Wavelet analysis of characteristic length scales and orientation of two-dimensional heterogeneous porous media

The global wavelet energy spectrum (GWES) has been used on one-dimensional transects of porous media to identify the characteristic length scale of a material property. The characteristic length scale is a measure of the distance over which property values are correlated. We extend the wavelet analysis technique to two-dimensional stationary porous media for identifying both the characteristic length scales and orientation of heterogeneity. We theoretically develop and evaluate the GWES for isotropic and anisotropic random fields using the two-dimensional Mexican hat wavelet. The relationship between the wavelet scale and the characteristic length scale of a two-dimensional random field is investigated for different covariance structures. The ability of the GWES to identify multiple characteristic length scales and orientations in a random field with nested covariance structures is also investigated. We use the technique to identify the characteristic length scales of a log hydraulic conductivity field used in a laboratory experiment.