grazing_function should be a string that is converted to int type

[mnlevy1981]

Currently, changing the grazing function for a particular grazer requires knowing that grazing_function = 1 refers to Michaelis-Menten and grazing_function = 2 refers to a sigmoidal function. Instead, users should specify a grazing function via a string (michaelis or sigmoidal) that is then converted to an integer to avoid string comparisons inside the computation stage. See treatment of caco3_bury_thres_opt in marbl_settings_mod.F90.

[mangelesg]

Im trying to understand the MARBL code, but I have a couple of questions about the subroutine compute_grazing.

1- Why does the subroutine loop over the vertical level k? Could this subroutine be vectorized?

2- In the part of the code shown below, Im not sure if the loop is supposed to sum over the Pprime of all the three autotrophs groups. For example for prey =1 and predator =1, it has auto_ind_cnt =1 and auto_ind2 =1. So if I understand correctly, there is no sum over all the autotrophs but just the corresponding one to the prey?
!-----------------------------------------------------------------------
! compute sum of carbon in the grazee class, both autotrophs and zoop
!-----------------------------------------------------------------------
work1 = c0 ! biomass in prey class prey_ind
do auto_ind2 = 1, grazing_relationship_settings(prey_ind, pred_ind)%auto_ind_cnt
auto_ind = grazing_relationship_settings(prey_ind, pred_ind)%auto_ind(auto_ind2)
work1 = work1 + Pprime(auto_ind,k)
end do

I would appreciate any help in this matter

[mnlevy1981]

1. We loop over k as part of the clean-up addressing migrate k loop inside of subroutines called from marbl_set_interior_forcing #53; we plan to investigate vectorization once MARBL v1.0.0 has been released (it is unclear if we will see much speed-up with vectorization in models that let the number of active levels vary from column to column)
2. This is a confusing portion of the code that will be cleaned up when we address grazing type should be a subset of zooplankton type #320. auto_ind2 is looping over all autotrophs that are preyed upon by a given grazer, but the way the code is currently written that will typically correspond to a single autotroph (because of the prey_ind dimension of grazing_relationship_settings(prey_ind, pred_ind)). But the general idea is that pred_ind grazes on auto_ind_cnt different autotrophs in a manner determined by the parameters in the grazing relationship datatype, and auto_ind is the index of autotroph_settings for each different prey. work1 is used to determine graze_rate for the given predator-prey pair, and is the sum of Pprime at a given level across all the autotrophs being grazed in a given grazing relationship. Note that each "grazing relationship" is defined by a unique entry in grazing_relationship_settings