[Question]: Code review for growth

[Rick-Methot-NOAA]

work on issue #155 exposed a few questionable code lines and possible discrepancies between the various approaches to growth. A code review of get_growth2 and get_growth3 seems warranted.

• With Linf_decay == -999, the plus group growth uses a decay rate of 0.2 and goes out to 1 * nages. With a positive value for Linf_decay, SS3 uses that value and goes out to 3 * nages.
• clarify usage of Linf_decay and its effect on updating size in the plus group in time series; add option -997.
• find that the value of t was not properly updated before calling growth3, which caused incorrect values for size in the plus group
• document where season duration (VBK_seas) is changed and test by comparing annual, to two 6-month season, to season as year with 6 month season. Do that test for vonB and Richards
• run models with null param values to match vonB
• find small discrepancy in plus group with seas-as-years, after year with time-vary Linf and only for growth1 model.
• demonstrate reasonable range of param values
• clean-up and improve echoinputs in biofxn routine
• consider making Gompertz growth a separate function
• Add FATAL warning if AFIX2 input value is > nages. Otherwise, incorrect Linf values are calculated and used.

[Rick-Methot-NOAA]

@e-perl-NOAA
We'll need a set of test control files for each of:
1 # GrowthModel: 1=vonBert with L1&L2; 2=Richards with L1&L2; 3=age_specific_K_incr; 4=age_specific_K_decr; 5=age_specific_K_each; 8=growth cessation

The internal code for 3, 4, 5 is mostly the same. So most of the testing can just be for:
1, 2, 3, 8.

From the manual, Arni's work says:
As special cases of the Richards growth model, b=1 is von Bertalanffy growth and b near 0 is Gompertz growth. To use a Gompertz growth curve, the b parameter can be fixed at a small value such as 0.0001. In SS3 input files, b =s labelled as "Richards".

After creating the above growth cases, then we want to run with:

1. one 12 month season per year
2. two 6 month seasons per year
3. years as seasons with 1 six month season

Note that this code seems superfluous and never implemented:

ss3-source-code/SS_readcontrol_330.tpl

Lines 942 to 947 in 9e4a8f8

	else if (Grow_type == 6) // not implemented
	{
	N_growparms = 2; // for the two CV parameters
	k1 = N_GP * gender; // for reading empirical length_at_age
	}