Short-term rainfall characteristics dependence on the length of processed period

Characteristics of short-term rainfalls, especially their intensity and design values, are very important for the technical (e.g., sewage system) and hydrological practice. However, measurement and data processing of these characteristics are rather complicated. The rainfall intensity has been regularly measured in the Czech Republic since 1898 with a few manual pluviographs. Automatic rain gauges started to replace them in the late 1990s. Nowadays, a large part of historical pluviographs records has been digitalized. Both the digitalized pluviographs records and the automatic rain gauge measurement provide rainfall series in temporal resolution of 1 minute. Concerning the joined series of these two types of measurement, the relatively long series of rainfall intensity are available. The series longer than 40 years are available for almost 50 stations of the Czech Hydrometeorological Institute. In addition, the indirect rainfall data of high temporal resolution can be also derived from radar measurement. The radar measurement has been in operation for a relatively short period (since 2002) in the Czech Republic but it provides detailed indirect information on a spatial distribution of rainfall.

In our contribution, characteristics of short-term rainfalls (duration 10 – 360 min) for selected stations with the longest series of measurement are studied with regard to different extent of processed period. Especially, return levels of short-term rainfalls estimated for time series of different lengths and beginning are compared. The sensitivity of short-term rainfalls characteristics to the choice of processed period is analyzed in order to assess a possibility to use relatively short rainfall series (e.g., 20-year or shorter) for estimation of return levels with regard to the return period. Thus, a possibility to use return levels estimated from radar data combined with the estimates from longer station series in order to improve information about their spatial distribution is discussed.