mesmer.create_emulations.create_emus_lt

mesmer.create_emulations.create_emus_lt(params_lt, preds_lt, cfg, concat_h_f=False, save_emus=True)

Create local trends (i.e., forced response) emulations for given parameter set and predictors.

Parameters
  • params_lt (dict) – dictionary with the trained local trend parameters

    • [“targs”] (emulated variables, str)

    • [“esm”] (Earth System Model, str)

    • [“method”] (applied method, str)

    • [“method_each_gp_sep”] (states if method is applied to each grid point separately, bool)

    • [“preds”] (predictors, list of strs)

    • [“scenarios”] (emission scenarios used for training, list of strs)

    • [xx] (additional params depend on method employed)

    • [“full_model_contains_lv”] (whether the full model contains part of the local variability module, bool)

  • preds_lt (dict) – nested dictionary of predictors for local trends with keys

    • [pred][scen] (1d/2d arrays (time)/(run, time) of predictor for specific scenario)

  • cfg (module) – config file containing metadata

  • concat_h_f (bool, optional) – determines if historical and future time period is concatenated into a single emulation or not, default = False (must be set to False if no historical data provided)

  • save_emus (bool, optional) – determines if parameters are saved or not, default = True

Returns

emus_lt (dict) – local trend emulations nested dictionary with keys

  • [scen][“targ”] (2d array (time, gp) of local trend emulations)

Notes

  • Assumptions:
    • same predictors for each target

    • if historical time period is included in predictors, it has its own dictionary key

    • if historical time period was included in training, it has its own scenario

    • either historical period is included for every scenario or for no scenario

  • Potential TODO:
    • evaluate if really need / want concat_h_f or if I want output to be determined by shape predictors