loading page

Comment on “Biases in Estimating Long-Term Recurrence Intervals of Extreme Events Due To Regionalized Sampling” by El Rafei et al. (2023)
  • Chi-Hsiang Wang,
  • John D Holmes
Chi-Hsiang Wang
Commonwealth Scientific and Industrial Research Organisation

Corresponding Author:chi-hsiang.wang@csiro.au

Author Profile
John D Holmes
JDH Consulting
Author Profile


The ‘super-station’ approach has been adopted since 1980s as a pragmatic method of improving extreme‑value predictions by grouping short-length datasets from several measurement stations to become a larger dataset to reduce uncertainties due to random sampling variation. El Rafei et al. (2023, https://doi.org/10.1029/2023GL105286) analyzed reanalysis and randomly generated wind extremes datasets and claimed that this technique can introduce unexpected biases in typical situations. We demonstrate by Monte-Carlo simulation, assuming the same number of grouped stations and data lengths used, that applying the grouping technique to samples from homogeneous datasets does not lead to biased prediction of extremes. In addition, the grouping technique effectively reduces the uncertainty and sampling errors that result from short-length datasets from individual stations of consistent meteorology.
06 Sep 2023Submitted to ESS Open Archive
11 Sep 2023Published in ESS Open Archive