This is the discussion thread for planning the merge of the phydro branch into the main model.

TODO before the merge:

• publish rsofun documentation paper as is

TODO for the merge:

• fix phydro: le is not equal to aet #220
• fix phydro issues #218
• calibration: reduce redundancies in cost-functions and streamline two methods of 'GenSA' or 'BayesianTools'
• update rsofun::p_model_drivers with readRDS(file = here::here("data/p_model_drivers_newformat.rds")), as well as rsofun::p_model_drivers_vcmax25
  • recompute forcing, forcing_24h, forcing_daytime, forcing_3hrmax for the drivers
• switch handling of whc from a params_modl to a site parameter
  • e.g. in runread_pmodel_f.R that sends the multi-site data.frame to individual simualtions?
  • or alternatively remove it form par: here in R, here in C and here in F90.
• include "regression tests" into the test suites to detect alteration of results, e.g. extend current test by verifying results remain similar
• design and document a robust, user-friendly interface for the different model flags
• ...

Other suggestions:

• reduce interface complexity:
  • does run_pmodel_f_bysite() need to be an exported (and documented) function? Or will it anyway always be called from runread_pmodel_f, and we can thus keep it internal (and undocumented)?
  • remove argument parallel form the runread_pmodel_f() and runread_biomee_f() functions since it is equivalent with setting ncores = 1
  • ...