| Abstract [eng] |
Extreme Tornado events in Lithuania are not common, but can have serious and damaging effects on humans’ society and infrastructure, as well as on ecosystems and wildlife. Furthermore tornado is the local extreme meteorological event, which can cause huge losses and significant destruction in a particular territory. Also, tornado could pose a risk for energy objects, such as nuclear power plant, hydro power plant or wind power plant. Due to the following reason estimating the probabilities of extreme weather events is crucially important for avoiding weather-related disasters. However, in general, mathematical models can not completely or without any error reflect the reality; therefore it is important to perform the uncertainty and sensitivity analysis of the model results. The main aim of the research work is to analyze the characteristics of extreme tornado probabilistic assessment methods, tornado hazard probability models, identify uncertainties of model inputs noting their distribution and variation range. Also, apply uncertainty and sensitivity analysis to the model. For this purpose SIMLAB and GLUE software packages were used. In this research the statistical data regarding the Lithuania tornadoes documented cases during the period of 1950 – 2013 was analyzed. Using this data a map of tornadoes events in Lithuania was composed. In addition, a methodology of probabilistic models application for tornado impact assessment in a local region was developed and used in the analysis. Then, the methodology of uncertainty and sensitivity analysis was applied for results of this model. Appling uncertainty analysis the tolerance limits for tornado model results, when different thresholds of tornado wind speeds are taken into account, were given. Using SIMLAB and GLUE software packages the tolerance interval for probability estimate that tornado wind speed equal or higher 70m/s is: (5.60E-10; 1.61E-06). Also Bayesian method based GLUE (Generalised Likelihood Uncertainty Estimation) uncertainty estimation technique was applied and outcomes were presented. Sensitivity analysis was also used in order to determine the parameters that have the greatest influence on the model results. In addition variance based methods were applied. This analysis enabled to determinate that the uncertainty of data regarding the tornado damage area has the greatest influence on tornado hazard probability assessment estimate. |
