11:15 〜 11:30
[AAS03-09] Analysis of Environmental Winds Determining Tornado Movement in Japan Using Hourly Reanalysis Data
キーワード：竜巻、再解析データ、台風
A tornado is a violent atmospheric vortex generated by an updraft within a cumulus or cumulonimbus cloud. In Japan, tornado-related damage is particularly frequent along the Pacific coast, and their overall occurrence is not negligible. To mitigate damage to humans and infrastructure, this study explores the potential for predicting tornado tracks after their formation and examines related meteorological factors.
According to Niino et al. (1997), more than half of the tornadoes that occurred in Japan between 1961 and 1993 moved toward the northeast quadrant. Nevertheless, as this data was based on visual observations, the reported tornado movement directions were biased toward only 8 of the 16 directions. To address this issue, this study utilizes the Japan Meteorological Agency (JMA) gusty winds database to objectively collect tornado movement data, extend the observation period, and analyze preferences in movement directions. Furthermore, we examined the relationship between large-scale wind fields and tornado movement, discussing the factors that determine tornado tracks and the predictability of their movement.
First, tornado movement directions are calculated from the latitudes and longitudes of the starting and ending points of the damage path. These calculations reveal that approximately 70% of tornadoes move toward the northeast quadrant. Despite changes in tornado data collection methods over time, the consistent preference for the northeasterly direction suggests that this preference is not coincidental.
The preference toward the northeast quadrant results from overlapping peaks in the eastward and northward directions. The eastward preference indicates that cumulonimbus clouds are primarily transported by westerly winds in the mid-troposphere, a pattern observed across all seasons. In contrast, tornadoes occurring in summer (June to August) and autumn (September to November) exhibit a stronger northward peak, likely due to the influence of typhoon-associated tornadoes.
Moreover, a statistically significant correlation of 0.6 to 0.8 is found between tornado movement directions and large-scale hourly mean wind directions, supporting the idea that tornadoes are generally transported by prevailing winds along with cumulonimbus clouds. The highest correlation and degree of coincidence are observed around the 700 hPa level, suggesting that horizontal winds at this altitude play a crucial role in tornado movement. However, the method of Bunkers et al. (2000) for estimating supercell movement does not closely align with observed tornado movement directions in Japan. It implies that tornado movement indicators developed in the U.S. may not be directly applicable to tornadoes in Japan.
According to Niino et al. (1997), more than half of the tornadoes that occurred in Japan between 1961 and 1993 moved toward the northeast quadrant. Nevertheless, as this data was based on visual observations, the reported tornado movement directions were biased toward only 8 of the 16 directions. To address this issue, this study utilizes the Japan Meteorological Agency (JMA) gusty winds database to objectively collect tornado movement data, extend the observation period, and analyze preferences in movement directions. Furthermore, we examined the relationship between large-scale wind fields and tornado movement, discussing the factors that determine tornado tracks and the predictability of their movement.
First, tornado movement directions are calculated from the latitudes and longitudes of the starting and ending points of the damage path. These calculations reveal that approximately 70% of tornadoes move toward the northeast quadrant. Despite changes in tornado data collection methods over time, the consistent preference for the northeasterly direction suggests that this preference is not coincidental.
The preference toward the northeast quadrant results from overlapping peaks in the eastward and northward directions. The eastward preference indicates that cumulonimbus clouds are primarily transported by westerly winds in the mid-troposphere, a pattern observed across all seasons. In contrast, tornadoes occurring in summer (June to August) and autumn (September to November) exhibit a stronger northward peak, likely due to the influence of typhoon-associated tornadoes.
Moreover, a statistically significant correlation of 0.6 to 0.8 is found between tornado movement directions and large-scale hourly mean wind directions, supporting the idea that tornadoes are generally transported by prevailing winds along with cumulonimbus clouds. The highest correlation and degree of coincidence are observed around the 700 hPa level, suggesting that horizontal winds at this altitude play a crucial role in tornado movement. However, the method of Bunkers et al. (2000) for estimating supercell movement does not closely align with observed tornado movement directions in Japan. It implies that tornado movement indicators developed in the U.S. may not be directly applicable to tornadoes in Japan.