Estimating Trout Abundance with Cataraft-Mounted Dual-Frequency Identification Sonar: a Comparison with Drift Diving

We investigated the potential of dual-frequency identification sonar (DIDSON) deployed from a drifting cataraft for estimating abundance in rivers of Brown Trout Salmo trutta larger than 20 cm. We compared triplicate trout density estimates made by DIDSON with drift-diving density estimates in three reaches of a clear-water river in New Zealand. DIDSON density estimates were much lower (∼22% of drift-dive estimates, range = 7–33%) and less precise than drift-dive estimates (DIDSON CV = 0.13–0.47; drift diving CV = 0.15–0.17). Variation in detecting fish in the DIDSON field survey contributed substantially more (95%) to DIDSON count variability than did fish detection in the image files. Highest precision with DIDSON was achieved in the reach with the least shallow habitat and most uniform channel. Fewer person-hours were required to undertake the field component of DIDSON surveys than the drift dives (5 versus 8.3 h), but the substantial time spent on image review (3.3 h) made DIDSON surveys 34% more costly than drift dives in terms of overall effort. Despite observed shortcomings, cataraft-mounted DIDSON has utility as a noninvasive survey method for estimating abundance of large (>20-cm) fish, particularly in situations where turbidity is too high for visual counting methods to be effective. Received September 22, 2014; accepted January 8, 2015