climpred.classes.PerfectModelEnsemble.bootstrap

PerfectModelEnsemble.bootstrap(metric='pearson_r', comparison='m2e', sig=95, bootstrap=500, pers_sig=None)[source]

Bootstrap ensemble simulations with replacement.

Parameters:
  • metric (str, default 'pearson_r') – Metric to apply for bootstrapping.
  • comparison (str, default 'm2e') – Comparison style for bootstrapping.
  • sig (int, default 95) – Significance level for uninitialized and initialized comparison.
  • bootstrap (int, default 500) – Number of resampling iterations for bootstrapping with replacement.
  • pers_sig (int, default None) – If not None, the separate significance level for persistence.
Returns:

Dictionary of Datasets for each variable applied to with the following variables:

  • init_ci: confidence levels of init_skill.
  • uninit_ci: confidence levels of uninit_skill.
  • pers_ci: confidence levels of pers_skill.
  • p_uninit_over_init: p value of the hypothesis that the
    difference of skill between the initialized and uninitialized simulations is smaller or equal to zero based on bootstrapping with replacement.
  • p_pers_over_init: p value of the hypothesis that the
    difference of skill between the initialized and persistence simulations is smaller or equal to zero based on bootstrapping with replacement.
Reference:
  • Goddard, L., A. Kumar, A. Solomon, D. Smith, G. Boer, P. Gonzalez, V. Kharin, et al. “A Verification Framework for Interannual-to-Decadal Predictions Experiments.” Climate Dynamics 40, no. 1–2 (January 1, 2013): 245–72. https://doi.org/10/f4jjvf.