Development of a General Jitter Function

Commentary

1 Palau test (2023-08-21)

Initial test and result

At SC19, John asked Arni if the current jitter() function can be used to jitter ALB par files. The initial test involved the jitter.R file that was used for the YFT 2023 assessment and a 09.par file from a single-region ALB model, provided by John:

• jitter.R
• 09.par

The result from the test script,

• 01_yft_2023_assessment.R

was an error message:

Error in if (flagval(par, -Selsfish[1], 48)$value == 1) { :
  argument is of length zero

Small improvement while in Palau

The error turned out to be caused by the line

uniqueSels <- max(flagval(par,-1:-nFish,24)$value)

which assumes that some fishery selectivities are grouped, but that's not the case for the single-region ALB model. Therefore, a new and slightly improved jitter() function was developed that handles the case when no fishery selectivities are grouped:

• jitter.R

The result from the test script,

• 02_palau.R

was the successful creation of 20 jittered par files. However, the jittered par files turned out to be truncated, containing too few parameters. We were of course hoping/expecting that the jittered par files would have the same number of parameters as the original par file with slightly different values.

That is where the development of a general jitter function was put on hold during SC19 in Palau (2023-08-21) as we had to turn our attention to SC19 tasks related to the BET and YFT assessments.