efforts will
be directly related to the amount of time and effort
required to conduct these surveys.
4. A combination of volunteers, agency biologists,
and personnel paid to specifically conduct marsh
bird surveys will probably be needed to implement a
continental monitoring program. The paid personnel
may be needed in the more remote areas where few
qualified volunteers are available, but their use
should be kept to a minimum. Training materials will
have to be developed for the volunteers, to ensure
that the methodology is consistently followed across
sites. These materials may also be helpful to recruit
new volunteers who may currently lack the
necessary expertise to conduct marsh bird surveys.
The necessity for use of tape playbacks should be
carefully considered, because many significant
logistical problems would be associated with
providing tapes and broadcast equipment to the
survey participants.
5. To associate marsh bird population data with
habitat changes, wetland habitat databases should
be developed for relatively large geographic scales.
Available sources such as the National Wetlands
Inventory and GAP projects will provide the initial
information for these databases, but the reliability of
these data for the regional monitoring of habitats
remains uncertain. Resources may be necessary for
ground-truthing of available habitat data or to
obtain data from other sources.
6. Outreach and communication strategies should be
developed at the same time as the marshbird
monitoring program is being developed. Marsh
birds are currently overlooked members of
communities in many areas, and their visibility
within bird conservation and monitoring efforts has
to improve in order to develop all of the financial and
volunteer resources needed to implement a
continental Marsh Bird Monitoring Program.
Information Needs
The workshop participants agreed that specific
research should be undertaken as soon as possible to
further the development of a Marsh Bird Monitoring
Program. The following topics were identified by the
participants as being important to the success of a
Marsh Bird Monitoring Program and ranked
according to priority of which should be done first
(Table 4).
First Priority
1. Statistical issues using call-playback techniques.
The Field Protocol Group recommended using call-playback
techniques because of the belief that using
call-playbacks would increase detectability of certain
species, leading to a reduced number of zero counts
and an assumed decrease in variance. However, the
Statistical Design group argued that call-playbacks
should not be used because an unknown area is
being sampled, birds most likely move into an area
in response to the call-playback, and bird movement
is dependent on a variety of currently unknown
habitat (e.g. size of marsh, vegetation type) and
biological factors (e.g. bird density). Some members
14 Proceedings of the Marsh Bird Monitoring Workshop
of the Statistical Design Group argued that call-playbacks
could lead to an increase in variance
rather than a decrease. Therefore, an information
need identified was the following: In comparing
passive and call-playback observations, (a) is there
an improvement of detectability with call-playback?
(b) is there a change in the variance and in what
direction? If there is a reduction, how much of a
reduction is seen? and (c) What are the implications
for required sample sizes in a power analysis?
2. Evaluation of a sampling frame/scheme.
The workshop participants agreed that local
protected areas such as refuges and state protected
areas could form their own stratum and be
completely or almost completely sampled without
impact to the validity of the Program. The major
concern was how to pick samples from the non-protected/
non-priority sites. Two suggested schemes
were proposed though there is no reason to think
there are only these two alternatives. The first
alternative was a habitat-based approach that is
dependent on existing maps for choosing wetlands
according to some statistically-based scheme. A
second alternative is an area-based approach where
areas are picked by some statistically-based scheme,
these areas are evaluated for the existence of
wetlands, and then those wetlands are sampled by
some statistically-based scheme. The workshop
participants agreed that a high priority research
need is to evaluate different sampling schemes for
use in a Marsh Bird Monitoring Program.
Second Priority
3. Development of habitat information.
The workshop participants agreed that the habitat
classification of Cowardin et al. (1979) should be used
for site selection. An alternative, the
hydrogeomorphic scheme, may be useful at
individual points. But it was acknowledged that a
status review of current databases was needed.
Some databases to consider are the National
Wetlands Inventory and GAP databases. Concerns
were raised about how current the databases were
and at what scale(s) they would be helpful. There
was no information on potential Canadian databases
nor what databases were currently being developed.
Another aspect to investigate is how well the habitat
of priority species would be covered by the existing
databases. There was some debate about whether or
not this would be covered under the Evaluation of a
Sampling Frame/Scheme. This need may or may not
be met, depending on the scope of the proposed
sampling frame project.
4. Documenting habitat change over time (including
what type of habitat data should be collected).
An acknowledged difficulty was the fact that
wetlands change over time. This means that habitat
information needs to be collected at the survey
points as well as documenting the existence of the
wetland (which may disappear due to development
or succession). This information will be important in
understanding trends seen in the data as well as
impacting the design of the program (e.g., fixed
points versus new points every year versus some
kind of mixture). So the information need is to
determine how wetland/habitat change will be
accommodated in the program. Some of these issues
may be addressed in the Evaluation of a Sampling
Frame/Scheme, depending on the proposed project.
5. Timing of counts (a.m. versus p.m.).
An issue that was brought up was the timing of the
counts from a biological point of view and from an
operational point of view. Some of the priority
species are not detected during morning surveys
while timing may not make a difference for other
priority species. Secondary species, though, may be
detected at much lower rates during evening
surveys. An operational problem is that if the
Proceedings of the Marsh Bird Monitoring Workshop 15
Table 4. Information needs and importance as ranked by participants in the Marsh Bird Monitoring Workshop.
(Not all information needs were ranked by all participants, therefore, row sums will not be equal.)
Information need Very Important Secondary Importance Can be done later
Statistical issues using call-playback techniques 31 8 0
Evaluation of a sampling frame/scheme 31 6 1
Development of habitat information 16 15 6
Documenting habitat change over time 11 18 8
Timing of counts (morning vs. evening) 14 16 7
Assessing what the index means 13 15 9
Phenology or geographical differences that 11 15 9
influence the timing of surveys
Data quality issues 5 13 19
program is to be volunteer-based and volunteers are
only available during the evenings, then should an
evening survey period be used? Some workshop
participants indicated they thought that the data
probably exist to answer the biological questions
and that all that needs to be done is to evaluate this
information.
6. Assessing what the index means.
The issue of what the index means in terms of
population density or size is a critical one and is an
issue for all monitoring programs. Most of the
programs that are considered weak are those whose
index cannot be linked to some measure of
population size. However, trying to answer this
question will be difficult to study.
7. Phenology or geographical differences that
influence the timing of surveys.
When to do the surveys will vary by region and the
workshop participants discussed the need to
understand these differences to provide guidance to
the monitoring program. There was some
agreement that the databases exist to answer this
question and that the information just needs to be
evaluated. One participant suggested a national
team could be used to do this.
Third Priority
8. Data quality issues.
The participants agreed that evaluating the quality
of the data collected by any program is important.
In particular, we need to understand the costs,
benefits, and trade-offs between using volunteers
(which can include professional biologists) and using
a dedicated set of technicians to collect the
monitoring data. Other issues, such as hearing
ability and song identification, were also discussed.
Topics Needing Further Discussion
Scale of the program.
Though the participants agreed that the program
should focus on trends at multiple geographic scales,
there was little discussion in the breakout groups
about the implications of multi-scale monitoring on
the design and objectives of the program. The
Statistical Design Group explicitly suggested a
multi-scale program at local and regional scales,
noting that national trends would come from the
regional programs. However, other than pulling the
local protected areas out as an individual stratum,
little was done to address design at multiple-scales.
Another suggestion was to consider all scales
suitable to establish hunting regulations, assess
threatened and endangered species and species of
management concern, and to address other
legislative requirements. Clearly, more discussion is
needed on the objectives and scale of the program.
2. What trend and other details.
One aspect of the program that was not discussed in
detail was the time-scale over which true changes
needed to be detected. One group suggested that
detecting differences in trend on an annual basis
was necessary. But there was little discussion in the
other groups. Most concurred that, at this stage, one
scenario was unrealistic and that research could
address different scenarios of trend/Type I error/
power that would be helpful in setting realistic goals.
3. Evaluation of habitat relationships/management
impacts.
Though identified as one of the goals of a program,
no discussion of how these questions might be
addressed in a monitoring program was made.
Clearly, specific local goals and how they might
interact with a monitoring program need further
discussion and development.
16 Proceedings of the Marsh Bird Monitoring Workshop
Objective 4: Enhance communication among
biologists and managers interested in marsh bird
monitoring issues
This workshop was the first step in forming a
network of professionals interested in marsh birds
and monitoring. It was recognized that this was only
the first step.
It was agreed that the research priorities identified
in the workshop should be used to keep the
momentum going in the development of the Marsh
Bird Monitoring Program. The participants also
agreed that a pilot program building on Bird Studies
Canada’s Marsh Monitoring Program, and having a
heavy research emphasis, is needed. However, long-term
strategies are needed for obtaining
institutional support for such a program. The U.S.
Fish and Wildlife Service and the USGS Biological
Resources Division were identified as the
appropriate agencies for developing a strategy for a
Marsh Bird Monitoring Program. A new Steering
Committee composed of representatives from U.S.
Federal and state agencies, and Canadian
organizations was established to go forward with the
development of a Marsh Bird Monitoring Program.
The Committee consists of the following people:
Marshall Howe, USGS-BRD (Chair)
Robert Blohm, USFWS, Office of Migratory Bird
Management
Peter Blancher, Canadian Wildlife Service
Russell Weeber, Bird Studies Canada
Scott Melvin, Massachusetts Division of Fisheries
and Wildlife (representing the International
Association of Fish and Wildlife Agencies)
Bruce Peterjohn, USGS-BRD
Jon Bart, USGS-BRD.
Activities of this committee will be communicated at
the following web site:
http://www.mp1-pwrc.usgs.gov/marshbird/
Proceedings of the Marsh Bird Monitoring Workshop 17
Statement of Support from the Deputy Chief Biologist,
USGS Biological Resources Division
August 12, 1998
The Biological Resources Division (BRD) of USGS
recognizes secretive marsh birds as one of several
guilds of North American birds whose populations
are not effectively monitored by existing methods.
The April 1998 marsh bird workshop, co-sponsored
by the US Fish and Wildlife Service (FWS) and
USGS, has successfully summarized the existing
state of knowledge about marsh bird populations
and has defined the information and research needs
that will enable a marsh bird monitoring program to
be implemented. USGS is fortunate to have a
considerable body of expertise in sampling design
and in development of population monitoring
protocols. USGS also places a high priority on
achievement of these goals for marsh birds and is
committed to contributing its expertise accordingly.
We propose to incorporate eventual databases
derived from a monitoring program into the national
bird population data center being developed at
Patuxent Wildlife Research Center. We look forward
to working cooperatively with the FWS and other
entities working for marsh birds in reaching these
goals.
Susan D. Haseltine
Deputy Chief Biologist for Science
Biological Resources Division
Statement of Support from the Assistant Director,
Refuges and Wildlife, U.S. Fish and Wildlife Service
The U.S. Fish and Wildlife Service acknowledges
secretive marsh birds as an integral component of
North America’s avifauna. Yet, little is known about
their abundance, distribution, and population status.
The Service considers the acquisition of scientific
information to enhance management decisions
extremely important to its mission and fully
supports research, monitoring and management
efforts to remedy these limitations. Consequently,
high priority will be placed on development of
standardized survey protocols for monitoring marsh
birds, through future budget initiatives and other
opportunities for cooperative support. The Service
looks forward to working with the Biological
Resources Division of the U.S. Geological Survey,
States, and other agencies and organizations to
share research and management expertise and
improve our collective ability to manage this
historically-overlooked group of birds.
Daniel W. Ashe
Assistant Director—Refuges and Wildlife
18
Letters of Support
Cowardin, L.M., V. Carter, F.C. Golet, and E.T.
LaRoe. 1979. Classification of wetlands and
deepwater habitats of the United States. U.S. Fish
and Wildlife Service FWS/OBS-79/31.
Gibbs, J. P. and S. M. Melvin. 1993. Call-response
surveys for monitoring breeding waterbirds.
Journal of Wildlife Management 57: 27-34.
19
Literature Cited
Bob Adamcik
U.S. Fish and Wildlife Service (USFWS)
Division of Refuges
Room 670
4401 N Fairfax Drive
Arlington, VA 22203
Phone: 703-3587-2359
Fax: 703-358-1826
bob_adamcik@mail.fws.gov
Paul Adamus
Department of Fisheries and Wildlife
Oregon State University
Corvallis, OR 97330
Phone: 541-745-7092
Fax: 541-753-4507
adamusp@ucs.orst.edu
Stanley Anderson
Wyoming Cooperative Fish
and Wildlife Research Unit
University of Wyoming
Biological Science Bldg
Room 419 Box 3166
Laramie, WY 82071-3166
Phone: 307-766-5415
Fax: 307-766-5400
anderson@uwyo.edu
Jon Bart
U.S. Geological Survey (USGS), Biological
Resources Division (BRD)
Snake River Field Station
970 Lusk Street
Boise, ID 83706
Phone: 208-331-5216
Fax: 208-331-5210
jbart@eagle.idbsu.edu
Robert Blohm
USFWS
Office of Migratory Bird Management
634 Arlington Square
4401 N Fairfax Drive
Arlington, VA 22203
Phone: 703-358-1966
Fax: 703-358-2217
robert_blohm@mail.fws.gov
David Brinker
Maryland Department of Natural Resources
1200 Frederick Rd.
Catonsville, MD 21228
Phone: 410-744-8939
Fax: 410-260-8596
dbrinker@dnr.state.md.us
John Bruggink
USFWS
Office of Migratory Bird Management
11500 American Holly Drive
Laurel, MD 20708-4016
Phone: 301-497-5627
Fax: 301-497-5871
john_bruggink@mail.fws.gov
Mike Cadman
Canadian Wildlife Service
75 Farquhar St.
Guelph, Ontario
CANADA N1H 3N4
Phone: 519-826-2094
Fax: 519-826-2113
Mike.Cadman@ec.gc.ca
Jennifer Casey
Lake Umbagog National Wildlife Refuge
Route 16, PO Box 240
Errol, NH 03579
Phone: 603-482-3415
Fax: 603-482-3308
jennifer_casey@mail.fws.gov
Courtney Conway
Division of Biological Sciences and
Montana Cooperative Wildlife Research Unit
University of Montana
Missoula, MT 59812
Phone: 406-243-5137
Fax: 406-243-6064
CONWAY@selway.UMT.edu
20
Appendix 1. Participant List for the
Marsh Bird Monitoring Workshop
National Wildlife Visitor Center, Laurel, Maryland, 26-28 April 1998.
David Dolton
USFWS
Office of Migratory Bird Management
PO Box 25486 DFC
Denver, CO 80225-0486
Phone: 303-275-2388
Fax: 303-275-2384
david_dolton@mail.fws.gov
Kevin Downs
Wyoming Cooperative Fish
and Wildlife Research Unit
University of Wyoming
Biological Science Bldg.
Room 419 Box 3166
Laramie, WY 82071
Phone: 307-766-4515
Fax: 307-766-5400
downsk@uwyo.edu
Sam Droege
USGS, BRD
Patuxent Wildlife Research Center
12100 Beech Forest Drive
Laurel, MD 20708
Phone: 301-497-5840
Fax: 301-497-5784
FROG@usgs.gov
Erica Dunn
Canadian Wildlife Service
National Wildlife Research Centre
100 Gamelin Blvd.
Hull, Quebec
CANADA K1A 0H3
Phone: 819-994-0182
Fax: 819-9543-6612
Erica.Dunn@ec.gc.ca
William Eddleman
Department of Biology-MS6200
Southeast Missouri State University
Cape Girardeau, MO 63701-4799
Phone: 573-651-2364
Fax: 573-651-2223
weddlema@biology.semo.edu
Charles Francis
Bird Studies Canada
PO Box 160
Port Rowan, Ontario
CANADA N0E 1M0
Phone: 519-586-3531
Fax: 519-586-3532
cfrancis@bsc-eoc.org
James Gibbs
State University of New York
College of Environ. Science and Forestry
350 Illick Hall
1 Forestry Drive
Syracuse, NY 13210
Phone: 315-470-6764
Fax: 315-470-6934
jpgibbs@syr.edu
Michael Haramis
USGS, BRD
Patuxent Wildlife Research Center
11410 American Holly Drive
Laurel, MD 20708-4015
Phone: 301-497-5651
Fax: 301-497-5624
Michael_Haramis@usgs.gov
Douglas Helmers
USDA, National Conservation Service (NRCS)
601 Business Loop I-70W
Parkade Center, Suite 250
Columbia, MO 65203
Phone: 573-876-0900
Fax: 573-876-0914
doug.helmers@mo.nrcs.usda.gov
William Hohman
USDA, NRCS
Iowa State University
124 Science II
Ames, IA 50011-3221
Phone: 515-294-8591
Fax: 515-294-2234
Whohman@iastate.edu
Marshall Howe
USGS, BRD
Patuxent Wildlife Research Center
12100 Beech Forest Drive
Laurel, MD 20708
Phone: 301-497-5858
Fax: 301-497-5784
Marshall_howe@usgs.gov
Douglas Johnson
USGS, BRD
Northern Prairie Wildlife Research Center
Route 1, Box 96C
Jamestown, ND 58401
Phone: 701-253-5539
Fax: 701-253-5553
Douglas_H_Johnson@usgs.gov
Proceedings of the Marsh Bird Monitoring Workshop 21
Greg Kearns
Patuxent River Park
1600 Croom Airport Road
Upper Marlboro, MD 20772
Phone: 301-627-6074
Fax: 301-952-9754
Eileen Kirsch
USGS, BRD
Upper Mississippi Science Center
2630 Fanta Reed Rd.
P O Box 818
LaCrosse, WI 54602-0818
Phone: 608-783-6451 Ext. 226
Fax: 608-783-6066
Eileen_Kirsch@usgs.gov
David Klute
Pennsylvania Cooperative Fish and Wildlife
Research Unit
Pennsylvania State University
107 Merkle Building
University Park, PA 16802
Phone: 814-865-3972
Fax: 814-863-4710
dsk5@psu.edu
Hal Laskowski
USFWS
2145 Key Wallace Drive
Cambridge, MD 21613
Phone: 410-221-1836
Fax: 410-228-3261
Hal_Laskowski@mail.fws.gov
Steve Lewis
USFWS
Bishop Henry Whipple Federal Bldg
1 Federal Drive
Fort Snelling, MN 55111-4056
Phone: 612-713-5473
Fax: 612-713-5286
steve_j_lewis@mail.fws.gov
Soch Lor
Cornell University
Department of Natural Resources
Fernow Hall
Ithaca, NY 14853-0188
Phone: 607-255-2606
Fax: 607-255-1895
SKl11@cornell.edu
Richard Malecki
USGS, BRD
New York Cooperative Fish
and Wildlife Research Unit
Fernow Hall, Room 202
Cornell University
Ithaca, NY 14853-0188
Phone: 607-255-2839
Fax: 607-255-1895
RAM26@cornell.edu
Jim Mattsson
C/o Mark Twain NWR
1704 N. 24th Street
Quincy, IL 62301
Phone: 217-224-8580
Fax: 217-224-8583
Jim_mattsson@mail.fws.gov
Scott Melvin
Massachusetts Division of Fisheries and Wildlife
Westboro, MA 01581
Phone: 508-792-7270 ext. 150
Fax: 508-792-7275
scott.melvin@state.ma.us
Kelly Mizell
Department of Wildlife and Fisheries Sciences
Texas A & M University
College Station, TX 77843-2258
Phone: 409-845-5777
Fax: 409-845-3786
Michael Monahan
Co-Director of Environmental Science
University of Denver
Denver, CO 80208
Phone: 303-871-3540
Fax: 303-871-3471
mmonahan@du.edu
Charles Paine
Max McGraw Wildlife Foundation
PO Box 9
Dundee, IL 60118
Phone: 847-428-6331
Fax: 847-428-8157
CRPaine@aol.com
Bruce Peterjohn
USGS, BRD
Patuxent Wildlife Research Center
12100 Beech Forest Road
Laurel, MD 20708-4037
Phone: 301-497-5841
Fax: 301-497-5784
Bruce_Peterjohn@usgs.gov
22 Proceedings of the Marsh Bird Monitoring Workshop
Dan Petit
USFWS
Office of Migratory Bird Management
4401 N Fairfax Drive
Arlington, VA 22203
Phone: 703-358-1824
Fax: 703-358-2272
daniel_petit@mail.fws.gov
Christine Ribic
USGS, BRD
Wisconsin Cooperative Wildlife Research Unit
University of Wisconsin
Department of Wildlife Ecology
226 Russell Labs
1630 Linden Drive
Madison, WI 53706-1598
Phone: 608-263-6556
Fax: 608-263-4519
caribic@facstaff.wisc.edu
John Sauer
USGS, BRD
Patuxent Wildlife Research Center
11510 American Holly Drive
Laurel, MD 20708-4017
Phone: 301-497-5662
Fax: 301-497-5666
John_R_Sauer@usgs.gov
Mark Shieldcastle
Crane Creek Wildlife Research Station
Ohio Division of Wildlife
13229 W St., Rt. 2
Oak Harbor, OH 43449
Phone: 419-898-0960
Fax: 419-898-4017
Mark.shieldcastle@dnr.state.oh.us
David Shuford
Point Reyes Bird Observatory
4990 Shoreline Highway
Stinson Beach, CA 94970
Phone: 415-868-1221, Ext. 14
Fax: 415-868-1946
dshuford@prbo.org
Jennifer Skoloda
University of Wisconsin
Department of Wildlife Ecology
226 Russell Labs
1630 Linden Drive
Madison, WI 53706-1598
Phone: 608-263-7595
Fax: 608-262-6099
jskoloda@students.wisc.edu
Graham Smith
USFWS
Office of Migratory Bird Management
11500 American Holly Drive
Laurel, MD 20708-4016
Phone: 301-497-5860
Fax: 301-497-5871
graham_smith@mail.fws.gov
Kelli Stone
USFWS
Midcontinent Ecological Science Center
4512 McMurray Avenue
Fort Collins, CO 80525-3400
Phone: 970-226-9408
Fax: 970-226-9230
kelli_stone@mail.fws.gov
Glenn Therres
Maryland Department of Natural Resources
Wildlife and Heritage Division
580 Taylor Avenue, E-1
Annapolis, MD 21401
Phone: 410-260-8572
Fax: 410-260-8596
gtherres@dnr.state.md.us
John Trapp
USFWS
Office of Migratory Bird Management
4404 N. Fairfax Drive ARLSQ-634
Arlington, VA 22203
Phone: 703-358-1821
Fax: 703-358-2217
john_trapp@mail.fws.gov
Cecilia Walther
Louisiana State University
227 Forestry-Wildlife-Fisheries Bldg.
Baton Rouge, LA 70803-6202
Phone: 504-388-4131
Fax: 504-388-4227
cwalthe@tiger.lsu.edu
Russ Weeber
Bird Studies Canada
PO Box 160
Port Rowan, Ontario
CANADA N0E 1M0
Phone: 519-586-3531
Fax: 519-586-3532
rweeber@bsc-eoc.org
Proceedings of the Marsh Bird Monitoring Workshop 23
Discussion Group I:
Field Protocols for Marshbird Monitoring
Leader: Scott Melvin
Recorder: Marshall Howe
(a) Develop a list of species (or groups of species)
covered by marsh monitoring. Describe any
techniques or protocols already in place for specific
species or species groups.
(b) What time of day should surveys be conducted?
The participants may want to come up with a list of
species or species groups that should be monitored
at dawn, dusk, day, or at night, and any species or
groups that could be monitored at a variety of
times).
(c) What weather restrictions are there for doing
surveys?
(d) Where should the sampling points be located? On
roads, off roads by walking into the habitat, in the
water done by boat? This will bring in practicality
issues. The group may want to discuss the optimum
location but consider the implications if the optimum
cannot be reached. Are the alternatives acceptable?
One of the criticisms of the BBS is that the surveys
are all done from roads. Can this criticism be
avoided at the outset? Is it an important
consideration?
(e) For call-playback surveys, what quality, loudness,
length, and order of play should tapes be?
(f) Can multiple survey methodologies be combined
into one survey? What are the implications of doing
that? What are the drawbacks? What are the
benefits?
(g) Consider logistical restrictions when sampling at
a refuge level, on a statewide level, and on a range-wide
level that might impact recommended
protocols.
(h) Should habitat information be collected every
time a survey is done? If so, what habitat variables
should be measured and what protocols should be
used?
(i) What are the highest priority research needs in
the area of field protocols, e.g. where should we be
putting our limited research dollars?
Discussion Group II:
Statistical Design and Sampling Considerations
Leader: Jon Bart
Recorder: Chris Ribic
(a) How would design and sampling change for
surveys done at various spatial scales:
1. Local (e.g., refuge) need information on
presence, abundance (maybe in just a gross
sense), habitat associations, and responses to
management.
2. State need information on abundance and trends,
especially for state-protected rare species. For
hunted species, also need information for setting
hunting seasons.
3. National need information on status and
landscape relationships. For hunted species, also
need information for setting hunting seasons.
4. North America (e.g. rangewide) need
information on status, trends and landscape
relationships.
(b) What are the implications for the sampling unit
when monitoring at these different scales?
(c) Would stratification be an important tool to
consider? What would be some possible stratification
variables?
(d) What data are necessary to do a power analysis
for a multi-species monitoring program at the
different scales?
(e) For call playback surveys, what is the
relationship of call-count indices to true population
levels?
(f) What are the highest priority research needs in
the area of sampling and statistical considerations
(e.g., where should we be putting our limited
research dollars)?
24
Appendix 2. Topics for the Discussion Groups at
the Marsh Bird Monitoring Workshop.
Discussion Group III:
Implementation of a Marshbird Monitoring Program
Leader: Bruce Peterjohn
Recorder: Glen Therres
(a) What species and geographic areas (e.g., refuge,
states, U.S.) should be focused on?
(b) What are the agencies’ needs?
(c) What role should each agency and organizations
(e.g., states, FWS, BRD, Flyway Councils,
NGOs) play in monitoring marshbirds?
(d) Is a central database desirable? If so, can BRD
oversee this?
(e) To what degree can volunteers be used in
implementing a monitoring program?
Proceedings of the Marsh Bird Monitoring Workshop 25
Group/Species Latin Name
Grebes Family Podicipedidae
Bitterns and herons Family Ardeidae
Rails, gallinules, moorhens, coots Family Rallidae
Cranes Family Gruidae
Snipe Order Charadriiformes
Pied-billed grebe Podilymbus podiceps
American bittern Botaurus lentiginosus
Least bittern Ixobrychus exilis
Yellow rail Coturnicops noveboracensis
Black rail Laterallus jamaicensis
Clapper rail Rallus longirostris
King rail Rallus elegans
Virginia rail Rallus limicola
Sora Porzana carolina
Purple gallinule Porphyrula martinica
Common moorhen Gallinula chloropus
American coot Fulica americana
Common snipe Gallinago gallinago
Franklin’s gull Larus pipixcan
Forster’s tern Sterna forsteri
Black tern Chlidonias niger
Belted kingfisher Ceryle alcyon
Alder flycatcher Empidonax alnorum
Willow flycatcher Empidonax traillii
Sedge wren Cistothorus platensis
Marsh wren Cistothorus palustris
Common yellowthroat Geothlypis trichas
Red-winged blackbird Aegelaius phoeniceus
Yellow-headed blackbird Xanthocephalus xanthocephalus
LeConte’s sparrow Ammodramus leconteii
Saltmarsh sharp-tailed sparrow Ammodramus caudacutus
American Ornithologists’ Union (AOU). 1998. Check-list of North American birds, 7th ed. Allen Press, Lawrence, KS.
26
Appendix 3. Common and Latin Names of Bird
Groups and Species in Report.
Latin names are taken from AOU (1998).
Considerations Regarding Refuge Participation in
Landscape Scale Monitoring
Robert S. Adamcik, Division of Refuges, U.S. Fish and Wildlife Service,
4401N. Fairfax Drive, Arlington, VA 22203
The National Wildlife Refuge System includes the single most extensive network of wetland reserves in the
United States. Hence, continental, national, or regional marshbird monitoring initiatives will likely include
refuges as sampling units. Gaining the participation of refuges in such initiatives requires a firm
understanding of the ecological and administrative context in which refuges function, the factors driving
refuge survey activities, and the manner in which refuge managers choose those surveys in which to
participate. Surveys on refuges can contribute to biological objectives at three spatial scales: (1) refuge-specific
surveys driven by specific refuge objectives, such as assessment of management activities; (2)
landscape scale surveys that address issues relevant to both the refuge and surrounding landowners; and
(3) regional, national, or continental monitoring programs. Refuge priorities are focused on Category 1,
which are derived from specific management objectives for each refuge. These surveys often provide critical
information for adaptive management. Refuge participation in the other two categories may be related to
relevance of the project to the refuge, as well as availability of funding and staff.
Recruitment of refuges into broader geographic monitoring initiatives will depend heavily upon whether the
data gathered are relevant to ongoing or proposed refuge management, and whether the sampling
technique and intensity are simple enough that a manager will not have to forego other management
priorities.
27
Appendix 4. Abstracts of Invited Presentations
at the Marsh Bird Monitoring Workshop
26-28 April, 1998.
Proposed Applications of Data from a Marshbird Monitoring Network
Paul R. Adamus, Department of Fisheries and Wildlife,
Nash Hall, Oregon State University, Corvallis, OR 97330
If implemented, a continental or regional marshbird monitoring program would serve mainly to detect
temporal trends in populations of individual species. However, with modification such a program could also
analyze spatial trends in collected data to identify areas where the biological integrity of wetlands and their
associated landscapes is being degraded. Especially when used with data from monitoring of other taxa,
knowledge gained from a marshbird monitoring program could form a basis for remedial action, such as
selection of particular wetland sites for restoration or implementation of watershed management plans.
Conceptually, one tool for interpreting and presenting the data from a marshbird monitoring program might
be the “Index of Biotic Integrity” (IBI). Aquatic biologists have developed and validated IBI’s regionally,
using fish and macroinvertebrates, to indicate the condition of streams and lakes. Developing and applying
wetland bird IBI’s probably would require identification of all bird species at each monitoring site, and
assignment of the species to guilds or assemblages known to be sensitive to pollution and other human
disturbance (as distinct from extreme naturally-occurring conditions). Pilot projects in riparian habitats in
Pennsylvania, Maryland, the mid-Atlantic highlands, and the Southwest have supported the usefulness of
bird communities as indicators of landscape disturbance, so continued development and validation of bird
IBI’s is warranted. Further research should focus on (1) defining the domain of regions, landscapes, wetland
types, and disturbance types over which wetland bird IBI’s can provide accurate information on
environmental impairment, (2) identifying which wetland bird assemblages are most sensitive to human-caused
disturbances, and (3) identifying the metrics (combinations of variables) most able to separate
human-related from natural, macroscale spatial variation in bird community structure.
Since 1996 the USEPA has sponsored a Biological Assessment of Wetlands Work Group (BAWWG),
consisting of scientists from state and federal agencies and universities, to organize and communicate
information on protocols for monitoring and analyzing data on wetland birds, amphibians,
macroinvertebrates, algae, and plants. BAWWG has recently prepared information for state agencies to use
in developing and testing bird IBI’s in wetland and riparian habitats.
28 Proceedings of the Marsh Bird Monitoring Workshop
Evaluation and Development of Survey Techniques for Common Snipe
Kevin D. Downs and Stanley H. Anderson, Wyoming Cooperative Fish and Wildlife Research Unit,
University of Wyoming, Box 3166, University Station, Laramie, WY 82071
Common snipe (Gallinago gallinago delicata) numbers in North America are unknown. No survey
technique to monitor their numbers has been developed and doubts exist about using the Breeding Bird
Survey (BBS) to monitor snipe trends. In order to help solve these problems, this research project focused
on four major objectives: (1) evaluate existing labor intensive techniques such as tape playback, territorial
mapping, territorial/spot mapping, flush counts and nest searches; (2) develop and evaluate indices (call,
winnow and call/winnow counts) that could monitor snipe numbers on a continental scale; (3) identify
important habitat characteristics; (4) evaluate the effectiveness of the BBS for common snipe. Four labor
intensive techniques were evaluated for effectively estimating the number of snipe or territories. Territorial/
spot mapping was found to be the most effective technique because the results obtained accurately
represented the number of territories present. Flush counts were the second best technique because they
did not accurately estimate the number of snipe present in all snipe habitats. Flush counts worked best in
long narrow strip habitats that could be easily covered by a few people. In large continuous habitats flush
counts did not accurately estimate snipe numbers because snipe were harder to flush. Tape playback
territorial mapping and nest searches did not accurately estimate snipe numbers due to varying snipe
responses and difficulty in finding nests. An index that counts the number of different calling and winnowing
snipe heard most accurately monitored snipe numbers because the probability of detection, percentage of
snipe detected and the estimate of snipe numbers was the greatest, most consistent and most accurate. The
index should be conducted during the 20 day time period before snipe begin to incubate and after the arrival
of females. This 20 day time period usually falls after the full moon in April and before the first quarter in
May. Transects should be established in the habitat with points spaced 1 km apart. minimum of seven points
should be used to conducts surveys from in order to obtain good results. The survey is conducted at dusk
approximately 5 minutes after most of the snipe become active and should be discontinued when snipe cease
or decrease activity. The survey period is short (only usually lasts 30-45 minutes), so only 3 points or less
may be surveyed in one night. Each point should be visited at least 3 times during the 20 day time period.
There are six factors (season, time of day, lunar cycle, wind speed, solar radiation, and temperature) that
effect snipe winnowing and calling and should be considered or at least measured when conducting snipe
surveys. Driving roads the season prior to actual surveying will help identify possible snipe areas. Those
areas should also be searched for snipe nests the season prior to surveying, in order to have good
phenological information about the snipe area in question, which will help identify appropriate survey times
the following season. Surveyors should be standardized prior to conducting actual surveys if more than one
surveyor is used. For more detailed information about surveying for snipe refer to Downs (1998). From our
results, we identified four important characteristics (distance to nearest shrubby cover, water depth, soil
moisture, the dominance of sedges and sedge heights) and used them to construct a ranking criteria for
categorizing snipe habitats. We found the BBS is not conducted at the right time of day nor does it consider
the variables (season, lunar phase, solar radiation and temperature) that affect snipe activity to be effective
in detecting snipe trends. As a result, the numbers of snipe the BBS detects is low and the snipe activity
detected could be related to nest success more than abundance. A final report by Masters candidate Kevin
D. Downs will be completed during the spring of 1998. These results are from the last year of a 3-year study
funded by the 1995 Webless Migratory Game Bird Research Program (U.S. Fish and Wildlife Service and
the U.S. Geological Survey-Biological Resources Division), U.S. Fish and Wildlife Service (Migratory Bird
Management Office), Wyoming Game and Fish Department and the Wyoming Cooperative Fish and Wildlife
Research Unit.
Downs, K. D. 1998. Common snipe surveys, habitat and evaluation of the breeding bird survey. M.S. thesis,
University of Wyoming, Laramie. 91pp.
Proceedings of the Marsh Bird Monitoring Workshop 29
A 12-Step Program for Creating a Monitoring Program:
An Extended Abstract Written in the Everyday Language of the Wildlife Biologist
Sam Droege, U.S. Geological Survey, Patuxent Wildlife Research Center, 11410 American Holly Drive,
Laurel, MD 20708-4015
The traditional approach to designing a new program is to pick the technique your buddies use, put in as
many samples as looks good on a map, locate those samples so that you get to hang out in your favorite
habitats, and hope that 20 years down the road someone will analyze and interpret it all for you.
In this document I propose an alternative model, one hopefully superior to that of the traditional one. For
those of you still captivated by the traditional approach, consider this a 12-step plan to freedom. When
examples are needed they will be for the situation of setting up a monitoring program for marsh inhabiting
birds.
Step 1.
First comes the question of which species are being targeting for monitoring. All birds that live in and
around marshes? Some ranking of primary species, secondary species, and collateral? What of things like
frogs, toads, and other non-birds too?
Step 2.
Over what time periods do you want your monitoring program to report on trends of marshbirds?
2, 3, 5, 10, 25, 100?
Step 3.
What MINIMUM magnitude of change or trend in marshbirds should this program be able to detect over
the time period stated in Step 2?
10%, 50%, 100%, ….- 3% per year, 5% per year?
Step 4.
Over what geographic region do you want to talk about changes in marshbird populations?
A refuge? A province? A country? A continent? Now we need to work through the whole notion of how
precisely you might want to talk about these trends.
Step 5.
Please designate your willingness to cry wolf, done via the setting of alpha (Type I error)
levels. That is, how often are you willing to claim that a significant population trend has occurred when, in
actuality, none has. For example, alpha (Type I error) can be 5%, 10%, or 20%. Hint: It is better to cry wolf
20% of the time and spend a little bit of time following up false leads than it is to set you alpha level so small
that populations have all gone to Hell while you waited for everything to be extremely statistically
significant.
Step 6.
Designate what proportion of the time the monitoring program should detect a trend if one
was really going on. This is known as setting the power of the program. For example, power can be 80%,
90%, 95%, or even, 100%. Hint: Detecting trends is the whole point of a monitoring program;. No?, so set
your power high.
Now we come to the phase of inspecting the monitoring technique you propose to use track population
trends and ask the question: Can it really deliver the goods?
WARNING: Pilot data required.
30 Proceedings of the Marsh Bird Monitoring Workshop
Step 7.
Either find someone who has used the methodology you are considering for your monitoring program (and
is willing to share their data with you) or go out and collect some pilot data on your own.
Step 8.
Ask your pilot data the following questions (Hint: These are simple calculations of means and sample
standard deviations and you can do them in any spreadsheet program).
What are the mean counts on plots or routes that get your target species? Low mean counts (<10) are
undesirable because the smaller the mean the greater your likelihood of getting a slug of zeros on a
significant number of your counts. Because the existence of negative birds is still a hotly debated topic
among biostatisticians (they really should spend more time out in the field) you have effectively run into a
wall with your monitoring program (that is, once your reach zero birds, no more information is available
regarding status) and the mathematics become difficult.
What is the spatial variation, that is, how variable are counts geographically (among plots or routes) for each
species?
How many completely zero plots or routes turn up (that is, plots that never get the species) when you set
your samples out? Sites that do not have the target species obviously tell you nothing about the trend in that
species, so, when you estimate how many samples you need to monitor a species you must adjust the sample
size upwards to account for the fact that a certain proportion of those plots will be always be zero.
How variable are counts for each species at a single site? The best way to answer this question is by looking
at long-term datasets, so that you can look at how the numbers vary over a period of years. If you only have
one year to collect pilot data and no existing datasets are available to mine, then you can take repeated
counts at a single site to get some minimum idea about sampling variation. However, be warned that
repeated counts WITHIN a year underestimate the overall count variation which must account for within
and among year variation in counts (this captures the variation due to the slop in your technique AND due
to normal fluctuations in numbers of critters among years).
Step 9.
Can a poorly performing technique be saved by covariates? If you find an ungodly amount of variation in
your counts, it may be possible that you can account for some of that noise (and decrease your needed
number of samples while increasing your power to detect trends) by accounting for the effects that
phenology, weather, and observers have on your system.
Step 10.
So that is all well and good, now how do I figure out how many samples I need?
What follows is a simple recipe for calculating that sample size. First go to the power analysis cookbook web
site at:
http://WWW.MP1-PWRC.USGS.GOV/powcase/powcase.html
(note: the occurrence of the letter one in the domain name, that is not the letter L)
And download the MONITOR.EXE sample size ez-baker and read over the enclosed directions for proper
use of this product. Then add the following ingredients into the MONITOR.EXE program:
  Take the mean and standard deviations of counts and make some CVs.
  Calculate the proportion of zero routes.
  Mix well with your targets for:
  Length and steepness of trend.
  Precision of the estimates targets.
  Bake your worst species CV first.
  Double the number of samples to make sure that you have enough.
Now we come to the very hardest part of the process, the part where most of us fall from grace.
Proceedings of the Marsh Bird Monitoring Workshop 31
Step 11.
Prove that the method you are using yields an unbiased view of trends in the critters you are monitoring.
Several paths are open to you, pilgrim.
High Road:
Empirically prove that there is a relationship between the index you propose and the real number of birds
out there.
Road to Nowhere:
Heuristically prove that the likelihood of bias affecting the counts or affecting the sign of the slope is low to
nil.
See http://WWW.MP1-PWRC.USGS.GOV/…. for an example using counts of frogs and toads.
Road to Ruin:
This would be the attitude expounded by Baron Ernest Rutherford: “If your result needs a statistician then
you should design a better experiment”
In general, discovering the effects of bias is a difficult and expensive path to travel upon. But by avoiding
that path you take the risk of your program painting a false picture of the trends in your target animals. At
minimum, an attempt should be made to use the technique in conjunction with a set of mark-recapture
populations to investigate the relationships between count and population over several potentially biasing
situations.
Step 12.
Ok, now that you have reached sainthood by passing through the previous 11 steps you are ready to figure
out where you are going to put out your samples. Keep in mind the following: All things being equal all
things should be equal, that is, plot and route placement should be random or systematic. However,
completely random placement of plots is rarely efficient and you will likely need to think about setting up a
set of decision rules (e.g., only wetlands near roads will be surveyed) along with some stratification rules
(e.g., we will put in a higher density of samples in these areas because they are more important or they have
more potential observers).
Congratulations, you have made it through the 12 Step Monitoring Program. Now go check in with your
neighborhood statistician and have him check under the hood of your sampling design just to make sure you
have really been cured of your old habits.
32 Proceedings of the Marsh Bird Monitoring Workshop
Statistical Issues Related to Monitoring Marsh Birds
Charles M. Francis and Russ C. Weeber, Bird Studies Canada,
P.O. Box 160, Port Rowan, Ontario N0E 1M0, Canada
We examined a number of statistical issues related to monitoring marsh birds that arose from a scientific
evaluation of the Marsh Monitoring Program (MMP) that has been run in the Great Lakes basin by Bird
Studies Canada (Long Point Bird Observatory) in collaboration with Environment Canada since 1995. Some
of these issues are discussed in the companion presentation by Weeber et al. Here, we focus on three
themes: route selection, sampling intensity (number of surveys per year), and power for trend detection.
1. Route Selection
The MMP was developed to sample marsh habitats throughout the Great Lakes basin, with particular
emphasis on marshes in “Areas of Concern” (AOC’s). The sampling procedure for AOC’s is different from
that outside of these areas. Within AOC’s, the survey goal has been to sample every marsh. Although this
has not been completely achieved, all AOC’s and most marshes within them are covered by at least one
route. As such, to the extent that stations are suitably positioned along these routes, the routes should
provide a fairly good sampling of AOC marshes (although sampling intensity is generally not adequate to
characterize individual AOC’s well). Outside of AOC’s, routes have been primarily picked by volunteers, and
we can be much less confident that they are representative of non-AOC marshes within the survey area (the
Great Lakes basin).
There are many challenges to developing a more representative (i.e. stratified randomized) sample of
marshes with a volunteer-based survey such as this one. One approach would be to select marshes based
upon a GIS database of all marshes/wetlands in the region. Selection could be based on strata determined
by criteria such as marsh characteristics (e.g. size, adjacent land use), ecoregions, and location (e.g. distance
from potential sources of surveyors). This method could provide known sampling probabilities for all
marshes in the study area, but has several potential difficulties. In practice, the sampling would have to be
further constrained by accessibility (not too far from a road and accessible by foot or possibly by canoe).
Discrete marshes are not necessarily easy to delineate. Problems could arise if the database is incomplete
(e.g. if some types or sizes of marshes are inadequately represented), if it does not accurately differentiate
all marshes from other types of wetlands, or if accessibility cannot readily be determined from the database.
At present, sampling for the whole Great Lakes Basin would require GIS databases from several sources
that probably differ in their characteristics.
An alternative sampling procedure would be to select random positions in the landscape, with appropriate
stratification, in a similar fashion to route selection for the Breeding Bird Survey, then survey the nearest
accessible marsh. This approach would be cheaper and easier to implement (GIS technology not required),
but will not lead to equal selection probability for all marshes: marshes with few neighboring marshes will
have a higher probability of being selected than marshes with many neighbors, and large marshes will be
more likely to be picked than small marshes. Stratification of the landscape by marsh density (if suitable
data are available to do this) could reduce but not eliminate this problem.
Regardless of the selection procedure, there are further constraints associated with matching routes and
volunteers. If routes are pre-selected, and volunteers allowed to choose from among the selected routes,
then the sample will no longer be strictly random if any of the selected routes are not covered—those
farthest from population centers or least accessible may be less likely to be picked. If routes are selected
upon demand for each volunteer, selections will need to be constrained to within a reasonable radius from
the volunteer’s home base. Furthermore, regardless of initial assignments, volunteers may be less likely to
continue surveying routes that are difficult or otherwise unattractive.
In view of the preceding limitations, it is quite possible that population trends on our volunteer-picked
routes will not be any less representative of the basin as a whole than those on randomly picked routes. For
this reason, we are currently encouraging volunteers to continue to survey existing routes, to ensure
adequate continuity for trend analysis using route regression. However, we plan to adopt some sort of
randomization procedure for selection of new routes as they are required in the future.
Proceedings of the Marsh Bird Monitoring Workshop 33
2. Number of Surveys
The MMP protocol for sampling birds requests two surveys per year on each route, at least 10 days apart
between May 20 and July 5 each year. Two surveys were chosen to reduce the effects of seasonal variation in
detectability of some species, and to increase the number of species detected at each station. However,
because the second survey carries a cost of reduced total numbers of routes surveyed, it is appropriate to
consider in some detail the actual benefits of two surveys based upon data from 1995-97. In each year, 70-
80% of stations were surveyed twice within the desired time frame. Of these, a mean of 0.5-1.1 (depending
upon the year) more species were detected on the first survey (mean date June 5) than on the second survey
(mean date June 25). On average, only 40% of species were detected on both surveys—the second survey
increased the number of species detected at each station from a mean of 17 to a mean of 24. Considering only
marsh nesting birds detected at 10 or more stations each year (22 species), 11 species were significantly
more likely to be detected on the first survey, but 4 species were more likely to be detected on the second
survey, highlighting the fact that the optimal survey time differs among species. For 4 of these species, there
were significant (p <0.05) differences among the three years in the relative value of the first and second
survey. This indicates that a single survey might confound annual variation in seasonal detectability
(possibly due to variation in vegetation phenology) with annual variation in populations. Thus the second
survey clearly does have benefits especially for evaluation of individual routes. From the perspective of
trend estimation, we do not know how many additional routes volunteers might be able to run if only a
single survey were required each year (it would probably be much less than double), and hence whether this
would outweigh the benefits of the second survey.
3. Survey Power for Estimating Trends.
We estimated the power of the survey to detect long-term population trends in mean numbers of 11 species
of marsh-nesting birds considered of particular conservation concern. We assumed analysis using log-linear
route regression models, and extrapolated variance estimates from changes on routes run in both 1995 and
1996. For 10 of the 11 species, a sample of 100 routes per year (on which the species was detected) was
estimated to be sufficient to detect a consistent annual change of 1-3% over 10 years (at p <0.05, power =
0.80). However, of 91 routes run in both years, most of the target species were detected on fewer than half of
the routes. Thus, for some of these species, the survey with its current sampling intensity could not be
expected to detect changes less than 4% per year over a 10-year period.
34 Proceedings of the Marsh Bird Monitoring Workshop
Techniques for Monitoring Inconspicuous Waterbirds:
Results from Studies in Maine and Massachusetts
James P. Gibbs, College of Environmental Science and Forestry, State University of New York,
1 Forestry Drive, Syracuse, NY 13210
Scott M. Melvin, Massachusetts Div.of Fisheries & Wildlife, Rte. 135, Westborough, MA 01581
Shawn K. Crowley, Department of Forestry and Wildlife Management, Holdsworth Hall,
University of Massachusetts, Amherst, MA 01003
We broadcast vocalizations of pied-billed grebe (Podilymbus podiceps), American bittern (Botaurus
lentiginosus), least bittern (Ixobrychus exilis), Virginia rail (Rallus limicola), and sora (Porzana carolina)
at 60 wetlands in Maine in 1989 and 1990 to derive standardized methods to monitor breeding populations of
these species. All wetlands contained 0.1-125.0 ha of emergent vegetation. Broadcasts of tape-recorded calls
at 1-10 stations per wetland between 1May and 15 July improved detectability for all species by 93 to 1,320%
over 15-min passive observation periods at the edge of each wetland. Detection rates at wetlands where
target species were known to occur ranged between 0.56 (least bittern) and 0.86 (pied-billed grebe) per
survey visit. Three visits to a wetland were adequate to determine presence or absence of all species with
90% certainty. Least bitterns, soras, and Virginia rails were detected primarily within 50 m of observers,
while pied-billed grebes and American bitterns were detected up to 500 m away. Most detections were of
birds heard and not seen. Responsiveness of each species varied in relation to seasonal chronology, time of
day, wind, precipitation, and cloud cover.
We surveyed 123 emergent wetlands in Massachusetts and Maine during the 1991 and 1992 breeding seasons
to determine if broadcasts of tape-recorded calls increased detection rates of pied-billed grebes, American
bitterns, least bitterns, green herons (Butorides striatus), Virginia rails, and soras. Following a 3-minute
passive listening period at each station, we broadcast 1 minute of calls for each species at 1-10 survey
stations per wetland. Detection rates of Virginia rails and soras were greater during the conspecific call
portion of the broadcast sequence than during passive listening periods. In addition, both rails exhibited a
peak of response within the broadcast sequence during broadcast of conspecific calls. In contrast, both
green heron and American bittern had higher detection rates during passive listening periods than during
broadcasts of calls. However, due to small sample size and potential bias, we recommend further study of
responses of American bitterns to broadcasts of taped vocalizations. Sample sizes of pied-billed grebe and
least bittern were too small for analysis.
We suggest that regional populations of marsh birds can be monitored using standardized surveys along
mini-routes comprised of 10-15 point-count stations per route. From 10-15 survey points can be done in a
morning, either within 1 large or several smaller wetlands. In the northeastern United States, we
recommend that broadcasts of tape-recorded vocalizations be used to increase detection rates of pied-billed
grebes, least bitterns, Virginia rails, soras, and, perhaps, American bitterns. Except for wetlands that are
small and easily accessible, surveys should be conducted by canoe. Surveys should be conducted between
0430 and 1000 h on days with calm or light winds and no steady precipitation. Randomly stratifying sampling
effort between large, species-rich wetlands and wetlands that are small, of marginal habitat quality, were
historically occupied by target species, or have been recently created, for example by beavers, may be
desirable. This design would balance the need to maximize detection rates while surveying populations over
a representative sample of available habitats. Finally, monitoring habitat changes along mini-routes would
be important for interpreting population trends of marsh birds.
Proceedings of the Marsh Bird Monitoring Workshop 35
Designing Effective Regional Monitoring Programs for Marshbirds
James P. Gibbs, State University of New York, College of Environmental Science and Forestry,
350 Illick Hall, Syracuse, NY 13210
Scott M. Melvin, Massachusetts Division of Fisheries and Wildlife, Field Headquarters,
Westborough, Massachusetts 01581
In developing regional monitoring programs, biologists should retain a focus on the major challenge that we
face: to devise sampling methods that permit unbiased and statistically powerful surveys of marshbird
populations in a logistically feasible manner. Clear articulation of monitoring objectives (time frame, trend
strength sought for detection, alpha, and beta) is the first step in this process. Based on a decade of
experience with marshbird monitoring in Maine and Massachusetts, we highlight several considerations for
the effective design of regional marshbird monitoring programs. These include whether call-response
surveys adequately “index” marshbird abundance, use of waterbird mini-routes’ as the unit for sampling
marshbird populations, and tools for estimating the trade-offs between survey effort and power to detect
trends in marshbird populations.
36 Proceedings of the Marsh Bird Monitoring Workshop
Sora Rail Studies on the Patuxent River, Maryland
Michael Haramis, U.S. Geological Survey, Patuxent Wildlife Research Center, 11410 American Holly
Drive, Laurel, MD 20708-4015
Gregory D. Kearns, Maryland National-Capital Park and Planning Commission, Patuxent River Park,
16000 Croom Airport Road, Upper Marlboro, MD 20772-8395
The freshwater marshes of the tidal Patuxent River are well known for their annual fall concentration of
migrant soras (Porzana carolina) and were formerly the most famous rail hunting grounds in the
Chesapeake Bay region. Because of concern over the apparent long-term decline in number of soras and the
decline in the quality of the Patuxent marshes, especially the loss of wild rice (Zizania aquatica), the
Maryland National-Capital Park and Planning Commission (MNCPPC), co-steward of the Jug Bay National
Estuarine Research Reserve, sponsored rail-related research beginning in 1987. Past efforts focused on
developing efficient trapping techniques, age and sex criteria, and monitoring body mass dynamics. Noted
progress was made in developing digital playback systems and trap improvements to enhance sora captures.
These improvements increased capture success by over an order of magnitude and resulted in capture of
2,315 soras and 276 Virginia rails (Rallus limicola) in the 5 year period, 1993-97. Although these methods
demonstrate the efficacy of banding large numbers of soras on migration and possibly winter concentration
areas, captures at the Patuxent River site have been 70-90% hatching-year birds and recoveries and
recaptures have been virtually nonexistent. With the present effort, this outcome precludes population
parameter estimation using traditional capture-recapture or recovery model methodologies.
In 1996, studies were initiated to employ radio telemetry methods to investigate length of stay, habitat use,
survival, and migration characteristics of fall migrant soras. These studies are ongoing and will be continued
through 1998 with a grant from the U.S. Fish and Wildlife Service’s Webless Migratory Game Bird
Research Program and support from the U.S. Geological Survey’s Patuxent Wildlife Research Center.
Supplemental funding has also been provided by MNCPPC, FWS Region 5, the Maryland Ornithological
Society, Quail Unlimited, and Prince Georges Community College.
During 1996-97 we developed a successful radio transmitter attachment technique to secure 1.8g radio
transmitters over the synsacrum of migrant soras. We modified Rappole and Tipton’s (1991) leg-loop
attachment method by addition of a waist loop to prevent soras from slipping transmitters over their short
tails. Thin gauge (0.6mm) elastic thread proved ideal for attachment and allowed for girth expansion
associated with fattening during stopover. Sixty instrumented soras have been monitored in two years of
study from early September until early November. Only a single mortality was recorded and 41 (68%) were
confirmed and another 13 (total 90%) were believed to have migrated from the study area. Only a single bird
slipped a radio transmitter. Most birds demonstrated a sedentary nature in the marsh throughout stopover.
Average length of stay was 44 days in 1997 (n = 29) with peak departure occurring 20-24 October.
Departing migrants were detected using a receiver/data-logger monitoring system placed 4 miles down
river from the study site. Thirty-six of 37 (97%) soras departed in a 2-hour window of time, beginning 1 hour
after sunset. Departure was synchronized with cold fronts on clear, starlit nights. Twenty-five soras were
monitored on migration from 8 km to as far away as 770 km. Findings indicate migration flight speeds of 40
to 50mph and a direct southward orientation from the study site at least until contact with the Atlantic
Ocean west of Cape Lookout, North Carolina. We also attempted to monitor passage of migrant soras radio
tagged at Iroquois National Wildlife Refuge in western New York by biologist Soch Lor. None were
recorded passing a monitoring unit at Haldeman Island in the middle of the Susquehanna River 10 miles
north of Harrisburg, Pennsylvania. This further corroborates our findings that when migrating overland,
soras move in a direct southward orientation and are not following land features such as major rivers. We
will attempt to expand our tracking effort of migrant soras especially in the deep South in fall 1998.
In 1997 we further examined the effect of our audio lure on the age, sex, and capture rate of soras. We
alternated 5 trap lines with and 5 lines without playback every 2 days for a period of about 5 weeks. The
results indicated a 2 x capture rate with playback (p <0.001), with no effect related to age (p >0.1), or sex
among AHY birds (p >0.9). Sex effect among HY birds was less conclusive because of sexing error. Using
DNA finger printing techniques on a grab sample of 81 HY soras, we determined our sexing error to be
about 13% in favor of males. Applying this correction, our test was marginally significant (p = 0.06, n = 622).
The DNA sexing results indicate that it is unlikely that HY soras can be sexed at a confidence level of 90%
or better. This finding is preliminary and we intend to reexamine this issue in 1998.
Proceedings of the Marsh Bird Monitoring Workshop 37
Comparison of Point Count and Play Back Survey:
Techniques for Detecting Breeding Wetland Birds
Eileen Kirsch, Upper Mississippi Science Center,
2630 Fanta Reed Rd. PO Box 818, LaCrosse, WI 54602-0818
Tape play back can be effective for detecting inconspicuous wetland birds. However, in a broader study
focusing on breeding passerines, I wanted to determine if playing the tape affected detectability of wetland
passerines, and if tape play back could increase detectability of the inconspicuous species although the
survey design was not optimal for them. Therefore, both survey types were conducted consecutively during
a single visit at each point. Within a 50 m radius of an observer, neither survey method was better than the
other (paired t-test) for detecting total numbers of birds, numbers of species, and individuals of most
species. The exception was that more Virginia rail (Rallus limicola) were detected with play back than point
count. At an unlimited distance, however, more birds, species, and individual Marsh Wrens (Cistothorus
palustris), Song Sparrows (Melospiza melodia), Swamp Sparrows (M. georgiana), Willow Flycatchers
(Empidonax trailii), Yellow Warblers (Dendroica petechia), and Red-winged Blackbirds (Agelaius
phoeniceus) were detected with point counts. At an unlimited distance, play back was more effective for
detecting Virginia Rails and Soras (Porzana carolina), but not Least Bitterns (Ixobrychus exilis), Common
Moorhens (Gallinula chloropus), and Pied-billed Grebes (Podilymbus podiceps). Survey order did not
affect comparisons between the 2 methods for either distance. If all wetland birds are the target for study or
monitoring, an unlimited detection distance is desired, and only a single visit to survey sites is possible, both
methods should be used.
38 Proceedings of the Marsh Bird Monitoring Workshop
Vocalization Behavior and Response of Black Rails
Michael L. Legare, Dynamac Corporation, DYN-2, Kennedy Space Center, FL 32899
William R. Eddleman, Department of Biology, MS6200, Southeast Missouri State University,
Cape Girardeau, MO 63701
P. A. Buckley, U.S. Geological Survey, Patuxent Wildlife Research Center,
11410 American Holly Drive, Laurel, MD 20708-4015
Colleen Kelly, Department of Mathematics, San Diego State University,
5500 Campanile Drive, San Diego, CA 92182-7720
We measured the vocal responses and movements of radio-tagged black rails (Laterallus jamaicensis) (n =
43, 26 males, 17 females) to playback of vocalizations at 2 sites in Florida during the breeding seasons of
1992-95. We used regression coefficients from logistic regression equations to model the probability of a
response conditional to the birds’ sex, nesting status, distance to playback source, and the time of survey.
With a probability of 0.811, non-nesting male black rails were most likely to respond to playback, while
nesting females were the least likely to respond (probability = 0.189). Linear regression was used to
determine daily, monthly, and annual variation in response from weekly playback surveys along a fixed route
during the breeding seasons of 1993-95. Significant sources of variation in the linear regression model were
month (F = 3.89, df = 3, p = 0.0140), year (F = 9.37, df = 2, p = 0.0003), temperature (F = 5.44, df=1, p =
0.0236), and month*year (F = 2.69, df = 5, p = 0.0311). The model was highly significant (p <0.0001) and
explained 53% of the variation of mean response per survey period (R2 = 0.5353). Response probability data
obtained from the radio-tagged black rails and data from the weekly playback survey route were combined
to provide a density estimate of 0.25 birds/ha for the St. Johns National Wildlife Refuge.
Density estimates for black rails may be obtained from playback surveys, and fixed radius circular plots.
Circular plots should be considered as having a radius of 80 m and be located so the plot centers are 150 m
apart. Playback tapes should contain one series of Kic-kic-kerr and Growl vocalizations recorded within the
same geographic region as the study area. Surveys should be conducted from 0-2 hours after sunrise or 0-2
hours before sunset, during the pre-nesting season, and when wind velocity is <20 kph. Observers should
listen for 3-4 minutes after playing the survey tape and record responses heard during that time. Observers
should be trained to identify black rail vocalizations and should have acceptable hearing ability.
Given the number of variables that may have large effects on the response behavior of black rails to tape
playback, we recommend that future studies using playback surveys should be cautious when presenting
estimates of “absolute” density. Though results did account for variation in response behavior, we believe
that additional variation in vocal response between sites, with breeding status, and bird density remains in
question. Playback surveys along fixed routes providing a simple index of abundance would be useful to
monitor populations over large ge