P5.8
A new tornado simulator reproducing flow fields under supercell
Koji Sassa, Kochi University, Kochi, Japan; and S. Takemura and A. Yamasaki
Only 20 % of the mesocyclones detected by Doppler radars cause tornado as stated by Wakimoto and Cai (2000). If we understand the flow condition of tornado genesis in a supercell, we can improve the reliability of nowcast for tornado outbreak. The present laboratory experiment aims to clear the fluid dynamical conditions of tornadogenesis in a supercell. We tried to realize the rotational updraft of mesocyclone and the gust front due to the rear flank downdraft in our newly designed simulator as shown in Fig.1. The mesocyclone simulator is composed of 48 guide vanes, a convection chamber of D = 900 mm in diameter and 4 fans generating updraft. It resembles conventional tornado simulators made by Snow and Lund (1997), Monji and Mitsuta (1985) and so on. It has, however, open space between the floor and it. Then rotating updraft similar to the actual mesocyclone developed in mid air is generated in the mesocyclone simulator. The gust generator supplies dense mist made from dry ice and simulates the gust front corresponding to the rear flank downdraft of supercell. The aspect of tornadogenesis was filmed with a digital video camera. Depending the attack angle of vanes of the mesocyclone simulator and the height of it, 4 different patterns were observed. Fig. 2 shows the multiple vortex structure having 3 suction vortices. Such flow pattern was actually observed by Wurman (2002). A pair of clockwise and counter-clockwise vortices is also observed as shown in Fig. 3. Such flow pattern has never been found in the traditional tornado simulators (e.g. Church et al. 1977). But, Fujita (1981) and Bluestein et al. (2007) observed a pair of cyclonic and anti-cyclonic tornadoes with Doppler radars. We also observed single tornado-like vortex and turbulent flow without tornado. We think that such flow patterns restage the actual flow fields in supercells.
Poster Session 5, Novel Instrumentation and Data Processing Techniques Posters
Tuesday, 12 October 2010, 3:00 PM-4:30 PM, Grand Mesa Ballroom ABC
Previous paper Next paper
Extended Abstract (2.7M)