Assessing the effectiveness of sustainable land management policies for combating desertification: A data mining approach.

This study investigates the relationship between fine resolution, local-scale biophysical and socioeconomic contexts within which land degradation occurs, and the human responses to it. The research draws on experimental data collected under different territorial and socioeconomic conditions at 586 field sites in five Mediterranean countries (Spain, Greece, Turkey, Tunisia and Morocco). We assess the level of desertification risk under various land management practices (terracing, grazing control, prevention of wildland fires, soil erosion control measures, soil water conservation measures, sustainable farming practices, land protection measures and financial subsidies) taken as possible responses to land degradation. A data mining approach, incorporating principal component analysis, non-parametric correlations, multiple regression and canonical analysis, was developed to identify the spatial relationship between land management conditions, the socioeconomic and environmental context (described using 40 biophysical and socioeconomic indicators) and desertification risk. Our analysis identified a number of distinct relationships between the level of desertification experienced and the underlying socioeconomic context, suggesting that the effectiveness of responses to land degradation is strictly dependent on the local biophysical and socioeconomic context. Assessing the latent relationship between land management practices and the biophysical/socioeconomic attributes characterizing areas exposed to different levels of desertification risk proved to be an indirect measure of the effectiveness of field actions contrasting land degradation.

Evaluation and selection of indicators for land degradation and desertification monitoring: methodological approach.

An approach to derive relationships for defining land degradation and desertification risk and developing appropriate tools for assessing the effectiveness of the various land management practices using indicators is presented in the present paper. In order to investigate which indicators are most effective in assessing the level of desertification risk, a total of 70 candidate indicators was selected providing information for the biophysical environment, socio-economic conditions, and land management characteristics. The indicators were defined in 1,672 field sites located in 17 study areas in the Mediterranean region, Eastern Europe, Latin America, Africa, and Asia. Based on an existing geo-referenced database, classes were designated for each indicator and a sensitivity score to desertification was assigned to each class based on existing research. The obtained data were analyzed for the various processes of land degradation at farm level. The derived methodology was assessed using independent indicators, such as the measured soil erosion rate, and the organic matter content of the soil. Based on regression analyses, the collected indicator set can be reduced to a number of effective indicators ranging from 8 to 17 in the various processes of land degradation. Among the most important indicators identified as affecting land degradation and desertification risk were rain seasonality, slope gradient, plant cover, rate of land abandonment, land-use intensity, and the level of policy implementation.

Evaluation and selection of indicators for land degradation and desertification monitoring: types of degradation, causes, and implications for management.

Indicator-based approaches are often used to monitor land degradation and desertification from the global to the very local scale. However, there is still little agreement on which indicators may best reflect both status and trends of these phenomena. In this study, various processes of land degradation and desertification have been analyzed in 17 study sites around the world using a wide set of biophysical and socioeconomic indicators. The database described earlier in this issue by Kosmas and others (Environ Manage, 2013) for defining desertification risk was further analyzed to define the most important indicators related to the following degradation processes: water erosion in various land uses, tillage erosion, soil salinization, water stress, forest fires, and overgrazing. A correlation analysis was applied to the selected indicators in order to identify the most important variables contributing to each land degradation process. The analysis indicates that the most important indicators are: (i) rain seasonality affecting water erosion, water stress, and forest fires, (ii) slope gradient affecting water erosion, tillage erosion and water stress, and (iii) water scarcity soil salinization, water stress, and forest fires. Implementation of existing regulations or policies concerned with resources development and environmental sustainability was identified as the most important indicator of land protection.