ABSTRACT
A numerical solution points out that instead of differential equations in the partial derivatives of the second order, an analogous system of finite difference equations can be used. The numerical solution always contains errors related to the transform from differential presentation of the balance of element of the groundwater current to finite differential analogue. This evinces the need for investigating the so-called conditions of similarity of solution of a finite difference equation to the analogue obtained by the analytical method. Numerical solutions, analogous in content to the analytical, often contain these errors because they are related to the substitution of a differential equation by its finite difference analogue. The stability and convergence of numerical solutions depend on the ratio of the intervals along co-ordinates of time and space. Hence the specified boundaries, for example criterion, are superposed on the ratio of the intervals.
