Reassessing acoustic target strength to improve remote classification of Baltic Sea small pelagic fish

Abstract Remote target classification in fisheries acoustics surveys requires knowledge about the frequency response and/or target strength (TS) of key targets, features that are known to vary with species and region. In Baltic Sea hydro-acoustic surveys, no frequency response information has been available, and a single TS–Length relationship, derived from North Sea herring, is currently applied to the three main pelagic species: herring (Clupea harengus), sprat (Sprattus sprattus), and three-spined stickleback (Gasterosteus aculeatus). Here, we combine morphometric measurements, fish shape modeling and Kirchoff Ray Mode estimations to derive refined, species-specific TS–Length relationships for the three species. Our new estimates for 38 kHz are the following: TSC harengus = 21.92 log10(L)–66.55, TSS sprattus = 26.47 log10(L) -72.60, TSG aculeatus = 40.93 log10(L)–84.13. These estimates were validated using in situ TS data obtained during 61 single species trawl hauls (2019–2024). Species-specific TS–Length relationships significantly reduced estimation errors compared to the generic TS–Length relationship currently used, which was found to overestimate abundance by an average of 263%. Further, the new TS–Length relationships enabled the correct remote classification of the dominant species in 85% of validation hauls based on their observed TS distributions. These findings demonstrate that adopting species-specific TS models is essential for providing unbiased abundance measurements and facilitating the transition toward nonlethal ecosystem monitoring using acoustics.