Estimating Occupancy and Detection Probability of Juvenile Bull Trout Using Backpack Electrofishing Gear in a West-Central Alberta Watershed

Occupancy modeling is suited to quantitative assessment of native trout distribution at multiple scales. Threatened in Alberta, bull trout are secretive, patchily distributed and often occur in low numbers.  We used models to estimate occupancy of juvenile bull trout (≤150 mm fork length) in a west-central Alberta watershed.  Based on a backpack electrofishing survey of 92 sites, we assessed the relative importance of stream habitat on detection probability (p) and potential for false absences to bias occupancy estimates.  Median distance to first bull trout detection was 16 m (range 0 – 289 m).  Models including ambient water conductivity as a covariate of detection probability were most supported with an 85 µS∙cm-1 increase resulting in over a tenfold increase in detection. Conditional detection probability approached 95% in the first 200 m of effort in streams with a conductivity around 200 µS·cm-1.  The potential for false absences in our study area was relatively low.  Modeled site (θ= 0.53; SE= 0.13) and patch-scale (Ψ= 0.47; SE= 0.12) occupancy closely corresponded to naive (i.e., assuming p = 1) estimates (0.47 and 0.43 respectively).  Our results highlight the efficiencies of an occupancy modeling approach when assessing fish distribution but consideration of model assumptions is necessary.