Model diagnostics have been applied to CASAL-based assessments of toothfish populations (Ziegler et al., 2015; SC-CAMLR-XXXIV/05, Appendix D). Similar diagnostics can be applied to any integrated model. The purpose of this note is to use some of these methods to explore aspects of the assessment being developed to set catch limits for Antarctic krill in Subarea 48.1.
A "base-case" configuration of the krill model was presented to EMM in 2015 (Kinzey et al. 2015a) along with seven alternative configurations based on different data weightings. These alternative configurations were to evaluate the influence of different data sources on the model estimates. The work described here augments the previous description of the base case and differently-weighted configurations with a table of likelihood components for the configurations and additional diagnostic methods for the base case. An additional diagnostic based on using simulated data as a "self-test" of model consistency for the estimates of derived parameters will also be provided.