Soil erosion is widely recognized as one of the most significant forms of soil degradation worldwide. The Mediterranean region is particularly vulnerable due to a confluence of natural and anthropogenic factors. Natural drivers such as sparse vegetation cover, low soil structural stability, steep slopes, and intense rainstorms are compounded by human activities including land cover change, forest fires, intensive grazing, and soil tillage practices, all of which exacerbate erosion risks.

Among Mediterranean agricultural systems, olive groves (Olea europaea) stand out both economically and culturally, with more than 95 % of global olive production concentrated in the Mediterranean basin. This importance is visually represented in Fig. 1, which maps the distribution of olive orchards across the European Mediterranean, highlighting the crop's dominance in southern Spain, Italy, and Greece. However, these groves are frequently situated on marginal, low-fertility, and steeply sloping land, where soil erosion constitutes a major threat to their long-term sustainability (Gómez et al., 2009a; Vanwalleghem et al., 2010). It is essential to recognize that the primary driver of this degradation is not the olive tree itself or the local conditions, but the conventional soil management practices associated with its cultivation. Both traditional and modern olive farming has been characterized by the systematic removal of competing vegetation through frequent mechanical tillage (Fig. 2), which degrades soil structure and leaves the ground surface bare and vulnerable (Álvarez et al., 2007; Gómez et al., 2004). For instance, in Spain, despite the promotion of conservation agriculture, over 50 % of the olive growing area still lacks vegetation cover, with the majority of protected land relying on spontaneous rather than cover crops (MAPA, 2024). A similar pattern is observed in Italy, where tillage remains the predominant practice in conventional orchards (ISMEA, 2025), with permanent vegetation cover largely restricted to the ∼ 25 % of the olive-growing area managed under organic farming protocols or specific agri-environmental schemes. This practice of maintaining bare soil is deeply rooted in a cultural identity where a “clean” tilled, weed-free field is perceived as signs of diligent farming, while the presence of groundcover is seen as neglect (Rodrigo-Comino et al., 2020; Sastre et al., 2017). When this practice is combined with the common siting of olive groves on steep, erodible slopes, the conditions for severe soil erosion are perfected.

Accurate, evidence-based knowledge of erosion rates is essential for defining effective soil conservation policies. However, the scientific literature on soil erosion in olive groves is marked by significant debate and seemingly contradictory findings. A frequently cited soil loss estimate of 80 t ha⁻¹ yr⁻¹ for south Spain groves is based on USLE model estimates (López-Cuervo, 1990). This very high soil loss rate is supported by long-term estimates based on soil truncation methods, in particular, on tree mound measurements in Jordan, 132 t ha⁻¹ yr⁻¹ (Kraushaar et al., 2014), and in South Spain, 184 t ha⁻¹ yr⁻¹ (Vanwalleghem et al., 2010) and on fallout radionuclides in Spain, 75 t ha⁻¹ yr⁻¹ (García-Gamero et al., 2024a) and runoff plot studies in Greece, 56 t ha⁻¹ yr⁻¹ (Koulouri and Giourga, 2007) and in Spain 60 t ha⁻¹ yr⁻¹ (Gómez et al., 2017). These figures paint an alarming picture of an agroecosystem in crisis. In contrast, (Fleskens and Stroosnijder, 2007) argue that average rates rarely exceed 10 t ha⁻¹ yr⁻¹. In response, Gómez et al. (2008) criticized Fleskens and Stroosnijder's (2007) incomplete interpretation and conclusions drawn from short-term plot-scale experiments. In any case, erosion rates far exceed natural soil formation rates, depleting this vital resource (Huber et al., 2008)

Understanding soil erosion in olive groves is complicated by the wide array of methods used for its quantification, each with inherent strengths, limitations, and, most critically, different spatial and temporal scales of operation. It must be noted that most, if not all, of these methods are standard approaches widely applied for soil erosion assessment, not only in olive groves but also across a broad range of agricultural systems. These methods can be broadly classified into field measurements and predictive models (Table 1). Field measurements provide direct empirical data but vary significantly in what they measure. A crucial distinction must be made between methods that estimate gross soil loss, i.e. the total amount of soil detached and transported from a specific area, and those that estimate net soil loss, which accounts for both erosion and deposition within a larger landscape unit. Additionally, some methods focus on sediment yield, which quantifies the amount of eroded soil that actually exits the catchment or watershed, typically measured at the outlet. Field measurements can be divided into runoff simulations (Palese et al., 2015; Repullo-Ruibérriz De Torres et al., 2018), runoff plots (Espejo-Pérez et al., 2013), soil truncation studies based on fallout radionuclides (FRN) (Gdiri et al., 2024; Mabit et al., 2012) and tree mound measurements (Kraushaar et al., 2014; Vanwalleghem et al., 2010), and sediment yield measurements at the catchment outlet (Gómez et al., 2014; Taguas et al., 2013). Small spatial scale studies, such as rainfall simulation and runoff plots, tend to miss key catchment erosion processes such as rill and gully formation, tillage erosion, and sedimentation within fields. Nevertheless, runoff plots are uniquely capable of capturing temporal variability at a high resolution, enabling the detailed analysis of erosive responses to individual storm events. A significant constraint, however, is their typically brief experimental duration – often under a decade – however, due to their typically limited experimental duration (often less than 10 years), these methods generally fail to capture the cumulative impacts of long-term land management changes. Long-term historical methods, such as tree mound measurements and fallout radionuclide-based estimates, can capture catchment erosion processes and cumulative effects, such as the long-term effects of land use change or soil conservation practices, but are inadequate for capturing the temporal variability and episodic high-intensity events typical of Mediterranean climates. No single method provides a complete picture; rather, each offers unique insights depending on the scale and timeframe of analysis (Fig. 3).

Models allow for long-term, large-scale predictions with resolutions that exceed the limitations of field experiments. However, their reliability is strongly tied to input data quality. Improper calibration and evaluation, or application beyond a model's original scope, often lead to misleading results, highlighting the “garbage in, garbage out” principle. Some examples of models applied to simulate soil erosion in olive groves are AnnAGNPS (Bingner and Theurer, 2001), WaTEM/SEDEM (Van Oost et al., 2000), SEDD (Ferro and Minacapilli, 1995) and the Revised Universal Soil Loss Equation (RUSLE) (Renard, 1997). RUSLE is the most widely used model, however it was designed to be applied at the plot scale and hence, it only estimates gross soil loss. When its application is upscaled to larger areas without accounting for deposition, the accuracy of its predictions can be significantly compromised (Boix-Fayos et al., 2006; De Vente and Poesen, 2005). Model validation/evaluation is particularly challenging due to scale mismatches and the scarcity of long-term, high-quality data. Moreover, the inconsistency in field measurement results significantly limits the evaluation of models, making it difficult to ascertain if a model behaves as expected or if its outputs align with real-world observations under comparable conditions. Therefore, investing in long-term monitoring and consistent data is not just an academic pursuit but a prerequisite for developing reliable, policy-relevant modelling tools.

This review analyses existing studies on soil erosion in Mediterranean olive groves, grouping them by measurement methodology, and hence in similar spatial and temporal scales. The aim is to discern clearer patterns and trends that are often obscured by the overall heterogeneity of the available data, thereby addressing a significant challenge in the current scientific literature. This will provide general findings to evaluate model performance and assess the effectiveness of current management practices, ultimately contributing to more robust conservation strategies. This systematic approach will also address key research questions, including: What are the typical soil loss rates and runoff ratios in Mediterranean olive groves? What is the influence of factors such as topography, soil, vegetation, and climate on soil loss and runoff generation? What is the impact of soil conservation practices?

Soil erosion is an inherently scale-dependent process, and no single method or metric can capture its full complexity. The ongoing debate between “alarmist” and “non-alarmist” interpretations (Fleskens and Stroosnijder, 2007; Gómez et al., 2008) of erosion severity is, at its core, a debate about the scale of truth. Different measurement methods target different parts (Fig. 6) of the erosion–transport–deposition continuum, leading to conflicting figures that are, in fact, complementary. To develop a realistic and comprehensive understanding of soil erosion, particularly in agricultural landscapes, it is essential to adopt a multi-method, multi-scale approach. Each method provides a partial view, thus reveals a different “truth”, emphasizing different spatial and temporal aspects of erosion dynamics:

Runoff plots measure what's being mobilized. Runoff plot studies are ideal for capturing gross soil loss in upslope areas where contributing areas are small and deposition minimal (Francia Martínez et al., 2006; Gómez et al., 2004). These plots are valuable for comparing land management practices and assessing soil susceptibility to detachment. However, they only represent the initial phase of the erosion process and typically underestimate the cumulative effects of long-term processes like tillage erosion or gully expansion.

Models, when properly calibrated and validated, serve as critical tools for bridging these scales and for conducting future scenario analysis. However, their reliability depends on the availability of multi-decadal observational data, which is essential to ensure observational records of sufficient duration to support distinct calibration and verification phases and to capture the long-term influence of land management on erosion dynamics. Soil truncation estimates, derived from Fallout Radionuclides (FRNs) or tree mound measurements, provide this necessary long-term data, yet they carry inherent uncertainties that are frequently overlooked (see Sect. 3.2). These inconsistencies not only reduce the precision of the erosion estimates themselves but also propagate uncertainty into the model calibration process, potentially undermining the reliability of long-term simulations. This creates a challenging feedback loop: models cannot be reliably validated without robust, long-term field data, and the available long-term field data remain uncertain. Therefore, the scientific community must prioritize generating a larger volume of long-term estimates while rigorously standardizing these methodologies to reduce uncertainty. Only by securing “more and better” calibration data can we ensure that model projections accurately reflect future erosion risks.

Sediment fingerprinting (Davis and Fox, 2009) and radiometric dating of deposited sediments (Smith et al., 2018) are methodologies not applied in the reviewed studies, but which could nonetheless provide valuable long-term information. Sediment fingerprinting primarily yields estimates of relative sediment source contributions rather than absolute soil loss rates; however, when combined with radiometric dating of depositional archives, it has the potential to link erosional processes across different spatial and temporal scales. Despite this potential, their application in Mediterranean olive groves is subject to important limitations. A key constraint is the scarcity of stable, long-term sedimentary archives: with the exception of water reservoirs, permanent water bodies are rare in these semi-arid landscapes, while alternative depositional environments, such as footslopes or alluvial fans, are frequently disturbed by tillage or re-incised by active gullying (Leenman and Eaton, 2022). Moreover, the widespread occurrence of gully erosion complicates the use of methods such as fallout radionuclides (FRNs). Gullies mobilize deep subsoil that is depleted in FRNs, generating a “dilution effect” that can obscure the signal of topsoil erosion in depositional zones. Consequently, observed sedimentation rates may be dominated by episodic channel or gully incision rather than reflecting diffuse hillslope erosion processes. Further research is therefore required to evaluate the feasibility, reliability, and methodological adaptations needed for applying sediment fingerprinting and radiometric approaches in Mediterranean olive grove catchments.

Ultimately, embracing a multi-scale, integrated approach is not just a methodological choice, it's a necessity. It allows us to capture both the localized detachment and the landscape-level sediment delivery and is critical for the design of effective soil conservation strategies and for the calibration and validation of erosion models. For example, in upslope areas close to the catchment boundaries, where deposition and the upslope contributing area are minimal (net soil loss closely approximated gross soil detachment), we could consider combining runoff plot estimates (representing gross soil loss due to water erosion) with soil truncation estimates (representing net soil loss due to both water and tillage erosion). The difference between the net soil loss (derived from soil truncation methods) and the gross soil loss (derived from runoff plots) in the same location could provide an inference of the tillage erosion contribution. This is because soil truncation methods inherently capture the cumulative effects of both water and tillage erosion over longer timescales, while runoff plots primarily isolate the detachment and transport by water. This can be conceptually represented as: Tillage Erosion Contribution≈Net Soil Loss (Soil Truncation)-Gross Soil Loss (Runoff Plots) However, comparing multi-decadal ST data with short-term RP data carries a significant temporal mismatch, primarily because short-term plots often miss the extreme events that log-term ST captures. Nevertheless, this conceptual comparison becomes quantitatively more robust in specific scenarios where timescales align – for instance, when mound measurements are taken on younger trees (e.g., <20 years) that match the duration of long-term runoff studies. This would help to disentangle the significant, yet often overlooked, role of tillage in overall soil displacement within agricultural landscapes as well as to calibrate and evaluate tillage erosion models. Only by acknowledging the complexity and scale-dependency of soil erosion can we resolve inconsistencies in the literature and move toward more sustainable land management.

Yet, this scientific understanding uncovers a critical paradox in olive cultivation, a “silent crisis” where olive yields have increased, mainly driven by the mechanization of cultivation and the increase of tree density (Amate et al., 2013), despite increasing long term soil erosion and ongoing degradation (Peñuela et al., 2023). The ability of deep soils to buffer initial losses, combined with management practices such as enhanced fertilization, pruning, and pest control, has effectively masked the unsustainability of current practices (Tubeileh et al., 2014). This absence of a negative impact on yield has meant there's been no immediate or direct incentive for farmers to adopt soil management practices that prioritize conservation, allowing prolonged unsustainable practices to continue. The current system is drawing down a vital natural capital, soil, without immediate visible consequences, but with severe long-term implications for future generations. This situation urgently calls for a fundamental paradigm shift in how agricultural success is defined and measured, emphasizing long-term ecological resilience alongside productivity. Therefore, policy and farmer education must move beyond short-term yield metrics to incorporate and prioritize long-term soil health indicators such as the Soil Footprint (García-Gamero et al., 2024b).

The evidence supports that increasing vegetation/ground cover between olive trees as the most effective strategy for erosion control in olive groves. Their multifaceted benefits, including significant erosion reduction, enhanced infiltration, increased organic matter content, and improved soil aggregation, simultaneously address several key drivers of degradation (Gómez et al., 2009a, b; Márquez-García et al., 2024; Repullo-Ruibérriz De Torres et al., 2018). Despite the acknowledged challenge of water competition, the magnitude of erosion reduction achieved suggests that the benefits often outweigh the risks, especially with careful species selection and adaptive management (Gómez et al., 2009b). Even the “non-alarmist” average gross soil loss rates measured in runoff plots (5.5 t ha⁻¹ yr⁻¹) exceed the upper limit of tolerable loss (0.3–1.4 t ha⁻¹ yr⁻¹; Verheijen et al., 2009) by nearly 400 %. This implies that soil conservation practices such as cover crops are not merely “an option” but represent a fundamental requirement for achieving long-term sustainability in Mediterranean olive groves (Bombino et al., 2021). Therefore, policy incentives and research efforts should prioritize the widespread adoption and optimization of soil conservation strategies, including the development of drought-tolerant cover crops species and adaptive management strategies designed to minimize water competition during critical dry periods.

Further research integrating multi-scale or multi-proxy field monitoring with robust model calibration and validation across a wider range of environmental and management conditions is essential to accurately quantify erosion risks and develop effective and sustainable soil management strategies for Mediterranean olive groves. The path to sustainable olive cultivation lies in a paradigm shift towards evidence-based management strategies. Prioritizing soil conservation strategies, minimizing intensive tillage and maximizing vegetation/ground cover are paramount. These practices not only effectively reduce soil and nutrient loss but also enhance soil health and resilience. However, successful adoption hinges on addressing socio-economic barriers, including perceived water competition, management costs, and traditional biases (Rodrigo-Comino et al., 2020; Sastre et al., 2017).

In this context, policy support becomes a decisive factor in shifting current management paradigms toward more sustainable practices. Conservation Agriculture (CA) (Gonzalez-Sanchez et al., 2015), through its emphasis on permanent groundcovers and no tillage (FAO, 2022), offers a robust framework for mitigating erosion and restoring soil functionality in Mediterranean perennial systems such as olive groves. Despite the clear effectiveness of cover crops in reducing erosion rates (often by an order of magnitude, see Table 2), widespread adoption remains limited, with over 50 % of the olive-growing area in major producing countries like Spain still maintained as bare soil (MAPA, 2024). Consequently, successful implementation requires adequate training, cross-compliance mechanisms, incentives, regulatory support, and integration into agricultural subsidy frameworks. The successful implementation of CA requires not only technical knowledge but also institutional alignment and policy coherence at local and national levels. Cross-compliance serves as the necessary regulatory baseline, ensuring that eligibility for public support is conditional upon avoiding the most harmful practices, such as maintaining bare soil on steep slopes. This “compliance push” creates a universal minimum standard. Simultaneously, financial incentives – such as eco-schemes under the Common Agricultural Policy – provide the necessary “economic pull”, compensating farmers for the opportunity costs and technical risks associated with adopting active conservation strategies. In regions where olive cultivation is dominant, promoting CA through targeted programs can serve as a powerful lever to reduce erosion, combat desertification, and strengthen the resilience of rural landscapes.

This literature review has synthesized a broad spectrum of research spanning from plot-scale field experiments to catchment-level monitoring and long-term soil truncation estimates. These are the main conclusions:

While natural factors such as topography, rainfall and soil properties establish a baseline risk, the evidence is unequivocal that agricultural management is the pivotal factor controlling soil degradation.