Found some bounding issues in my explorations. These are updated within data_loading.R
[x] Turn off the survey age comp likelihood component and estimation of survey selectivity if the survey fleet mirrors a fishery fleet. Currently this only works if have one survey fleet (as it turns off all or none) but shouldn't double count the same data. If have survey parameters estimated, they dont affect model results because they mirror a fleet, so they are free parameters and tend to go to bound.
[x] Adjust Rdevs_start to be 0.18 which is square(sigma_recruit)/2 to offset bias correction and recruitment deviations are at 1
[x] set upper and lower bounds for init_abund to be symmetric. Otherwise, if N_init is used but fixed at starting conditions the values are fixed at a value other than 1 and therefore affect model results. Currently this does not affect our runs because we decay from R0 (i.e. init_abund_switch = 1).
[x] lower catchability bound from -10 to -15. Bound was being reached.
[x] Adjust bounds for selectivity. lb_sel_betaX and ub_sel_betaX are on the log scale whereas fixed values for sel_betaX are not. Therefore when estimating, we are exploring extreme spaces. Scale appropriately.
Found some bounding issues in my explorations. These are updated within data_loading.R