Session: 12-02 Risk Assessments and Management - Session 2

Paper Number: 135030

135030 - Impact of Synoptic Weather Patterns Along the Pacific Coastline of Japan on Tornado Wind Hazard Curves 

Abstract: 

Determining the return period of wind speeds is essential for assessing tornado risks at nuclear sites. This process, known as tornado wind hazard analysis, involves developing a hazard curve using historical tornado data. The analysis model requires to be developed in consideration of characteristics surrounding the region of interest. Approximately 54 tornadoes, including waterspouts, are observed annually across Japan (2007–2022). Many tornadoes occur along the coastline, which presents unique challenges. The occurrence of tornadoes over the sea is a challenge for the collection of tornado statistics, which can lead to incomplete or uncertain data. Moreover, the region of interest becomes restricted to a narrow strip along the coastline, creating the possibility of under- or over-sampling issues, particularly when tornadoes cross the boundary. To address these challenges, Central Research Institute of Electric Power Industry developed the TOrnado Wind speed hazard for Limited Area (TOWLA) model.

When conducting tornado wind hazard analysis, it is crucial to objectively review historical tornado data while taking into account regional climatology. It is necessary to identify the primary factors that cause tornado outbreaks in a particular area. For example, in Japan, tornadoes can result from typhoons, extra-tropical cyclones, monsoons, fronts, and other factors. However, it is unclear how sensitive these factors are to the return period of wind speeds. This study categorizes tornado data by synoptic weather patterns along the Pacific coastline of Japan and examines their impact on the return period through a case study.

Tornado data along the coastal region of Japan were classified based on seven synoptic weather patterns, following the method proposed by Kobayashi et al. (2007), which was further extended based on additional observations. A 5 km strip from the coastline was selected as the data acquisition region for the period 1961 to 2016, limiting the tornado classification to the F-scale era. To address the undercounting of weak tornadoes in older data, tornado data were restricted to those occurring after 2007 for F0 tornadoes and 1991 for F1 tornadoes. Various statistics such as mean, standard deviation of tornado counts per year, wind speed, length, and width were calculated for each synoptic weather pattern.

The analysis showed that typhoons and extra-tropical cyclones were the predominant causes of tornadoes in the region, and tornado counts per year showed good agreement with Poisson and Polya distributions. These findings were further incorporated into our TOWLA model, demonstrating that differences in synoptic weather patterns have a significant effect on the return period of wind speeds.

Presenting Author: Kota Fujiwara Central Research Institute of Electric Power Industry

Presenting Author Biography: Kota Fujiwara is currently employed at the Sustainable System Research Laboratory, Central Research Institute of Electric Power Industry (CRIEPI), Japan as research scientist. His major research interest is meteorological fluid dynamics, multiphase flow including bubbles, drops, aerosols, measurement methods, tracking algorithms, and the risk assessment of nuclear powerplants.

Authors:

Kota Fujiwara Central Research Institute of Electric Power Industry
Daisuke Nohara Central Research Institute of Electric Power Industry
Yuzuru Eguchi Central Research Institute of Electric Power Industry
Yasuo Hattori Central Research Institute of Electric Power Industry
Hiromaru Hirakuchi Central Research Institute of Electric Power Industry