/wgsam-skill-assessment

WGSAM skill assessment project

Primary LanguagePostScript

WGSAM skill assessment

Gadget models for the WGSAM multispecies skill assessment project

Implement single and multispecies Gadget models using internally consistent multispecies data with known observation error characteristics from an end-to-end ecosystem operative model based on the Norwegian and Barents Sea (NOBA) Atlantis model by Hansen et al. (2016, 2019). Gadget is one of several estimation models used in the WGSAM multispecies skill assessment project which aims to model performance testing, alternative models comparison, and performance testing of model ensembles. The objectives link to the WGSAM ToR c "Establish and apply methods to assess the skill of multispecies models intended for operational advice".

How to use this repository

  1. build an mfdb database and import the OM simulations mfdb/import_noba_simulations.R available via ms-keyrun
  2. start building single species models. Each model is built with models/xx.sppName/setup.R and fitted to data with models/xx.sppName/run.R