Landscape indices behavior: A review of scale effects

Various landscape metrics have been used in the past to provide a quantitative description of landscape structure. Many authors have pointed out that the scale of the input data and the scale of the analysis must be coherent in order to calculate and interpret landscape metrics correctly. The issue of scaling is related to the grain size (pixel size), or the minimum mapping unit, and the extent of the study area. Recently, it has been demonstrated that the thematic resolution also influences the values of landscape metrics. The responses of landscape metrics to changes in the spatial extent, grain size and thematic resolution of landscapes has therefore become an important topic in landscape ecology studies in recent decades due to advances in computer technology. Several reviews have discussed the utilization of indices in landscape ecology and ecological research. However, there is no overview of the sensitivities of the various indices to either spatial or thematic scaling, and there are no recommendations available concerning the use of indices with reference to these types of scaling. In this study, we review the behavior of the indices commonly used for quantifying landscape structure. We assess the effects of changes in the spatial and thematic scales on the values of these indices and suggest directions for further research on the behaviors of these indices. Based on this review, we conclude that simple and easily interpretable indices with predictable reactions to changes in scale, e.g., the number of patches (NP), patch density (PD), edge density (ED), patch richness (PR), and mean patch size (MPS) of the landscape, are the most suitable indices to use for the assessment of landscape structure. ► We have reviewed papers on behavior of diverse landscape indices. ► We focused on the indices behavior in terms of spatial and thematic resolution. ► We found that the authors differ in the results achieved. ► We bring an overview of these differences. ► We highlight simple indices with predictable behavior.