Cumulonimbus clouds and tornadoes detected by MODIS (Why the tornadoes were not predicted?)

     We express our heartfelt sympathy to those who injured and damaged by the tornadoes in May 6, 2012.

     The Tokyo University of Information Sciences (TUIS) captured the views of the cumulonimbus clouds and tornadoes over the Kanto area in Japan by the Moderate Resolution Imaging Spectroradiometer (MODIS) on Terra and Aqua on May 6, 2012.  The Terra and Aqua are polar orbit satellites flying about 800 km over the surface and observe the surface with an interval of 100 minutes.  As a result, the MODIS on Aqua and Terra provides the views of surface once or twice in the day time over the same region and it is not an appropriate to observe the local phenomena in high frequency.  Usually, the geostationary satellites provide a high frequent view of the surface.  The cumulonimbus clouds and tornadoes were detected accidentally on May 06, 2012.

     In the morning of May 6, 2012, the MODIS on Terra observed the cumulonimbus clouds at 10:50 JST over Yamanashi area.  Then, the MODIS on Aqua observed the tornadoes as super cells at 12:30 JST over Ibaragi and Tochigi area.  The 4 cumulonimbus clouds observed in the morning developed to the 4 super cells by 80 minutes.  Exactly at12:30, the tornado damaged the towns in Ibaragi and Tochigi of the Kanto area.

1.Images observed by MODIS(RGB:143)

  Fig.1-a and Fig.1-b are the cloud distribution observed at 10:50JST and 12:30JST on May 6, 2012 respectively. The detailed images are available on clicking each image.

Fig.1-a Cloud distribution observed by MODIS at 10:50 JST on May 6, 2012. Fig.1-b Cloud distribution observed by MODIS at 12:30 JST on May 6, 2012.

2.Estimated location of the cumulonimbus clouds and tornadoes

 Fig.2-a and Fig.2-b indicate the estimated locations of the cumulonimbus clouds and tornadoes.  On Fig.2-a, the locations of four developing cumulonimbus clouds are estimated from the cloud distribution.  The diameters of cumulonimbus clouds are 5 to 10 km.  On Fig.2-b, the locations of four cumulonimbus clouds showing eddy patterns are estimated from the cloud distribution.  The diameters of two eddies of cumulonimbus clouds in south, which were corresponding to tornadoes, are about 20 km, and two eddies of cumulonimbus clouds in north, which didn't leave severe damages, are about 10 km.

Fig.2-a Locations of cumulonimbus clouds at 10:50 JST on May 6, 2012. Fig.2-b Locations of tornadoes and cumulonimbus clouds at 12:30 JST on May 6, 2012. 

3.Estimated courses of cumulonimbus clouds to Tornadoes

 The cumulonimbus clouds observed at 10:30 JST on May 6, 2012 over the Yamanashi disappeared by the noon and shifted to Ibaragi and Tochigi.  According to the disaster report by the Japan Metrological Agency, the active cumulonimbus clouds passed over the Kanto plane.  The four red lines are estimated on Fig.3-a and 3-b with connecting the cumulonimbus clouds and the tornadoes.  As the MODIS didn't provide any images between two observations, these four courses are the best estimate of tornadoes.

Fig.3-a Estimated routes from the cumulonimbus clouds to Tornadoes. Fig.3-b Plots of estimated routes.

4.Estimated speed from the cumulonimbus clouds to tornadoes

 The estimated speeds of the cumulonimbus clouds to tornadoes are listed on Table 1.  The estimated speeds are calculated based on the distances between two locations of the cumulonimbus clouds and tornadoes and the time difference of 80 minutes.

Table 1.   Estimated distance and speed from the cumulonimbus clouds to tornadoes

ID Estimated distance(km) Estimated speed(km/h)

Ta

130

76

Tb

148

87

Tc

132

78

Td

132

78

5.Enlarged image of tornadoes

 Fig.4-a shows the distributions of tornadoes and cumulonimbus clouds at 12:30 JST.   Fig.4-b shows the routes from the cumulonimbus clouds to tornadoes with polygons of local governments.  Although the estimated locations of tornadoes from images are slightly different from actual paths of Tornadoes, at 12:30 JST.

  1) the tornado in the south (Ta1) was located in the northwest end of Bando-city, and moved to Joso-city and Tsukuba-city,

  2)  the tornado next to Ta1 (Tb1) was located on the boundary between Shimotsuke-city and Kamimikawa-town, and moved from Maoka-city to Masuko-town, and from Tsukusei-city to Sakuragawa-city,

  3) the two northern cumulonimbus clouds didn't exhibit any damages but the flurry over Okunikko around 22 m/s was reported by the Metrological Agency.

 

Fig.4-a tornadoes at 12:30 on May 6, 2012. Fig.4-b Estimated routes from cumulonimbus clouds to tornadoes over polygons of local governments.

6.Why the tornadoes were not predicted?

     A difficulty to predict tornadoes goes to less experiences in Japan.  Numerical models were proposed but its operational use was difficult in the real-time.  The enlarged cumulonimbus clouds were observed in the north of Mt. Fuji in the morning of May 6, 2012.  While the entrainments of cumulonimbus clouds to the northeast area, JMA observed flurries or hails over the metropolitan Tokyo, but those observations were not utilized to predict the tornadoes.  Although the tornadoes are enforced very quickly and a difficulty of forecasting is obvious,  the authorities may pay attentions to the cloud distributions from satellite and should find a possibility to predict tornadoes with the knowledge that the Kanto plane was heated and the strong south wind were dominant.

Ichio ASANUMA, The Tokyo University of Information Sciences