1. Organisms associated with lotic systems rank among the most threatened because of global change. Although translocation is being increasingly applied as a conservation strategy, most studies have focused on survival and recruitment of individuals, and few have attempted to identify how habitat attributes influence short-term settlement of animals during the critical post-release period.
2. We demonstrate the application of resource selection modelling in an information theoretic framework to identify release-site characteristics that will increase the likelihood of settlement for a fully aquatic benthic stream salamander, the Ozark hellbender (Cryptobranchus alleganiensis bishopi). We fit discrete choice models using data from 29 radio-tagged hellbenders that were translocated to two sites in the North Fork of the White River (NFWR), Missouri (U.S.A.). We defined resource availability at two spatial scales (stream reach and home range) and quantified abiotic habitat attributes at 3181 salamander locations and 6329 random available locations collected between May 2008 and August 2009.
3. At both sites and spatial scales, a single model received substantially greater support (0.96–1.00 of total model weight) than all other models, and top-ranked models were similar in form and predictive ability. At both spatial scales, selection was positively influenced by the presence of cobble-boulder substratum relative to bedrock and finer substrata. We also noted a negative interactive effect between distance to the nearest substratum particle large enough to provide cover (i.e. at least one axis ≥15 cm in length) and an increase in either a direct or relative (i.e. pool, run, and riffle) measure of water velocity.
4. Collectively, salamanders released in our study selected resources indicative of long-term benthic microhabitat stability. However, despite strong selection of cobble-boulder substratum, 8% (282 of 3181) of captive-reared hellbender locations occurred in bank crevices and root masses. Although several studies have reported the importance of near bed hydraulics in determining occurrence of stream macroinvertebrates, our findings are the first to indicate that spacing among cobble-boulder substrata may be important for hellbenders.
5. To increase the likelihood of short-term settlement of captive-reared hellbenders in the wild, we recommend prioritising release sites where the average distance between cobble-boulder particles within habitat patches is minimised. In general, average spacing among cobble and boulder substrata should be <1 m in habitat patches where mean benthic water velocity exceeds 0.1 m s−1, and <0.5 m where water velocity approaches 0.30 m s−1. Based on home range sizes of captive-reared Ozark hellbenders, the collective extent of suitable cobble-boulder habitat patches within release sites should approximate at least 10 m2 per salamander released.