| Summary: | Precipitation amount and duration measured at the Hamburg Weather Mast with 1-minute time resolution during the 17‑year period 1997–2014 are used for a statistical analysis of precipitation extremes for various aggregation times (AT) from minute to month. Extremes are defined here by the 95th, 99th percentile and the absolute maximum of the PDF and are calculated for the total 17‑year period and data subsets such as same months and same hours of the day to determine annual and diurnal cycles. Mean monthly precipitation amount (duration) has a maximum in July (December) and a minimum in April (April). Monthly extremes of precipitation amount for AT = 1d, 1 h and 1 min follow the mean annual cycle only partly with their maxima also in July/August but their minima in February. Hourly extremes of precipitation amount for AT = 1 h and 1 min show no diurnal cycle in winter but a clear one in summer where the corresponding maxima all occur in the afternoon time window 14–18 CET. The 95th and 99th percentile values for the total 17‑year period follow power laws over the entire AT range from 1 min to 1 month. This does not hold for the maxima; the well-known Jennings power law underestimates the maxima in the sub-daily and particularly sub-hourly AT range. Our results can be used to estimate sub-daily (‑hourly) extremes for stations with e.g. only daily records.Extreme precipitation needs favourable large-scale conditions. We test the potential of the DWD (Deutscher Wetterdienst) weather type classification as downscaling predictor of extreme precipitation. All days (hours) exceeding the 95th (99th) percentile of 13.8 mm/d (5.2 mm/h) are inspected. Daily (hourly) extremes show up in 22 (18) of the 40 weather types. While in winter one north-westerly weather type occurs outstandingly frequent, several south-westerly weather types are equally frequent in summer. However, the probability that these weather types, when they occur, lead to extreme daily (hourly) precipitation is not higher than 16 % (19 %). The higher number of weather types involved in summer than winter extremes can be attributed to the modes of precipitation (stratiform, convective) classified from the amount-duration relationship. From composites of the leading weather types for days with and without extreme precipitation and from a separate (stratiform, convective) inspection of the large-scale flow situation, we identify overarching synoptic features which occur with extreme days but are not captured by the indices of the DWD classification. The potential of the classification as downscaling predictor of extreme precipitation is limited but may be improved by some simple extensions.
|
|---|
