The use of acoustics to characterise mid-trophic levels of the Southern Ocean pelagic ecosystem

Abstract

Mid-trophic level (MTL) organisms link primary and tertiary consumers and so play a key role in pelagic open-ocean marine ecosystems. Despite their importance, our knowledge of MTL organisms is still very limited. Estimates of total biomass are uncertain, and basic ecological information, such as spatial distribution, abundance indices, and spatio-temporal variability are not yet established. Sampling difficulties posed by the vast areas of open-ocean and the biology of MTL organisms have led to initiatives to collect acoustic data opportunistically. This approach can provide a starting point for the foundation of more advanced ecosystem indicators and models, and consistent monitoring strategies to generate long time series for detecting and interpreting changes in the ecosystem properties due to fishing and climate change. A unique 7-year time series provided the opportunity to study the MTL of the pelagic open-ocean marine ecosystem in the New Zealand (NZ) sector of the Southern Ocean (SO). Thirty three transects of mainly single frequency (38 or 18 kHz) but also multi-frequency (18, 38, 70, 120 and 200 kHz) acoustic data, were collected opportunistically during the transit of toothfish fishing and research vessels between NZ and the Ross Sea from 2008 until 2014. Acoustic data was post-processed following the Integrated Marine Observing System (IMOS) protocols, with some modifications that improved signal processing. Processed acoustic measurements were used to thoroughly characterise the distribution and behaviour of the MTL of the NZ sector of the SO, and to assess the seasonal and interannual variability. This information was also used to develop predictive models for acoustic backscatter (sa) in the epi- and mesopelagic zones. Finally, sa was converted into biological density of MTL to provide estimates for the NZ sector of the SO, and evaluate whether biological density patterns correlate with those observed in backscatter. Mean sa at 38 kHz varied between years, but overall was reasonably stable, and of the same order of magnitude across all transects. Vertical distribution showed clear diurnal vertical migration patterns, and seasonal differences in concentration and behaviour with four distinctive bands of backscatter (epipelagic, transition, mesopelagic, deep mesopelagic). Large-scale horizontal distribution patterns of vertically summed backscatter were analysed and a consistent and significant decrease from north to south was detected. The deep scattering layers (DSL) detected in acoustic transects stopped north of the Ross Sea, which may be related to the temperature tolerance of DSL organisms. Satellite-derived sea surface temperature and time of the day were successfully used for developing an explanatory and predictive model for sa, a proxy for MTL organisms, in the epipelagic zone. Sea surface temperature and depth were used for developing explanatory and predictive models for sa in the mesopelagic zone. These models were tested in the NZ sector of the SO and also in an independent dataset in the Indian Ocean sector of the SO, and showed to perform well. Catch information from research voyages was used to characterise MTL species’ distribution in the NZ sector of the SO. This information and available literature on speciesspecific target strength-length relationships were used to convert sa into biological density. Results suggested that changes in backscatter with latitude reflected different species’ composition, rather than changes in biological density. This research provides the first time series of acoustic data on MTL organisms in the pelagic open-ocean marine ecosystem of the SO. The findings establish an excellent baseline for detecting and monitoring future changes in the SO ecosystem and its MTL component, and for understanding their implications, and cascade effects, for lower and higher trophic levels. Predictive models provide a tool for inferring abundance and distribution of MTL in the NZ sector and other parts of the SO. This information can assist studies of top-predator population dynamics by providing insights into prey abundance. Additionally, this information can be used for ecosystem model validation and parametrisation, and ecological studies. Because of the vast extent of the open-ocean and scarcity of observations, the development of models for explaining and predicting MTL provide a valuable tool for filling current data gaps. The continuation of the ongoing data collection program and its processing is recommended. Consistent acoustic monitoring of the MTL can contribute useful ecosystem Essential Ocean Variables to observe and understand the status of the SO.