Reporting land degradation sensitivity through multiple indicators: Does scale matter?

Abstract

This work provides a multi-scale, multi-temporal assessment of the robustness of 6 indicators of land degradation aggregated at various spatial domains relevant to environmental reporting. Based on the Environmentally Sensitive Area (ESA) approach – widely used for environmental reporting of land degradation in Europe – we tested six indicators including (i) the average ESA score, (ii) the maximum ESA score, (iii) the coefficient of variation in the ESA scores, (iv) the normalized range in the ESA scores, as well as the extent of (v) ‘fragile’ and (vi) ‘critical’ areas based on a standard land classification developed on behalf of the ESA framework. Statistical robustness and intrinsic stability of these indicators were verified at six spatial domains (administrative regions, provinces, elevation belts, homogeneous economic districts, rural districts, municipalities) separately for three time points (early-1960s, early-1990s, and early-2010s). Results of a mixed parametric/non-parametric correlation analysis indicate that pair-wise relationships between indicators were mostly linear. A Principal Component Analysis identified two non-redundant dimensions associated with the average level of land degradation sensitivity and its intrinsic variability over space; indicators resulted to be associated exclusively with one of these two dimensions for all study years. Average level of sensitivity and variability over space provide, together, a comprehensive and statistically robust assessment of land degradation at vastly different planning levels, irrespective of the territorial domain adopted for environmental reporting.

Keywords

• Statistical robustness
• Multi-scale validation
• Environmental reporting
• Southern Europe