Quantifying the effectiveness of mountain terraces on soil erosion protection with sediment traps and dry-stone wall laser scans

Mountain depopulation in the Mediterranean region over the past decades has led to a decline in the use and maintenance of agricultural terraces and consequently the collapsing of dry-stone walls, which can increase soil erosion rates and downstream sedimentation. A field experiment has been set up on a degrading terraced hillslope in the Troodos Mountains of Cyprus, to quantify the effectiveness of terrace maintenance on protecting cultivated land against soil erosion. The monitored site is cultivated with grapes. The terrace riser (22 m long) that forms the linear outlet of the hillslope has 11.4 m of standing dry-stone wall and 10.6 m of collapsed wall. It has been instrumented with seven 1 m wide sediment traps, three on standing sections of the wall and four on collapsed sections. When dry, sediment was collected from the traps after rainfall events, from December 2015 to November 2017. Uncertainties in the drainage areas of the 31.5-m long slope were quantified both for the terrace wall and for the individual traps through hydrologic delineations based on a detailed topographic survey. The sediment data were complemented by laser scanner surveys that were conducted in November 2015, May 2016 and April 2017, on a dry-stone terrace wall upslope from the outlet section. Wall degradation was assessed from the consecutive 3D model reconstructions. Rainfall was 469 mm in the first year and 515 mm in the second year and the average erosivity was 1148 MJ mm ha−1 h−1 y−1. The average soil erosion rate was 2.4 Mg ha−1 y−1, when linear drainage areas are considered (693 m2), 3.2 Mg ha−1 y−1 when the borders are delineated with the topographic data (520 m2). Nearly half of the soil erosion (43%) occurred during two very intense rainfall events (maximum 30-min intensity exceeding 35 mm h−1), out of the 34 monitored events. Erosion from standing terrace sections was 3.8 less than the erosion from the collapsed sections. For the scanned terrace wall, soil erosion from the standing sections was 2.2 lower than from the degraded sections. The laser scanner surveys identified some preferential erosion paths, but failed to recognize single stone collapses, whereas possible wall displacement was masked by scanning artifacts. The sediment traps were found to be an effective method for understanding and quantifying soil erosion in terraced mountain environments, while further research is needed to develop a more rigorous acquisition procedure for laser scanner surveys to derive useful information on wall degradation.