Doyle L. Crosswhite1,4, Stanley F. Fox1, David M. Leslie, Jr.2, and Mark S. Gregory3
1 Department of Zoology,
2 Oklahoma Cooperative Fish and Wildlife Research Unit, USGS Biological Resources Division,
3 Department of Agronomy, Oklahoma State University, Stillwater, OK 74078
4Current address: Natural Sciences Division, Saint Gregory's University, Shawnee, OK 74801.
We surveyed the status and distribution of the Ouachita Dusky Salamander (Desmognathus brimleyorum) in southeastern Oklahoma. To predict occurrence of D. brimleyorum, we used Geographic Information System (GIS) techniques to identify regions with appropriate habitat conditions. During June 1993-September 1994, 130 animals were observed at 18 of 57 sites in five counties. The species was absent from seven of 13 historic collection sites examined. Desmognathus brimleyorum was locally abundant given certain habitat conditions, such as, in and around seeps, springs, and high gradient streams. This species is likely sensitive to disturbances that disrupt the hydrology of these habitats. ©1998 Oklahoma Academy of Science
Stejneger (1) first described Desmognathus brimleyorum as a relatively large and robust plethodontid salamander that is often aquatic or semiaquatic. This salamander generally inhabits areas near streams, springs, and seeps where it is often found under submerged rocks, woody debris, or buried in wet gravel. Although D. brimleyorum is restricted to the Ouachita Mountains (2), there has been limited study of its status (3-5). Until recently, the range of this species was defined poorly, and its abundance was virtually unknown. Aside from our study, the most recent collections were made from 1977 to 1984 when Karlin et al, (6) obtained 861 individuals from 49 locations, 11 of which were in Oklahoma.
One factor of concern regarding D. brimleyorum is the overlap of its range with that of an intensive timber industry. Several timber companies in the Ouachita Mountains practice clearcutting, which may potentially alter hydrologic features around springs and seeps that are primary habitats for the species. For example, clearcutting can cause changes in soil structure, hydrology, and both horizontal and vertical vegetation structure that subsequently affects temperature and moisture regimes. These changes in turn affect microhabitats that are important to salamanders (7-13). The water quality of streams may be degraded by increased sedimentation. Both a deterioration of microclimatic conditions on the forest floor and a decrease in stream quality are facilitated by the removal of canopy, the elimination of the moisture-retaining forest floor litter, and soil compaction (8,14,15).
Our objectives were to determine the current status and distribution of D. brimleyorum in Oklahoma by reviewing museum records and sampling historic locations, and to document any changes in its distribution or abundance.
To evaluate the historical distribution of D. brimleyorum in Oklahoma, we compiled museum records from Oklahoma State University, University of Oklahoma, University of Kansas, University of Texas-Austin, and the American Museum of Natural History. From June 1993 to September 1994, streams, springs, and seeps were surveyed in Choctaw, Latimer, LeFlore, McCurtain, and Pushmataha Counties in southeastern Oklahoma. Thirteen of the 57 sample sites were selected based on historical collections of D. brimleyorum; the
other 44 sites were selected opportunistically within the historic range for the species. Presence or absence of D. brimleyorum was determined by sampling salamanders at each site using aquatic dip nets, 3.6 × 1.8-m nylon seines (3.2-mm mesh), and by hand.
Each of the sample locations was plotted on 1:100,000 scale United States Geological Survey topographic quadrangle maps, which were digitized into a Geographic Information System (GIS) using GRASS 4.0 software (Geographical Resources Analysis Support System, U.S. Army Corps of Engineers, Construction Engineering Research Laboratory, Champaign, Illinois) and an Altek graphic digitizer. The GIS was then used to compare the presence or absence of D. brimleyorum with ancillary data layers to identify areas with appropriate habitat conditions. The ancillary data layers included land cover/use, soil mapping units, elevation, and percent slope. Data on land cover/use, soil mapping units, and percent slope were obtained from the Natural Resources Conservation Service (MIADS) database. Elevational data were derived from USGS digital elevation models. All data were georeferenced to a Universal Transverse Mercator projection, zone 15, North American Datum 1927.
One hundred thirty D. brimleyorum were observed at 18 of 57 localities in five counties in southeastern Oklahoma (Table 1, Fig. 1). D. brimleyorum was absent from seven of 13 historic collection sites examined. At least one range extension was noted: we found D. brimleyorum at Goat Creek (T3N,R20E,S3), an isolated, upland stream in northern Latimer County. We believe this is the northwestern-most record for the species. We located populations at six sites in McCurtain County, nine in LeFlore County, one in Latimer County, and two in Pushmataha County. Populations were most dense in LeFlore County along the north-facing slopes of Winding Stair and Rich Mountains within the Ouachita National Forest. D. brimleyorum was most abundant at Horsethief Spring, LeFlore County.
Overall, we found D. brimleyorum in riparian zones adjacent to and within high gradient streams, first- and second-order streams with depths ranging from 3 to 20 cm and widths from 0.2 to 5.0 m. Salamanders seemed to be associated with streams having a well developed, deciduous canopy. Water temperatures at sites where D. brimleyorum were present ranged from 4 to 21 °C. D. brimleyorum were most often found in streams that contained coarse, rocky or gravel substrates
with at least some rocks > 10 cm in width, and where both pools and riffles existed. We never found D. brimleyorum in or near streams with sand or silt substrates. We found most adult salamanders under rocks in flowing water, and most larvae were observed on the bottom of partially shaded pools. Salamanders typically did not occur in streams that supported populations of predatory fish such as sunfish (Lepomis), bass (Micropterus), and catfish (Ictalurus).
Based on our GIS analysis, D. brimleyorum is most likely to be found within shortleaf pine-oak mixed forests with elevations of 152-666 m, and slopes ranging from 1 to 21%. D. brimleyorum was associated with the following soil mapping units: Carnasaw-Goldston association, Carnasaw-Octavia complex, Ceda-Rubble land complex, Clebit-Carnasaw-Pirum complex, Goldston-Carnasaw-Sacul association, Kenn-Ceda complex, and Octavia-Carnasaw complex. These soils tend to be well drained, shallow, and stony with relatively low fertility and are associated with steep upland areas (16,17). Using the results of the above analyses, we refined our predictive habitat criteria and used the GIS to produce a set of maps that estimate the occurrence of suitable habitat for the species. Our GIS analysis estimated the percentage of the total area of each county that was suitable habitat for D. brimleyorum: McCurtain, 10.7%; Pushmataha, 5.6%; LeFlore, 4.9%; Latimer, 1.8%; and Choctaw, 0.9%.
In Oklahoma, the Ouachita Dusky Salamander can be locally common under certain conditions. Although the species may attain relatively high densities in and around seeps, springs, and high gradient streams, we believe that it is sensitive to certain disturbances that could disrupt the hydrology of these habitats. We found the most productive sites within the Ouachita National Forest, places protected from intensive agriculture and silviculture. We found D. brimleyorum almost exclusively in upland areas that were geographically isolated by expanses of lowland habitat and river flood plains.
Our GIS analysis produced maps that predict the occurrence of D. brimleyorum in southeastern Oklahoma, but there are limitations to this methodology. Although the land use/cover categories produced by GIS were coarse, these categories possessed useful predictive power when linked to ancillary information such as soils and topography. The habitat may be more extensive than the metapopulation. That is, even though the correct set of habitat conditions may exist at a site, geographic isolation of this relatively immobile salamander may result in the local absence of a population, or the habitat patch may be too small or patches too fragmented to sustain a viable population. Extant populations should be monitored to insure minimal disturbance and their long-term viability.
This project was funded under project E-24 of the Federal Aid Endangered Species Program of the Oklahoma State Department of Wildlife Conservation, with additional support from the Oklahoma Cooperative Fish and Wildlife Research Unit, Oklahoma State University, Oklahoma Department of Wildlife Conservation, U.S. Biological Resources Division, and Wildlife Management Institute, cooperating. We thank P. Shipman, J. Yoshioka, M. Crosswhite, M. Tiernan, M. Howery, J. Evans, and C. Fincher for their assistance with data collection.
1. Stejneger L. A new salamander from Arkansas with notes on Ambystoma annulatum. Proc US Nat Mus 1895; 17(1023):597-599.
2. Conant R, Collins JT. A field guide to reptiles and amphibians of eastern/central North America. Boston (MA): Houghton Mifflin Co; 1991. 450 p.
3. Burt CE. Further records of the ecology and distribution of the amphibians and reptiles in the middle west. Amer Midl Nat 1935 Jan;16(1):311-336.
4. Grobman AB. The distribution of the races of Desmognathus fuscus in the southern states. Nat Hist Misc (Chic.) 1950 Oct;70:1-8.
5. Means DB. The status of Desmognathus brimleyorum Stejneger and analysis of the genus Desmognathus. Flor Bull Fla State Mus Biol Sci 1974;18(1):1-100.
6. Karlin AA, Guttman SI, Means DB. Population structure in the Ouachita Dusky Salamander, Desmognathus brimleyorum. Southwest Nat 1993 Mar-38(1):36-42.
7. Heatwole H, Lim K. Relation of substrate moisture to the absorption and loss of water by the salamander . Ecology 1961;42(4):814-819.
8. Bury RB. Differences in amphibian populations in logged and old-growth redwood forests. Northwest Sci 1983;57:167-178.
9. Ash AN. Disappearance of salamanders from clearcut plots. J Elisha Mitchell Sci Soc 1988;104:116-122.
10. Pechman JK, Scott DE, Semlitsch RD, Caldwell JP, Vitt JL, Gibbons JW. Declining amphibian populations: the problem of separating human impacts from natural fluctuations. Science 1991 Aug;253(1):892-895.
11. Matlack G. Plant demography, land-use history, and the commercial use of forests. Conserv Biol 1994 Mar;8(1):298-299.
12. Pough FH, Smith EM, Rhodes DH, Collazo A. The abundance of salamanders in forest stands with different histories of disturbance. For Ecol Manag 1987 July; 20 (1-2):1-9.
13. Heatwole H. Environmental factors influencing the local distribution and abundance of the salamander Plethodon cinereus. Ecology 1962;43(3):460-472.
14. Raymond LR, Hardy LM. Effects of a clearcut on a population of the mole salamander, Ambystoma talpoideum, in an adjacent unaltered forest. J Herpetol 1991 Dec;25(4):509-512.
15. Bratton SP. Logging and fragmentation of broadleaved deciduous forests: are we asking the right ecological questions? Conserv Biol 1994 Mar;8(1):295-297.
16. Abernathy EJ, Olszewski KM, Peters R. Soil survey of LeFlore County, Oklahoma. Washington (DC): National Cooperative Soil Survey, US Department of Agriculture; 1983. 99p. Available from US Government Printing Office, Washington DC.
17. Reasoner RC. Soil survey of McCurtain County, Oklahoma. Washington (DC): US Department of Agriculture and Soil Conservation Service; 1974. 211p. Available from US Government Printing Office, Washington DC.
Received: 1998 Feb 25; Accepted: 1998 Jul 08.