A factor analysis of 67 landscape pattern metrics was performed to quantify the ability of landscape pattern metrics to explain land cover pattern, and to report individual landscape pattern metric values that are statistically independent. This land cover pattern is measured from 7.68 x 7.68 [km] GeoTiff image tiles of the conterminous United States Geological Survey (USGS) 1992 National Land Cover Dataset (NCLD). Using factor analysis to rank independent landscape pattern information, each landscape pattern metric produces the explanatory power of that landscape pattern metric amongst the other 66 landscape pattern metrics—any landscape pattern metrics that report similar values contribute redundant information. The metrics that contribute the most information are Jackson’s Contagion statistic (P005), typically contributing to 97 % of the explained variability; the Fourier Metric of Fragmentation (FMF), typically contributing to 65 % of the explained variability; and average LCLU class lacunarity (TLAC), typically contributing to 62 % of the explained variability. Two other Fourier-based landscape pattern metrics we tested, the Least Squares Fourier Transform Fractal Dimension Estimation (LsFT) and the Fourier Metric of Proportion (FMP), contributed 50 % and 12 % to the explained variability, respectively. In addition, the values reported by each of the Fourier metrics are revealed to be relatively independent amongst commonly used landscape pattern metrics and are thus demonstrated to be appropriate for explaining general landscape pattern variability.
