Nat. Hazards Earth Syst. Sci., 12, 1969–1981, 2012 www.nat-hazards-earth-syst-sci.net/12/1969/2012/ Natural Hazards doi:10.5194/nhess-12-1969-2012 and Earth © Author(s) 2012. CC Attribution 3.0 License. System Sciences Severe wind gust thresholds for Meteoalarm derived from uniform return periods in ECA&D A. Stepek, I. L. Wijnant, G. van der Schrier, E. J. M. van den Besselaar, and A. M. G. Klein Tank Climate Services Dept, KNMI, P.O. Box 201, 3730 AE De Bilt, The Netherlands Correspondence to: A. Stepek ([email protected]) Received: 23 August 2011 – Revised: 14 March 2012 – Accepted: 1 May 2012 – Published: 25 June 2012 Abstract. In this study we present an alternative wind gust We are aware of other, sometimes binding factors (e.g. warning guideline for Meteoalarm, the severe weather warn- laws) that prevent participating counties from implementing ing website for Europe. There are unrealistically large dif- this climatology based warning system. ferences in levels and issuing frequencies of all warning levels currently in use between neighbouring Meteoalarm countries. This study provides a guide for the Meteoalarm community to review their wind gust warning thresholds. 1 Introduction A more uniform warning system is achieved by using one pan-European return period per warning level. The associ- There is a growing demand for accurate severe weather warn- ated return values will be different throughout Europe be- ings. Worldwide the recorded number of so called “great cause they depend on local climate conditions, but they will natural catastrophes” (events with thousands of casualties, not change abruptly at country borders as is currently the hundreds of thousands made homeless and substantial eco- case for the thresholds. As return values are a measure of nomic losses) has gradually increased over the past 60 yr the possible danger of an event and its impact on society, (Munich Re, 2009). Of the weather related catastrophes, 7 % they form an ideal basis for a warning system. Validated wind are caused by extratropical storms and in 2009 alone, such gust measurements from the European Climate Assessment storms caused losses equivalent to 81 billion US Dollars. and Dataset (ECA&D, http://www.ecad.eu) were used to cal- Extreme weather alarms are an important part of efforts to culate return values of the annual maximum wind gust. The curb these losses. On the European scale, Meteoalarm plays current thresholds are compared with return values for 3 dif- a leading role in answering the demand for severe weather ferent return periods: 10 times a year return periods for yel- warnings. low warnings, 2 yr periods for orange and 5 yr periods for Meteoalarm is an initiative by EUMETNET, the European red warnings. So far 10 countries provide wind gust data Weather Services Network within the World Meteorologi- to ECA&D. Due to the ECA&D completeness requirements cal Organization (WMO), to present all weather warnings and the fact that some countries provided too few stations to for Europe on one website: http://www.meteoalarm.eu. Since be representative for that country, medians of the return val- the website became fully operational in 2007, the site has be- ues of annual maximum wind gust could be calculated for come increasingly popular with more than 1 billion hits since 6 of the 10 countries. Alternative guideline thresholds are its start and a doubling of the number of visitors since the in- presented for Norway, Ireland, The Netherlands, Germany, troduction of the “Europe of regions” map. In one overview the Czech Republic and Spain and the need to distinguish this map presents the status of all weather warnings with between coastal, inland and mountainous regions is demon- up-to-date reports in about 650 areas of the 30 participat- strated. The new thresholds based on uniform return peri- ing countries. Warnings for as many as 10 weather param- ods differ significantly from the current ones, particularly for eters are provided by the National Meteorological Service coastal and mountainous areas. (NMS) of each country, for up to 24–48 h ahead. The sever- ity of the warning is related to a colour code with green for no warning and yellow, orange and red for increasing levels Published by Copernicus Publications on behalf of the European Geosciences Union. 6 1970 A. StepekStepek et et al.: al.: Severe Severe wind wind gusts gusts for for Meteoalarm Meteoalarm 6 Conclusionsof exceptionally severe weather. Although Meteoalarm pro- issuing frequency codes green, yellow, orange and red vides warning threshold guidelines for the different kinds and 100 Meteoalarmlevels of is severe the EUMETNET weather, setting initiative the country-specific for a pan-European thresh- 98 severe weather warning system. Although Meteoalarm pro- olds is the responsibility of the individual NMS’s. This is of- 96 videsten guidelines an elaborate for process warning involving thresholds various for stakeholders, the different gov- kindsernment of severe officials weather, and the meteorologists. NMSs of the The member major countries drawback 94 issuing frequency are responsibleof the exclusive for choosing national the focus thresholds, in this process often inand an the ela non-b- 92 compulsory nature of the Meteoalarm guidelines is that it orate process involving various stakeholders, government of- 90 leads to questionable differences in the frequency and level ficials and meteorologists. The main drawback of the ex- 90 clusiveof warningsnational focus between in neighbouringthis process is countries. that the issuing fre- quencyMeteoalarm need not be became similar aware to neighbouring of this problem countries. and instigated As a 80 this study of a warning system that both reflects the spe- result of determining warning thresholds at a national level, 70 cific national needs while maintaining a pan-European uni- questionable differences in the frequency and level of warn- form issuing frequency for severe weather. The need for in- 60 ings between neighbouring countries are observed. creased uniformity is illustrated in Fig. 1, which shows the 50 InMeteoalarm order to assist wind the gust Meteoalarm warnings issued community in the period of NMSs from with the ongoing optimization of their warning systems for 1 August 2007 to 30 June 2009. This figure shows issu- 40 extremeing wind frequencies gusts, ranging a methodology from asis much proposed as 62 % and to demon- as few as issuing frequency 30 strated0.15 that % forcan all be warnings used to combined help alleviate (yellow, some orange of the and dif- red). ferencesFor betweenthe orange neighbouring warning level, countries the issuing and frequency regions. This varies 20 methodology is based on pan-European uniform return peri- between 15.3 % and 0 %. 10 ods. InThe this aim study of we this chose study tois to use develop return alternative periods of Meteoalarm 10 times 0 a yearguidelines for code for yellow the thresholds warnings, of once wind in gust 2 years warnings. for code These Meteoalarm participating countries orangealternative and once guidelines in 5 years arefor code based red on which one return roughly period corre- per spondwarning to the Meteoalarm level for the wholeguideline of Europe for a median thus providing sized coun- a pan- Fig. 1. Histogram of the issuing frequency of wind gust warnings for the participating Meteoalarm countries, based on 666 days be- try. TheEuropean return uniformvalues that warning correspond system. to Thresholds these uniform based retu onrn the Fig. 1. Histogram of the issuing frequency of wind gust warnings associated return values are representative of the local wind tween 1 August 2007 and 30 June 2009. The colours in the diagram periods are different for each individual country. They are a forcorrespond the participating to the colour Meteoalarm coding of countries, the warning based (yellow, on 666 orange days and be- climate and consequently give a good indication of the pos- measure of how extreme an event is compared to the local tweenred are 1 August increasing 2007 levels and of 30 warning June 2009. and greenThe colours means inno the warning). diagram climatesible and danger are a of powerful the event means and its of impact predicting on society. the possible correspondSource: Meteoalarm, to the colour EMMA coding EUMETNET. of the warning (yellow, orange and danger ofThe an return event periods and its that impact are onused society. to represent the warning red are increasing levels of warning and green means no warning). levels are loosely based on the current Meteoalarm guide- The results of this study indicate that warnings for haz- Source: Meteoalarm, EMMA EUMETNET. lines for median sized countries and therefore include past ardousdamage wind experiencegusts should of thebe regionalisedNMS’s as explained for coastal, in Sect. inlan 3.1.d In ECA&D currently (January 2012) receives data from and mountainous areas. The median return values and the this study we will derive return values associated with these time62 countries series may throughout no longer Europe be representative and the Mediterranean, for future con- cli- NMS thresholds concur reasonably well for the inland ar- return periods for station data from 6 European countries re- mate.tains The daily climate data from change more may than impact 4800 stations on the regionalization and serves as eas but less well for coastal and mountainous areas. This gionalised into coastal, inland and mountain areas. In doing ofa severe data portal weather and thresholds climate monitoring as well. Recently, tool for the McVicar climate etre- al. makes sense because NMSs generally do not regionalise their so we derived warning thresholds that depend on local wind (2010)search observed community. a divergence New stations in the are speed regularly of decline added in to near- the warning thresholds and base them on past experience of dam- climate conditions that not only vary between countries, but surfacedatabase.