How many of Australia's stock assessments can be conducted using stock assessment packages?

Abstract

Most of the stock assessments conducted in the USA and in New Zealand are based on packages that have been developed for generic use, are well documented, and have been tested using simulation. However, this is not the case for assessments conducted in Australia and many other countries. This paper reviews all of the model-based stock assessments for Australian fisheries to evaluate how many of these assessments could have been conducted using the publicly-available stock assessment packages used widely in the USA and New Zealand. The 76 model-based assessments reflect 37% of the 2013 catch recorded in Australia's Status for Key Australian Fish Stocks Reports (or 34% of the total catch in 2013). All but 18 (or 24 if full rather than approximate age-size-structured models need to be used) of the stock assessments could have been conducted using stock assessment packages used in the United States and New Zealand. Adoption and use of packages for more stocks in Australia should increase the likelihood that results are based on correctly-coded models whose estimation performance is widely understood, reduce the time needed to conduct assessments, and speed up the peer-review process. The availability of training, manuals, and example data sets for stock assessment packages should partially address their additional complexity. Additional benefits, in terms of numbers of assessed stocks, could occur if Australian stock assessment scientists develop a forum to collaborate and share methods. These results are applicable to many other jurisdictions that undertake stock assessments.