Christopher M Brierley, Michael Simpson, Indrani Roy , Christopher Nankervis

The return periods for extreme events are estimated from observational datasets. Often those datasets are relatively short in comparison to timescales of natural climate variability; potentially introducing a systematic bias into the extreme estimates. Here we combine observations with global climate models to show that this bias is statistically insignificant for the case of extreme UK-wide rainfall estimates. This is unlikely to hold for other locations and spatial scales, yet the methodology we have developed provides a simple approach to quantify the bias for other cases. [Funded by NERC NE/P000703/1; Determining Extreme Values for the Insurance Sector]