Simulation of spatially correlated data in two dimensions

Martheron's turning bands method for the simulation of spatially correlated data in two or three dimensions requires that the relationship between the covariance obeyed by the realizations first generated on lines, and the covariance of the two or three dimensional process must be solved. In three dimensions the solution is immediate, but in two dimensions the integral equation relating the covariances of the one and two dimensional processes is more complex. This equation was developed and solved numerically by Brooker and Paul (1982). In this paper the analytic solution is presented and applied to the most common model used in ore deposit description.