1: \begin{abstract}

2: Numerical solutions of

3: stationary diffusion equations on the unit sphere with isotropic 

4: lognormal diffusion coefficients are considered.

5: H\"older regularity in $L^p$ sense for isotropic Gaussian random fields is obtained and 

6: related to the regularity of the driving lognormal coefficients. This yields regularity in $L^p$ sense 

7: of the solution to the diffusion problem in Sobolev spaces.

8: Convergence rate estimates of 

9: multilevel Monte Carlo Finite and Spectral Element discretizations 

10: of these problems are then deduced. 

11: Specifically, a convergence analysis 

12: is provided with convergence rate estimates in terms of

13: the number of Monte Carlo samples of the solution to the considered diffusion equation

14: and 

15: in terms of the total number of degrees of freedom of the spatial discretization,

16: and with bounds for the total work 

17: required by the algorithm in the case of Finite Element discretizations.

18: The obtained convergence rates

19: are solely in terms of the decay of the angular power spectrum

20: of the (logarithm) of the diffusion coefficient.

21: Numerical examples confirm the presented theory.

22: \end{abstract}

23: