Abandonment of soil and water conservation structures in Mediterranean ecosystems A case study from south east Spain

Traditional rainfed agriculture in semi-arid regions heavily relies on soil and water conservation (SWC) structures to supplement the sparse rainfall. As referring to the ecosystem functions of these constructions, when extensive such systems prevent any runoff into the fluvial system. The extent to which these dams and terraces resist major events is variable, and earthen dams can be major sediment sources. Extensification and increasing mechanization of rainfed agriculture in marginal areas have led to a change in cropping systems. Large-scale almond and olive plantations with widely spaced trees do not rely on runoff water, but draw the soil water from a large soil volume of bare soil maintained by regular shallow tillage. The high density of terraces has now become a nuisance to the farmers. The aim of this paper is to i) demonstrate the degradation of SWC structures and the relative importance of the driving forces, ii) assess the limits of the protection that earthen dams can provide by surveying their resistance during a heavy storm (return period 8.2 years) and iii) demonstrate the implications of the abandonment of SWC structures over the period 1956 2005 for the hydrological connectivity between croplands and the ephemeral rivers system. The headwaters of a marl catchment with a continuous area treated with SWC structures in Murcia region (Spain) already had a very high density of step terraces and check dams (182 m ha(-1)) in 1956. This density decreased by 27% in the period 1956-2005. Furthermore, many terraces have not been maintained and flow traces indicate that they no longer retain water. This is particularly true for the check dams in abandoned lands. The distance between the step terraces has increased over time, making them vulnerable to erosion. The mean drainage area of the dams that failed during the heavy storm was significantly (3.16 ha) higher than that of the ones that remained intact (1.11 ha). The probability of failure increases with drainage area from P=0.16 for an area of 1.8 ha to P=0.8 for an area of 20 ha. The percentage of cropland draining directly on the river system without interference of a check dam has increased from just 9% in 1956 to 31% in 2005 and 40% after the storm in November 2006. (C) 2008 Elsevier B. V. All rights reserved