Wind storms are among the most important natural hazards in our country. In December 1999, storm Lothar caused considerable damage, with costs amounting to about 1.8 million Swiss francs. Some 80 persons lost their lives across Europe, 14 of whom in Switzerland. Wind gusts reached peak speeds of 110 to 150 km/h in most lowland regions north of the Alps. In general, however, even markedly lower wind speeds may suffice to induce significant damage. In Switzerland, several damaging storms have been observed in the last few decades.
How often are extreme wind speeds likely to happen in Switzerland? In statistical terms, the frequency of an extreme wind event can be expressed as a so-called “return period”, measured in years. The return period is a measure of the expected time span between two events of a given intensity. Thus, it is a function of wind intensity, i.e. of wind speed: higher wind speeds correspond to larger return periods.
Statistical analysis methods make it possible to express wind gust speeds as a function of the return period. The relation between the two parameters can be represented graphically. As an example, the results of the analysis for the station Zurich-Fluntern are shown in Fig. 1. The vertical axis indicates wind gust speeds, while the return period is depicted in the horizontal axis (logarithmic scale). The blue curve is the result of the statistical analysis of the data and represents the relation between wind gust intensity and return period. Further, the wind gusts observed during several strong wind events are indicated by black dots.
Fig. 1: Wind gusts as a function of return period in Zurich-Fluntern. The blue curve represents the function calculated by means of statistical analysis. Wind gusts measured over the period 1981-2007 are depicted by black dots. See text for further details.
figure1.png, 50 KBFrom Fig. 1, the return period corresponding to any wind gust speed may be determined, and vice-versa. Thus, the return period of storm Lothar (based on the peak wind gust speed) amounts to 44 years approximately (red arrow). It must be noted, however, that there is a strong uncertainty on this estimation (see Figure 2). Conversely, a 2-year return period corresponds to a wind gust speed of about 127 km/h. In other words, wind gusts of 127 km/h or stronger are expected to happen every 2 years on average in Zurich-Fluntern.
Figure 2 summarises the results of the statistical analysis for a chosen set of stations. In addition to the parameters represented in Fig. 1, the limits of the 95% confidence interval are indicated by green curves. This interval measures the uncertainty of the estimated function (blue curve), and describes the range within which the “true” function is situated with a probability of 95%. It can be seen that for increasing wind gust speeds the confidence interval grows rapidly, i.e. the uncertainty of the fitted function becomes larger.
Fig. 2: Results of the statistical analysis for six measurement stations. For legend, see Figure 1. Confidence intervals are represented by green curves. The three strongest wind gusts measured over the period 1981-2007 are labelled with their respective measurement date.
figure2.png, 229 KBThe differences in wind gust climatologies between the stations considered in this analysis are particularly striking. Stations located on wind-exposed mountain peaks (e.g. La Dôle, Pilatus) feature markedly higher wind speeds than other, less wind-exposed stations. Nevertheless, significant differences may also be observed between lowland stations. Such variations can mostly be explained by differences in geographical position as well as in topography. Thus, the Zurich-Fluntern measurement station, situated on the west side of a hill, generally records stronger wind gusts than other Swiss lowland stations. Stations located in the Ticino lowlands feature noticeably lower wind gusts than lowland stations on the north side of the Alps. As an example, no wind gusts above 100 km/h have been recorded in Stabio since the beginning of wind speed measurements in 1981.
Estimating the frequency of extreme wind speeds allows evaluating the spatial distribution of wind-storm-related risks, which makes it possible to set appropriate warning thresholds. Such information is of invaluable importance in a densely populated country like Switzerland.
Reference:
Technical Report 219, « Extreme value analysis of wind observations over Switzerland »
Contact:
Paulo Ceppi, MeteoSwiss Climate Services


