Project Summary:
Amphibian declines world-wide have created a need for more extensive
and standardized monitoring of amphibian populations and for elucidating
underlying causes of amphibian declines. In response to concerns about
amphibian populations, a study evaluating and validating amphibian survey
techniques was initiated in Shenandoah and Big Bend National Parks under
the auspices of PRIMENet.
The goals for the project are to document spatial and temporal variation
in amphibian populations and associations between amphibian populations
and natural and anthropogenic environmental factors in the Parks. We
will (1) evaluate bias, precision, and efficiency for several sampling
methods for amphibians, (2) conduct validation studies to determine
the relationship between capture indices and adjusted population estimates
using capture-recapture and other techniques, 3) evaluate temporal and
spatial variation in amphibian populations in association with environmental
features, and (4) implement park-wide monitoring programs that will
provide baseline data on amphibian populations.
In order to design appropriate monitoring programs, it is necessary
to first develop sampling frames for the species of interest, then validate
and estimate variances associated with site-specific population estimates
within the sampling frames. Several different sampling frames and estimation
procedures will be needed to ensure that amphibian groups are adequately
sampled. In the first year, emphasis will be placed on studies designed
to validate population indices and estimate variation in the indices
spatially and in relationship to environmental variables. In the second
year, the information from the first year will be used to develop extensive
surveys based on the tested indices (or visibility-adjusted population
estimates) within sampling frames based on information on habitats and
on spatial variation in environmental variables. In the third year,
which will be continued through a series of university contracts, further
development and testing of methodological and spatial components will
occur.
Sampling frames must be developed for terrestrial salamanders, stream
and streamside salamanders, and frogs and toads. For each group, habitats
will be identified and preliminary strata will be developed using GIS
information from the parks. Within these preliminary strata, a variety
of validation studies will be conducted to assess appropriate sampling
methods in the first year of the project. In the second year, the GIS-based
sample frames will be used to select random samples within strata and
sampling (and additional validation) conducted within the strata. Based
on results from the second year, the sampling frames will be reassessed
and additional sampling and validation will be conducted in the third
year with the intent of better documenting associations between environmental
features and amphibian populations.
Central to the development of monitoring methods is validation of population
indices. In general, only a portion of animals in a given sampling area
are generally encountered, and investigators must either use the count
data as an index to abundance or use estimation procedures such as capture-recapture
to estimate the proportion of animals missed. Indices to abundance never
provide estimates of absolute abundance unless the proportion of animals
sampled (the detection rate) is estimated; differences in indices can
not be used to estimate population change unless the detection rates
are the same in the indices that are compared. Consequently, before
indices are used in a monitoring program or any analysis of population
change, it is necessary to estimate detection rates to determine their
magnitudes and evaluate whether the rates change over time or space.
If the indices are not consistent (i.e., the detection rates differ
over time or space), estimation of detection rates must be an implicit
component of the monitoring program.
With amphibians, further complications are introduced because a single
sampling method cannot be used to monitor all the species. Instead,
several methods must be used, each of which samples a group of species
with common life history attributes. Each of these methods must be evaluated
with regard to variation in detection rates for each group of species.
We will implement a variety of validation studies for preliminary indices
to amphibian abundance. Additional studies in progress at the USGS Patuxent
Wildlife Research Center and other sites will provide information that
will also be used to estimate detection rates for indices used in the
PRIMENet monitoring. The validation studies have three primary objectives:
(1) estimation of detection rates for selected indices of amphibian
abundance, (2) estimation of measurement error associated with indices,
and (3) estimation of spatial and temporal variation in indices, and
of relevant environmental features likely to be associated with variation
in indices. Where possible, these validation studies will be placed
in experimental contexts where these estimates of population size and
variation can be used to test for relationships with natural and anthropogenic
stressors.
In spring/summer 1999, additional tasks of sampling amphibians along
a Pb gradient in Acadia National Park will be coordinated with the K.
Weathers investigation of deposition gradients.
PRIMENet
amphibian publications
2001 PRIMENet Meeting Presentation:
PRIMENet
Amphibian Project (Robin Jung, Sam Droege, John Sauer)
2000 PRIMENet Meeting
Presentations: PRIMENet
Amphibian Project (Robin Jung, Sam Droege and John Sauer); PRIMENet
Amphibian Project 2 (Robin Jung, Sam Droege and John Sauer)
For more information
about the PRIMENet Amphibian Monitoring Program visit: http://www.mp2-pwrc.usgs.gov/amphibian_monitoring.html
