Roland Proud presenting at the Hobart MESOPP meeting, November 2017.

Roland Proud presenting at the Hobart MESOPP meeting, November 2017.

An important objective of MESOPP is to build the link between acoustic data and the mesopelagic fish component of ecosystem models. Roland Proud from St Andrews University and Nils Olav Handegard from IMR visited Rudy Kloser at CSIRO to work with ecosystem modellers and acousticians on this topic. The objective was to initiate the development of an acoustic observation model that can translate predicted biomass in the ecosystem model to corresponding predicted acoustic observations, with uncertainty.
The dominant scattering mechanisms that drives acoustic observations for the deep scattering layer was reviewed. Some organisms contribute relatively more to the total backscattered energy than others. The effect is that the acoustic energy will not be linear to biomass, and if these groups are not separated in the model, the uncertainty in the conversion using the observation model will be high.
The conversion from model variables to the corresponding simulated observations depends on how the different groups are represented in the model. If the groups in the model contain groups of similar scattering properties, the conversion can be done with reasonable uncertainty, whereas if the groups in the model contain a mix of organisms, with vastly different scattering properties, the conversion from model variables to acoustic observations is associated with major uncertainty.
During the visit there was also meetings with Martin Cox from the Australian Antarctic Division (AAD acoustics) and several ecosystem modelers representing different models, including Beth Fulton, Andrew Constable, Jessica Melbourne-Thomas and Heidi Pethybridge for Atlantis, and Julia Blanchard and Rowan Trebilco for size-based models, both models planned to be in MESOPP. The objective was to map out how the models represent the mesopelagic community and how well they translate into acoustics.
This work is ongoing and part of deliverable 3.6 and 4.2 in the MESOPP project.