In this paper, we describe developments towards a set of standard diagnostic principles and tools used to characterise assessment models. CCAMLR has yet to develop a set of recommended statistical diagnostics tools to support their evaluations. While there are a number of standard visual and statistical diagnostics available for integrated assessments, their sufficiency to evaluate if a model is well specified and fits the data adequately needs to be considered. This paper provides a platform for discussions about describing the characteristics and diagnostic information needs that are common to integrated stock assessments used by CCAMLR. We further identify evaluations that could be undertaken when performing an assessment, as well as discuss statistical tools that will enable a more systematic evaluation of stock assessment models in the future.
Dr Philippe Ziegler (Австралия)
Пункт(ы) повестки дня