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.

Mediterranean ecosystems are commonly vulnerable to wildfires. Accelerated erosion processes due to wildfires in those environments constitute a major restrictive factor in their sustainability. This study aims at evaluating the use of the Revised Universal Soil Loss Equation (RUSLE) and the Pan-European Soil Erosion Risk Assessment (PESERA) models in predicting the changes in spatial variability of soil erosion following a wildfire event. A site in Greece on which a wildfire occurred in the summer of 2007 is used as a case study. Soil erosion rates for the site before and after the fire outbreak were estimated by the two models. Inputs for both models included climatic, land-use, soil type, topography, Earth Observation (EO) as well as management and other ancillary data. Both models showed a substantial increase of soil erosion rates in the affected area as a result of the fire, particularly towards the steeper slopes and on areas of high burning severity. Yet, there were noticeable differences in the predictions between the 2 models in terms of absolute estimates of soil erosion rates before and after the fire event. Mean pre-fire erosion rates from RUSLE were~2.5 times higher than those from PESERA. RUSLE predicted considerably higher mean erosion in comparison to PESERA for the post-fire conditions, yet of much less spatial variability. Average post-fire soil loss value, compared to pre-fire, was about nine and six times greater when using the RUSLE and the PESERA model respectively. RUSLE predictions were most sensitive to topographic and rainfall erosivity factors. PESERA showed high sensitivity to the vegetation coverage as well as to the soil characteristics inserted as crusting and erodibility variables. To our knowledge, this is the first study performing an intercomparison of soil erosion rate predictions between these two models, particularly so in the context of the influence of wildfire events. This study provides a key contribution towards our ability to better understand the effect of fire on soil erosion in the Mediterranean and elsewhere, as well as the ability of those widely used models as efficient tools to be used for this purpose. The latter is of key significance and practical value for research and policy decision making purposes alike, where information on spatiotemporal estimates of soil erosion rates may be required.

The need for reliable estimates of soil loss under different land management practices (LMPs) is becoming imperative in the Mediterranean basin to inform decisions on more effective strategies for land management. The effect of LMPs on soil erosion and land degradation has been investigated using experiments from November 2008 to November 2011 in an olive grove in central Crete (Greece). The study area was on sloping land with soils formed on marl deposits which are vulnerable to desertification because of surface runoff and tillage. The experimental design included three treatments with two replicates (3 × 5 m experimental plots) corresponding to the following LMPs: (i) no tillage–no herbicide application, (ii) no tillage–herbicide application and (iii) ploughing to 20 cm perpendicular to the contours. The following variables were monitored: surface water runoff, sediment loss, soil temperature at 10 cm, soil moisture content at depths of 20 and 50 cm, as well as selected climatic variables. The results show that the no tillage–no herbicide management practice gave the lowest sediment loss (1.44–4.78 g/m2/yr), the lowest water runoff (1.8–11.5 mm/yr), the greatest amount of water stored in the soil, the lowest soil temperature and the lowest desertification risk compared with the other treatments. Tillage resulted in the greatest sediment loss (13.6–39.2 g/m2/yr) and surface runoff (16.5–65.0 mm/yr), and an intermediate amount of water stored in the soil. In addition, this treatment led to the loss of soil thickness of 3.7 mm/yr because of ploughing. The results demonstrate the high risk of desertification in the investigated region and the methodology can be used in other Mediterranean areas as an assessment framework for evaluating land degradation and the impact of land management on soil erosion.

No abstract

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.