Simulating with a combination of RUSLE GIS and sediment delivery ratio for soil restoration

Erosion by water is the main cause of land degradation. Landscapes degraded by erosion need to be restored in many respects, and particularly in terms of ecosystem services. From an economic and management perspective, care is needed to select priority areas and determine the means to be applied to restore them. Globally, the model most commonly used to produce scenarios for the prevention of soil losses is the Revised Universal Soil Loss Equation (RUSLE). This study of the subbasin of the Sulakyurt Dam Basin in Turkey aims (1) to identify the distribution of soil losses over time and by location, and (2) to grade the priority areas for the prevention of soil losses by means of a simulation. The average potential soil losses in the area under study are estimated at 42.35 t ha−1 year−1, and the average actual losses at 39.49 t ha−1 year−1. According to the simulation, 27.61% of the study area (2782 ha) is of the highest priority for soil restoration. In our study, forests have the highest soil losses, which is contrary to the natural protection that forests provide against erosion. The high rates are due to the slope, the forest area is very steep. So it is the slope factor that outweighs the vegetation cover factor. Of the forest areas, 41.74% (1766 ha) falls within the areas of highest priority. The study serves as a guide for landscape planning and the determination of erosion risk in restoration efforts, and for identifying the methods to be adopted during the restoration work to reduce the loss of soil.


Introduction
Erosion has many effects on the world, while the tolerable erosion (geological erosion) has a positive effect on natural processes such as shaping the landscape and material transport, the accelerated erosion causes soil degradation (Lal, 2001). According to Ruhe (1969), the most fundamental finding is that most land formations are produced by erosion. Consequently, the shaping of the landscape is largely designed by erosion. However, erosion has numerous negative impacts as well. Today, the discussion is not only about the carrying away of the soil but also about the displacement of the nutritional elements, chemicals and fertilizers which the soil contains (Morgan, 2009). This process causes the fertility of agricultural lands to decline (Bogunovic et al., 2018). It also constitutes a threat to aquatic life in the lakes and pools where the sediment accumulates. The accumulation of eroded soil in dams built upon rivers reduces their economic lifetimes and reduces the amount of energy generated (Aga et al., 2018;Özcan & Ediş, 2016;Saygın et al., 2014). The degree of soil erosion depends on the wind, rainfall and the related flow processes, on the propensity of the soil to erosion and on the nature of the land cover and the way the land is managed (Aksoy & Kavvas, 2005;Panagos et al., 2015a). The spatial distribution and amount of soil loss around the world is changing due to the landscape processes and soil management. For example, using the RUSLE model, Borrelli et al. (2017) estimated potential soil losses worldwide as 2.8 t ha −1 year −1 , Panagos et al. (2015a)  The control of erosion is just as important for ecosystem services as it is for landscape planning. In the categorization of ecosystem services, it is considered one of the regulating services (De Groot, 2006). Landscape planning has been proven to be an effective way of controlling erosion, and requires the formulation of an institutional framework on the main physical plan of an area subject to environmental management (Owonaiye, 2012). Where landscapes have been degraded, landscape planning and management (Pena et al., 2020) and soil protection measures (Panagos et al., 2015b) need to be used in conjunction with one another. The monitoring of soil erosion is an important tool for planning the protection of the soil. However, calculating soil losses is expensive and time-consuming, even for trial plots (Mohammed et al., 2020). Land can also be managed by understanding how these erosion processes come about and which areas are vulnerable to soil loss. A variety of empirical models have been developed for estimating water erosion so as to be able to give priority to those areas where the erosion is most intensive and to make projections for areas for which limited data is available (Aksoy & Kavvas, 2005;Issaka & Ashraf, 2017). The most widely used model for the determination of soil losses is the USLE/RUSLE model (Wischmeier & Smith, 1978;Renard et al., 1996). The RUSLE model is used not only to estimate the amounts of soil eroded and their distribution by location but also to indicate what kind of intervention needs to be made in the pattern of the landscape in order to reduce the erosion or protect the soil (Benavidez, 2018;Özcan & Aytaş, 2020;Pınar et al., 2020;Zare et al., 2017). In the model, the cover management factor (the C factor) relates to the conditions that can be most easily controlled in reducing erosion for the purpose of landscape restoration, and can therefore be described as one of the most important planning instruments for policies and land use decisions (Biddoccu et al., 2016;Maetens et al., 2012;Mukharamova et al., 2021).
In the 1950s, the US Department of Agriculture (USDA), accepting the potential for accelerated erosion under modern industrial agriculture, determined T-values for use in evaluating the soil loss tolerance levels or the "acceptable" rates of soil erosion (NRCS, 2003). Schertz (1983) took the T-values to be equivalent to erosion of 5-12 t ha −1 per year or 0.4-1 mm year −1 (assuming the weight of the soil to be 1200 kg m −3 ). In land use planning, the aim is to reduce the erosion to this level. In Turkey, the erosion limit has generally been taken as 10 t ha −1 (Madenoğlu & Erpul, 2018;Özcan & Aytaş, 2020;Pınar et al., 2020). Both the C factor and the P factor can be used in scenario analyses to understand how the loss of soil can be reduced or exacerbated through the use of varying means of soil protection (Benavidez, 2018). Özcan and Aytaş (2020) used the RUSLE model to simulate alternative soil protection practices to prevent Lake Bakkal Dolin (Çankırı/Turkey) from being filled with sediment, and demonstrated which method would potentially reduce the most soil losses. Panagos et al. (2015b) used the P factor alongside the C factor on the scale of Europe, while Pınar et al. (2020) and Madenoğlu and Erpul (2018) used the C factor in their scenario approach for sustainable basin management.
Sulakyurt (Kırıkkale/Türkiye), where the present study was conducted, has a fragile, semi-arid ecosystem. The aims of the study were (1) to assess the soil losses in the dam basin in terms of time and location, (2) to use erosion reduction simulations for soil restoration to identify areas with different degrees of priority, and (3) to provide a guide for determining the erosion function and the risky areas in landscape planning. The results of the study will contribute to the understanding of the relationship between landscape patterns and soil losses, the promotion of sustainable landscape management, the availability of alternatives for landscape planning in areas where the landscape is being degraded, and the availability of new planning methods.

Study area
The study area is the basin of the Sarıkızlı brook, located on the North-facing slopes of the Karagüney Mountains in the district of Sulakyurt/Kırıkkale, which is one of the most important tributaries attached to the Sulakyurt Dam, and covers an area of about 106 km 2 (Fig. 1). The basin has an altitude of between 730 and 1370 m (Mean: 1061.69 ± 139.23 m) and a gradient of between 0 and 78° (Mean: 11.17 ± 8.38 degree). Although the dominant climate in Central Anatolia is a semi-arid climate, the climate in the basin is better described as semi-humid, due to the effect of the Karagüney Mountains (Yılmaz & Çiçek, 2018). The annual average precipitation is 389 mm, most of which falls in the winter and spring. The average annual temperature is 12.7 °C, varying from 21 °C in the hottest month (July) to 0.8 °C in the coldest month (January) (MGM, 2020). The Kızılırmak Formation, consisting of an intercalation of marl, conglomerates, mudstone and sandstone, prevails over most of the study area. However, the Faraşlı volcanics between the villages of Faraşlı and Kalekışla also contain seams of basalt (Evcimen, 2011). No specific study has been made of the flora of the basin, but there are studies of the Karagüney Mountains within which the basin is located (Dönmez, 2002;Özcan, 2021). According to Dönmez (2002), 378 species and 868 taxons belonging mostly to the Asteraceae, Fabaceae, and Poaceae families have been identified in the Karagüney Mountains, which fall under the Irano-Turanian phytogeographical region. The South of the area between 900 and 1350 m is made up of forests. The basin contains areas of turkey oak and downy oak forest, and there is a Valonia oak forest in the agricultural areas that takes the form of a low-density (10-40%) wooded area (Özcan, 2021). Stands of downy oak (at low altitudes) and turkey oaks (from 1150 m upwards) form forests up to 10 m tall with dense canopies. The most significant element of the flora in the study area is the existence of the largest remnant Valonia oak forest outside its natural distribution area (Özcan, 2021). Figure 2 shows the flowchart of the methodology. The USLE/RUSLE estimation model was used to determine the risk of erosion in the study area (Wischmeier & Smith, 1978;Renard et al., 1996) (Eq. (1)).

Soil losses
In Eq. (1), A represents the average soil loss due to water erosion (t ha −1 year −1 ); R, the rainfall and runoff erosivity factor (MJ mm ha −1 h −1 year −1 ); K, the soil erodibility factor (t ha hour ha −1 MJ −1 mm −1 ); L, the slope-length factor; S, the slope-steepness factor; C, the cover management factor; P, the support practice, and SDR, the sediment delivery ratio. The values of the factors were converted into SI units (Système International d'Unités) for the study (Foster et al., 1981).
The erosivity factor is defined as the erosive force of rainfall and is calculated as a product of the annual total energy of rainstorm (EI) and the maximum 30-min intensity (I30) (Foster et al., 1981;Wischmeier & Smith, 1978). For the study area, the R factor value calculated monthly by Erpul et al. (2016) for the Çankırı Meteorology Station (Table 1) has been used.
In Eq.
(2), E represents the annual total energy of rainstorm (MJ ha −1 month −1 ) and I the amount of rainfall (mm hour −1 ) at the maximum 30-min intensity. The use of a single value to generate the R factor map may lead to erroneous results, since it does not take into account the impact of altitude on the precipitation conditions (Toy & Foster, 1998). For this reason, a map of the rainfall in the study area for each month was first developed using WorldClim average monthly rainfall data (Fick & Hijmans, 2017), and then the R factor values for the entire area of the basin were established using Eq. (3) (Erdogan et al., 2007).
In Eq. (3), Ry represents the monthly R factor value for the pixel being estimated, Rt the monthly R factor value for the meteorology station, Py the monthly average amount of rainfall for each pixel, and Pr the monthly average amount of rainfall for the meteorology station. The total of the R factors for all the months gives the annual R factor value.
The K factor is defined as the resistance of the soil to erosion. The K factor map was obtained from the 1:25,000 scale Turkey Erosion Data Base . In the database, the K factor is calculated with Torri et al. (1997) Eq. (4) using 23,000 soil profile data points.
In Eq. (4), K represents the soil erodibility factor (t ha hour ha −1 MJ −1 mm −1 ); DG is the geometrical average diameter of the primary soil components Environ Monit Assess (2023) 195:719 Page 5 of 17 719 Vol.: (0123456789) (mm); OM is organic matter, and C is the percentage of clay. The topography is the most important factor that controls the risk of erosion. The LS factor, which is a product of the length and steepness of the slope, has been calculated using Eq. (5) in Digital Elevation Model (DEM) and ArcGIS.  In Eq. (5), X is the surface flow condensation value, n is the size of the cells on which the calculation is performed, and θ is the steepness of the slope. The gradient in the study area was calculated using DEM and the length of the slope was taken to be a fixed value of 15 m for each pixel (Ogawa et al., 1997).
The plant cover is the primary factor that controls the risk of soil erosion. In general, the plant cover reduces the kinetic energy of the raindrops before they strike the surface of the soil and prevents the dispersal of the soil by holding it together. This factor takes a value of between 0 and 1. In order to be able to calculate the C Factor for this study, a Sentinel 2B image was obtained from the Copernicus Data and Information Access Services for each month between January and December 2018 taking the cloud ratios into account (USGS, 2020). The NDVI values obtained were used to calculate the RUSLE C factor values by location using the exponential Eq. (6) developed by Van der Knijff et al. (1999).
In Eq. (6), α and β are regression coefficients. Van der Knijff et al. (1999) state that a value of 2 for α and a value of 1 for β could be a suitable result. Since no soil preservation measure has been taken in the study area, the value of P was taken to be 1.
A basin may be studied in two main parts: the accumulation area and the erosion area. However, the USLE/RUSLE model calculates the basin as an erosion area. Not all of the sediment carried off may reach the accumulation area; some of it accumulates in areas where the slope becomes less steep or is interrupted. The proportion of sediment that accumulates in such places without reaching the main accumulation area can be called the sediment delivery ratio (SDR) (Yılmaz, 2006). In this study, the SDR was calculated using Eqs. (7) and (8) (Ferro et al., 2001).
NDVI − NDVI of hydrological units evaluated; lp,i and Sp,i, the length and steepness of the slope of the hydrological path, and β, the assumed value of the coefficient for the study area (m −1 ). The coefficient β was obtained by means of Eq. (8).
In Eq. (8), (RUSLE-R) i is the value of the RUSLE R factor for the study area. Finally, the potential soil losses were obtained by multiplying the SDR map by the RUSLE soil losses map.

Simulation for mitigating soil losses
In our scenario approach, the soil loss tolerance (T) level is taken as the threshold value for conservation measures in soil erosion. The value of T is generally taken to be between 5 and 12 t ha −1 (Schertz, 1983). Consequently, the permissible soil loss for the basin is set at 10 t ha −1 year −1 . The soil loss tolerance level could be reduced further. However, considering that greater interventions in natural processes would have both ecological and economic costs, this level of soil loss tolerance can be considered appropriate. First the areas where there is a problem of erosion were identified and the CP factor required to reduce the loss of soil in these problem areas to the threshold level was calculated for each pixel using Eq. (9). To avoid duplication, the map showing the distribution of the C factor by location was divided by the soil losses map and removed from the calculation.
In Eq. (9), CP represents the factor value required to reduce the soil loss for each pixel to less than 10 t ha −1 year −1 ; T, the soil loss tolerance (10 t ha −1 year −1 ), and A, the soil loss (t ha −1 year −1 ). In classifying the CP factor by location for the purposes of soil restoration, this factor, which takes a value of between 0 and 1, was divided into five restoration In Eq. (7), SDRi represents the sediment delivery ratio; (SA)t, the area of the study area; N, the number classes at intervals of 0.2. Finally, the maps were prepared indicating the priorities for the restoration Page 7 of 17 719 Vol.: (0123456789) of the landscape for agricultural, pasture, and forest land.

Soil losses
The soil losses in the study area were determined using the RUSLE model. The most important advantage of the RUSLE model is that it reveals not only the locations of the areas at risk of erosion and the levels of soil erosion by location but also the measures to be taken. Each of the variables that constitute the model was analysed separately. In particular, the C factor and R factor, which depend on climatic variables, were calculated on a monthly basis.
The R factor maps for the basin were calculated on a monthly and annual basis ( Table 2). The annual value of the R factor in the basin was found to vary between 252.34 and 320.70 MJ mm ha −1 year −1 (mean: 284.98 ± 15.17) (Fig. 3a). The southern and south-eastern parts of the basin have higher rainfall erosivity values than the northern and north-western parts, and there is a clear upward trend in erosivity from North to South within the basin, in line with the differences in altitude. In addition to the spatial distribution of the RUSLE-R results, the distribution over time is of great importance in the semi-arid regions of Central Anatolia, where the occurrence of precipitation events is rather irregular (Bayramin et al., 2006;Özcan et al., 2008). A rainstorm may result in a serious loss of soil in the fallow period, whereas the same field might hardly be damaged at all throughout the growing season. In Ethiopia (Nyssen et al., 2005) and Switzerland (Mabit et al., 2013), monthly estimates of rainfall and rainfall intensity have been used to assess temporal variations in erosivity. An examination of the R factor by months puts rainfall in May at 48.90-71.90 MJ mm ha −1 month −1 .
In seasonal terms, the R factor reaches its highest value in the summer months: the values of the R factor average 56.02 MJ mm ha −1 month −1 in June, 22.73 MJ mm ha −1 month −1 in July and 44.21 MJ mm ha −1 month −1 in August. While precipitation is low in summer, the main reason for these high values is the incidence of sudden rainstorms. As for the winter, the value of the R factor averages 3.30 MJ mm ha −1 month −1 in November, 5.29 MJ mm ha −1 month −1 in December and 12.82 MJ mm ha −1 month −1 in January. The low value of the R factor for the winter months may be attributable to the gentle nature of the precipitation and the fact that it generally falls as snow. Panagos et al. (2017) show that rainfall coefficients in Greece are more "sensitive" between May and December than between January and April. With global warming, it is predicted that the annual decline in rainfall in the Mediterranean region and Anatolia as a function of the increase in average annual surface temperatures globally will be more than 50 mm/K (Lionello and Scarascia, 2018). Panagos et al. (2017) have calculated that the average R factor in Europe will increase by 16% by 2050 under the RCP 2.6 climate change scenario (the optimistic scenario) and by 27% under the RCP 8.5 climate change scenario, and that 81% of the area of Europe will be subject to increases in the R factor. For Turkey, there is no estimate of how the R factor will develop in the future. However, it is clear that global warming will be accompanied by many climatic events such as changes in temperature and precipitation, shifts in seasonal rainfall patterns and violent storms. Consequently, it will be more useful to calculate the R factor on a monthly basis when adopting soil conservation measures. It is not possible to change or reduce the R factor, but this will make an important contribution to the management of the C factor.
The K factor maps (Fig. 3b) are obtained from erosion databases (Erpul et al., 2016). The K factor in the basin varies from 0.019 t ha hour ha −1 MJ −1 mm −1 to 0.021 t ha hour ha −1 MJ −1 mm −1 , with an average value of 0.02 t ha hour ha −1 MJ −1 mm −1 . The soils in the basin are predominantly reddish brown. These reddish-brown soils are low in organic matter, alkaline and rich in lime. The soils formed beneath the mixed forests contain greater amounts of organic material than the soils under the meadows. Research has shown that organic matter has certain roles in regulating numerous physical qualities of the soil such as its bulk density, its hydraulic conductivity and its aggregate stability, and that it greatly affects the proclivity of the soil to erosion (Cerda, 1996;Wu & Tiessen, 2002;Başaran et al., 2008). The LS factor determines the spatial resolution (cell dimensions) of the soil erosion model and expresses the potential for soil erosion stemming from surface flows (Panagos et al., 2015c). The L factor indicates the effect of the length of the slope and the S factor the effect of its steepness. The LS factor is non-dimensional. Its value for the basin is between 0 and 293 (mean: 21.24 ± 23.79) (Fig. 3c). The slope becomes steeper towards the South of the basin. In areas where the gradient and the length of the slope are greater, the value of the LS factor is higher. In the cliff-like areas in the forests in the South of the basin, where the gradient is highest, the value of the LS factor is particularly high.
The C factor and P factor are acknowledged to be dynamic factors that change over time. Traditionally, a single empirical value is attached to the C factor based on the tables in the USLE/RUSLE guidebook (Wischmeier & Smith, 1978;Renard et al., 1996) or the classification of the land cover (Ganasri & Ramesh, 2016;Kavian et al., 2017). This method is relatively easy, but fails to capture the real variations in the plant cover over space and time, and therefore leads to errors in the calculation of C values (Zhao et al., 2013). Simply attributing RUSLE-C values to all the classes of land cover does not allow for a very accurate estimate of the growth and decline of the vegetation in the course of the year (Van der Knijff et al., 1999). In the present study, C factor maps were obtained for each month using NDVI and regression equations (Fig. 3d). When evaluated together with the monthly erosivity maps, this makes it possible to take account of the interaction between the rainfall and the growth of the plant cover. The month in which the C factor values (Table 3) for the basin are lowest is June, with an average of 0.24, and the month in which they are highest is December, with 0.75. The average values of the C factor in agricultural areas are between 0.23 and 0.73. Here, the C factor is at its lowest between March and July, and is just 0.23 in June. The agricultural land within the study area is generally used to produce grains. The C factor values here are consistent with the harvest season in the region, which is the end of June, and with the C factor values for agricultural areas given in a study of the same basin (Özcan & Aytaş, 2020). Another consideration that helps to explain the high average C factor values in the agricultural areas is the use of a fallow system. In fact, one of the most important findings of the study was obtained from the agricultural areas where the Valonia oak grows at a density of 10-40%. In these areas, the value of the C factor was between 0.23 ± 0.09 and 0.73 ± 0.12. Compared to the agricultural areas where the oak trees grow, the C factor values of the other agricultural areas are higher and more varied. The protective influence of trees growing in and around areas of dry agriculture against erosion is quite distinctive, particularly when the fallow system is considered. From the point of view of soil conservation, these results underline the impact which the planting of trees within and around areas of dry agriculture can have. In the pastures, the value of the C factor turned out to be lowest in June, at 0.23, and highest in December, at 0.81. The pastures within the study area are not steep-sloping areas. Moreover, these areas are faced with a serious conservation problem due to grazing. The vegetation cover in the pastures has fallen to less than 30%. In the forest areas, the C factor is lowest in June with a value of 0.064 and highest in December with 0.85. The main reason why the C factor turns out to be high in December regardless of the form of land use is that most of the surface of the basin is covered in snow.
The potential soil losses in the study area were obtained by location by multiplying the six factors. The potential soil losses in the area vary between 0 and 618.19 t ha −1 year −1 (mean: 42.32) (Fig. 4a). The forest areas in the study area show the highest degree of erosion, with soil loss values ranging from 0 to 610.31 t ha −1 year −1 (mean: 56 ± 55.13 t ha −1 year −1 ). The soil losses in the forests were well above the average for the basin. This contrasts with what happens generically with forests, since they have low erosion rates due to the protection they provide to the soil; however, the explanation for this is to be found in the other factors that influence soil losses, especially in topographical factors. The average soil loss worked out at 25.77 ± 32.05 t ha −1 year −1 (0-513.77 t ha −1 year −1 ) in the agricultural areas and 32.50 ± 35.70 t ha −1 year −1 (0-560.13 t ha −1 year −1 ) in the pastureland. The parts of the study area where the potential soil losses are the lowest are the valley floors, where the land is almost flat and is used for agriculture. Very low values were also recorded in the extensive flat areas in the upper regions of the forest land in the South. The highest values within the study area were found to be in the scattered oak forests in the cliff-like areas in the North and South where the gradient is very steep. The results obtained differ from the findings of studies of soil erosion conducted in nearby areas with respect to some forms of land use (Yılmaz, 2006;Erdoğan et al., 2007;Özcan et al., 2008;Saygın et al., 2014;Özcan & Ediş, 2016;Özcan & Aytaş, 2020). In the literature, agricultural land such as orchards, vineyards and arable land is frequently said to have higher rates of soil erosion than forests (Cerdan et al., 2010). Özcan et al. (2008) and Saygın et al. (2014) estimated lower potential soil losses for layered black pine and oak forests with high canopy cover.
According to the calculations made for each month (Table 4), the potential soil losses were below average in January, February, March, July, September, October, November, and December. In the remaining months, the average soil losses were above average. The least soil loss was found to occur in November (mean: 0.58 t ha −1 month −1 ). Although the C factor in November, at 0.40, was higher than the average of 0.37, the fact that the average R factor was very low, at 3.30 MJ mm ha −1 month −1 , resulted in a low soil loss estimate. By contrast, May was the month when the greatest soil loss was observed (8.68 t ha −1 month −1 ). Approximately 21% of all the potential soil loss is estimated to occur in May. Similarly, 72% of the soil losses turned out to occur in the 5-month period from April to August. The main reason for this is that the R factor, which represents the erosivity of the rainfall, is high in these months: the total of the R factor values calculated for these months makes up 76% of the total for the entire year. Unlike the R factor, the C factor records its lowest average values during these months. In a remote-sensing study, Yang et al. (2020) showed that the rates of sediment loss approximately doubled between summer and winter due to the influence of the vegetation growth cycle or its seasonal effect.
The soil that is eroded is not all carried to the accumulation areas, but may accumulate in various other places, such as points where the gradient becomes less steep or the slope is interrupted (Yılmaz, 2006). The sediment delivery ratio (SDR) values for the basin are between 0 and 1, with an average of 0.86 (Fig. 4b). The main reason for the high SDR in the basin is the structure of the slopes, which are predominantly convex. In addition, the SDR is known to fall as the area of the subsidiary basins within a basin increases (Yılmaz, 2006). In this case, another reason for the high SDR is that while there are a relatively large number of subsidiary basins in the study area, they are small in area. The highest ratios were obtained for the forest areas. These ratios are proportionate to the LS factor values. The actual soil losses were estimated by multiplying the potential soil losses in the study area with the SDR value. The topographical structure of the Sarıkızlı Basin resulted in a high SDR. For this reason, there was not a great difference between the potential soil losses and the actual soil losses. The actual soil losses for the study area were calculated to be between 0 and 610.31 t ha −1 year −1 (mean: 39.49 t ha −1 year −1 ) (Fig. 4c). By comparison with the potential soil losses, the actual soil losses were 2.86 t ha −1 year −1 or 10% lower on average.

Simulation for mitigating soil losses
The P factor is calculated for controls to be applied to the flow concentrations and speeds of the types of drainage and to the surface of the soil (Renard et al., 1991). Since no erosion control work is being carried out in the study area, the P factor was taken as 1. The C factor represents the average of the soil loss ratios (SLR) as they change over time, each one weighted by a part of the rain erosivity during the same period (Napoli et al., 2016). When modelling the P factor on the scale of Europe, Panagos et al. (2015b) did so in conjunction with an estimate of the C factor. In the same way, Madenoğlu and Erpul (2018) and Pınar et al. (2020) used the C factor in their scenario approach to sustainable basin management in Turkey. In the present study, the C and P factors were assessed simultaneously for the reduction of soil losses.
The loss of soil in approximately 6576.06 ha of the basin (62.88% of the total area of the basin) is greater than the value of T (t ha −1 year −1 ) ( Table 5).
The simulation values of the CP factor calculated for these areas range from 0.01 to 0.99, with an average of 0.35 ± 0.24 (Fig. 5). In the simulation conducted by Özcan and Aytaş (2020) in order to reduce the sediment reaching Lake Bakkal Dolin, the average CP factor obtained was 0.59 ± 0.24. In the remaining 3882.42 ha (37.12%) of the basin, there is no need to take any action for the loss of soil. The CP values for the basin were quite low. In about 48.13% of the study area, the CP factor needs to be reduced to less than 0.4. In classifying the CP factor by location for the purposes of soil restoration, this factor, which takes a value of between 0 and 1, was divided into five restoration classes at intervals of 0.2. In the areas falling within TRI, the class with the greatest priority, the CP factor is between 0 and 0.2. These are the areas where the risk of erosion is highest. This top priority area for land restoration comprises 27.61% of the entire area including 41% of the forests, 14.71% of the agricultural areas, and 20.02% of the pastureland. In 37.5% of the agricultural areas, 24.9% of the pastures and 16.60% of the forests, there is no risk in terms of soil losses.
The results obtained allow validating the methodology used, as well as the combined use of RUSLE GIS and SDR for the quantification of soil degradation by erosion, with the objective of soil and vegetation restoration, as already done in some previous works (Galdino et al., 2016;Sun et al., 2014).

Conclusions and suggestions
Landscape planning and management comprises future-oriented actions for developing, restoring or creating landscapes to ensure sustainable development, which can be achieved by ensuring sustainable ecosystem services. The primary and common reason for land degradation is the improper use of land and erroneous land management practices. Therefore, a dynamic stability must be aimed at in terms of time and scale for landscape planning in order to be sustainable. Eliminating soil erosion also ensures that the quality of the soil improves and thereby serves as a guarantee for food and water security. Estimates of soil loss in the study region have been developed, with data broken down by the different land uses and locational distribution. In wooded areas, erosion is most severe. The most important reason for this is that the forest areas are on sloping land, and the plant cover is particularly low in these areas. These areas have very high CP value according to the simulation results. In these areas, the economic impact of plantation techniques needs to be considered. Reducing slope is expensive and time-consuming. But the slope length can be reduced with terraces. It would be appropriate to plant afforestation with one-meter-wide water collection terraces (which also encourages the p value) in forest areas with low cover. However, in areas where it is necessary to reduce the p value, it may be necessary to take measures such as improvement benches and stone walls.
Agricultural areas are subject to the lowest level of soil erosion, since they are situated on alluvion beds in the valleys and flatter lands within the forests. Low CP values of farmland are due to more vegetation in the crop rotation process, especially the fallow of grass-dominated areas triggers erosion. Erosion control measures, which require minor adaptations to traditional farming practices, appear to be the most cost-effective. It shows that the implementation of tillage measures in these areas is an effective way to control soil loss. In these areas, contour farming, mulching, buffer strips, high-density planting are the most cost-effective erosion control measures for agricultural areas. The long-term value of the C factor is influenced by the crop rotation in the arable area. Agricultural species that are resistant to semi-arid climate and cover the soil should be preferred. Finally, in the grassland, the vegetation dries quickly with decreasing rainfall and increasing temperatures. We think that as a drastic solution in the pasture areas, which are the root of the issue, is to temporarily stop animal grazing on the slopes. Stone walls can contribute to the reduction of erosion, especially on steep slopes.
According to the results obtained, there is generally a very high level of soil erosion within the basin. The purpose of the simulation model is to decide (1) which areas are to be prioritised in landscape restoration and (2) what form and method of restoration is to be adopted. The level of intervention has been determined for each type of land use. The factor that most influences decision-making in landscape restoration is the economic value of the intervention. Taking this cost into account, the first rule of landscape restoration is to monitor the land before it becomes degraded. The simulation model also provides a map of the areas that need to be monitored for the basin. Recognizing traditional practices, adopting suitable local policies, using a participatory approach, and being mindful of the cost-benefit ratio are all necessary for landscape restoration and rehabilitation projects to increase sustainable land management.