Does not show complete result - Public Forums - Stock Synthesis

I am not sure, but it looks like you are attemtping to use mean weight at age data in the data file, which should go to the wtatage.ss file instead. However, I also notice that the mean weight at age data are weird (all last age classes have weight of 20 compared to something between 0 and 1 for the others).

Max

Dear

I have made some modifications to the files and now the model runs and converges without major issue, fitting the data reasonably (even if I dont even know which stock is this:))

Major changes

1. The starter file was modified to get the full results, I made also some small changes to it so you need to verify is what you intend to do here. Anyhow, at least now, when plotting, you can see the model results in full version.

2. Your model is an age based model (no lenght info used), so you dont need to estimate growth.

3. R0 was set too low, I have changed it. Also, generally it is hard to estimate steepness, so best for now to fix it to a reasonable value but you might profile it together with sigmaR.

4. I have used the weight at age option instead of adding weight at age data in the data file, I found it more easy to handle but up to you. The weight at age file I have created contains your data but the fecundity slot (-2) needs to be amended, for now is only the weight at age but it should be weight at age x maturity.

5. I have also changed selectivity to random walk (option 17). You seems to have rather high quality age data so option 17 is generally superior in this situation. However, you might want to estimate not all parameters although they seem to be all rather well defined in the model. Or of course, you might revert to spline.

6. I have changed to the float option for the survey CPUE so that the model has one less parameter to estimate without impacting the model results (did some trial).

7. I havent changed recdev part so this needs attention as well. 

Cheers

Max

Also, I have not changed the forecast file for now but this might also need some modifications although I would need to know what you intend to do with the model in order to do that properly.

Cheers

Max

Great response Max.  Let me just add some explanation regarding body weight-at-age.

Max got the model working by setting it up like many age-structured integrated models.  This approach reads the body weight-at-age as empirical inputs  from the file wtatage.ss.  SS knows to read this file by setting this flag at the top of the control file:

1  # 0 means do not read wtatage.ss; 1 means read and use wtatage.ss and also read and use growth parameters
When this flag is set, SS ignores body weight information generated from the growth curve.  As Max said, selectivity should be specified as age-base; length-selectivity will probably produce spurious results.  Alas, we have not yet automated a change to the report settings such that growth based information is not displayed in the output, nor have we removed the reading of growth parameters from the input files.  That's on the wishlist.

0  # 0 means do not read wtatage.ss; 1 means read and use wtatage.ss and also read and use growth parameters
When the flag is not set, then is not read and SS will generate body weight-at-age information for the population from the growth parameters and, for the fleets, taking into account length selectivity.

This second approach requires accurate growth parameters.  These can be determined external to SS, or estimated in SS if informative data are provided.  Some types of informative data are mean body length-at-age, mean body weight-at-age, conditional age-at-length (which is the age composition for a specified range of lengths, e.g. each row of the age-length key).  Modal progression in length composition data is also somewhat informative.  On the wishlist is growth increment data from tag-recapture studies.

I hope this helps add to the good info Max provided..

Rick

Here are a couple of suggestions:

First is to check the units in your files relative to SS conventions.  SS expects body wt in kg, catch weight in tonnes, numbers of fish in 1000s.

If those are correct, then it is possible that the overall trends in your data are not informative regarding the scale of the population.  One investigative approach is to profile over a range of fixed values for the model parameter ln_B0, which sets the scale of recruitment.

Rick

I also suspect that the units of the landings are most likely expressed in kg and not in tonnes. The WAA seems more reasonable. By the way, which stock is this one?

Cheers

Max

In that case, it is really a large stock as it has catches over 4 millions ton, I have rarely read about such huge stocks except the classical Peruvian anchoveta and Alaskan Pollock. Just by curiosity, could you tell us which stock is this? 

I am note sure which is the issue but surely you should include the other 40% of the catches from the other fleets if you want to have a reasonable assessment of the stock.

Max

Sorry for being imprecise in my message.  I was referring to SR_LN(R0) parameter as a good candidate for profiling.

Here is another possible cause for the model result to have estimates of stock biomass that are much higher than the catch:  If the selectivity for a fleet is dome-shaped or strongly shifted to older ages, the catch will be coming from a subset of the ages and the high biomass could be in other ages. 

Rick

I have looked over your input files and find that they have some complexities that are unnecessary and will cause conflicts.  In particular, the use of length data and length-selectivity in combination with use of empirical weight-at-age input is not advised.  Also, you commented that you are only using the catch data from one fleet, but the input files have catch for four fleets.  It will take me a couple of days to get back to working on your request, but I will try to get in configured to work properly.   In the meantime, I suggest that you look at the empirical wtatage example in our examples folder:  https://vlab.noaa.gov/web/stock-synthesis/document-library/-/document_library/0LmuycloZeIt/view/11684751?_com_liferay_document_library_web_portlet_DLPortlet_INSTANCE_0LmuycloZeIt_redirect=https%3A%2F%2Fvlab.noaa.gov%3A443%2Fweb%2Fstock-synthesis%2Fdocument-library%2F-%2Fdocument_library%2F0LmuycloZeIt%2Fview%2F11684741