Indian Ocean Tuna Commission yellowfin tuna stock synthesis assessment grid

The Tuna Regional Fisheries Management Organisations commonly develop uncertainty grids to condition models in integrated stock assessments to account for uncertainties in parameters that cannot be estimated from the data. For example, by constructing multidimensional grids that consist of different plausible combinations of assumptions, fixed parameter values, and data sets. It is not always clear, however, whether this is intended as an uncertainty analysis or a sensitivity analysis. The choice of assessment model scenarios and ways of estimating uncertainty has an impact on the risk of exceeding the limit and missing target reference points. Therefore, to better understand the impact of uncertainty on stock assessment advice and the risk of failing to meet conservation and sustainability objectives, we used the uncertainty grid developed by the Indian Ocean Tuna Commission (IOTC) for albacore tuna (Thunnus alalunga), a full factorial design with 1,440 model configurations (IOTC, 2019..., It was compiled by the Indian Ocean Tuna Commission. There are two parts i) a grid of input datasets comprising catch, catch per unit effort, size data, and biological parameters with associated control file; and ii) the fits to the data using stock synthesis.
Using i) a variety of runs can be conducted, i.e., historical fits to the data, retrospective analyses, hindcasts, projections, bootstraps, and MCMC. Under ii) summaries as used in \"Empirical Validation and Risk Equivalence: A Pathway to Resilient Fisheries Management\" Kell et al., 2024 are provided as R data.frames., , # Indian Ocean Tuna Commission yellowfin tuna stock synthesis assessment grid

[https://doi.org/10.5061/dryad.v6wwpzh3s](https://doi.org/10.5061/dryad.v6wwpzh3s)

## Description of the data and file structure

The analysis was based on running 1440 different stock assessment scenarios. This requires SS3 input and control files for each scenario. The outputs from SS were then processed into summary R data.frames. There are therefore 4 sets of files i) SS3 inputs and control files, ii) summary R data.frames, iii) R scripts for running the analyses, i.e., historical assessments and hindcasts, including as array job on a cluster, and iv) post-processing, i.e., combining outputs into common files.

## i) Stock synthesis inputs

The stock synthesis control and data files are provided along with the outputs

## ii) Summary data sets

### Time series

ts.RData

### Summary of reference points and likelihood components

smry.RData

### Prediction skill, i.e., MASE

mase.RData

### Retrospective ...