Modeling Mother Nature

07.09.13 -

What occurs within a thunderstorm that leads to the formation of destructive weather events such as tornadoes and downbursts? Leigh Orf of Central Michigan University, NCSA’s Robert Wilhelmson, and Eric Savory of the University of Western Ontario are utilizing an idealized cloud model designed specifically for massively parallel architectures (CM1) to model storms and hoping that advances in supercomputing power will bring them closer to answers.

Running CM1 on the Kraken supercomputer, an XSEDE resource located at the National Institute for Computational Sciences (NICS) at the University of Tennessee, Orf and his colleagues simulated storms in environments known to be conducive to creating downbursts, which happen 10 times more often than tornadoes.

A downburst is created by a column of sinking air that, after hitting ground level, spreads out in all directions and is capable of producing damaging straight-line winds in the range of 80-150 mph. The debris pattern is what distinguishes a downburst from a tornado. Downburst debris will be laid out in straight lines parallel to the outward wind flow as opposed to the rotational damage pattern seen with a tornado. Orf and Savory are exploring the specific stresses that downbursts present to structures, such as power transmission lines. Their approach produces a more physically realistic wind field not found in simpler engineering models that do not include clouds and precipitation.

NCSA visualization expert Rob Sisneros used the VisIt software on NCSA’s Blue Waters to visualize Orf’s results. Here we see a snapshot in time of potential temperature perturbation (roughly, the dense, cold air) of the downburst-producing air mass thunderstorm simulation referenced above. Downbursts descend from the cloud base primarily forced by the negative buoyancy induced by evaporation, melting, and sublimation.

National Science Foundation

Blue Waters is supported by the National Science Foundation through awards OCI-0725070 and ACI-1238993.

National Science Foundation

XSEDE is supported by National Science Foundation through award ACI-1053575.