This page contains information regarding the data used for the extreme value analyses, the available precipitation durations, the range of return periods provided, and the precipitation intensity diagrams.
Good to know
The extreme value analyses on the web platform are based on two time periods.
The choice of the appropriate time period depends on the user’s needs:
- Standard Period: For users who need to evaluate the extremeness of a precipitation event or compare the results at different stations (e.g. annual reports on the return period of events in the vicinity of a barrage, reports on precipitation events with impacts at regional or national level, expert opinions, claims), the web platform provides return values and return periods estimated for the standard period. A common time period is necessary for comparing return periods of events taking place at different times or different locations, regardless whether they took place during the standard period. It is also necessary for comparing return values at different stations. The standard period has a length of 50 years, and is shifted forward in time by 5 years every 10 years.
- Longest Period: For users who need a return value at a specific location or from a specific data series, the web platform provides return values estimated for the longest possible period with available data.
Estimating return levels with respect to different time periods inevitably leads to different results. In addition, both the return values and the goodness-of-fit vary from standard period to standard period in case (a), and from year to year in case (b) due to the annual update. At the Lugano station (LUG), for example, the 100-year return value for 1-day precipitation events is 5 mm lower if the standard period is used instead of the longest possible period, and at the Schaffhausen station (SHA) it is 20 mm higher. These differences are not statistically significant, however. To ensure consistency, MeteoSwiss will use the results estimated for the standard period to put precipitation events into context.
Different analysis periods are only available for (multi-)day precipitation events. For sub-daily precipitation, the time period used is prolonged by one year every year until it covers 40 years and can be considered as a standard period. Until then, “Standard Period” and “Longest Period” will display the same results for sub-daily precipitation events.
As part of the Digihom project, daily precipitation sums for selected stations were digitised and quality-checked. These are used in the extreme value analyses based on the longest possible period. On the other hand, no digitization was carried out for sub-daily precipitation measurements (e.g. records on graph paper from integrating rain gauges).
The time series of 10-minute precipitation data have now been quality checked over the entire period of analysis. As of 2018, the extreme value analyses for 10, 20 and 30 minute sums are therefore available on the web platform.
Extreme value analyses for 30-day totals will not be made available on the web platform, as the time period (number of years available) is not sufficiently long for a reliable estimation. Evaluating the extremeness of a 30-day event can be done by listing the five to ten largest events. For such a comparison, no extreme value statistics are required, however, and other locations on the MeteoSwiss website are being examined for better suitability.
Unfortunately, MeteoSwiss does not have quality-checked precipitation measurements for durations shorter than 10 minutes (e.g. 2 and 5 minutes), and is therefore unable to provide the corresponding extreme value analyses.
The return values for return periods of up to 300 years can be either downloaded or viewed via the search tool above the return level plot. As 1,000-year return values cannot be reliably estimated for time series of 30 to 150 years, these are not provided on the platform. Return periods of more than 300 years must be prepared individually. A cost estimate for such calculations can be obtained from MeteoSwiss Customer Service (email@example.com).
The web platform does not provide Intensity Duration Frequency (IDF) curves at the daily stations, as the estimated return values for short-duration precipitation events are prone to serious errors. Instead, it provides extreme value analyses of the precipitation intensities at durations from 1 hour to 7 days at stations with sub-daily data. These can be found by clicking on the radio button “intensity diagram” under “Standard Period”.
The results of the present analyses can be expected to differ, sometimes substantially, from those published by Swiss Federal Institute for Forest, Snow and Landscape Research WSL (Zeller, J, Geiger, H. & Röthlisberger, G.: 1976-1992, Starkniederschläge des schweizerischen Alpen- und Alpenrandgebietes, Eidgenössische Anstalt für das forstliche Versuchswesen, Birmensdorf) [ZGR]. A comparison has been carried out at a selection of stations, and will be added in the form of a new chapter to the 2nd edition of the Technical Report “MeteoSwiss extreme value analyses: User manual and documentation”.
For specific stations, a quantitative comparison of the values estimated by ZGR on the one hand, and by MeteoSwiss over the Standard Period (1966-2015) on the other hand can be ordered from the MeteoSwiss customer service. These comparisons can be obtained for precipitation durations of 10 minutes, 1 hour, 1 day, and 5 days. The available return periods are 2.33, 10, 20, 30, 50, and 100 years. The comparisons are documented in 2 page documents in pdf-format (in English). An example of such a comparison for the station Zürich /Fluntern (SMA) can be found here.
Can I search by zip code?
No. MeteoSwiss is planning, however, to provide station search via zip code in the future.
Will a spatial interpolation of results be available one day?
In the future, MeteoSwiss plans to provide extreme value analyses for catchments as well as extreme value analyses interpolated on a grid.