Comparing Direct Field Measurements of Soil Erosion with RUSLE Model Estimates in Mediterranean Olive Orchards ^†

Abstract

Soil erosion is a major threat to land productivity and environmental sustainability in Mediterranean regions, where sloping terrain, intense seasonal rainfall, and traditional agricultural practices accelerate soil loss. Olive orchards, which dominate much of the Mediterranean landscape, are particularly vulnerable. As climate change increases the frequency of extreme weather events, understanding and controlling erosion becomes even more critical. This study investigates soil erosion dynamics in a representative olive-growing watershed in Messenia, Greece, by combining field monitoring with erosion modeling using the Revised Universal Soil Loss Equation (RUSLE). A field experiment was carried out during the 2024–2025 wet season, using runoff plots installed on a 16% slope to directly measure sediment loss from natural rainfall events. The observed erosion data served as a basis for calibrating a GIS-based RUSLE model applied across the 60 km² watershed. Model predictions showed strong agreement with field measurements, with estimated soil loss closely matching the observed seasonal total (~0.6 t/ha). This consistency demonstrates the reliability of the RUSLE model when supported by localized data. The spatial analysis further revealed that erosion risk varies widely across the landscape, with steep, poorly vegetated areas being most at risk. The results highlight the importance of local field measurements for improving model accuracy and guiding sustainable land management. Continuous monitoring and targeted erosion control strategies are essential to protect soil resources, maintain agricultural productivity, and reduce downstream environmental impacts under increasing climate pressures.

1. Introduction

Soil erosion is a critical environmental challenge in Mediterranean regions [1], where steep terrain, intense seasonal rainfall, and unsustainable farming practices combine to accelerate soil loss [2]. Climate change is expected to compound this problem by increasing the frequency of extreme weather events with prolonged droughts followed by intense rainfalls that increase runoff and erosion. Olive orchards on sloping lands are particularly vulnerable [3] and traditional practices such as frequent plowing [4] or herbicide use [5] leave the soil exposed, leading to reduced fertility, lower productivity, and greater flood risk. In contrast, conservation-oriented practices like maintaining vegetative cover can significantly mitigate erosion by protecting the soil surface and improving infiltration. Many studies in Mediterranean olive groves have found that introducing cover crops can reduce annual soil loss compared to bare, tilled plots, without notably harming olive yields in many cases while enhancing soil health [6,7]. These findings underscore the importance of sustainable land management to preserve soils in olive-growing regions.

Quantifying soil erosion is therefore essential for assessing the scale of the problem and designing appropriate conservation strategies. Several approaches are available like field measurements that provide direct and reliable evidence of soil loss or empirical models that can extend estimates across larger areas [8,9]. The Revised Universal Soil Loss Equation (RUSLE) [10] is widely used to estimate long-term soil erosion rates by integrating key factors of erosion: rainfall erosivity (R), soil erodibility (K), topography (slope length and steepness, LS), land cover (C), and conservation practices (P). RUSLE provides a convenient empirical framework and has been applied at catchment scales using GIS data for each factor [11]. However, applying RUSLE with generic parameters may yield uncertain results under Mediterranean conditions, making calibration with local observations an important step [8].

This study evaluates the performance of RUSLE in a Mediterranean olive orchard watershed in Messenia (Peloponnese, Greece). The study area is the Xerias River watershed in the Municipality of Pylos-Nestor, covering a typical hilly terrain in NW Peloponnese with elevations up to ~300 m and an average hillslope of about 16%. Messenia has a Mediterranean climate, characterized by mild, wet winters and dry summers [12]. Olive cultivation is widespread on the slopes, and conventional management often involves bare soil under trees [13]. A field experiment was implemented across three plots (each 100 m²) within an olive orchard on a representative 16% slope. To directly measure soil loss, runoff collection systems were installed in the three plots, capturing runoff and sediment yield from natural rainfall events. By integrating empirical field observations with spatial RUSLE modeling, the study provides a comprehensive understanding of how slope, land cover, and agricultural practices interact to influence soil erosion in Mediterranean olive orchards. The findings are expected to enhance the reliability of RUSLE-based erosion estimates under Mediterranean conditions, contributing to the broader goal of mitigating land degradation and preserving productive soils in the face of climate variability and increasing human pressures.

2. Materials and Methods

The research took place in the Xerias River watershed (about 60 km²) in the Pylos-Nestor region of Messenia, Greece (Figure 1). The area is a typical Mediterranean landscape with rolling hills and streams carved into the land that flow west to the Ionian Sea. Soils in the watershed are mainly calcareous Mediterranean soils (loams and clay loams) developed on marl and flysch parent material, with moderate to high erodibility. The region’s climate has mean annual rainfall on the order of 600–700 mm falling mostly between October and April. Summers are dry, which means vegetation cover plays a crucial role in protecting soils during winter rains. Land cover is dominated by olive orchards alongside patches of shrubland and sclerophyllous forests. Traditional olive groves here often have widely spaced trees with little ground cover. Significant portions of the sloping farmland are terraced or contoured, reflecting past efforts to control runoff.

Within an olive grove on a ~16% slope, three experimental plots (10 m × 10 m each) were established to monitor soil erosion. Each plot was equipped with a runoff collection system at its downslope edge, consisting of a channel connected to a storage tank to capture runoff and sediment during natural rainfall events. Monitoring was carried out during the 2023–2024 wet season (October–April), with runoff volume and dried sediment mass measured after each major rainstorm. These measurements were then converted into soil loss rates, expressed in tonnes per hectare per event and for the seasonal total. The resulting dataset provided a direct empirical basis for calibrating the RUSLE model and assessing erosion under representative orchard conditions.

RUSLE model equation was then implemented in a GIS environment to produce a soil erosion risk map for the entire Xerias watershed. RUSLE computes average annual soil loss A (t/ha·yr) as:

A = R × K × LS × C × P

where:

• R is the rainfall erosivity factor,
• K is the soil erodibility factor,
• LS is the topographic slope length-steepness factor,
• C is the cover-management factor,
• P is the support practice factor.

Each factor was derived as a raster map at 10 m resolution (projection: WGS84/UTM) and multiplied to obtain the spatial distribution of A. Workflow of RUSLE model is presented in Figure 2.

The factor maps were prepared as follows:

2.1. Rainfall Erosivity (R)

Rainfall erosivity expresses the effect of raindrop impact and the contribution of rainfall to runoff generation. In principle, the R factor is derived from the maximum 30 min rainfall intensity [14], but in Greece and many other countries, such high-resolution records are not widely available. As an alternative, several studies have estimated R from long-term average annual precipitation using empirical equations developed for Mediterranean conditions [15]. In this study, daily precipitation data from three nearby meteorological stations were processed to obtain the mean annual rainfall for the period 2017–2024, which was 695 mm. Applying the regional equation:

R = 0.83 N − 17.7

where N is the mean annual rainfall in mm, the rainfall erosivity factor for the study area was calculated as 559 MJ·mm/(ha·h·yr). This value was then used as the R factor input for the RUSLE model.

2.2. Soil Erodibility (K)

Physicochemical analyses were carried out on soil samples collected from several olive orchards located across the main soil types of the watershed. The laboratory results provided information on particle size distribution (percentages of sand, silt, and clay) and organic matter content, which were then used to calculate the soil erodibility factor (K) following the Wischmeier and Smith nomograph equation. This approach accounts for the influence of texture, structure, permeability, and organic matter on soil susceptibility to erosion. The resulting K values in the study area ranged from 0.018 to 0.028 t·ha·h/(ha·MJ·mm), indicating moderate erodibility. Higher values were associated with fine-textured soils with low organic matter content, while lower values corresponded to coarser soils or soils with slightly better aggregation. Although the variation is relatively narrow, these differences contribute to the spatial distribution of predicted erosion risk in the watershed.

2.3. Topographic Factor (LS)

The LS factor was derived from a Digital Elevation Model (DEM) with 5 m resolution, resampled to 10 m for use in RUSLE. Slope steepness (in degrees) and flow accumulation were calculated for each grid cell, and the following equation [16] was applied to estimate LS:

LS = ((flow accumulation × cell size)/22.13)^0.4 × ((sin (slope × 3.14/180))/0.0896)^1.3

This formulation incorporates both slope length (represented by flow accumulation) and slope steepness, providing a dimensionless index of topographic influence on erosion. In the study area, LS values range from nearly 0 on flat valley bottoms to more than 20 on long, steep hillslopes. The results indicate that slope is a key driver of spatial variation in soil loss, as erosion potential increases sharply on steep and elongated terrain compared to gentler slopes.

2.4. Cover-Management Factor (C)

The C factor was derived from a land cover map prepared using supervised classification of Sentinel-2 satellite imagery (10 m resolution), supported by field observations. Major land cover classes identified in the watershed included olive orchards, annual crops, shrub/grassland, and dense natural vegetation such as sclerophylls’ forest. C values were assigned to each class according to canopy cover, understory conditions, and seasonal vegetation dynamics, drawing on values reported in relevant Mediterranean studies [17]. For olive orchards managed under conventional bare-soil practices, a C factor of approximately 0.25 was applied, reflecting the partial protection offered by tree canopies and litter compared to fully bare ground. Orchards with spontaneous grass cover for part of the year were assigned lower values (~0.1), while dense shrub and forest areas received very low values (~0.01–0.05), and tilled annual crops were assigned higher values (~0.3). Using literature-based coefficients ensured consistency with established RUSLE applications in Mediterranean environments, while also allowing for adaptation to the local land cover characteristics identified in the study area.

2.5. Support Practice Factor (P)

The P factor accounts for erosion control measures that reduce runoff velocity or promote deposition, such as terracing, contour plowing, strip cropping, or contour hedgerows. In our case, the most relevant support practice is the presence of terraces or contour banks in olive groves. Through field mapping and aerial photo interpretation, we identified terraced sections of the hillside orchards. We assigned a P value of 0.4–0.5 for well-terraced olive fields (meaning erosion is reduced to ~40–50% of what it would be with straight downslope cultivation), based on values from the literature [8] for contouring on moderate slopes. Non-terraced or negligibly managed areas were given P = 1 (no erosion control).

After generating the R, K, LS, C, and P rasters, these were multiplied to compute the annual soil loss map for the watershed (Figure 3). For evaluation, the RUSLE outputs were compared with the soil loss measurements obtained from the experimental plots on the representative 16% slope. The modelled values for these plots fell within the same range as the observed erosion (~0.6 t/ha), indicating that the RUSLE factors used were appropriate for the local conditions. No calibration of the model parameters was performed; rather, the comparison served to assess the model’s performance and consistency with field evidence. The agreement between the model and the measurements supports the reliability of the RUSLE-derived erosion risk estimates for the wider Xerias watershed.

3. Results and Discussion

Field monitoring in the olive orchard plots during the wet season (October 2024–April 2025) recorded a cumulative soil loss of approximately 0.6 t/ha. The RUSLE model, applied across the Xerias watershed with locally derived factors, produced estimated annual erosion rates in the range of 0–0.6 t/ha/yr, placing the study area within the same order of magnitude as the field observations. This agreement indicates that, even with limited field data, the model provides a realistic representation of erosion dynamics under local conditions.

The spatial results (Figure 3) also reveal that erosion risk is not uniformly distributed across the catchment. Gentle slopes and areas with natural vegetation or terraces show very low soil loss, whereas certain steep olive groves without ground cover are more vulnerable, potentially exceeding tolerable erosion limits during intense storms. This spatial heterogeneity is important because even if the watershed average is modest, localized hotspots can drive significant soil degradation. On bare or poorly managed slopes, soil fertility is gradually degraded through the removal of fine particles and organic matter, reducing productivity and increasing reliance on external inputs. In addition, eroded sediments are transported downslope, where they may block drainage channels and increase the risk of localized flooding, while also degrading water quality. Such impacts extend the consequences of erosion well beyond the farm scale, turning soil loss into both an agricultural and an environmental problem. The results therefore underscore the need for sustainable practices, such as maintaining vegetative cover, adopting cover crops, or restoring terraces, to safeguard not only the long-term viability of olive orchards but also the health of downstream ecosystems.

For the moment, the performance of RUSLE appears satisfactory, however in future research we plan to establish additional experimental plots throughout the watershed. Expanding in situ measurements will allow for more robust calibration and full validation of the model, further strengthening its reliability for conservation planning.