The effect of the duration of data on the return value estimates was investigated and the results suggest that using short periods of data(e.g.lessthan7years)may yield inaccurate results.
Using data with a duration of more than 10 years is likely to produce
more reliable return value estimates.
100-yr return values were estimated by fitting a GPD to all excesses above a high threshold, selected as the 95% quantile.
These values were compared with previously published estimates