Refining acoustic estimation of fish biomass in the ocean and a lake using broadband sampling and target strength modelling

Abstract

Acoustic technology has been widely accepted as one of the most efficient methods for underwater biological investigation. The paucity of information on the target strength of fish can curb the understanding of their species composition, biomass and distribution, during acoustic studies. In addition, the limitations, regarding the resolution, of traditional continuous wave (CW) echosounders also restricted the acoustic species classification. In this thesis, the more recent broadband (or frequency modulated, FM) echosounders, which offer better range resolution and signal-to-noise ratio (SNR) than traditional echosounders, were used in research in Lake Victoria, East Africa, and the ocean's mesopelagic zone (200-1000 m depth) in order to tackle several key challenges in these fields.

This PhD project uses a combination of broadband echosounding, theoretical modelling, and horizontal acoustic beaming techniques to address the key challenges surrounding fish biomass estimation in different environments. The main themes of the thesis are: 1) developing a Kirchhoff-ray mode (KRM) TS model for dagaa (Rastrineobola argentea), a key species in Lake Victoria, East Africa, to improve its lake-wide biomass estimates, 2) using a horizontal beaming method to investigate the fish distribution and biomass in shallow water (<10 m depth) of Lake Victoria, to aid the lake-wide fisheries management in Lake Victoria, and 3) improving knowledge of the biomass, species composition, and distribution pattern of fish in mesopelagic layers, and aiding in the development of a lowered echosounder suitable for investigating mesopelagic layers.

Key findings from the TS modelling (Chapter 2) indicate that the swimbladder (which comprises 2.6 to 8.2% of body volume) of dagaa accounts for 65% to 90% of the total backscattering intensity at 120 kHz. The KRM model-informed relationship will lead to a substantial reduction (by ca. 30%) in lake-wide estimated biomass. Shallow water (<10 m depth) surveys in Lake Victoria in the Tanzanian (Chapter 3 and Chapter 4) sector suggest that from 2019 to 2023 fish biomass in these shallow margins ranged from 0.15 to 0.36 million tonnes (MT), which equates to between 9.1% and 22.6% of the fish biomass in the Tanzanian open lake (based on the 2020 lake-wide acoustic survey, that observed fish biomass of 1.6 MT). Avoidance behaviour of fish was observed and can result in a 41.5% reduction in the fish density estimation. GAMM (generalized additive mixed effect model) showed the highest fish densities in the major rainy season (February to April). During the dry season (June to August), shallow-water fish density was lower and more variable between regions. Chapter 5 revealed that mesopelagic populations in Mediterranean waters were predominantly comprised of numerous small fish with low individual biomasses. This differs from the situation in Atlantic waters, which exhibit higher biomass but lower biodiversity. Further, estimated fish abundance from trawl nets, ship-borne echosounders, and DSV (Deep-water Sonar and Visual sampler) indicated that DSV reported higher fish density compared to ship-borne echosounders and trawl nets, by factors of up to 3.04 and 20.85, respectively. This suggests that DSV may be less prone to avoidance and sampling biases than ship-borne echosounders and trawl nets, thereby advocating for the use of DSV in surveying the mesopelagic zone.

The TS modelling and broadband acoustic sampling strategies derived in this PhD serve to advance the field of active acoustic observation and enhance our understanding of underwater biological resources. The consequent results will aid both fisheries and conservation management, which now requires more holistic investigation methods for monitoring and assessing the ecosystem.