“…Observed habitat associations collected during diel tracking were ordinated by means of PCA and diel times, mean principal component scores, and collective variation (i.e., the 68% confidence interval, or 61 SD) included in the PCA biplot to show habitat associations during the day and at night. For PCA, qualitative variables (reproductive season, cover, mesohabitat, substrate, and river) were represented by dummy variables, and quantitative variables (temperature, day number, flow, distance to cover, depth, and velocity) were z-score transformed (Krebs 1999;Williams and Bonner 2006). Univariate statistical analyses were used to assess cover, substrate, and mesohabitat preference or avoidance as well as habitat suitability (i.e., the relative use of available habitats) for current velocity and depth by comparing the observed usage with the total available habitat.…”