I am a senior scientist at NCAR, and section head, Turbulence Numerics Team (TNT). The team comprises today Duane Rosenberg (Software Engineer IV) and Pablo Mininni (0.25 FTE, and also at Universidad de Buenos Aires), together with Amrik Sen (graduate student, Applied Mathematics, CU), Josh Stawarz (graduate student, Space Physics, CU), Cecilia Rorai (graduate student, Trieste and U. Maryland).
I am also the Director of the Geophysical Turbulence Program (GTP), see its web page athttp://gtp.ucar.edu/.
In TNT, we develop programs for the community that possess spectral accuracy, either with periodic boundary conditions (GHOST) or with adaptive mesh refinement (GASPAR) and that parallelize efficiently (up to ~ 40,000 processors for GHOST). We do this in order to integrate numerically the equations that govern the dynamics of fluids as encountered in the atmosphere and the oceans, and in the solar environment as well, in simplified settings.
With these codes, we perform large numerical simulations, with several world-first, on our way to peta-scale computing.
In forced rotating turbulence (15363 grid points, and 30723), we study the interactions of turbulent eddies and waves, the return to isotropic motions in the small scales, and the role that helical motions (think of a corkscrew), as observed in supercell storms and hurricanes, play in the formation and further development of large-scale coherent structures.
In unforced magnetohydrodynamics (MHD), when coupling the dynamics of the fluid velocity to a magnetic field, we used 15363 and 20483 grid points; these runs led us to the discovery of the roll-up of current sheets and of their bursting, as observed for example in the plasma of the Solar Wind.
Furthermore, we implemented numerically a four-fold symmetry that is allowing us to compute on an equivalent grid of 61443 points in ideal MHD (work performed in collaboration with a team in France). We also look at the origin of such magnetic fields, the so-called dynamo problem, and have recently examined the role helical motions (again) play in the formation of large-scale structures in MHD, as observed for example in stars and Galaxies.
The (large) data sets thus produced in all these runs are stored, visualized and analyzed by us; they are also made available to the community.
We then use the data to test and improve models of parametrization of the small-scales. Such models can be incorporated in codes built to examine specific problems of interest to the community, linked for example to the development of extreme events in weather, space weather and climate.