The Evolving Land-Use Distribution Along Urban-Rural Gradients: Statistical Metrics Tell You More

Metropolitan expansion in Mediterranean regions has intensively transformed mixed urban–rural landscapes, fueling the expansion of dispersed settlement patterns, land take, and environmental degradation at large. The present study develops an exploratory framework assessing long-term landscape change in metropolitan Athens, Greece. It implements an indicator-based analysis of the statistical distribution of three land-use classes (built-up areas, agricultural land, and forests) at three observation times (1960, 1990, 2020). Distributions were investigated along a gradient of 115 municipalities moving from strictly urban (>10,000 inhabitants/km2) to strictly rural (<100 inhabitants/km2) locations, thus reflecting specific processes of urbanization-suburbanization and landscape transformation. Based on such assumptions, the study incorporates a dashboard of twenty (descriptive, distributional, and diversity) metrics initially analyzed through descriptive statistics and then summarized with multivariate approaches (namely, a Principal Component Analysis (PCA) identifying latent, non-redundant patterns of settlement development and landscape restructuring over time). A descriptive analysis of statistical metrics revealed a substantial increase in built-up land shares along the density gradient—shifting from 37.1% to 56.5% in total municipal area between 1960 and 2020, on average. A reverse, marked decline in forest land (from 21.8% to 4.7%, on average) and relatively stable agricultural land proportions were also observed in the municipal sample. Over time, built-up land became more homogeneous and evenly distributed across Athens’ municipalities, possibly reflecting (i) compact metropolitan expansion (1960), (ii) suburbanization (1990), and (iii) more mixed growth patterns in-between counter-urbanization and a weak re-urbanization (2020) with economic re-polarization in central and peripheral locations. In contrast, forest land exhibited increasing spatial concentration, variability, and statistical asymmetry, suggesting a progressive loss of natural areas with habitat fragmentation along the gradient, in line with the empirical results of earlier studies. Agricultural land maintained a comparatively stable statistical distribution across municipalities. This setting, however, is not enough to exclude the existence of mixed (and, possibly, non-linear) processes of cropland consumption over time. By explaining 88% of the total variance, PCA identified a dominant gradient discriminating balanced and mixed (agricultural and forest) landscapes with rather homogeneous statistical configurations (mainly representative of rural municipalities) from highly heterogeneous statistical configurations reflecting urban municipalities. Taken together, these findings highlight the coexistence of compact and dispersed processes underlying settlement expansion, demonstrating that metropolitan transformations are frequently driven by multiple spatial logics rather than individual (mostly stable over time) urbanization-suburbanization trajectories.