Include catches in weight in the likelihood

[ejardim]

It would allow to fit the model to time series with longer series of catches than abundance. Currently these fits end up with very low biomass and catches, mainly the stock and the fishery disapears, which doesn't make much sense. An alternative would be to include somekind of initial depletion rate ... not my favorite option though.

• code likelihood for total catch observation
• ensure only either total catch likelihood, or age structured likelihood is used (switch)
• sanity checks using test data

[colinpmillar]

This is a special case of a more generic idea where we can supply a list describing the ages to sum over. e.g. total catch is a sum over all ages caught, while another scenario would be if we have catch weight by age category (0-3, 4+)...

not sure how common this is though