Gully erosion is an important erosive process in Mediterranean basins. However, the long-term dynamics of gully networks and the variations in sediment production in gullies are not well known. Available studies are often conducted only over a few years, while many gully networks form, grow, and change in response to environmental and land use or management changes over a long period. In order to clarify the effect of these changes, it is important to analyse the evolution of the gully network with a high temporal resolution. This study aims at analysing gully morphodynamics over a long timescale (1956–2013) in a large Mediterranean area in order to quantify gully erosion processes and their contribution to overall sediment dynamics. A gully network of 20 km² located in southwestern Spain has been analysed using a sequence of 10 aerial photographs in the period 1956–2013. The extension of the gully network both increased and decreased in the study period. Gully drainage density varied between 1.93 km km⁻² in 1956, a minimum of 1.37 km km⁻² in 1980, and a maximum of 5.40 km km⁻² in 2013. The main controlling factor of gully activity appeared to be rainfall. Land use changes were found to have only a secondary effect. A new Monte Carlo-based approach was proposed to reconstruct gully erosion rates from orthophotos. Gully erosion rates were found to be relatively stable between 1956 and 2009, with a mean value of 11.2 t ha⁻¹ yr⁻¹. In the period 2009–2011, characterized by severe winter rainfalls, this value increased significantly to 591 t ha⁻¹ yr⁻¹. These results show that gully erosion rates are highly variable and that a simple interpolation between the starting and ending dates greatly underestimates gully contribution during certain years, such as, for example, between 2009 and 2011. This illustrates the importance of the methodology applied using a high temporal resolution of orthophotos.

Land use change (LUC) is identified as one of the main drivers of soil erosion in the Mediterranean. However, very little information exists regarding the relationship between land use and erosion over longer time periods and on regional scales. We quantified the LUC in Southern Spain between 1956 and 2018, examining its effect on soil erosion and assessing the mitigation role of the permanent grassland (PG). The land use influence on erosion is represented by the RUSLE's C-factor, which was modelled using the Monte Carlo Method (MCM) based on historical LUC. Moreover, future LUC scenarios by 2038 were developed by binary logistic model (scFS) and by a complete conversion of PG to cropland (scPC), permanent crop (scPP) and forest and natural (scFP). Historically, Southern Spain has experienced an impressive intensification of its agricultural system. While soil loss variation is noted within the classes, no big variation is observed in cumulative erosion on a regional scale. The underlying reasons for this resilience are multifold, but mainly attributed to the fact that a small fraction of the total surface (20%), dominates total erosion (67%). The C-factor decrease in this area displays a LUC towards forest and natural area, suggesting an agriculture abandonment. On the other hand, the agricultural intensification that has taken place in the remainder of the area, contributes much less to overall soil erosion. Future LUC scenarios illustrate the importance of PG for erosion mitigation. scFS scenario does not project major changes. However, scCP and scPP, show an abrupt increase in regional erosion by 13% and 14%, while scFP shows a negligible reduction of erosion close to 0%. This allows to quantify the erosion mitigation offered by maintaining the PG and should be taken into account for future agricultural policy.