| Summary: | It is known that many fisheries data sets contain several different sources of error. It is likely that models with different structures will, in general, be affected in different ways by these errors. It is therefore important to examine what those differences are, both in order to select models of “appropriate” complexity, and to interpret discrepancies in situations where multiple models are run on the same datasets. This paper examines the Northeast Arctic Cod (Gadus morhua). A series of experiments have been conducted introducing additional, known, errors into some of the survey data sets. These errors have been selected in order to mimic problems known to occur in practice. The modified data sets have then been used as inputs into two different models. The two models are currently both run as part of the assessment process for Northeast Arctic Cod. One is a relatively simple, XSA age-structured model, and one a more complex biologically-detailed age-length based model constructed using Gadget. The effects on the model results of these known forced errors are then compared.
|
|---|