Spatiotemporal prioritization of soil erosion risk using the RUSLE model and CMIP6 projections under future climate scenarios in a Mediterranean watershed

Projected shifts in the Eastern Mediterranean rainfall regime suggest fewer but more intense precipitation events, which are expected to alter soil erosion dynamics. This study evaluates the spatiotemporal variations in rainfall erosivity (R-factor) and soil loss in the Küçük Menderes River Basin, Turkey, using the Geographic Information System (GIS) based Revised Universal Soil Loss Equation (RUSLE). The most appropriate R-factor formula was selected through a comparative evaluation of ten empirical approaches. Projections were based on downscaled rainfall data from 13 General Circulation Models (GCM) under two Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5) scenarios from 2021 to 2100. The baseline period (1970–2000) corresponded to a mean soil loss of 29.56 t ha⁻¹ yr⁻¹ and a sediment yield of 6.81 t ha⁻¹ yr⁻¹. Under SSP2-4.5, soil loss exhibited small fluctuations forming a subtle U-shaped pattern, while SSP5-8.5 projected reductions driven by decreased precipitation. While climate projections influenced temporal variations in soil loss magnitudes, the spatial distribution of high-risk zones remained predominantly controlled by the basin’s steep topography. These persistent erosion hotspots highlight the need for targeted conservation strategies, aligning with efforts to mitigate land degradation Sustainable Development Goal 15.3 (SDG 15.3).

1 Introduction

Soil erosion constitutes a major environmental problem worldwide (Chappell et al., 2016). This process results in soil nutrient depletion, accelerated land degradation, deterioration of water quality, and increased sediment delivery to aquatic systems (Amundson et al., 2015; Ghosh et al., 2022). Water-induced erosion is initiated by the detachment of soil particles due to raindrop impact or surface runoff, after which the mobilized sediments are transported by rainfall or runoff processes (Deviren Saygin and Erpul, 2019). Soil erosion and sedimentation arising from hillslope and channel processes pose serious environmental concerns by contributing to water pollution in rivers and lakes and by reducing water storage capacity in reservoirs (Lal, 2001; Özşahin, 2023). Globally, the average annual soil erosion rate is estimated at 2.4 t/ha/yr, whereas erosion rates in Turkey substantially exceed this value (Wuepper et al., 2020). Turkey is highly susceptible to water erosion due to its semiarid climate, steep topography, and shallow soil characteristics, with intense rainfall events in transitional zones between continental and Mediterranean regimes, particularly in Central Anatolia and the Aegean region, further intensifying erosion risk. This vulnerability results in an average annual soil loss of 8.24 tons per hectare and an estimated 642 million tons of soil displaced nationwide (Erpul et al., 2018; 2020; Mohammadi et al., 2025).

Future soil erosion dynamics are shaped by the combined effects of climatic conditions, land use patterns, socio-economic development, farmers decision-making processes, and agro-environmental policies (Panagos et al., 2021). Warming climate conditions intensify water erosion with negative implications for water security and ecological conservation (Teng et al., 2018), and human induced activities accelerate erosion processes, demanding evaluation across past and future scenarios at the catchment level (Behera et al., 2020). Climate change significantly influences precipitation characteristics (Kahya et al., 2016; Şen and Kahya, 2021), while increased greenhouse gas emissions result in global temperature rise and subsequent alterations in regional climate patterns (Tavakoli and De Smedt, 2012; Vicuna and Dracup, 2007). The IPCC (2013) projects rising temperatures and declining precipitation across the Mediterranean Basin throughout the 21st century. Moreover, significant decreases in precipitation observed in western Turkey, particularly in the Aegean and Thrace regions highlight the necessity of evaluating the impacts of shifting rainfall regimes on soil erosion in the Küçük Menderes River Basin (Tayanç et al., 2009). Similarly, Kilic and Gunal (2021) projects a significant increase in rainfall erosivity for the eastern Black Sea region by 2050 and 2070 under both Representative Concentration Pathways (RCP 2.6 and RCP 8.5) scenarios, while most other parts of Turkey, including the Aegean region, are expected to experience declining precipitation and a corresponding reduction in rainfall induced erosion risk.

In recent years, erosion models have been employed to assess the impacts of climate change across various catchments. Global scale studies indicate that approximately 73% of watersheds are likely to experience intensified soil erosion under future climate conditions, while the remaining 27% may exhibit a decline in erosion and sedimentation rates (Abeysingha and Ray, 2025). Although the general trend points to increasing erosion, region specific projections highlight the potential for localized reductions. These contrasting patterns have led researchers to explore region specific drivers more closely.

Revised Universal Soil Loss Equation (RUSLE) has been extensively applied across various basins to evaluate erosion dynamics under diverse climate and land use projections. For instance, Teng et al. (2018) projected soil erosion in the Tibetan Plateau by 2050 using multiple linear regression (MLR) to estimate future rainfall erosivity based on Coupled Model Intercomparison Project Phase 5 (CMIP5) data, and reported increasing erosion under both RCP2.6 and RCP8.5 scenarios. In the Lake Tana Basin, Ethiopia, Getachew et al. (2021) integrated land use projections via the CA-Markov model and climate simulations from CanESM2, revealing up to a 10% increase in soil loss by 2080, driven primarily by climate variability rather than land use change. Similarly, Hateffard et al. (2021) investigated erosion in a semi arid region of Iran and found increased erosion in rangelands under RCP2.6 and RCP8.5, though decreases were noted under RCP4.5 in some areas. Panagos et al. (2021) conducted projections for agricultural areas in the EU and UK using RUSLE and Common Agricultural Policy Regionalised Impact (CAPRI) models with CMIP5 based climate and land use data, estimating a 13%–22.5% increase in erosion by 2050. However, their maps reveal significant regional variations. While most of Europe shows increases of up to 100% in soil erosion rates, certain areas within southern Spain and southern Italy demonstrate substantial decreases, with reductions of up to 100%. These reductions are particularly notable in Mediterranean regions such as Andalusia, Calabria, highlighting both regional climate variability and the potential role of conservation practices in offsetting erosion risks. Patriche (2023) simulated future erosion for Romania under RCP4.5 and RCP8.5 scenarios and predicted slight increases across 84%–90% of the country during 2041–2080. In Turkey, Özşahin (2023) applied RUSLE and Modified Universal Soil Loss Equation (MUSLE) to estimate erosion and sedimentation in the Naip Dam Basin, reporting values between 0.31 and 0.36 t ha⁻¹ y⁻¹ under current conditions and validating results with reservoir data. Most recently, Gezici et al. (2025) compared 2023 baseline and future SSP scenarios in the Oltu Basin, northeastern True, and found that although average erosion remained stable, erosion risk became more spatially concentrated due to uneven changes in rainfall erosivity. Other models have also been employed; for example, in the mountainous Portaikos catchment in Central Greece, Stefanidis and Stathis (2018) applied the Erosion Potential Model (EPM) and projected a −4.9% reduction in soil erosion potential by the late 21st century, due to a −21.2% decrease in annual precipitation and a 3.6 °C increase in mean temperature.

Rainfall erosivity is widely used in erosion modeling and also functions independently as an indicator of soil loss potential (Eekhout and de Vente, 2022). Similar to RUSLE based projections, global assessments of rainfall erosivity indicate an overall increase, yet with regional variability, while many areas are projected to face increased erosion risk, others, such as Mediterranean regions, may experience declines. Similar to RUSLE based projections, global assessments of rainfall erosivity indicate an overall increase under climate change; however, regional variability remains significant. While many areas are projected to face increased erosion risk, others such as Mediterranean regions may experience declines. According to recent Coupled Model Intercomparison Project Phase 6 (CMIP6) based projections (Chen et al., 2024), global rainfall erosivity is expected to rise moderately by midcentury (2041–2060) and more substantially by the end of the century (2081–2100), particularly under the SSP5-8.5 scenario. Mediterranean regions, including coastal Turkey, are among the few areas projected to experience substantial declines, with relative reductions exceeding 15% by 2100. Declines are also projected in parts of Western Europe and Australia. Panagos et al. (2022) compiled 20 regional studies and developed global rainfall erosivity projections for 2050 and 2070 based on 19 CMIP5 models. Using 2010 as the baseline, the regional analyses indicated substantial increases in parts of Asia, Europe, and Africa, while regions such as Greece, Crete, the Tibetan Plateau, and Turkey exhibited variability with both increases and decreases. At the global scale, the projections estimated an increase of approximately 26%–29% by 2050 and 27%–34% by 2070, with higher RCP scenarios associated with stronger increases. However, under the RCP 8.5 scenario, notable decreases of up to 15% by 2070 were projected for Mediterranean coastal zones, including Turkey’s Aegean and Mediterranean coasts, southern Spain, and Italy, underscoring the region’s deviation from global trends due to climatic variability. Grillakis et al. (2020) developed a daily-timescale rainfall erosivity formula for Crete and projected moderate near-future changes in rainfall erosivity (R factor), while far-future projections under RCP2.6 indicated further increases and under RCP8.5 a significant decrease, highlighting that both the frequency and intensity of erosive events determine the net change in rainfall erosivity.

While many studies propose simple empirical equations to estimate R factor, only a few have evaluated RUSLE’s performance using different R factor formulations (Sakhraoui and Hasbaia, 2023). For instance, Efthimiou (2018) compared nine different empirical R factor equations in the Venetikos River catchment in Greece to identify the most suitable approach for the region. The RUSLE outputs were indirectly validated using sediment yield data, and the Renard and Freimund (1994) formula was found to provide the best agreement with observed values.

Although global and regional studies have examined soil erosion under climate change, projection-based assessments at the basin scale in Turkey remain limited. This study addresses this gap by integrating several methodological advances. Owing to the absence of sub-hourly rainfall data in climate projections, ten empirical R-factor formulations were systematically evaluated against observed high temporal resolution rainfall erosivity values to identify the most suitable approach for Mediterranean climate conditions. To reduce model-related uncertainty, ensemble projections derived from 13 CMIP6 Global Circulation Models (GCM) under the SSP2-4.5 and SSP5-8.5 scenarios were employed. Furthermore, the study implements subbasin-scale erosion risk prioritization across 60 subbasins, enabling an integrated evaluation of both temporal changes in erosion magnitude and the spatial persistence of high-risk zones under future climate conditions. Accordingly, this study aims to quantify spatiotemporal variations in soil loss, assess erosion risk distribution at the subbasin scale, and examine driving factors influencing erosion patterns.

2 Materials and methods

2.1 Study area

The Küçük Menderes River Basin is located along the Aegean coast of western Türkiye and encompasses an area of approximately 352,871 ha (Figure 1). Agricultural land constitutes the dominant land use within the basin, followed by forested and semi-natural areas (GDWM, 2018). The basin exhibits a pronounced topographic gradient, transitioning from steep mountainous terrain in the eastern parts to low gradient plains toward the west. Approximately one-third of the basin consists of an east-west elongated plain, which is bordered by steeply rising mountain ranges and the Aegean Sea (Yagbasan, 2016). Within the basin, agricultural activities on sloping lands where forests have been converted to cropland particularly mixed and cultivated areas with low vegetation density are associated with moderate to high erosion risk under current conditions (Gökçe Gündüzoğlu and Gündüzoğlu, 2025).

Figure 1

Figure 1. Location of the Küçük Menderes river basin and meteorological stations.

The region is characterized by a Mediterranean climate, with hot, dry summers and mild, rainy winters. The mean annual precipitation is 622 mm, corresponding to an average annual discharge of 11.45 m³ s⁻¹ at the basin outlet (GDWM, 2016). The combination of pronounced topographic variability, seasonal rainfall concentration, and the predominance of agricultural land use suggests that runoff generation and soil erosion processes may vary spatially across the basin. These characteristics make the Küçük Menderes River Basin a suitable case study for evaluating basin-scale soil erosion dynamics and their potential response to future climate change.

2.2 Climate change scenarios

To project rainfall erosivity for future time horizons, this study utilized precipitation datasets from the WorldClim database, which provides downscaled climate projections derived from the Coupled Model Intercomparison Project Phase 6 (CMIP6) framework (Fick and Hijmans, 2017). The analysis employed gridded precipitation data at a spatial resolution of 2.5 arcminutes (∼5 km).

Rather than relying on a single climate model, an ensemble mean approach was adopted, incorporating outputs from 13 Global Circulation Models (ACCESS-CM2, BCC-CSM2-MR, CMCC-ESM2, EC-Earth3-Veg, FIO-ESM-2-0, GISS-E2-1-G, HadGEM3-GC31-LL, INM-CM5-0, IPSL-CM6A-LR, MIROC6, MPI-ESM1-2-HR, MRI-ESM2-0, UKESM1-0-LL). This multi-model ensemble methodology enhances projection reliability by reducing uncertainties inherent in individual climate models and provides a more robust representation of potential future conditions (Yılmaz et al., 2025; Ahmed et al., 2020; Panagos et al., 2022).

Projections were generated for four time periods (2021–2040, 2041–2060, 2061–2080, and 2081–2100) under two Shared Socioeconomic Pathways; SSP2-4.5, representing moderate mitigation, and SSP5-8.5, depicting a high-emission trajectory with limited climate action (O'Neill et al., 2016).

2.3 Revised Universal Soil Loss Equation (RUSLE)

The Revised Universal Soil Loss Equation (RUSLE) is an empirical model designed to estimate the long term average annual soil loss from rainfall and associated surface runoff (Renard et al., 1997). In its standard form, annual soil loss ($A$, t ha⁻¹ yr⁻¹) is computed as the multiplicative product of five main factors. Mathematically, the RUSLE is expressed as (Equation 1):

$A=R\times K\times LS\times C\times P(1)$

The rainfall erosivity factor ($R$, MJ mm ha⁻¹ h⁻¹ y $r^{-1}$), which represents the impact of raindrop energy and rainfall intensity; the soil erodibility factor ($K$, t h MJ⁻¹ mm⁻¹), which quantifies the susceptibility of soil to detachment and transport; the slope length and steepness factor ($LS$, dimensionless), accounting for topographic effects on erosion processes; the cover management factor ($C$, dimensionless), representing the influence of vegetation cover and land use; and the support practice factor ($P$, dimensionless), reflecting the effectiveness of erosion control measures such as contour farming or terracing (Wischmeier and Smith, 1978; Renard et al., 1997). RUSLE calculations and raster processing were performed using Python libraries including rasterio, GDAL, and geopandas.

The RUSLE-based analysis followed a systematic workflow consisting of sequential processing steps. (1) RUSLE factor–related datasets, including climate projections, soil data, topographic data, and vegetation indices, were preprocessed to ensure consistency in projection, extent, and data quality; (2) all input layers were resampled to a common spatial resolution to enable pixel-wise calculations; (3) spatial maps of the individual RUSLE factors (R, K, LS, C, and P) were generated based on their respective methodologies; (4) spatial maps of annual soil loss were generated through raster-based multiplication of the RUSLE factors, followed by sediment yield estimation; and (5) the resulting soil loss maps were aggregated at the subbasin scale to support erosion risk prioritization.

2.3.1 Rainfall erosivity (R) factor

The rainfall erosivity factor (R factor) represents the erosive potential of rainfall to detach and transport soil particles from an uncovered surface (Renard et al., 1997), which is defined by Renard and Freimund (1994) as the product of rainfall kinetic energy (E, MJ/ha) and the maximum 30-min intensity (I₃₀, mm/h). However, high-temporal-resolution rainfall data required for direct EI₃₀ calculation are often unavailable from meteorological stations, particularly in climate projection datasets. Consequently, numerous empirical formulas have been developed to estimate the R-factor from available rainfall data (Gezici et al., 2025). Due to the absence of sub-hourly rainfall data, the R-factor was estimated using empirical approaches. Based on the performance assessment conducted within this study, the Torri et al. (2006) formulation was used to estimate R values (Equation 2).

$\mathrm{R}=3.08\times \mathrm{P}-944(2)$

where R is the rainfall erosivity factor (MJ mm ha⁻¹ h⁻¹ yr⁻¹) and P is mean annual precipitation (mm).

2.3.2 Soil erodibility (K) factor

The K factor represents the soil’s susceptibility to erosion, which depends on its textural and structural composition as well as its organic matter content (Patriche, 2023). The soil erodibility (K) factor dataset was obtained from the Dynamic Erosion Modeling and Monitoring System (DEMIS), a nationwide monitoring framework (Erpul et al., 2018). This database, compiled by the General Directorate of Combating Desertification and Erosion in Turkey, integrates essential soil properties required for K factor estimation and is provided at a spatial resolution of 100 m. It is based on 22,000 soil profiles collected across Turkey. K factor values were calculated using the empirical equation proposed by Torri et al. (1997), Torri et al. (2002), a method validated in Turkish case studies and used in the development of the national soil erosion map (Akgoz et al., 2022) (Equation 3).

$K=0.0293\left(0.65-D_G+D_G^2\right)\times e^{\left[-0.0021\left(\frac{OM}{C}\right)-0.00037{\left(\frac{OM}{C}\right)}^2-4.02C+1.72C^2\right]}(3)$

In this formulation, K denotes the soil’s susceptibility to erosional processes, DG is the geometric mean particle diameter, C is the clay fraction percentage, and OM refers to the organic matter content.

2.3.3 Topographic (LS) factor

The LS factor represents the combined influence of slope length (L) and slope steepness (S) on soil erosion, capturing the effect of topography through increased runoff accumulation along longer slopes and higher runoff velocity on steeper terrain (Ghosal and Das Bhattacharya, 2020). The LS factor dataset was obtained from the Dynamic Erosion Modeling and Monitoring System (DEMIS) (Erpul et al., 2018). It was generated using a 10 m resolution Digital Elevation Model (DEM) and the method proposed by Moore and Burch (1986a), Moore and Burch (1986b) (Equation 4).

$LS={\left(\frac{\chi \eta }{22.13}\right)}^{0.4}·{\left(\frac{\sin \theta }{0.0896}\right)}^{1.3}(4)$

where, $\chi$: flow accumulation derived from DEM, $\eta$: cell size, and $\theta$: slope steepness in degrees.

2.3.4 Cover management (C) factor

The C factor reflects the influence of vegetation and land cover on soil erosion and is defined as the ratio of soil loss from managed cropland to that from bare, continuously tilled land (Wischmeier and Smith, 1978). By intercepting raindrop impact, vegetation reduces erosion, with C values controlled by vegetation type, growth stage, and ground cover extent (Ghosal and Das Bhattacharya, 2020).

Given the strong control of vegetation cover on the C factor, satellite-derived vegetation indices are used for its spatial estimation. Normalized Difference Vegetation Index (NDVI) quantifies the normalized difference between near infrared (NIR) and red (RED) reflectance bands, with values ranging from −1 to +1, widely used as an indicator for tracking vegetation dynamics (Rouse et al., 1974). Moderate Resolution Imaging Spectroradiometer (MODIS) MOD13Q1 16-day composite NDVI data with a spatial resolution of 250 m were employed to calculate the C factor for RUSLE analysis. A total of 20 images from the year 2000, obtained via NASA Earthdata’s AppEEARS service (NASA, 2024) and subjected to geometric and atmospheric corrections, were processed to capture the complete annual vegetation cycle. The use of composite images throughout the year ensured comprehensive representation of annual vegetation patterns while minimizing seasonal biases. C-factor calculations employed the Van der Knijff et al. (1999) formula using recommended α = 2 and β = 1 coefficients (Equation 5).

$C=e^{\left[-\alpha \times \left(\frac{NDVI}{\beta -NDVI}\right)\right]}(5)$

where, NDVI is the Normalized Difference Vegetation Index.

2.3.5 Conservation practice (P) factor

The P factor represents the effect of conservation practices on reducing surface runoff and ranges from 0 to 1, with a value of 1 indicating the absence of such practices. It reflects the relative effectiveness of support practices in reducing soil loss compared to slope-parallel cultivation (Wischmeier and Smith, 1957; Renard et al., 1997). In this study, the P factor was set to 1 due to the lack of reliable information on conservation practices. Soil conservation practices are present only in limited areas in Turkey (Altıkat et al., 2018), and therefore the omission of this factor may not significantly influence basin-scale erosion estimates.

2.4 Sediment delivery ratio (SDR)

RUSLE estimates soil erosion but does not calculate sediment yield at the outlet. A sediment delivery ratio (SDR) is required to determine the portion of eroded soil that reaches the watershed outlet. The SDR depends on watershed characteristics including area, topography, land use, and soil properties (Ghavami et al., 2024). In this study, the Boyce (1975) empirical method was applied to estimate sediment yield for the Küçük Menderes River Basin. This power function approach represents one of the most common empirical SDR methods (Woznicki and Nejadhashemi, 2013). The Boyce formula is expressed as (Equation 6):

$\text{SDR}=0.5656A^{-0.11}(6)$

where SDR is the sediment delivery ratio (dimensionless) and A is the watershed area (km²).

3 Results

3.1 Selection of the rainfall erosivity equation

High temporal resolution sub-hourly precipitation records are often absent from observational datasets, as noted by Biasutti and Seager (2015), and this limitation also extends to widely used climate projection datasets. This makes it necessary to select an appropriate empirical formula to estimate the rainfall erosivity factor (R) from available coarse resolution data.

A total of 17 meteorological stations, located within and around the Küçük Menderes River Basin, were selected based on their ability to represent the basin’s climatic conditions and the availability of sufficiently long precipitation series and precalculated EI₃₀ values (Figure 1). EI₃₀ data were obtained from the General Directorate for Combating Desertification and Erosion (Erpul et al., 2018). For each station, R-factor values were calculated using 10 different empirical formulas and compared against EI₃₀ derived R values (Table 1). Here, P represents the mean annual precipitation (mm) and Modified Fournier Index (MFI) which was developed by (Arnoldus, 1977). Only common years were used to ensure consistency across models and stations.

Table 1

Table 1. Empirical equations used for estimating the R-factor.

Model performance was evaluated using Percent Bias (PBIAS) as the metric, which quantifies the average tendency of predicted values to exceed or fall short of observed measurements (Gupta et al., 1999) and is recommended for quantitative performance evaluation (Moriasi et al., 2007). Although correlation and correlation-based metrics evaluate linear relationships, they tend to be overly influenced by extreme values and fail to reflect additive or proportional discrepancies between predictions and observations (Legates and McCabe, 1999). Given that R-factor represents long-term average rainfall erosivity rather than year-to-year fluctuations, PBIAS was selected to effectively capture systematic bias in long term mean values. To avoid geographical bias caused by unequal data availability among stations, a station-based averaging approach was adopted. In this method, the mean R-factor value for each station was first calculated over the available data period (2004–2023), and then PBIAS was computed using these 17 station averages. This approach ensures equal weighting of all stations regardless of differences in data completeness (Table 2).

Table 2

Table 2. Performance of empirical equations for estimating the R-factor.

The results indicated that the Torri et al. (2006) model achieved the best overall performance with a PBIAS of −8.23%, indicating minimal bias. Originally developed for the Tuscany, Abruzzi, and Lucania regions, this model demonstrated a strong linear relationship between annual precipitation and erosivity under a Mediterranean climate regime (r² = 0.768, n = 88). The Küçük Menderes River Basin exhibits similar climatic characteristics, with wet winters, dry summers, and intense erosive rainfall events in autumn.

The 17 station average R-EI₃₀ value was 956 MJ·mm·ha⁻¹·h⁻¹·yr⁻¹, while the Torri et al. (2006) model yielded a closely matching mean of 881 MJ·mm·ha⁻¹·h⁻¹·yr⁻¹. Compared to other formulas, the Torri et al. (2006) model provided the most consistent results and the closest agreement with observed EI₃₀ values. Considering its statistical performance with minimal bias, climatic compatibility, and practical applicability to low temporal resolution climate projection data, the Torri et al. (2006) model was selected for calculating future R-factor values in this study.

3.2 Evaluation of RUSLE model parameters

The RUSLE model provides a framework for quantifying soil loss and delineating erosion-prone areas at the watershed scale. Following the methodology outlined in Section 2.3, all five RUSLE input factors (R, K, LS, C, and P) were derived for the Küçük Menderes River Basin.

Rainfall erosivity (R-factor) maps were generated following the methodology described in Section 2.3.1, using the Torri et al. (2006) formula identified as the most suitable approach for the basin in Section 3.1. R-factor maps were produced for four distinct future periods (2021–2040, 2041–2060, 2061–2080, and 2081–2100) under two Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5) and subsequently integrated into the RUSLE framework (Figure 2).

Figure 2

Figure 2. Spatial distribution of R-factor for the Küçük Menderes River Basin under different scenarios and periods.

The temporal variations in the basin’s rainfall erosivity potential (R factor) were analyzed by comparing future scenarios with the reference period of 1970–2000, for which the average R factor of 1250.4 MJ mm ha⁻¹ h⁻¹ yr⁻¹ was calculated using the original gridded WorldClim v2.1 climate data instead of station-based observations to avoid potential bias. Under the SSP2-4.5 scenario, R-factor values exhibited small overall fluctuations of +1.7 to −5.8% throughout the century. A slight near-term increase to 1276.1, was observed in 2021–2040, followed by a gradual decline to 1177.9 in 2061–2080, and a modest increase to 1205.3 by 2081–2100. These variations indicate relatively stable rainfall erosivity with a brief enhancement in the early decades and moderate decreases thereafter. In contrast, the high-emission SSP5-8.5 scenario showed a more pronounced decreasing trend, primarily associated with the reduction in the number of erosive rainfall events. R-factor values reached 1236.8 in 2021–2040, then declined to 1177.8, 1082.2, and 986.6 in the later projection periods, representing reductions from 5.8% up to 21.1% compared to the baseline (Table 3).

Table 3

Table 3. Statistics parameters of RUSLE model.

These trends suggest that rainfall erosivity under SSP2-4.5 remains relatively stable, showing small near-term increases followed by moderate declines without a clear long-term trend. In contrast, the pronounced decrease under SSP5-8.5 could reflect the dominance of fewer erosive storms and reduced total precipitation over the basin. This interpretation is consistent with multiple GCM projections that indicate an overall reduction in annual rainfall. Therefore, while SSP2-4.5 may indicate near-steady erosivity driven by a balance between rainfall intensity and frequency, the decline under SSP5-8.5 suggests that decreasing rainfall frequency likely outweighs potential increases in event intensity. Additionally, rainfall erosivity exhibit a decreasing spatial gradient from coastal to inland areas, consistent across all projection maps.

The K-factor, obtained following the methodology described in Section 2.3.2 (Soil erodibility (K) factor), ranged from 0.02 to 0.04 t ha h ha⁻¹ MJ⁻¹ mm⁻¹, with a mean value of 0.03 (Figure 3). Lower K-factor values correspond to soils with reduced susceptibility to erosion, while higher values indicate greater vulnerability to rainfall and runoff-induced detachment. The narrow range of K-factor values reflects relatively uniform soil erodibility across the basin.

Figure 3

Figure 3. Spatial distribution of the K factor (soil erodibility).

The LS-factor was derived by adopting the methodology described in Section 2.3.3 (Topographic (LS) factor) and ranged from 0.00 to 49.48, with a mean value of 4.58. Topographic classification revealed that 44.3% of the basin comprises flat terrain (LS: 0–1), 19.8% gently sloping areas (LS: 1–5), 33.9% steep slopes (LS: 5–20), and 2.0% very steep slopes (LS > 20). Higher LS values (>20) were predominantly observed in the northern and southern mountainous regions, where steep gradients and extended slope lengths increase runoff velocity and erosion susceptibility. Conversely, the western plains and interior lowlands exhibited low LS values (<5), indicating reduced topographic contribution to erosion risk (Figure 4).

Figure 4

Figure 4. Spatial distribution of the LS factor (slope length and steepness).

The C-factor, calculated from NDVI using the Van der Knijff et al. (1999) approach (Section 2.3.4), exhibited spatial variation reflecting the basin’s land cover patterns. Values ranged from 0.001 to 0.94, with a basin-wide average of 0.18 (Figure 5). Lower C-factor values corresponded to forested areas in the northern and southern mountainous regions, indicating reduced erosion susceptibility, while higher values were observed in agricultural lands and sparsely vegetated areas within the central basin and in northwestern urban zones, consistent with land use patterns identified from Coordination of Information on the Environment (CORINE) Land Cover analysis (İpek and Kahya, 2026).

Figure 5

Figure 5. Spatial distribution of the C factor (cover management factor).

Following Section 2.3.5, the P-factor was uniformly assigned a value of 1 throughout the basin, reflecting the lack of reliable information on conservation practices.

3.3 Soil erosion projections and sediment yield estimation

Following the integration of all RUSLE factors described in Section 3.2, soil loss maps were generated for the baseline period and future climate scenarios (Figure 6). The mean annual soil loss during the baseline period (1970–2000) was estimated at 29.56 t ha⁻¹ yr⁻¹.

Figure 6

Figure 6. Spatial distribution of soil loss estimated by RUSLE scenarios.

Under the SSP2-4.5 scenario, a slight near-term increase of 1.7% was observed in 2021–2040, reaching 30.07 t ha⁻¹ yr⁻¹, followed by minor fluctuations in the subsequent periods ranging to −5.9%, with soil loss slightly decreasing to 28.45 t ha⁻¹ yr⁻¹ in 2041–2060, 27.80 t ha⁻¹ yr⁻¹ in 2061–2080, and then rising again to 28.46 t ha⁻¹ yr⁻¹ in 2081–2100, indicating relatively stable soil erosion potential over the century. In contrast, the SSP5-8.5 scenario showed a decrease, with soil loss dropping to 29.21 t ha⁻¹ yr⁻¹ in 2021–2040 and further declining to 27.79, 25.60, and 23.37 t ha⁻¹ yr⁻¹ in the later periods, representing reductions from 1.2% up to 20.9% relative to the baseline (Table 3).

For the Küçük Menderes River Basin, with a drainage area of 3528.71 km² and a mean annual soil loss of 29.56 t ha⁻¹ yr⁻¹ estimated by RUSLE, the sediment delivery ratio (SDR) was calculated as 0.230 (23.0%), indicating the proportion of eroded material delivered to the watershed outlet. Accordingly, the resulting sediment yield, representing the transported sediment reaching the outlet, was estimated at approximately 6.81 t ha⁻¹ yr⁻¹.

3.4 Subbasin risk prioritization

In this study, 60 subbasins were classified into five erosion risk categories based on annual average soil loss rates under multiple climate scenarios. These categories were defined as very low risk for rates less than 2 t ha⁻¹ yr⁻¹, low risk for rates between 2 and 5 t ha⁻¹ yr⁻¹, moderate risk for rates between 5 and 10 t ha⁻¹ yr⁻¹, high risk for rates between 10 and 20 t ha⁻¹ yr⁻¹, and very high risk for rates exceeding 20 t ha^-1 yr^-1 (Koralay and Kara, 2025) (Figure 7).

Figure 7

Figure 7. Distribution of subbasin erosion risk categories under different climate scenarios.

The subbasin-scale erosion risk classification indicates that the spatial distribution of risk classes remains largely stable across future climate scenarios, with only minor temporal variations. During the baseline period (1970–2000), the basin was characterized by severe erosion risks. While the majority of the 60 subbasins were classified as “High” at 30.0% and “Very High” at 56.7%, the remaining areas comprised “Moderate”, “Low-Moderate”, and “Low” risk classes with shares of 8.3%, 3.3%, and 1.7% respectively.

Under the SSP2-4.5 scenario, these proportions remained remarkably stable throughout the century across the entire risk spectrum. The “Low” and “Low-Moderate” classes remained constant at 1.7% and 3.3%, while the “Moderate” class exhibited minimal variation ranging from 8.3% to 10.0%. Similarly, the “Very High” risk category showed limited fluctuation oscillating between 53.3% and 56.7%.

Under the SSP5-8.5 scenario, the risk distribution remained consistent with the baseline and SSP2-4.5 patterns for the majority of the century (2021–2080). In these earlier periods, the combined “Low” and “Low-Moderate” classes remained stable at values between 1.7% and 3.3%, and the “Moderate” class fluctuated slightly between 10.0% and 15.0%. However, the late 21st century (2081–2100) emerged as a distinct exception to this general stability. Only in this final period did the spatial distribution diverge significantly, with the “Very High” risk class decreasing to 48.3%. This reduction was directly offset by a marked expansion in the “Moderate” risk class which reached 18.3%, marking the 2081–2100 period as the sole deviation from the otherwise stable risk trends observed across the basin.

Overall, despite the singular shift in the late SSP5-8.5 period, the relative hierarchy of risk remained largely unchanged. The combined “High” and “Very High” risk classes consistently encompassed the vast majority of the basin across all scenarios. These high-risk classes were predominantly concentrated in areas with steep slopes and high elevations, corresponding to regions with elevated LS factor values. This spatial pattern is consistent with the basin’s topography, which transitions from steep, mountainous terrain in the east to low-gradient plains toward the west, and it is also evident in the extensive mountainous areas in the northern and southern parts of the basin. This persistence confirms that topography plays the decisive role in the spatial distribution of erosion risk across the basin, overriding climate-induced variations in most periods.

4 Discussion

This study investigated erosion dynamics in the Küçük Menderes River Basin under CMIP6 climate change scenarios, focusing on the spatial and temporal variations in rainfall erosivity, soil loss and erosion risk across subbasins. Rainfall erosivity (R factor) assessment required empirical model evaluation due to the limited temporal resolution of available climate datasets. Among the ten empirical formulas tested, the Torri et al. (2006) model demonstrated better accuracy, confirming its suitability for Mediterranean climate conditions. R-factor findings are generally align with the 700–1150 MJ mm ha⁻¹ h⁻¹ yr⁻¹ range reported by Panagos et al. (2017) for the Aegean region of Turkey.

CMIP6 projections indicated a decreasing trend in total precipitation over the Küçük Menderes Water Basin, particularly becoming more pronounced toward the end of the century under the SSP5-8.5 scenario. This pattern suggests a general decline in rainfall erosivity potential. Substantial decreases in R factor were observed for the 2041–2080 and 2081–2100 periods, with reductions ranging from 3.6% to 21.1% depending on the scenario SSP2-4.5 and SSP5-8.5. These findings are consistent with prior R factor assessments covering the Aegean and western regions of Turkey. For instance, Chen et al. (2024) used CMIP6-based R factor projections and reported decreases generally below 15% for the periods 2041–2060 and 2081–2100, with reductions exceeding 15% under the SSP5-8.5 scenario in 2081–2100. Similarly, Panagos et al. (2022) projected decreases generally below 15% for the period 2050–2070 under both RCP4.5 and RCP8.5 scenarios. Kilic and Gunal (2021) projected decreases in rainfall erosivity for the Aegean region under RCP8.5 by 2070, −16.29% (MIROC5) to −33.97% (CCSM4).

Similar mechanisms have been reported for other Eastern Mediterranean regions. For example, Grillakis et al. (2020) projected a decrease in rainfall erosivity for Crete under certain future scenarios, with a reduction under RCP8.5, attributing the change to a combination of fewer erosive events and increased event intensity, where the relative magnitude of these factors determined the net change in R factor. In their detailed analysis, they found that in the near future the number of erosive events is expected to decrease by about 5% across all scenarios, while the intensity of these events shows a modest increase, particularly under RCP8.5. In the far future, the reduction in the number of erosive events becomes more pronounced, reaching −15% to −25% under RCP4.5 and RCP8.5, whereas changes in intensity remain minor, except for a slight 3%–5% increase under RCP4.5. These findings highlight how opposing trends in event frequency and intensity jointly shape the overall change in rainfall erosivity. Comparable patterns in the Küçük Menderes River Basin, where SSP2-4.5 exhibited relatively stable erosivity while SSP5-8.5 showed pronounced declines, suggest that similar mechanisms may operate across Eastern Mediterranean basins.

Although previous studies offer R factor projections for adjacent regions and Turkey in general, RUSLE-based soil loss projections specifically for the Küçük Menderes River Basin remain limited and should be expanded to support more robust, basin-specific erosion management strategies.

It is important to note that reductions in R factor or soil loss do not necessarily translate into changes in erosion risk classes. Despite projected decreases in soil loss under future climate scenarios, the spatial distribution of risk classes remained largely stable across subbasins. Even under the most extreme SSP5-8.5 scenario for 2081-2100, where R-factor declined by 21.1%, the very high-risk class decreased by only 8.4%, from 56.7% to 48.3%, while under other scenarios risk distributions remained nearly constant with variations under 3.4%. This may indicate that even with reduced rainfall erosivity, topographic and land-use factors continue to dominate the relative erosion risk levels within the basin, emphasizing the need for targeted conservation strategies in persistently high-risk areas. Previous studies indicate that erosion severity can vary with topographic factors such as slope aspect and position (Tuo et al., 2023; Ugese et al., 2022; Menghis et al., 2025; Liu et al., 2024). Consistent with these findings, our analysis revealed that high-risk classes were predominantly concentrated in areas with steep slopes and high elevations, corresponding to regions with elevated LS factor values. This further indicates that topography plays a decisive role in shaping the spatial distribution of erosion risk.

Based on RUSLE model calculations, the river basin experiences an average annual erosion rate of 29.56 tons per hectare, with a corresponding sediment delivery ratio of 23%. This result is consistent with sediment delivery ratios documented in previous research, which typically vary between 13% and 40% (Ouillon, 2018; Walling and Webb, 1996).

Finally, this study is subject to certain limitations. The C factor was assumed to be constant over time, which represents a limitation for predicting future conditions. Although several studies investigating the impacts of climate change on soil erosion have applied a similar assumption (Teng et al., 2018; Gezici et al., 2025; Hateffard et al., 2021), this simplification may overlook potential long-term changes in vegetation cover. Previous research has shown that future changes in soil erosion can be strongly influenced by variations in the R factor (Borrelli et al., 2020; Marcinkowski et al., 2022; Polykretis et al., 2020); however, changes in other factors may also affect erosion rates and should be considered in future studies.

RUSLE applications commonly integrate input factors derived from datasets with varying spatial resolutions, which are subsequently resampled to a uniform grid for analysis, consistent with many previous studies (Behera et al., 2020; Gezici et al., 2025; Özşahin, 2023). In this context, the spatial resolution of the climate projection data (∼5 km) used in this study represents a limiting factor for capturing fine-scale erosion patterns. Future studies employing higher spatial resolution climate projections may enable more detailed assessments of localized erosion risks under changing climatic conditions. At the same time, the reliance on monthly CMIP6 data necessitated the use of empirical relationships for R-factor estimation, as direct R-EI30 calculation requires sub-hourly resolution. Future availability of sub-hourly projections would significantly refine the assessment of rainfall intensity and its specific impact on sediment dynamics.

5 Conclusion

This study assessed the potential impacts of climate change on soil erosion dynamics in the Küçük Menderes River Basin by integrating the RUSLE model with CMIP6-based climate projections. Among the ten empirical models evaluated, the Torri et al. (2006) model demonstrated the superior performance with minimal bias. It proved particularly well-suited for Mediterranean climatic conditions, offering a robust solution for R factor estimation in the absence of sub-hourly precipitation data required for direct EI₃₀ calculation a common limitation in both observational datasets and climate projections.

The ensemble mean of 13 GCMs was utilized to project future climate variability under SSP2-4.5 and SSP5-8.5 scenarios across four periods: 2021–2040, 2041–2060, 2061–2080, and 2081–2100. The baseline R factor was estimated at 1250.4 MJ mm ha⁻¹ h⁻¹ yr⁻¹, corresponding to a mean soil loss of 29.56 t ha⁻¹ yr⁻¹ and a sediment yield of 6.81 t ha⁻¹ yr⁻¹ (SDR: 0.23).

Under the SSP2-4.5 scenario, both rainfall erosivity and soil loss exhibited only minor fluctuations with slight increases in the early and late century, forming a subtle U-shaped temporal pattern. This suggests that long-term variability, rather than a consistent directional trend, dominates the basin’s response under intermediate emissions. In contrast, the SSP5-8.5 scenario projected substantial declines, particularly toward the late century, driven by marked reductions in precipitation and the frequency of erosive events. The strong divergence observed in the late SSP5-8.5 period highlights the scenario’s representation of extreme climate change compared to the generally moderate fluctuations seen earlier. However, it is important to note that rising temperatures and associated shifts in evapotranspiration may negatively affect vegetation cover, indirectly heightening soil erosion risk despite the projected decrease in rainfall erosivity.

Despite projected reductions in soil loss magnitudes, the spatial distribution of erosion risk exhibited strong persistence, largely anchored by the basin’s mountainous topography. Across the majority of scenarios, the “Very High” risk class maintained a stable range of 53.3%–56.7%, while the ‘High’ risk class fluctuated between 26.7% and 31.7%. This analysis reveals that while climate projections influenced temporal variations in soil loss magnitudes, the spatial distribution of high-risk zones remained predominantly controlled by the basin’s steep topography. This persistence weakened only in the late 21st century under SSP5-8.5, where the share of “Very High” risk declined to 48.3%, suggesting that substantial transitions in risk classification occur only under extreme climatic forcing.

The persistence of high-risk zones in the steep, mountainous eastern subbasins, contrasted with lower risks in the western plains, supports the dominant role of the LS factor in determining spatial variability. Consequently, adaptive management strategies should not rely solely on favorable climate projections but should consider prioritizing structural interventions in these topographically constrained, erosion prone areas. Projecting future soil losses is of great importance from economic and environmental perspectives, particularly under climate change and other driving factors, in order to implement effective control measures in erosion-prone areas. Such projections may be carried out using the RUSLE model or developed models in future studies.

Data availability statement

The data analyzed in this study is subject to the following licenses/restrictions: The MODIS MOD13Q1 NDVI dataset is publicly available from NASA Earthdata’s AppEEARS service (https://appeears.earthdatacloud.nasa.gov/). The WorldClim dataset is publicly available at https://www.worldclim.org/data/index.html. Other datasets used in this study, including meteorological data from the Turkish State Meteorological Service and rainfall erosivity (R-EI₃₀), soil erodibility (K factor), and LS factor data from the Dynamic Erosion Modeling and Monitoring System (DEMIS) maintained by the General Directorate for Combating Desertification and Erosion, are under restricted access and are not publicly available. Requests to access these datasets should be directed to Turkish State Meteorological Service: cGF6YXJsYW1hQG1nbS5nb3YudHI=; General Directorate for Combating Desertification and RXJvc2lvbjpjZW1AY3NiLmdvdi50cg==.

Author contributions

Aİ: Conceptualization, Data curation, Formal Analysis, Visualization, Writing – original draft. EK: Conceptualization, Supervision, Validation, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This study was supported as a Doctoral Thesis Project by the Scientific Research Fund of the Istanbul Technical University. Project Number, MDK-2024-46198.

Acknowledgements

This paper builds upon the PhD thesis research conducted by Aİ under the supervision of EK. The authors would like to express their gratitude to the staff of the Republic of Türkiye Ministry of Environment, Urbanization and Climate Change, the Desertification and Erosion General Directorate, the Turkish State Meteorological Service, and the General Directorate of State Hydraulic Works for their invaluable support and for providing the essential datasets used in this research. We also acknowledge the World Climate Research Programme’s Working Group on Coupled Modelling, which is responsible for CMIP, and thank the climate modeling groups for producing and making available their model outputs, as well as the providers of the WorldClim dataset for making high resolution climate data accessible for research purposes.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Abeysingha, N. S., and Ray, R. L. (2025). Rivers at risk, soil erosion in a changing climate: a comprehensive review. Discov. Soil 2, 16. doi:10.1007/s44378-025-00040-9

CrossRef Full Text | Google Scholar

Ahmed, K., Sachindra, D. A., Shahid, S., Iqbal, Z., Nawaz, N., and Khan, N. (2020). Multi-model ensemble predictions of precipitation and temperature using machine learning algorithms. Atmos. Res. 236, 104806. doi:10.1016/j.atmosres.2019.104806

CrossRef Full Text | Google Scholar

Akgoz, R., Deviren Saygin, S., Erpul, G., and Tel, S. (2022). Monitoring seasonal and phenological variability of cover management factor for wheat cropping systems under semi-arid climate conditions. Environ. Monit. Assess. 194, 395. doi:10.1007/s10661-022-10064-1

PubMed Abstract | CrossRef Full Text | Google Scholar

Altıkat, S., Kuş, E., Küçükerdem, H. K., and Gülbe, A. (2018). The importance of conservation agriculture for Turkey. Iğdır Univ. J. Inst. Sci. Technol. 8 (2), 73–80. doi:10.21597/jist.427894

CrossRef Full Text | Google Scholar

Amundson, R., Berhe, A. A., Hopmans, J. W., Olson, C., Sztein, A. E., and Sparks, D. L. (2015). Soil and human security in the 21st century. Science 348 (6235), 1261071. doi:10.1126/science.1261071

PubMed Abstract | CrossRef Full Text | Google Scholar

Arnoldus, H. M. J. (1977). Methodology used to determine the maximum potential average annual soil loss due to sheet and rill erosion in Morocco. FAO Soils Bull. 34.

Google Scholar

Arnoldus, H. M. J. (1980). “An approximation to the rainfall factor in the universal soil loss equation,” in Assessment of erosion. Editors M. De Boodt, and D. Gabriels (Chichester: John Wiley and Sons), 127–132.

Google Scholar

Babu, R., Dhyani, B. L., and Kumar, N. (2004). Assessment of erodibility status and refined Iso-Erodent map of India. Indian J. Soil Conserv. 32, 171–177.

Google Scholar

Behera, M., Sena, D. R., Mandal, U., Kashyap, P. S., and Dash, S. S. (2020). Integrated GIS-based RUSLE approach for quantification of potential soil erosion under future climate change scenarios. Environ. Monit. Assess. 192, 733. doi:10.1007/s10661-020-08688-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Biasutti, M., and Seager, R. (2015). Projected changes in US rainfall erosivity. Hydrol. Earth Syst. Sci. 19, 2945–2961. doi:10.5194/hess-19-2945-2015

CrossRef Full Text | Google Scholar

Borrelli, P., Robinson, D. A., Panagos, P., Lugato, E., Yang, J. E., Alewell, C., et al. (2020). Land use and climate change impacts on global soil erosion by water (2015-2070). Proc. Natl. Acad. Sci. 117 (36), 21994–22001. doi:10.1073/pnas.2001403117

PubMed Abstract | CrossRef Full Text | Google Scholar

Boyce, R. C. (1975). “Sediment routing with sediment delivery ratios,” in Present and prospective technology for predicting sediment yields and sources (Washington, DC: U.S. Department of Agriculture), ARS-S-40, 61–65.

Google Scholar

Chappell, A., Baldock, J., and Sanderman, J. (2016). The global significance of omitting soil erosion from soil organic carbon cycling schemes. Nat. Clim. Change 6 (2), 187–191. doi:10.1038/nclimate2829

CrossRef Full Text | Google Scholar

Chen, Y., Wei, T., Li, J., Xin, Y., and Ding, M. (2024). Future changes in global rainfall erosivity: insights from the precipitation changes. J. Hydrol. 638, 131435. doi:10.1016/j.jhydrol.2024.131435

CrossRef Full Text | Google Scholar

Choudhury, M. K., and Nayak, T. (2003). “Estimation of soil erosion in sagar Lake catchment of central India,” in Proceedings of the International Conference on Water and Environment (Bhopal), 387–392.

Google Scholar

Deviren Saygin, S., and Erpul, G. (2019). Modeling aggregate size distribution of eroded sediment resulting from rain-splash and raindrop impacted flow processes. Int. J. Sediment Res. 34 (2), 166–177. doi:10.1016/j.ijsrc.2018.10.004

CrossRef Full Text | Google Scholar

Eekhout, J. P. C., and de Vente, J. (2022). Global impact of climate change on soil erosion and potential for adaptation through soil conservation. Earth Sci. Rev. 226, 103921. doi:10.1016/j.earscirev.2022.103921

CrossRef Full Text | Google Scholar

Efthimiou, N. (2018). Evaluating the performance of different empirical rainfall erosivity (R) factor formulas using sediment yield measurements. Catena 169, 195–208. doi:10.1016/j.catena.2018.05.037

CrossRef Full Text | Google Scholar

Erpul, G., Şahin, S., İnce, K., Küçümen, A., Akdağ, M. A., Demirtaş, İ., et al. (2018). Turkey water erosion atlas. Ankara: General Directorate of Combating Desertification and Erosion Publications.

Google Scholar

Erpul, G., İnce, K., Demirhan, A., Küçümen, A., Akdağ, M. A., Demirtaş, İ., et al. (2020). Water erosion provincial statistics - soil erosion control strategies. Ankara: General Directorate of Combating Desertification and Erosion Publications.

Google Scholar

Ferrari, R., Pasqui, M., Bottai, L., Esposito, S., and Di Giuseppe, E. (2005). “Assessment of soil erosion estimate based on a high temporal resolution rainfall dataset,” in Proceedings of the Seventh European Conference on Applications of Meteorology (ECAM) (Utrecht).

Google Scholar

Fick, S. E., and Hijmans, R. J. (2017). WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37 (12), 4302–4315. doi:10.1002/joc.5086

CrossRef Full Text | Google Scholar

GDWM (2016). Impact of climate change on water resources, final report, appendix 8 – küçük menderes river basin. Ankara: General Directorate of Water Management.

Google Scholar

GDWM (2018). Drought management plan for the küçük menderes basin. Ankara: General Directorate of Water Management Publications.

Google Scholar

Getachew, B., Manjunatha, B. R., and Bhat, G. H. (2021). Assessing current and projected soil loss under changing land use and climate using RUSLE with remote sensing and GIS in the Lake Tana basin, upper blue nile river basin, Ethiopia. Egypt. J. Remote Sens. Space Sci. 24, 907–918. doi:10.1016/j.ejrs.2021.10.001

CrossRef Full Text | Google Scholar

Gezici, K., Şengül, S., and Kesgin, E. (2025). Assessment of soil erosion risk in the mountainous region of northeastern Türkiye based on the RUSLE model and CMIP6 climate projections. Environ. Earth Sci. 84, 167. doi:10.1007/s12665-025-12184-6

CrossRef Full Text | Google Scholar

Ghavami, M. S., Ayoubi, S., Khaleghpanah, N., Mosaddeghi, M. R., and Gohari, A. (2024). Soil loss estimation using RUSLE model: comparison of conventional and digital soil data at watershed scale in central Iran. Soil Tillage Res. 244, 106238. doi:10.1016/j.still.2024.106238

CrossRef Full Text | Google Scholar

Ghosal, K., and Das Bhattacharya, S. (2020). A review of RUSLE model. J. Indian Soc. Remote Sens. 48 (5), 689–707. doi:10.1007/s12524-019-01097-0

CrossRef Full Text | Google Scholar

Ghosh, A., Rakshit, S., Tikle, S., Das, S., Chatterjee, U., Pande, C. B., et al. (2022). Integration of GIS and remote sensing with RUSLE model for estimation of soil erosion. Land 12 (1), 116. doi:10.3390/land12010116

CrossRef Full Text | Google Scholar

Gökçe Gündüzoğlu, H. A., and Gündüzoğlu, G. (2025). Assessment of potential erosion risk in küçük menderes basin according to land use/land cover classes. Artvin Coruh Univ. J. For. Fac. 26 (1), 90–103. doi:10.17474/artvinofd.1540609

CrossRef Full Text | Google Scholar

Grillakis, M. G., Polykretis, C., and Alexakis, D. D. (2020). Past and projected climate change impacts on rainfall erosivity: advancing our knowledge for the eastern mediterranean island of Crete. Catena 193, 104625. doi:10.1016/j.catena.2020.104625

CrossRef Full Text | Google Scholar

Gupta, H. V., Sorooshian, S., and Yapo, P. O. (1999). Status of automatic calibration for hydrologic models: Comparison with multilevel expert calibration. J. Hydrologic Eng. 4 (2), 135–143. doi:10.1061/(asce)1084-0699(1999)4:2(135)

CrossRef Full Text | Google Scholar

Hateffard, F., Mohammed, S., Alsafadi, K., Enaruvbe, G. O., Heidari, A., Abdo, H. G., et al. (2021). CMIP5 climate projections and RUSLE-based soil erosion assessment in the central part of Iran. Sci. Rep. 11, 7273. doi:10.1038/s41598-021-86618-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Hurni, H. (1985). “Erosion—productivity—conservation systems in Ethiopia,” in Proceedings of the 4th International Conference on Soil Conservation (Maracay), 654–674.

Google Scholar

IPCC (2013). Climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change. Cambridge: Cambridge University Press.

Google Scholar

İpek, A. F., and Kahya, E. (2026). Integrating spatio-probabilistic mapping and remote sensing for comprehensive drought risk assessment. Model. Earth Syst. Environ. 12, 43. doi:10.1007/s40808-025-02696-2

CrossRef Full Text | Google Scholar

Kahya, E., Mehr, A. D., Şen, O., Akçakaya, A., and Özger, M. (2016). “A comparative study on the current and future design storms over Rize province, Turkey,” in 12th International Congress on Advances in Civil Engineering (ACE 2016) (Istanbul).

Google Scholar

Kilic, O. M., and Gunal, H. (2021). Spatial-temporal changes in rainfall erosivity in Turkey using CMIP5 global climate change scenario. Arab. J. Geosci. 14, 1079. doi:10.1007/s12517-021-07184-2

CrossRef Full Text | Google Scholar

Koralay, N., and Kara, Ö. (2025). Risk assessment of soil erosion with RUSLE using geographic information system and organic carbon and total nitrogen loadings of suspended sediment in sogutlu stream watershed of Trabzon, Turkey. Environ. Earth Sci. 84, 258. doi:10.1007/s12665-025-12271-8

CrossRef Full Text | Google Scholar

Lal, R. (2001). Soil degradation by erosion. Land Degrad. and Dev. 12 (6), 519–539. doi:10.1002/ldr.472

CrossRef Full Text | Google Scholar

Legates, D. R., and McCabe, G. J. (1999). Evaluating the use of 'Goodness-of-Fit' measures in hydrologic and hydroclimatic model validation. Water Resour. Res. 35 (1), 233–241. doi:10.1029/1998wr900018

CrossRef Full Text | Google Scholar

Liu, P., Guo, B., Zhang, R., and Wang, L. (2024). Spatial–temporal evolution pattern of soil erosion and its dominant factors on the loess Plateau from 2000 to 2020. Land 13, 1944. doi:10.3390/land13111944

CrossRef Full Text | Google Scholar

Marcinkowski, P., Szporak-Wasilewska, S., and Kardel, I. (2022). Assessment of soil erosion under long-term projections of climate change in Poland. J. Hydrology 607, 127468. doi:10.1016/j.jhydrol.2022.127468

CrossRef Full Text | Google Scholar

Menghis, T. B., Zdruli, P., and Dobos, E. (2025). A GIS-based approach to soil erosion risk assessment using RUSLE: the case of the mai nefhi watershed, barka river basin, Eritrea. Earth 6, 58. doi:10.3390/earth6020058

CrossRef Full Text | Google Scholar

Mohammadi, A. A., Akgoz, R., Deviren Saygin, S., and Erpul, G. (2025). Validation and comparison of event-based sediment yield estimations using MUSLE and RUSLE/SDR models: the case of Çarşamba suyu sub-basin, southwest central Anatolia, Türkiye. Nat. Hazards 121, 13501–13542. doi:10.1007/s11069-025-07328-7

CrossRef Full Text | Google Scholar

Moore, I. D., and Burch, G. J. (1986a). Physical basis of the length-slope factor in the universal soil loss equation. Soil Sci. Soc. Am. J. 50, 1294–1298. doi:10.2136/sssaj1986.03615995005000050042x

CrossRef Full Text | Google Scholar

Moore, I. D., and Burch, G. J. (1986b). Modeling erosion and deposition: topographic effects. Trans. ASAE 29, 1624–1630.

Google Scholar

Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and Veith, T. L. (2007). Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans. ASABE 50 (3), 885–900. doi:10.13031/2013.23153

CrossRef Full Text | Google Scholar

NASA (2024). Application for extracting and exploring analysis ready samples (AppEEARS). Available online at: https://appeears.earthdatacloud.nasa.gov/(Accessed, 2024).

Google Scholar

O'Neill, B. C., Tebaldi, C., van Vuuren, D. P., Eyring, V., Friedlingstein, P., Hurtt, G., et al. (2016). The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geosci. Model Dev. 9, 3461–3482. doi:10.5194/gmd-9-3461-2016

CrossRef Full Text | Google Scholar

Ouillon, S. (2018). Why and how do we study sediment transport? Focus on coastal zones and ongoing methods [editorial]. Water 10, 390. doi:10.3390/w10040390

CrossRef Full Text | Google Scholar

Özşahin, E. (2023). Climate change effect on soil erosion using different erosion models: a case study in the naip dam basin, Türkiye. Comput. Electron. Agric. 207, 107711. doi:10.1016/j.compag.2023.107711

CrossRef Full Text | Google Scholar

Panagos, P., Borrelli, P., Meusburger, K., Yu, B., Klik, A., Jae Lim, K., et al. (2017). Global rainfall erosivity assessment based on high-temporal resolution rainfall records. Sci. Rep. 7, 4175. doi:10.1038/s41598-017-04282-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Panagos, P., Ballabio, C., Himics, M., Scarpa, S., Matthews, F., Bogonos, M., et al. (2021). Projections of soil loss by water erosion in Europe by 2050. Environ. Sci. Policy 124, 380–392. doi:10.1016/j.envsci.2021.07.012

CrossRef Full Text | Google Scholar

Panagos, P., Borrelli, P., Matthews, F., Liakos, L., Bezak, N., Diodato, N., et al. (2022). Global rainfall erosivity projections for 2050 and 2070. J. Hydrol. 610, 127865. doi:10.1016/j.jhydrol.2022.127865

CrossRef Full Text | Google Scholar

Patriche, C. V. (2023). Applying RUSLE for soil erosion estimation in Romania under current and future climate scenarios. Geoderma Reg. 34, e00687. doi:10.1016/j.geodrs.2023.e00687

CrossRef Full Text | Google Scholar

Polykretis, C., Alexakis, D. D., Grillakis, M. G., and Manoudakis, S. (2020). Assessment of intra-annual and inter-annual variabilities of soil erosion in Crete island (greece) by incorporating the dynamic 'nature' of R and C-Factors in RUSLE modeling. Remote Sens. 12 (15), 2439. doi:10.3390/rs12152439

CrossRef Full Text | Google Scholar

Renard, K. G., and Freimund, J. R. (1994). Using monthly precipitation data to estimate the R-factor in the revised USLE. J. Hydrol. 157, 287–306. doi:10.1016/0022-1694(94)90110-4

CrossRef Full Text | Google Scholar

Renard, K. G., Foster, G. R., Weesies, G. A., McCool, D. K., and Yoder, D. C. (1997). Predicting soil erosion by water: a guide to conservation planning with the revised universal soil loss equation (RUSLE). Washington, DC: USDA.

Google Scholar

Rouse, J. W., Haas, R. H., Schell, J. A., and Deering, D. W. (1974). “Monitoring vegetation systems in the great plains with ERTS,” in Proceedings of the third Earth resources technology Satellite-1 symposium (Washington, DC: NASA), 309–317.

Google Scholar

Sakhraoui, F., and Hasbaia, M. (2023). Evaluation of the sensitivity of the RUSLE erosion model to rainfall erosivity: a case study of the ksob watershed in central Algeria. Water Supply 23, 3262–3284. doi:10.2166/ws.2023.182

CrossRef Full Text | Google Scholar

Schwertmann, U., Vogl, W., and Kainz, M. (1990). Soil erosion by water. Stuttgart: Eugen Ulmer.

Google Scholar

Şen, O., and Kahya, E. (2021). Impacts of climate change on intensity–duration–frequency curves in the rainiest city (rize) of Turkey. Theor. Appl. Climatol. 144, 1017–1030. doi:10.1007/s00704-021-03592-2

CrossRef Full Text | Google Scholar

Stefanidis, S., and Stathis, D. (2018). Effect of climate change on soil erosion in a mountainous mediterranean catchment (central pindus, Greece). Water 10, 1469. doi:10.3390/w10101469

CrossRef Full Text | Google Scholar

Tavakoli, M., and De Smedt, F. (2012). Impact of climate change on streamflow and soil moisture in the vermilion basin, Illinois. J. Hydrol. Eng. 17, 1059–1070. doi:10.1061/(asce)he.1943-5584.0000546

CrossRef Full Text | Google Scholar

Tayanç, M., İm, U., Doğruel, M., and Karaca, M. (2009). Climate change in Turkey for the last half century. Clim. Change 94, 483–502. doi:10.1007/s10584-008-9511-0

CrossRef Full Text | Google Scholar

Teng, H., Liang, Z., Chen, S., Liu, Y., Viscarra Rossel, R. A., Chappell, A., et al. (2018). Current and future assessments of soil erosion by water on the Tibetan Plateau based on RUSLE and CMIP5 climate models. Sci. Total Environ. 635, 673–686. doi:10.1016/j.scitotenv.2018.04.146

PubMed Abstract | CrossRef Full Text | Google Scholar

Torri, D., Poesen, J., and Borselli, L. (1997). Predictability and uncertainty of the soil erodibility factor using a global dataset. Catena 31, 1–22. doi:10.1016/S0341-8162(97)00036-2

CrossRef Full Text | Google Scholar

Torri, D., Poesen, J., and Borselli, L. (2002). Corrigendum to “Predictability and uncertainty of the soil erodibility factor using a global dataset” [Catena 31 (1997) 1–22]. Catena 46, 309–310. doi:10.1016/S0341-8162(01)00175-8

CrossRef Full Text | Google Scholar

Torri, D., Borselli, L., Guzzetti, F., Calzolari, M. C., Bazzoffi, P., Ungaro, F., et al. (2006). “Italy,” in Soil erosion in Europe. Editors J. Boardman, and J. Poesen (Chichester: John Wiley and Sons), 245–261.

Google Scholar

Tuo, D., Lu, Q., Wu, B., Li, Q., Yao, B., Cheng, L., et al. (2023). Effects of wind–water erosion and topographic factor on soil properties in the loess Hilly region of China. Plants 12, 2568. doi:10.3390/plants12132568

PubMed Abstract | CrossRef Full Text | Google Scholar

Ugese, A. A., Ajiboye, J. O., Ibrahim, E. S., Gajere, E. N., Itse, A., and Shaba, H. A. (2022). Soil loss estimation using remote sensing and RUSLE model in koromi-federe catchment area of jos-east LGA, Plateau state, Nigeria. Geomatics 2, 499–517. doi:10.3390/geomatics2040027

CrossRef Full Text | Google Scholar

Van der Knijff, J., Jones, R., and Montanarella, L. (1999). Soil erosion risk assessment in Italy. Brussels: European Commission.

Google Scholar

Vicuna, S., and Dracup, J. A. (2007). The evolution of climate change impact studies on hydrology and water resources in California. Clim. Change 82, 327–350. doi:10.1007/s10584-006-9207-2

CrossRef Full Text | Google Scholar

Walling, D. E., and Webb, B. W. (1996). Erosion and sediment yield: a global overview, in Walling, D., and Webb, B. (Eds.), Erosion and Sediment Yield: Global and Regional Perspectives (Proceedings of the Exeter Symposium, July 1996) 236. Wallingford, United Kingdom: IAHS Publ, 3–19.

Google Scholar

Wischmeier, W. H., and Smith, D. D. (1957). Factors affecting sheet and rill erosion. Trans. Am. Geophys. Union 38 (6), 889–896.

Google Scholar

Wischmeier, W. H., and Smith, D. D. (1978). Prediction rainfall erosion losses: a guide to conservation planning. Agric. Handb. 537.

Google Scholar

Woznicki, S. A., and Nejadhashemi, A. P. (2013). Spatial and temporal variabilities of sediment delivery ratio. Water Resour. Manag. 27, 2483–2499. doi:10.1007/s11269-013-0298-z

CrossRef Full Text | Google Scholar

Wuepper, D., Borrelli, P., and Finger, R. (2020). Countries and the global rate of soil erosion. Nat. Sustain. 3 (1), 51–55. doi:10.1038/s41893-019-0438-4

CrossRef Full Text | Google Scholar

Yagbasan, O. (2016). Impacts of climate change on groundwater recharge in küçük menderes river basin in Western Turkey. Geodin. Acta 28 (3), 209–222. doi:10.1080/09853111.2015.1121802

CrossRef Full Text | Google Scholar

Yılmaz, M., Alemdar, K. D., and Tosunoğlu, F. (2025). Evaluation of NEX-GDDP-CMIP6 to simulate precipitation using multi-criteria decision-making analysis over Türkiye. Theor. Appl. Climatol. 156, 578. doi:10.1007/s00704-025-05805-4

CrossRef Full Text | Google Scholar