Use of simulated and real data to identify heterogeneity domains in scale-divergent forest landscapes

Abstract

It is widely recognized that there is no single scale at which a landscape or ecosystem should be studied, and forest planning and management require the definition of homogeneous units whose characteristics depend strongly on the scale at which their spatial arrangement occurs. Such “characteristic scales” range from fine to coarse, especially when planning efforts must be developed for broad spatial extensions, like forest areas. In such cases, the same analytical scale of analysis may be appropiate for some landscapes, while in others fine-scale features may be overlooked, or the scale may provide too much detail, depending on the characteristics of the spatial pattern. In this study, we aimed to develop a straightforward methodology to identify and discriminate among scale-divergent areas in landscapes represented by categorical maps. For this purpose, artificial landscapes were generated by use of a Modified Random Clusters method, and the Shannon-Wiener index was then applied at different scales by use of a moving-window approach. The results were analysed by contrasting the generation parameters with the statistical characteristics of the spatial pattern at each scale. This enabled us to relate the characteristic patterns to scale behaviour, and to define the minimum extension needed for a satisfactorily description of a landscape. The information obtained was then compared with real landscapes in order to validate the method, which we believe will facilitate the identification of homogeneous management units (in terms of spatial heterogeneity) and definition of the minimum area required for inclusion of essential descriptive elements.