Seasonal detection rates of river otters (Lontra canadensis) using bridge-site and random-site surveys

Abstract

Randomization of survey sites is generally desired because of its unbiased approach, but is often abandoned because of logistical constraints. This is true for river otters ( Lontra canadensis (Schreber, 1777)), with bridges commonly determining survey locations. We conducted seasonal sign surveys for river otters on two rivers in southern Missouri, USA, using randomized survey points and fixed bridge-crossing points in 2001–2003. Otter sign was more likely to be detected at randomized sites than at bridge sites in summer (P < 0.0001), with sign being detected on 68% of visits to random sites (n = 348) and on 40% of visits to bridge sites (n = 60). Scat abundance was higher (P = 0.0001) at random sites (8.82 ± 0.6, mean ± SE) than at bridge sites (3.96 ± 1.0) during the summer. Similar but nonsignificant trends were found during the winter. Detection probabilities were significantly higher at random sites than at bridge sites in both seasons. Our results indicate that surveys of bridge sites for river otters may yield inaccurate results for distribution and relative abundance, particularly if conducted during the summer.