Patuxent Wildlife Research Center
NAAMP III Archive - calling surveys
Home | Archive by Alphabetical Order | Archive by Category
Patuxent
Wildlife Research Center
NAAMP III Archive
- calling surveys
Home | Archive
by Alphabetical Order | Archive by Category
Leonard J. Shirose1, Christine A. Bishop1,
Cameron J. MacDonald2, Ronald J. Brooks2
[ Abstract ]
Extensive monitoring programs using volunteer observers in conjunction with
professional biologists are currently being used in Canada and the United States
in an attempt to detect fluctuations and trends in both the abundance and distribution
of amphibian populations using standardized techniques. Amphibian call count
surveys have been implemented in Quebec, Ontario, Manitoba, Saskatchewan, New
York, Pennsylvania, Ohio, Michigan, Indiana, Wisconsin, Illinois, Minnesota,
Missouri, and Iowa. The amphibian call count protocol used in Ontario is a point-count
method that relies primarily upon the efforts of volunteer observers. Three-minute
surveys are conducted along routes consisting of up to 10 permanently marked
stations. Because of the difficulty of estimating distance in the dark, no attempt
is made to limit the calls counted to a fixed distance around each survey point.
Stations are set at intervals of at least 500 m to minimize the risk of counting
individuals more than once. Surveys begin one-half hour after sunset and are
completed no later than midnight. Each route is surveyed three times during
the spring and early summer, with the dates of surveys depending on the latitude
of the route.
Ontario has been divided into north, central and south regions demarcated at the 43rd and 47th degrees of latitude. The three survey periods are 15 days in duration and are set a minimum of 15 days apart. Participants are asked to survey during specific weather conditions. Minimum air temperatures for surveying are 8C (first survey), 13C (second survey), and 21C (third survey). At each station, participants record observer name, date of survey, survey number, time, and weather conditions including air and water temperatures, precipitation type, cloud cover, and wind speed rated using the Beaufort Wind Scale (World Meteorological Organization,1970). For more details about the development of this protocol, see Bishop et al. (in press).
Observers estimate the relative abundance of calling species by rating the
density of individual calls as: 0 = no calls heard, 1 = individual calls not
overlapping, 2 = some overlapping calls, but the number of individuals calling
can be reliably estimated, 3 = a continuous chorus of calls in which individual
calls could not be discerned. If calling intensity is code 1, the number of
individuals calling is counted; if code 2, the number is estimated (Bishop et
al., in press).
During the summer of 1995, we looked at the presence/absence data to see how
well volunteers agreed with experts about what species were present, and to
see whether a 3-minute point count is long enough to hear most of the species
that are calling on any given night. We found that there was a high degree of
agreement between observers (although there was some bias associated with the
level of experience of the volunteers) and we found that most species calling
on a given night are heard within the first 2-minutes of a survey. This suggests
that these calling amphibian surveys could be useful for gathering information
on the presence of species over a fairly broad geographic range.
Another aim of calling amphibian surveys is to use calling intensity as an
index of relative abundance, and to make inferences about demographic trends
based on trends in relative abundance. In 1995 we combined intensive and extensive
monitoring and found a linear relationship between the number of individuals
heard calling and the number of males captured in choruses in Bullfrogs and
Fowler's Toads. We didn't find a relationship between estimated population size
and the number of individuals heard calling in either species, but we had only
a single population of each species to work with, so we didn't consider those
results to be conclusive.
Our objectives in the 1996 validation study were to:
1) quantify the relationships between number heard calling and number of males
captured in choruses, and between calling intensity code and the number of males
captured in choruses for the Wood Frog and the American Toad
2) quantify the relationships between number heard calling and estimated population
size (males only) and between the calling intensity code and estimated population
size for the Bullfrog, Wood Frog, and American Toad
In this study we report preliminary results of the 1996 validation study.
Methods
To investigate the relationship between relative abundance evaluated with auditory
surveys and actual numbers present, we coupled amphibian call count surveys
with intensive monitoring (mark-recapture) for 3 populations of Wood Frogs in
the Long Point area between 15 April-20 April, 3 populations of American Toads
at Long Point and Guelph between 8 May-22 May, and 5 populations of Bullfrogs
at Pelee Island, Lake St. Clair, Long Point, Jack Lake and Algonquin Park in
between 8 June-23 July.
Auditory surveys according to the protocol used by the volunteers. After the
auditory surveys, we captured and marked all the males we could find during
2 complete sweeps of the pond. Captured animals were marked with a toe-clip
(in the Wood Frogs and American Toads) or a pit-tag (in the Bullfrogs). They
were sexed, measured and released.
Population size estimates for each population were calculated using 3 methods: a Petersen Estimate (a very simple model for closed populations), Jolly's Estimate (a model for open populations), and a Frequency of Capture Method (which makes inferences about population size based on fitting a zero-truncated Poisson distribution to the frequency of capture data and estimating the number of individuals that were never caught at all) (Caughley, 1977).
We used regression analysis for comparisons involving the number of callers heard vs the number of males captured and for the number of callers vs the estimated population size.
For comparisons involving the calling intensity code vs the number of males
captured and vs the estimated population size, we used Analysis of Variance
and treated the calling intensity. code as a categorical variable.
Results
We found a significant linear relationship between the number of males heard
calling and the number of males captured for Rana sylvatica (n
= 9, r2 = 0.486, P = 0.037;
Fig. 1) and Bufo
americanus (n = 13, r2 = 0.411, P
= 0.018; Fig.
2). The slope of the regression line was well under 1 for both
species, so the number calling generally underestimates the number of males
actually present.
We found a significant positive relationship between the calling intensity
code and the number of males captured in choruses for the Wood Frogs (n
= 12, r2 = 0.713, P = 0.015; Table
1), but not for the American Toads (n = 12, r2
= 0.713, P = 0.015; Table 1).
The estimates of bullfrog population size calculated with the Frequency of
Capture method and the Petersen method were reasonably similar in size and in
the order in which the different populations were ranked according to size.
The estimates calculated with the Jolly method were consistently much lower
than those from the other 2 methods and produced a different ranking order (
Table 2).
There was a significant relationship between number of bullfrogs heard calling
and population size when population size was estimated with Petersen's Method
(n = 12, r2 = 0.34, P = 0.036) and the
Frequency of Capture method (n = 12, r2 = 0.307,
P = 0.05; Fig.
3). There was no significant relationship between the number of
bullfrogs calling and the population size estimated with Jolly's Method (n
= 13, r2 = 0.143, P = 0.187). There was a positive
relationship between calling intensity code and population size calculated with
the Petersen method (n = 13, r2 = 0.494, P
= 0.007) and with the Frequency of Capture Data method (n = 13, r2
= 0.448, P = 0.012). There was no significant relationship when Population
size was calculated with Jolly's method (n = 13, r2
= 0.143, P = 0.202).
There was no relationship between the number of males heard calling and the
size of populations of American toads or wood frogs when population size was
estimated with any of the 3 methods (American toads: Petersen- n =
10, r2 = 0.042, P = 0.569; Frequency of Capture-
n = 13, r2 = 0.054, P = 0.518 (
Fig.4); wood frogs:
Petersen- n = 17, r2 = 0.052, P = 0.381;
Frequency of Capture- n = 17, r2 = 0.052, P
= 0.380 ( Fig.
5); Jolly- n = 17, r2 = 0.051, P
= 0.384). We found no relationship between calling intensity code and population
size (calculated with any of the 3 methods) for the wood frogs ( Petersen- n
= 17, r2 = 0.087, P = 0.660; Frequency of Capture-
n = 17, r2 = 0.088, P = 0.659; Jolly-
n = 17, r2 = 0.087, P = 0.662) or the
American toads (Petersen- n = 10, r2 = 0.011, P
= 0.773; Frequency of Capture- n = 13, r2 = 0.031,
P = 0.626).
Discussion
Volunteer-based calling amphibian surveys in North America have been around
for several years now, and there is an expectation among both volunteers and
funding agencies that these programs should soon be providing answers to the
question of which species are in decline. However, we cannot begin to interpret
the data until we know what we are measuring with auditory surveys. Measures
of calling intensity, whether they are simple counts of numbers calling or estimates
of calling intensity level, will only be useful indices of abundance if there
is a significant statistical relationship between calling intensity and the
actual number of animals present.
To date, we have found a significant linear relationship between the number
of males heard calling and the number captured in choruses for four species:
Rana catesbeiana (Shirose et al. in press), Bufo fowleri (Shirose
et al. in press), Rana sylvatica (this study), and Bufo americanus
(this study). This suggests that the number heard calling can be used as an
index of chorus size in these species, although the shallow slope of the regression
line indicates that call counts consistently underestimate the number of males
present, and the relatively low correlation coefficient suggests that a large
sample size might be required to detect trends. More research is needed to determine
if the relationship is free of trends across years, which is necessary if call
counts are to be used to track chorus size. It is possible that the relationship
between call counts and chorus size is density dependent, with choruses becoming
saturated with callers as densities increase and additional males adopting alternative
breeding strategies that do not involve calling (Arak, 1983; Arak, 1988; Hoglund
and Robertson, 1988; Tejedo, 1993).
We found a significant positive relationship between the calling intensity
code and the number of males captured in choruses for Rana sylvatica,
but not for Bufo americanus, which suggests that the utility of calling
intensity codes as indices of chorus size depends on which species are present.
The lack of a relationship between calling intensity and chorus size in Bufo
americanus may be related to the characteristics of its call. The call
of this species is a long, loud trill that carries for a long distance. A chorus
of calling American toads may be easily resolved into individual callers at
a close distance (and therefore be recorded as code = 1), while a chorus of
the same size but much farther away are more difficult to resolve (and may therefore
be recorded as code = 2 or 3). Because volunteers are asked to record amphibians
calling at an unlimited distance, the spatial distribution and distance from
the observer of calling males may influence the calling intensity code assigned
for species whose calls carry a long distance.
Chorus size may not be a good indicator of population size because it is likely
that the proportion of adult males in a population is not constant across years,
and that calling activity varies from night-to-night according to present and
antecedent weather conditions. We found no significant relationship between
the number of males heard calling and the estimated population size in either
Rana sylvatica or Bufo americanus, although we found a relationship
between the number heard calling and chorus size in both species. Similarly,
we found no relationship between the calling intensity code and the estimated
population size in Rana sylvatica, although we did find a significant
relationship between calling intensity code and chorus size. Neither call counts
nor calling intensity codes appear to be useful indices of population size in
either of these species. We did find a significant positive relationship between
the estimated population size and both the number of males heard calling and
the calling intensity code for Rana catesbeiana (but only with 2 of
the 3 methods of population size estimation), which suggests that call counts
and calling intensity codes may be useful as indices of population size in this
species.
Because it appears that measures of calling intensity (as opposed to presence
data) are not useful indices of population size for all species of anurans,
validation experiments should be undertaken on a species-by-species basis as
a prelude to making any inferences based on such data. In cases where a relationship
between calling intensity and population size has been demonstrated, power calculations
based on nightly variation in calling activity for discrete populations should
be performed to ascertain whether trends can be identified with a reasonable
number of routes.
These are preliminary results. We have a small amount of data left to analyze
and some reservations about what we have so far. The strength (very existence!)
of the relationship between calling intensity and population size in Rana
catesbeiana depends on which method of population estimation is used and
we have not yet determined which method is most appropriate (there are many
more models to choose from). Any advice on selecting the most appropriate model
would be very welcome. Also, we are concerned about pseudoreplication in our
regression analysis because we have multiple observations from discrete populations.
In the most strict sense, we have demonstrated a relationship between calling
intensity and site, and we have assumed that differences among sites were due
primarily to differences in population size among the sites. We have not tested
this assumption, and we must acknowledge the possibility that differences in
acoustic properties, etc. among sites may have muddied the waters. We welcome
any comments or advice on how to deal with this.
Literature Cited
Arak, A. 1983. Male-male competition and mate choice in anuran
amphibians. Pp 181-210. In P. P. G. Bateson (Ed.), Mate Choice. Cambridge University
Press, Cambridge, U.K.
Arak, A. 1988. Callers and satellites in the natterjack toad: Evolutionary
stable decision rules. Anim. Behav. 36:416-432.
Bishop, C. A., K. E. Pettit, M. E. Gartshore and D. A. McLeod. Monitoring anuran
populations using call counts and road transects. In press. In D. M. Green (Ed.),
Amphibians in Decline. Reports from the Canadian Declining Amphibian Task Force.
Herpetological Conservation Vol. 1, Society for the Study of Amphibians and
Reptiles/ Canadian Association of Herpetologists, Montreal, Canada.
Caughley, G. 1977. Analysis of Vertebrate Populations. J. Wiley and Sons, Ltd.,
Toronto, Canada.
Hoglund, J., and J. G. M. Robertson. 1988. Chorusing behaviour, a density-dependent
alternative mating strategy in male common toads (Bufo bufo). Ethology 79:324-332.
Shirose, L. J., C. A. Bishop, C. J. MacDonald, and R. J. Brooks. in press.
validation tests of an amphibian call count survey technique in Ontario, Canada.
Tejedo, M. 1993. Do male natterjack toads join larger breeding choruses to
increase mating success. Copeia 1993:75-80.
World Meteorological Organization. 1970. The Beaufort Scale of wind force.
WMO Commission for Maritime Meteorology, Marine Sciences Affairs Report 3:1-22.
U.S. Department of the Interior
U.S. Geological Survey
Patuxent Wildlife Research Center
Laurel, MD, USA 20708-4038
http://www.pwrc.usgs.gov/naamp3/naamp3.html
Contact: Sam Droege, email: Sam_Droege@usgs.gov
Last Modified: June 2002