Final Report Summary - EWENT (Extreme weather impacts on European networks of transport) Project context and objectives: The goal of the EWENT project was to assess the impacts of extreme weather events on European Union (EU) transport system. These impacts were monetised. EWENT also evaluated the efficiency, applicability and finance needs for adaptation and mitigation measures which will dampen and reduce the costs of weather impacts. The methodological approach was based on generic risk management framework that follows a standardised process from identification of hazardous phenomena (extreme weather), followed by impact assessment and closed by mitigation and risk control measures. EWENT started this by identifying the hazardous phenomena, their probability and consequences and proceeded to assess the expected economic losses caused by extreme weather when it impacts the European transport system, taking also into account the present and expected future quality of weather forecasting and warning services within Europe. EWENT applied the IEC 60300-3-9 risk management standard framework all the way through its research process and the project's work breakdown also followed the standard structure. The modal coverage was quite wide, including all the modes, but some covered more specifically than others. The consequences and impacts were looked from three angles: infrastructure, operations and indirect impacts to third parties. Both passenger and freight transport were covered for all considered modes. The main idea of the research strategy was to split the work in clearly structured sections, while keeping the modes present in all stages of the work. One of the critical decisions on research strategy was the following of causality between extreme weather phenomena and final consequences. This proved to be a challenging task, but ultimately the right decision, which was shown by the results from the later work packages WP4 ('Costs') WP5 ('Risks'), which would have lacked the necessary data for adequate analysis. EWENT was organised in a traditional manner, with partner key researchers building the broader management team and coordinator's own team building a core management group. The project was further supported by consultative board, which was represented by European Investment Bank (EIB), Organisation for Economic Cooperation and Development (OECD) international transport forum, Finnish ministry of transport and communications, Allianz and Politecnico di Turin. Especially, consultative board proved to be extremely useful, although some more time for interactions between it and the research consortium would have been even more beneficial. For management purposes, risk and quality plan for the project was drafted during the first months. Also exploitation and dissemination plan served well the management of the project. The major risk identified in the risk and quality plan, which was in fact realised, was the time resources the key researchers were able to commit to the project. The project consortium was exceptionally cross-disciplinary, which was challenging at the beginning of the project, but later proved to be a particular strength. Project results: The first work package of EWENT provided the results concerning the relevant extreme weather phenomena: - a list of critical weather phenomena which, on the basis of literature and media mining, are clearly such that they have consequences on transport systems; - threshold values of parameters for the above phenomena which, if met or exceeded, indicate a high probability of measurable harmful impacts and consequences; - selected impact mechanisms (12) that indicate why certain impacts and consequences start to occur; the meaning of these causal maps is to help later in the identification of efficient mitigation and adaptation measures. The methodological approach was the following: Firstly, there was the traditional review of professional literature. Secondly, media mining was done in order to get more empirical data and to assess which modes in which parts of Europe seem to be affected the most. Thirdly, there was a compilation of specific case studies on past extreme incidents, helping to assess the specific consequences of certain phenomena. Relevant adverse and extreme weather phenomena were analysed by taking into account the ranking and threshold values defined from the viewpoint of different transport modes. The following phenomena were analysed, based on extensive literature review of more than 150 references (Leviäkangas et al., 2010): strong winds, heavy snowfall, blizzards, heavy precipitation, cold spells and heat waves. In addition, visibility conditions determined by fog and dust events, small-scale phenomena affecting transport systems such as thunderstorms, lightning, large hail and tornadoes. Events that damage the transport system infrastructure were also considered, but not included in quantitative data analysis. In order to assess the spatial and temporal variation of adverse weather conditions over the European continent, two gridded datasets were used: the E-OBS European high-resolution land-only gridded dataset (Haylock et al., 2008) produced through spatial interpolation of daily station data daily mean (TG), maximum (TX), minimum (TN) temperature and precipitation sum in a 0.25 degrees regular latitude-longitude grid were utilised for the time period 1971 - 2000. In addition, the ERA-Interim reanalysis dataset of the atmospheric state produced at the European Centre for Medium-Range Weather Forecasting (ECMWF) was used. ERA-Interim uses four-dimensional (4D) variational analysis on a spectral grid with T255 horizontal resolution (corresponds to approximately 80 km) and a hybrid vertical coordinate system with 60 levels (Simmons et al., 2006). We extracted for our analysis the 6-hour forecast of 10 m wind gust, 6-hour forecast of precipitation sum, and 6-hour reanalysed 2 m mean temperature in Gaussian grid with a spacing of about 0.7 degrees for the time period 1989-2010. The European severe weather database (Dotzek et al., 2009) contains an ever-growing collection of reports of individual severe weather events and is managed by the European Severe Storms Laboratory (ESSL). We presented maps for hail (2 cm or larger) and tornadoes. In addition, to calculate the frequency of freezing precipitation, the 16-year NOAA National Climatic Data Centre database of surface observations at airports covering the period 1982-1997 was used (Lott, 2000), and additional observations were used in the evaluation of fog and dust occurrence, as well as lightning frequency. Six high-resolution (approximately 25 x 25 km2) Regional climate model (RCM) simulations produced in the ENSEMBLES project were used to estimate future changes in adverse weather conditions for transport. All GCMs used to drive RCMs were forced with the A1B (medium, non-mitigation) emission scenario (van der Linden and Mitchell, 2009). The analyses compared the near-future (2011 to 2040) and far-future (2041 to 2070) time horizons with the present climate (1971-2000). The RCMs chosen were: a) SMHIRCA-ECHAM5-r3; b) SMHIRCA-BCM; c) SMHIRCA-HadCM3Q3; d) KNMI-RACMO2-ECHAM5-r3; e) MPI-M-REMO-ECHAM5-r3; f) C4IRCA3-HadCM3Q16. There are three main uncertainties in climate projections: internal variability of the climate system that exists even in the absence of any external forcing; uncertainty in radiative forcing due to future emissions of greenhouse gases and aerosols; and model uncertainty (Hawkins and Sutton, 2009). It was chosen to neglect the uncertainty due to emissions and focus on the A1B emission scenario, because the uncertainty in emission scenarios rivals model uncertainty only in the latter half of the 21st century. Based on the calculation of frequencies using the six RCMs, the multi-model mean of the change compared to the control period (1971-2000; for wind gusts and for blizzards 1989-2009), furthermore the upper and lower limits of the change has been calculated for each threshold of the adverse and extreme phenomena. The multi-model mean is the average change indicated by the six models giving each model equal weight. The range of changes is also indicated for each grid that describes well the inter-model variability, i.e. upper and lower limit. The upper limit (maximum) shows the 'most positive' change of any model, while the lower limit (minimum) indicates the most negative change. The results of WP3 summarise the added risks for delay and accidents in dependence from the expected climatic changes. As the expected changes, concerning heavy precipitation are negligible, the focus is laid on the other phenomena of extreme weather (wind gusts, snowfall, heat waves and cold waves). For the different climatic regions, the expected change from 2010 for each of those phenomena is given by a percentage rate. For example, the heat waves for the temperate region will increase between 0.1 and 7 % until 2040. The effect of the changes on different traffic modes is described for the indicators delay and accident rate in the same table. For the above mentioned example, the inland waterway transport will be slowed down by droughts. For all transport modes, the accident rate should be reduced or will stay on a low level in future, because it is expected that better technologies and higher safety standards (which today are best utilised in aviation), will influence the accident rate more than the expected weather changes. From an economic point of view, the former trend clearly outweighs the latter. EWENT results summarise the cost analysis findings at present and projected to 30 years in the future. Needless to point out, the figures are very rough benchmarks and bring in the magnitudes of different cost items resulted by extreme weather phenomena, but do not necessarily represent any specific contexts, regions or cases. What is evident, is that road sector costs dominate the picture quite clearly. This is because of one main reason only: most of the transport is done on roads. If relative cost analysis would have been used, the picture might be slightly different, though not too much. Roads are still today relatively unsafe and due to the nature of road network, the vulnerability is high: roads are everywhere and they are not managed as systematically as other networks. As to the future costs, there is an apparent trend in declining accident costs, first and foremost because of general trends, and secondly because the winters are getting shorter and warmer in the Northern hemisphere. Icy and slippery roads raise the accident risk up to two-three times higher than on dry roads. The winter maintenance operations costs are also expected to decrease throughout Northern Europe. But the actual impact of more frequent weather extremes remains still an open question. However, the magnitudes of that, even if these extremes would become more frequent, will not be that significant compared to the big picture. Natural catastrophes and extremes that bring societies to their knees are of course another chapter. In road transport, as the data on estimated future accident levels shows there will be considerable improvements in vehicle technologies that will contribute to greater safety for passengers. Thus, the scenarios take these developments into consideration as given baseline of future accident volume developments. The set of results attempt to capture the relevant changes in costs due to climatological changes. For many items, the changes are positive, but not all. In aviation, the trend is to see costs from delays to go up by 2040 from present mainly due to value of time changes but declining by 2070 as events become less frequent. The risk assessment is based on the definition of transport systems vulnerability to extreme weather events in different countries and on calculations of the most probable event chains, starting from adverse weather phenomena and ending with events that are harmful to the transport network in different climate regions. The probabilistic section is the hazard analysis. The vulnerability of a particular mode in a particular country is a function of exposure (indicated by transport or freight volumes and population density), susceptibility (infrastructure quality index, indicating overall resilience) and coping capacity (measured by GDP per capita, purchasing power adjusted). Hence, we defined the extreme weather risk as a product of probability of negative consequences and vulnerability assessment: risk = hazard × vulnerability = P(negative consequences) × f(exposure, susceptibility, coping capacity). Based on this analytical approach information, the risk indicators for each mode and country have been derived. Due to the techniques used in calculations, the risk indicator is by definition a relative indicator, and must not be considered as an absolute measure. It is a very robust ranking system, first and foremost. List of websites: http://www.ewent.vtt.fi
