A full-torus simulation of turbulent transport scaling shows that transport driven by microscopic-scale fluctuations (ITG modes) in present devices can change character and transition from Bohm-like scaling ~(rIvi) to Larmor-orbit-dependent "gyro-Bohm" scaling ~(rivi)(rI/a).

W. W. Lee, Z. Lin, W. M. Tang, J. L. V. Lewandowski, W. Wang, and S. Ethier, Princeton Plasma Physics Laboratory

Research Objectives
Our objective is improvement of the existing 3D Gyrokinetic Toroidal Code (GTC) by using split-weight df and hybrid schemes for the electrons for studying the trapped particle and finite-b effects. The improved GTC code will be used for studying neoclassical and turbulent transport in tokamaks and stellarators as well as for investigating hot-particle physics, toroidal Alfvén modes, and neoclassical tearing modes.

Computational Approach
The basic approach for transport studies in tokamaks and stellarators is the gyrokinetic particle-in-cell method. GTC, in general geometry with 3D numerical equilibria, is written in Fortran 90 with MPI/OpenMP and is scalable on MPP platforms. The inclusion of the vital electron dynamics makes the code computation-intensive because of the mass disparity between the electrons and the ions.

Accomplishments
We have carried out gyrokinetic particle simulation studies on turbulent transport using GTC and have gained substantial understanding of the zonal flow physics. For example, the interplay between the ion temperature gradient (ITG) driven turbulence, zonal flow generation, and the collisional effects in the simulation is shown to rise to the bursting behavior observed in Tokamak Fusion Test Reactor (TFTR) experiments. A full-torus simulation of turbulent transport in a reactor-sized plasma indicates a more favorable scaling of transport driven by ITG modes as the size of the plasma increases. Gyrokinetic calculations of the neoclassical radial electric field in stellarator plasmas have also been carried out with the GTC code, which is part of the PPPL's effort for designing the next generation of stellarators. In addition, we have developed several versions of the split-weight schemes for the electrons to account for the trapped-particle and finite-b effects.

Significance
A key issue in designing a fusion reactor is the realistic assessment of the level of turbulent transport for reactor-grade plasma conditions. Up until very recently, this has been done by extrapolating to larger reactors the transport properties observed in smaller experimental devices. This approach relies on models of transport scaling that have often stirred debates about reliability. Taking advantage of the power recently accessible in new supercomputer capabilities, we have been able to take a major step forward in understanding turbulent transport behavior in reactor-sized plasmas by using direct numerical simulations. These advanced simulations have just become feasible because of the recent development of better physics models and efficient numerical algorithms, along with the newly available 5 teraflop/s IBM SP at NERSC.

Publications
Z. Lin and L. Chen, "A kinetic-fluid hybrid electron model for electromagnetic simulations," Phys. Plasmas 8, 1447 (2001).

Liu Chen, Zhihong Lin, Roscoe B. White, and Fulvio Zonca, "Nonlinear zonal flow dynamics of drift and drift-Alfvén turbulences in tokamak plasmas," Nuclear Fusion 41, 747 (2001).

J. L. V. Lewandowski, A. H. Boozer, J. Williams, and Z. Lin, "Gyrokinetic calculations of the neoclassical radial electric field in stellarator plasmas," Phys. Plasmas 8, 2849 (2001).