BiomeE output contains repeated cID since splitting does not generate a new cID

[fabern]

The column cID in the BiomeE model output contains repeated values for single years.
I would expect each cID to appear only once for each year.

> rsofun::biomee_p_model_output$data[[1]]$output_annual_cohorts |> tibble() |> arrange(year)
# A tibble: 5 × 35
  cohort  year   cID   PFT layer density flayer    DBH   dDBH height   age        BA      dBA  Acrown   Aleaf
  <chr>  <dbl> <dbl> <dbl> <dbl>   <dbl>  <dbl>  <dbl>  <dbl>  <dbl> <dbl>     <dbl>    <dbl>   <dbl>   <dbl>
1 1        251    25     1     1   122.  0.206  23.7   0.429   17.5  80.9  0.0442     1.58e-3 17.3    54.6   
2 2        251    25     1     1    12.9 0.0130 16.9   0.391   14.8  80.9  0.0224     1.02e-3 10.4    32.8   
3 3        251    25     1     1  4783.  0.781   5.18  0.268    8.19 29.8  0.00211    2.13e-4  1.77    5.57  
4 4        251    25     1     2  1298.  0.0252  4.19  0.0599   7.37 29.8  0.00138   -5.15e-4  1.29    2.03  
5 5        251    25     1     2 37832.  0.0831  0.371 0.0300   2.19  2.85 0.0000108 -7.82e-9  0.0339  0.0366
# ℹ 20 more variables: nsc <dbl>, seedC <dbl>, leafC <dbl>, rootC <dbl>, sapwC <dbl>, woodC <dbl>, nsn <dbl>,
#   treeG <dbl>, fseed <dbl>, fleaf <dbl>, froot <dbl>, fwood <dbl>, GPP <dbl>, NPP <dbl>, Rauto <dbl>,
#   Nupt <dbl>, Nfix <dbl>, n_deadtrees <dbl>, c_deadtrees <dbl>, deathrate <dbl>

Below MWE indicates, that upon splitting of cohorts no new cIDs are generated.
Thus we end up with multiple times the same ID for distinct cohorts.

• check if the main culprit is in relayer_cohorts()
  • probably here it should attribute a new ccID and increase MaxCohortID, instead of copying over the old one (Should we rename this to nextFreeCcID?)
    

    rsofun/src/vegetation_biomee.mod.f90

    Line 1271 in c50d858

    newCC(i) = oldCC(idx(k))

• check what happens when vegn_mergecohorts() (can we make it so that ccID of the larger one remains?)
• in both the above functions what happens with cohort properties other than ccID
• side-quest: in below MWE why do we get extremely small numbers (1e-323) or 9999 when setting: spinup <- FALSE (Note that cohort is at 8188. One-off error?

[fabern]

Promised MWE:

MWE (for test-driven-development)

# Test gs_leuning simulation:
new_biomee_gs_leuning_output <- runread_biomee_f(
  biomee_gs_leuning_drivers,
  makecheck = TRUE,
  parallel = FALSE)
new_biomee_gs_leuning_output_cohort <- new_biomee_gs_leuning_output |>
  unnest_wider(data, simplify = FALSE) |>
  unnest(output_annual_cohorts)
new_biomee_gs_leuning_output_cohort |>
  group_by(year) |>
  summarise(cohort_IDs = paste0(cID, collapse=","),
            has_duplicated_cID = any(duplicated(cID))) |>
  tail()

# Test p_model transient simulation over 250 years (remove spinup to understand what happens):
new_biomee_p_model_drivers <- biomee_p_model_drivers
forcing_2009 <- new_biomee_p_model_drivers$forcing[[1]]
new_forcing <- forcing_2009
base_year <- lubridate::year(new_forcing$date[1])
for (it in 1:250){
  add_year_forcing <- forcing_2009
  lubridate::year(add_year_forcing$date) <- base_year + it
  new_forcing <- bind_rows(new_forcing, add_year_forcing)
}
new_biomee_p_model_drivers$forcing[[1]] <- new_forcing
new_biomee_p_model_drivers$params_siml[[1]]$spinup <- TRUE # NOTE if we set this to FALSE this gives gibberish output
new_biomee_p_model_drivers$params_siml[[1]]$spinupyears <- 0
new_biomee_p_model_drivers$params_siml[[1]]$nyeartrend  <- nrow(new_biomee_p_model_drivers$forcing[[1]])/365
new_biomee_p_model_output <- runread_biomee_f(
  new_biomee_p_model_drivers,
  makecheck = TRUE,
  parallel = FALSE)
new_biomee_p_model_output_cohort <- new_biomee_p_model_output |>
  unnest_wider(data, simplify = FALSE) |>
  select(sitename, output_annual_cohorts) |>
  unnest(output_annual_cohorts)
# new_biomee_p_model_output_cohort |> print(n=100)
# new_biomee_p_model_output_cohort |> group_by(cID) |> summarise(years = paste0(year, collapse=","))
new_biomee_p_model_output_cohort |> group_by(year) |> 
  summarise(cohort_IDs = paste0(cID, collapse=","),
            has_duplicated_cID = any(duplicated(cID))) |> print(n=50) # duplication starts in year 38
  # # A tibble: 251 × 3
  #     year cohort_IDs                      has_duplicated_cID
  #    <dbl> <chr>                           <lgl>             
  #  1     1 1                               FALSE             
  #  2     2 1                               FALSE             
  #  3     3 1                               FALSE             
  #  4     4 1                               FALSE             
  #  5     5 1                               FALSE             
  #  6     6 1                               FALSE             
  #  7     7 1                               FALSE             
  #  8     8 1                               FALSE             
  #  9     9 1                               FALSE             
  # 10    10 1,2                             FALSE             
  # 11    11 1,2                             FALSE             
  # 12    12 1,2                             FALSE             
  # 13    13 1,2,3                           FALSE             
  # 14    14 1,2,3                           FALSE             
  # 15    15 1,2,3,4                         FALSE             
  # 16    16 1,2,3,4                         FALSE             
  # 17    17 1,2,3,4,5                       FALSE             
  # 18    18 1,2,3,4,5,6                     FALSE             
  # 19    19 1,2,3,4,5,7                     FALSE             
  # 20    20 1,2,3,4,5,7,8                   FALSE             
  # 21    21 1,2,3,4,5,7,9                   FALSE             
  # 22    22 1,2,3,4,5,7,9,10                FALSE             
  # 23    23 1,2,3,4,5,7,9,11                FALSE             
  # 24    24 1,2,3,4,5,7,9,11,12             FALSE             
  # 25    25 1,2,3,5,7,9,11,13               FALSE             
  # 26    26 1,2,3,5,7,9,11,13,14            FALSE             
  # 27    27 1,2,3,5,7,9,11,13,15            FALSE             
  # 28    28 1,2,3,5,7,9,11,13,15,16         FALSE             
  # 29    29 1,2,3,5,7,9,11,13,15,17         FALSE             
  # 30    30 1,2,3,5,9,11,13,15,17,18        FALSE             
  # 31    31 1,2,3,5,9,11,13,15,17,19        FALSE             
  # 32    32 1,2,3,5,9,11,13,15,17,19,20     FALSE             
  # 33    33 1,2,3,5,9,11,13,15,17,19,21     FALSE             
  # 34    34 1,2,3,5,9,11,15,17,19,21,22     FALSE             
  # 35    35 1,2,3,5,9,11,15,17,19,21,23     FALSE             
  # 36    36 1,2,3,5,9,11,15,17,19,21,23,24  FALSE             
  # 37    37 1,2,5,9,11,15,17,19,21,23,24,25 FALSE             
  # 38    38 1,2,5,9,11,15,19,21,21,23,25    TRUE              
  # 39    39 1,2,5,11,15,15,19,21,23,25      TRUE    

new_biomee_p_model_output_cohort |> group_by(year) |> 
  summarise(cohort_IDs = paste0(cID, collapse=","),
            has_duplicated_cID = any(duplicated(cID))) |> tail()      # now only cohort 25 remains duplicated
  # # A tibble: 6 × 3
  #    year cohort_IDs        has_duplicated_cID
  #   <dbl> <chr>             <lgl>             
  # 1   246 25,25,25,25,25,25 TRUE              
  # 2   247 25,25,25,25,25    TRUE              
  # 3   248 25,25,25,25,25    TRUE              
  # 4   249 25,25,25,25,25    TRUE              
  # 5   250 25,25,25,25,25    TRUE              
  # 6   251 25,25,25,25,25    TRUE    

[fabern]

Below plot shows an animation of the DBH of distinct cohorts, labelled with their IDs. Towards the end (years ~110) we can see how cohort with cID 25 grows in the top layer and spawns subcanopy cohorts with the same cID.