Empirical mode decomposition (EMD) has been used to separate the spatial variability in soil properties at different scales in one dimension. The objective of this note is to illustrate the use of a two-dimensional extension of the EMD (known as bi-dimensional empirical mode decomposition or BEMD) to separate the spatial variability at different scales. A digital elevation model (DEM) was used as an example to demonstrate the method. The BEMD separated the overall variation into a number of components according to the dominant spatial scales. The variance contributions of each component towards the total variance indicated their importance which can help choose the important scale components for various applications including digital soil mapping.