climpred.bootstrap.bootstrap_perfect_model(init_pm, control, metric='pearson_r', comparison='m2e', dim=None, reference=None, resample_dim='member', sig=95, iterations=500, pers_sig=None, **metric_kwargs)[source]

Wrap py:func:bootstrap_compute for perfect-model framework.

  • initialized (xr.Dataset) – prediction ensemble.

  • verif (xr.Dataset) – Verification data.

  • hist (xr.Dataset) – historical/uninitialized simulation.

  • metric (str) – metric. Defaults to "pearson_r".

  • comparison (str) – comparison. Defaults to "m2e".

  • dim (str) – dimension to apply metric over. Defaults to: ["init", "member"].

  • reference (str, list of str) – Type of reference forecasts with which to verify. One or more of ["persistence", "uninitialized", "climatology"]. If None or [], returns no p value.

  • resample_dim (str or list) – dimension to resample from. Defaults to: "member".

    • “member”: select a different set of members from initialized

    • “init”: select a different set of initializations from initialized

  • sig (int) – Significance level for uninitialized and initialized skill. Defaults to 95.

  • pers_sig (int) – Significance level for persistence skill confidence levels. Defaults to sig.

  • iterations (int) – number of resampling iterations (bootstrap with replacement). Defaults to 500.

  • ** metric_kwargs (dict) – additional keywords to be passed to metric (see the arguments required for a given metric in Metrics).



(xr.Dataset): bootstrapped results for the three different kinds of


  • initialized for the initialized hindcast initialized and

describes skill due to initialization and external forcing

  • uninitialized for the uninitialized/historical and approximates skill

from external forcing

  • persistence for the persistence forecast computed by

compute_persistence or compute_persistence_from_first_lead depending on set_options(“PerfectModel_persistence_from_initialized_lead_0”)

  • climatology

the different results:
  • skill: skill values

  • p: p value

  • low_ci and high_ci: high and low ends of confidence intervals

based on significance threshold sig


Goddard et al. [2013]