2019 STAR Panel: Estimating length-based retention

[chantelwetzel-noaa]

Provide estimates of retention external to the model and evaluate if the approach of doing it within the model (and then fixing the values because of discard length composition data affecting model results other than the shape of the retention curve).

This was primarily an issue in the 2019 model since all selectivities were age-based, and there were only length data on discarded fish. This issue is connected to issue #4.