Patuxent Wildlife Research Center

NAAMP III Archive - community & regional
Home | Archive by Alphabetical Order | Archive by Category

Recommendations for Monitoring Amphibians at Multiple Spatial Scales

Karen M. Dvornich
Washington Cooperative Fish and Wildlife Research Unit
University of Washington
Box 357980
Seattle, WA 98195-7980

Kelly R. McAllister
Washington Department of Fish and Wildlife
600 Capitol Way N.
Olympia, WA 98501

[ Abstract ]

The Washington State Gap Analysis Project (WAGAP) is part of a national program which seeks to enhance conservation of biological diversity by identifying plant communities and terrestrial vertebrate species that are underrepresented on lands managed for biodiversity (Scott et al. 1993). Typically, species occurrence data and published information on habitat affinities are used to create models of each species' range so that species richness analyses can be completed over large geographic areas.

Historically, amphibian specimens collected and deposited in herpetological collections have been accompanied by very little documentation. A complete lack of habitat information has been the rule rather the exception. President Clinton's Forest Ecosystem Management Assessment Team (FEMAT, 1993) found few data available for amphibians, and what were available were distributed throughout museums, agencies, timber companies and private collections in a variety of formats. Species' based land management planning efforts must therefore depend upon general amphibian references for habitat information. These references (see for example, Corkran and Thoms 1996, Leonard et al. 1993, Nussbaum et al. 1983, Stebbins 1985) are usually replete with microhabitat descriptions but are often lacking in landscape level habitat descriptions. Therefore, modeling habitat relations for microhabitat associates using macrohabitat information was considered a daunting challenge.

To meet the challenge, WAGAP established a Herp Committee consisting of academicians and natural resource agency professionals. The Herp Committee started the data collection process with a nationwide search for distribution data based on records associated with voucher specimens available in herpetological collections. Since most of these data were relatively old (pre-1960), batches of more recent field inventory data were obtained and incorporated to the database. All data were reviewed for reasonableness. In some instances, specimens were examined to assess the accuracy of the identification. Field data that were not supported by voucher specimens were accepted if trained and experienced scientists were involved in the data collection. All of these data were used to create the best predicted vertebrate distribution maps possible. We attempted to capture the available geographic precision with expectations of conducting further analyses at a finer resolution


WAGAP characterized the entire state by ecological zones and predominate vegetative cover types (Cassidy, 1996). Sixteen scenes of 1991 Landsat Thematic Mapper (TM) satellite imagery covering Washington State were spectrally clustered into approximately 200 classes per scene. The classes were grouped by similar spectral values and areas of similar land cover types were delineated manually using the clustered TM imagery as a backdrop. Each resulting land cover polygon was given a label that included ecoregion, vegetation zone, and primary, secondary and tertiary land cover and the proportion of each polygon by land cover. An ecoregion is a contiguous geographic area of similar climate and geological history. A vegetation zone is an area in which moisture, temperature, elevation and other environmental parameters combine to create conditions that favor similar vegetation communities.

The primary land cover is the actual land cover that occupies the greatest proportion of the area in a polygon (e.g. closed canopy conifer forest or irrigated row-crop agriculture). Secondary and tertiary land covers are actual land covers that occupy the second and third greatest proportion of area in the polygon. Washington State's land cover map consisted of approximately 15,000 land cover polygons. There were 1,865 unique combinations of ecoregion/vegetation zone/land cover (habitat) within the State. Unique combinations of ecoregion and vegetation zone totaled 100.

Vertebrate Models

WAGAP vertebrate models were developed as matrices delineating ecoregion, vegetation zone and land cover into core and peripheral zones on a species by species basis. Core vegetation zones were identified by literature search, expert opinions and known breeding occurrence. Core zones composed the range where species were most commonly found. Peripheral zones were areas of infrequent occurrence where distribution appeared to become uncharacteristically patchy. Often, these areas were at the margins of a species' range limits.

The matrices also included a suitability code for each unique habitat combination. The minimum mapping unit of Gap was 100 ha for terrestrial cover types and 40 ha for wetlands. Therefore, it was not possible to produce maps based on detection of the microhabitats that amphibians are usually associated with.

The matrices were assigned codes indicating the suitability of the land cover polygon. Polygons were labeled as not suitable, good, adequate, contingent upon suitable habitats below the minimum mapping unit. The assignment of suitability codes was obviously subjective and further study is needed to refine such a scheme.

Sampling Schemes

The combination of ecological vegetation zones and land cover types produces polygons of similar character. The vertebrate models provide guidance for developing field survey protocols appropriate for the fauna predicted to occur in the area to be studied. We recommend that the WAGAP habitat characterization be used to develop sampling schemes for both inventory and long-term monitoring of amphibians on state and regional scales.

The WAGAP vertebrate models are well-suited for use in designing inventory and monitoring. The amphibian models benefited from a baseline dataset of 10,631 records. The Rana cascadae range model, for example, was based on 744 records from 411 separate locations. These records spanned a 76 years of zoological collection and observation (Dvornich et al, 1996). They clearly demonstrate an association with higher elevation vegetation zones in the Cascades and Olympic Peninsula ecoregions. Here, the species inhabits meadows, small ponds, and streamsides.

The vertebrate model for Rana cascadae indicated which vegetation zones were likely to have populations of this species. Table 1 illustrates the vegetation zones selected within each ecoregion. The total number of habitats in each vegetation zone were compared with the habitats selected. Each vegetation zone was identified as core or peripheral. A total of 754 ecoregion/vegetation zone/habitat combinations existed for the 6 ecoregions and 10 vegetation zones in Rana cascadae's predicted distribution, and 438 were selected by the habitat model matrix A total of 1,502 polygons were selected for its core habitat and 553 for peripheral. Inventory and monitoring protocols employed in these vegetation zones should include techniques known to be effective in detecting Rana cascadae and, for monitoring, the techniques should be effective for assessments of population size or relative abundance of the species.

A stratified random sampling scheme could be used to select polygons in which monitoring sites and protocols could be established. Within the zones inhabited by Rana cascadae, much of the land is roadless and very difficult to access. In such zones, we recommend stratified random sampling to allow greater sampling effort in the more accessible polygons. Stratified sampling should also utilize information on the location of streams, emergent marshes, and small ponds to ensure that polygons within these important habitats are well sampled.

It is our recommendation that inventory and monitoring programs treat each ecoregion/vegetation zone as a separate entity. Within each zone sampling should be adequate for assessing conditions throughout the zone. Field techniques should take advantage of WAGAP range models to design field techniques capable of detecting and/or enumerating all of the species predicted to occur in the zone.

Priorities for field inventories of ecoregion/vegetation zones are those zones known or predicted to have rare or declining forms. Rana pretiosa, Rana pipiens, Plethodon larselli, and Bufo woodhousei are well known for their rarity or the extent of their declines. Bufo woodhousei is an example of a species with a small predicted range which falls entirely within Rana catesbeiana's predicted range. Few Bufo woodhousei populations are known. For Bufo woodhousei, the WAGAP database contained only 25 historical locations, though a study on the Hanford Nuclear Reservation was responsible for adding most of the 42 new locations added in 1996. Sampling of sites with known populations and sampling within ecoregion/vegetation zones predicted to have Bufo woodhousei and other rare or declining species should be considered a priority.

As alluded to in the previous paragraph, there are other considerations in establishing the priorities for inventory and monitoring activities. One is to consider the likelihood of continued range expansion of introduced species such as Rana catesbeiana and to develop inventories that are able to track changes in their ranges. Rana catesbeiana may play a role in the decline of other amphibians. Predicting range was difficult for this species because field surveys have not kept up with the natural and human-assisted range increases of this introduced species. Additionally, the range of habitats used by Rana catesbeiana (and many other amphibians) in Washington is imperfectly known. Because of the suspected negative effects of Rana catesbeiana on native amphibians, sampling of areas adjacent to the currently known range of this species should be considered a priority.

Other Recommendations

It became clear during the development of the vertebrate range models that improvement of these models will be possible with finer scale habitat data as well as better information on the attributes of suitable habitat for each species. At a coarse scale, Dicamptodon copei, for instance, can be assigned to streams within conifer forests. However, eventually, resource inventories and more powerful computers will allow for distinctions between streams of different sizes and substrate types as well as the already existing data that allow us to model based on elevation, aspect, and slope. To better understand how these features relate to the habitat needs of reptiles and amphibians, field surveys should begin to measure and record these parameters.

We also recommend that observations from the general public, especially school children, be included in amphibian inventory and monitoring programs. While the public may not adhere to strict sampling protocols needed for rigorous statistical analysis, their data will provide important information on the status and habitat use of reptiles and amphibians. In the initial phases of a statewide NatureMapping program utilizing school teachers and their students, sixty school children reported Phrynosoma douglassii at locations known to them. These 60 locality records represent more than what has been reported historically for this species and these children have provided greater mapping precision and more detailed habitat information than was available from most other data sources.


Cassidy, K.M., 1996. Land cover of Washington State, Volume 1. In K.M. Cassidy, C.E. Grue, M.R. Smith, K.M. Dvornich, eds., Washington State Gap Analysis - Final Report, Washington Cooperative Fish and Wildlife Research Unit, University of Washington, Seattle, Volumes 1-5.

Corkran, C. C. and C. Thoms. 1996. Amphibians of Oregon, Washington, and British Columbia - A field identification guide. Lone Pine publ. 175pp.

Dvornich, K.M., K.R. McAllister, and K.B. Aubry, 1996. Amphibians and Reptiles of Washington State, Volume 2. In K.M. Cassidy, C.E. Grue, M.R. Smith, K.M. Dvornich, eds., Washington State Gap Analysis - Final Report, Washington Cooperative Fish and Wildlife Research Unit, University of Washington, Seattle, Volumes 1-5.

Forest Ecosystem Management Assessment Group, 1993. Forest ecosystem management: An ecological, economic and social assessment. Report of the Forest Ecosystem Management Assessment Team, Chapter IV, 211 pp.

Leonard, W. P., H. A. Brown, L. L. C. Jones, K. R. McAllister, and R. M. Storm. 1993. Amphibians of Washington and Oregon. Seattle Audubon Soc. Seattle. 168pp.

Nussbaum, R. A., E. D. Brodie, and R. M. Storm. 1983. Amphibians and reptiles of the Pacific Northwest. Univ. Press of Idaho, Moscow. 332pp.

Scott, J.M., F. Davis, B. Csuti, R. Noss, B. Butterfield, C. Groves, H. Anderson, S. Caicco, F. D'Erchia, T.C. Edwards, Jr., J. Ulliman, and R.G. Wright, 1993. Gap Analysis: A geographic approach to the protection of biological diversity, Wildlife Monographs, 123:1-41.

Stebbins, R. C. 1985. A field guide to western reptiles and amphibians, 2nd ed. Houghton Mifflin, Boston. 336pp.


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