Development of a Sonar System to Enumerate Sockeye Salmon Smolt in Bristol Bay, Alaska

Out-migrating sockeye salmon (Oncorhynchus nerka) smolts have been enumerated with nets, traps, photo counters, and/or acoustics in Bristol Bay Rivers on and off since 1955.  Smolt abundance estimates are used for pre–season forecasting of adult returns and for refining river-specific spawning escapement goals.  The Alaska Department of Fish and Game operated a sonar developed by the Bendix corporation on several Bristol Bay rivers from 1970 – 2001.  Smolt abundance estimates were used with other information to forecast adult salmon returns for nearly two decades.  Changes to sonar equipment in the 1980s, some high profile forecast errors in the mid 1990s, and subsequent funding pressures led to the termination of the sonar-based smolt projects by 2001.    In 2007 we began development of a modern up-looking digital sonar system to estimate out-migration of sockeye smolts, and tested the system in the Kvichak River (2008-2010), and the Ugashik River (2010).  The system is based on a design of independent pods, each with a digital transceiver and transducer deployed in series and connected to a shore-based power supply with network attached storage drives.  Very high passage rates (millions per day), near-surface orientation, daytime schooling behavior, and the cross-channel distributions of sockeye smolts were important considerations in our system design.  To quantify measurement error we operated two independent sonar systems 1.5 to 3 km apart and prepared two annual abundance estimates on the Kvichak River..  The acoustic data were echo integrated and scaled by the in situ sigma obtained from a sample of single targets.  Water velocity was measured daily and used to convert the acoustic smolt density estimates to estimates of smolt passage (number of smolts per hour and day).  Smolt passage in the unsampled area between pods was estimated interpolating between depth-and pod-specific passage rates.  In 2011 we incorporated a shore-based side-looking narrow-beam transducer to compare cross-channel distribution of smolts with that estimated from the up-looking pods. Furthermore, we have added a split-beam transducer to one up-looking pod at each site to compare split beam sigma estimates to EMS deconvolution values derived from the single beam pods.