1: \begin{abstract}

2: We propose a second-order temporally implicit, fourth-order-accurate

3: spatial discretization scheme for the strongly anisotropic heat transport

4: equation characteristic of hot, fusion-grade plasmas. Following {[}Du

5: Toit et al., \emph{Comp. Phys. Comm}., \textbf{228} (2018){]}, the

6: scheme transforms mixed-derivative diffusion fluxes (which are responsible

7: for the lack of a discrete maximum principle) into nonlinear advective

8: fluxes, amenable to nonlinear-solver-friendly monotonicity-preserving

9: limiters. The scheme enables accurate multi-dimensional heat transport

10: simulations with up to seven orders of magnitude of heat-transport-coefficient

11: anisotropies with low cross-field numerical error pollution and excellent

12: algorithmic performance, with the number of linear iterations scaling very weakly with grid resolution and grid anisotropy, and scaling with the square-root of

13: the implicit timestep. We propose a multigrid preconditioning strategy

14: based on a second-order-accurate approximation that renders the scheme

15: efficient and scalable under grid refinement. Several numerical tests

16: are presented that display the expected spatial convergence rates

17: and strong algorithmic performance, including fully nonlinear magnetohydrodynamics

18: simulations of kink instabilities in a Bennett pinch in 2D helical

19: geometry and of ITER in 3D toroidal geometry.

20: \end{abstract}

21: