PARENT SESSION
Oral Session # 31: Reserve Design and Planning.
Presiding: J Franklin
Tuesday, August 5. 1:30 PM to 5:00 PM, SITCC Meeting Room 105.

Predictors of habitat use by herpetofauna at multiple scales at the Savannah River Site.

LaBram, Jill^*,1, Allen, Craig¹, ¹ South Carolina Cooperative Fish and Wildlife Research Unit, Clemson, SC

ABSTRACT- Understanding patterns of habitat selection at different scales is important for managing and conserving species. It may be insufficient simply to maintain a landcover for a species if required microhabitat components (e.g., amount of woody debris, ground cover, litter depth) are not provided within that landcover. We sampled herpetofauna for three seasons (fall 2001, spring 2002, and fall 2002) using pitfall-drift fence arrays, funnel traps, PVC pipes, coverboards, and visual sightings at the Savannah River Site. We measured twenty-one habitat variables in 35 sites within seven landcovers to determine which variables at what scales best predicted the occurrence of herpetofauna. We used logistic regression to build models that predicted the presence or absence of eight abundant species (two salamanders, two anurans, three lizards, and one snake). Significant models included between two and ten variables and r² values ranged from 0.13-0.67. The percent litter type was positively associated with the presence of all species. At least one landcover type was useful in predicting presence or absence of all species except Tantilla coronata. The presence of the lizards, Sceloporus undulatus and Scincella lateralis, was associated with greater percentage of ground shade. Ambystoma opacum, Gastrophryne carolinensis, and Scincella lateralis were predicted by the presence of coarse woody debris. Models that incorporate broad-scale elements such as landcover, together with appropriate fine scale variables, better predict herpetofauna occurrence than models at single scales.

Key words: habitat use, scale, herpetofauna, modeling