Zonal Flows and Structure Formation in Turbulent Plasmas

The group studies the zonal flows with massively parallel computer simulations of plasma and planetary turbulence, with the goal to make predictions of their long-term evolution and experimentally observed switching effects between different flow patterns.

Zonal Flows in Jovian Atmospheres

Everybody has already observed how a flow — for example when pouring milk into a cup of coffee — decays turbulently into smaller and smaller vortices until it is completely consumed. A strikingly unfamiliar behaviour occurs on the giant gas planets Jupiter and Saturn, which are turbulent due to the temperature contrast between their hot interior and cold surface: instead of producing smaller and smaller eddies, the turbulence creates planet-spanning east and west flows. These "zonal flows" are very conspicuous on the planets from the dark and bright cloud bands that align with them along the lines of latitude.

Zonal Flows in Magnetised Plasmas

Zonal flows in a NLET tokamak turbulence simulation start to shear turbulence structures; inset: poloidal flow velocity

Convective turbulence also exists in the torus shaped fusion reactors of the tokamak or stellarator type due to the enormous temperature gradients therein. Analogous to the case of the gas planets it also gives rise to global flows, this time along the small torus circumference. The flows can be detected by sampling the electric potential, which shows pronounced bands, each extending on a complete flux surface, similar to the Jovian cloud bands. The flows in turn exert a decisive damping effect on the turbulence, which is favourable for the plasma confinement and greatly reduces the technical effort necessary to keep the plasma burning.

Massively Parallel Turbulence Simulations

Bands of zonal (azimuthal) flow velocity in a deep Jovian turbulence simulation with the NAN code. Bright colours indicate eastward flows, dark westward flows.

The exploration of the zonal flows necessitates the simulation of the underlying turbulence to sufficient precision and for so long that it requires to employ massively parallel computers with thousands of cores and the corresponding specialised computer codes. For this purpose the group develops the non-local two-fluid code NLET for the plasma turbulence as well as the anelastic Cartesian code NAN for the planetary turbulence.

The long-term prediction of the evolution of the zonal flows would on one hand provide a safer basis for extrapolations of current fusion experiments to larger machines, such as Iter. On the other hand this might open a way to manipulate the flows to improve the confinement, which could reduce drastically the cost and complexity of nuclear fusion.

For planetary flows, the expected results will help to explain the observed structure on the different planets and yield predictions for the corresponding flows on the extrasolar giant planets discovered in recent years.

Go to Editor View