age-length distributions

[Richard Methot - NOAA Federal]

Does SS output a Length-Age Key (i.e. the distribution of age conditioned on length)? Or do I need to make a dummy age-conditioned on length for each time period (and fishery if there is age based selectivity) to get the values?

 This would allow me to create catch-at-age data from the observed length compositions to use as a diagnostic.

[Richard Methot - NOAA Federal]

The age-at-length distribution depends upon the recruitment and F time series, so cannot be output as a generic time-invariant quantity.

You can get output of the catch-at-age distribution by creating a dummy observation, then the output will contain the expected value for that observation.  However, it would be based on the internally modeled age-length key and not the observed length composition.  It would be consistent with the expected length comp for that fleet x time.

Rick

[Mark Maunder]

Thanks Rick,

 

If I make a dummy age data set conditioned on a single length bin, that should give me the age distribution for that length for that time period. If there is no age-specific selectivity, then this should be the same for each fishery. I could then repeat it for each length bin and each time period to give me a time specific age-at-length key. This could then be applied to each length composition data set to provide an estimate of catch at age (conditioned on everything in the  model such as recruitment strength as you mentioned). At this stage I am not sure what it will tell me about model misspecification, but it may be easier to interpret. I have an idea for a depletion based cohort analysis diagnostic that I want to try. Another possibility is comparing the SS predicted catch at age not conditioned on length to the length-comp converted into catch-at-age in the way I am suggesting may provide some insights into age-specific selectivity (movement), cohort targeting, or age-specific M.

 

Regards,

 

Mark   

--
You received this message because you are subscribed to the Google Groups "SS3 - Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ss3-forum+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ss3-forum/d593c6ce-0ba3-4038-83b0-313db9999edfn%40googlegroups.com.

[Richard Methot - NOAA Federal]

That should work.

Another possibility might be to use the length-at-age matrix that is already in the output and the catch-at-age by fleet and time to expand that catch-at-age to catch at length x age x fleet x time.

Rick