Assessment of soil erosion response to climate change in the Sululta catchment, Abbay Basin, Ethiopia Open Access

Erosion of soils becomes a global problem that can cause a variety of issues, including decreased yield of soil, deteriorating water quality, reduced reservoir capacity for storage, as well as the disruption of the aquatic environment (Wang et al. 2020). Soil in good condition serves an extensive variety of critical functions, such as food and biomass production, water and carbon-storing and filtering, and carbon dioxide, as well as other emitted gas monitoring (Moges et al. 2020). As a consequence, determining the loss of soil is essential for the development of local ecological environments, and soils should be preserved and utilized in an environmentally friendly way.

The erosion of soil triggered by water movement is nowadays among the most severe problems worldwide (Chakrabortty et al. 2020). Removing nutrients from the top layer can reduce soil fertility and lead to soil degradation (Keesstra et al. 2018). The declining soil fertility is both environmentally and economically detrimental to the whole area. Deforestation, overgrazing, and intensive subsistence farming may become major contributors to soil erosion and degradation as population pressure rapidly increases (Blaikie 2016). Human activities in the environment as a whole, including modifications to methods of land use, can accelerate soil erosion (Pal & Shit 2017). The soil rate of erosion may decrease crop yields and enhance sedimentation rates (Wang et al. 2012).

The changing climate has the potential to exacerbate the erosion of soil, damage to crops, and the accumulation of excess water in plants' root zones, making farming challenging or even impossible (Girmay et al. 2021). Changes in the climate will cause changes in the rate of erosion of soil resulting from the enhanced erosive power of rainfall and changes in plant cover (Pruski & Nearing 2002). Erosion of the soil will be influenced by changes in plant cover and moisture content in the soil as temperatures rise (Wang et al. 2020). According to the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC), overall temperatures and precipitation are changing and are going to keep doing so in the 21st century (IPCC 2013). Precipitation and temperature changes have an impact on absorption rates, water content of the soil, the biomass of plant output, and the management of crops, as well as the erosion of soil and runoff (Li & Fang 2016). The impact of climate change may accelerate worldwide erosion from water from +30 to +66% in the coming 50 years (Borrelli et al. 2020). Precipitation exerts a direct impact on the loss of soil, whereas temperature has an indirect impact. Changes in precipitation are the primary cause of erosion of soil (Li & Fang 2016). Ethiopia is among those countries severely impacted by the increased erosion of soil resulting from changing climates, leading to a projected 23% rise in soil loss in 2050 (Moges et al. 2020). Due to high slopes as well as excessive precipitation, the Ethiopian highlands have a considerable loss of soil (Moisa et al. 2021). There are many studies quantifying soil erosion under the present climate in Ethiopia (Gelagay & Minale 2016; Haregeweyn et al. 2017; Gashaw et al. 2018). Few studies have been conducted to predict the potential threat of erosion of soil as a result of changes in the climate (Mengistu et al. 2015; Moges et al. 2020; Girmay et al. 2021). Moges et al. (2020) conducted research employing the Revised Universal Soil Loss Equation (RUSLE) model together with the geographic information system (GIS) to assess soil erosion using the RCP4.5 scenario from 2018 to 2050. Girmay et al. (2021), on the other hand, utilized the Universal Soil Loss Equation (USLE) model coupled with the GIS to project the erosion of soil under both RCP4.5 and RCP8.5 for the Agewmariam watershed from 2011 to 2100. This study utilized the RUSLE model in combination with the GIS to project the loss of soil under both RCP4.5 and RCP8.5 over the period from 2021 to 2080.

There are several methods for assessing the erosion of soil at different levels, such as globally, regionally, and locally (Poesen et al. 2003). Several models were developed to estimate the loss of soil rates (Poesen et al. 2003; Pan & Wen 2014). The following are among the most frequently employed models in soil erosion assessment: Water Erosion Prediction Project (WEPP) (Nearing et al. 1989), Chemical Runoff and Erosion for Agricultural Management System (CREMS) (Knisel 1980), European Soil Erosion Model (EuroSEM) (Morgan et al. 1998), Areal Non-point Source Watershed Environment Response Simulation (ANSWERS) (Beasley et al. 1980), Soil and Water Assessment Tool (SWAT) (Srinivasan et al. 1998), USLE (Wischmeier & Smith 1978), RUSLE (Renard 1997), and others. Due to its structural simplicity (Volk et al. 2010), adaptability, and integration with the GIS (Pandey et al. 2021), the RUSLE (Renard 1997; Ozsoy et al. 2012) is a widely employed empirical equation. It is also suitable with remote sensing and a digital elevation model (DEM) for the assessment of soil erosion (Kouli et al. 2009). The application of process-oriented approaches like WEPP is difficult owing to a shortage of suitable input information. The RUSLE model was selected in the current study because it is a commonly employed tool in assessing soil loss hazards owing to its minimal demand for data and simple model composition (Pan & Wen 2014). The objective of the research is to evaluate the soil loss response to changes in climate for the Sululta catchment employing the RUSLE equation coupled with a GIS tool. Assessing soil erosion under climatic conditions helps to quantify impacts and design suitable mitigation and adaptation actions at both regional and local levels.

The RUSLE model has some limitations. It might not provide reliable soil loss predictions when used in a larger watershed because it was designed at the field level. The model also considers only rill and sheet erosion types. The determination of the LS parameter using the GIS is best at 30 m DEM resolution. On the other hand, it is difficult to measure C and K factors in a large watershed. Because of the absence of observed soil erosion information, it is challenging to validate the RUSLE approach estimates.

Filling in missing observed rainfall data is critical before determining the value of the R factor and conducting the bias correction of future rainfall data. The missing observed rainfall data in this study were filled by XLSTAT. Using the inverse distance weight (IDW) interpolation technique in an ArcGIS environment, annual average rainfall was utilized to determine the R factor.

The K factor is derived using soil information. The K factor describes the resistance of soil to erosion caused by the raindrop effect, as well as the frequency and amount of runoff generated by the effect of rain in normal circumstances (Ghosal & Das Bhattacharya 2020). The soil's susceptibility to erosion is expressed by the K factor (Kayet et al. 2018). K values for different Ethiopian soil types depend on the color of the soil and are considered to represent good indicators of soil physical characteristics (Moges et al. 2020).

The C factor is directly affected by vegetation type, vegetation growth stage, and vegetation coverage. It refers to conditions where vegetation cover reduces soil loss. It describes how vegetation or plant coverage and management strategies affect soil loss rates (Renard 1997; Ozsoy et al. 2012). C values are closely related to land use (Ferreira et al. 2016). The C value varies from 0 to 1. The zero C value indicates no vulnerability to soil erosion, and 1 represents the vulnerability of land to soil erosion (Mohammed et al. 2020; Olorunfemi et al. 2020).

In this study, C values were obtained from the reclassified LULC. C factors related to each land use were chosen after reviewing several references or studies (Hurni 1985; Tiruneh & Ayalew 2015; Gashaw et al. 2018; Asmamaw & Mohammed 2019; Girmay et al. 2021).

The P parameter represents the rate of loss of soil for specific conservation practices available for the associated soil loss with the existence of up and downward gradients, contour cultivation, and tillage technique (Renard 1997; Ozsoy et al. 2012). The factor is mainly employed to examine the impact of agricultural techniques and water and soil protection policies on soil loss (Taye et al. 2018). Strip cropping, contour cultivation, and terracing are the most common soil erosion management strategies. P values vary from 0 to 1, with zero showing adequate protection and erosion management and 1 representing inadequate or poor erosion protection (Olorunfemi et al. 2020). The P factors were determined based on LULC, and slope class. To determine the value of the P factor, the study area was divided into agricultural and non-agricultural land (woodland, shrubland, and grassland). The cultivated land or agricultural land is categorized based on slope class. P factors for agricultural or cultivated land and non-agricultural or other land uses were assigned separately and combined using union in ArcGIS to create a P factor map. According to Wischmeier & Smith (1978), the P factor is allocated for agricultural land depending on slope classes, and a value of 1 is allocated for non-agricultural land for this study area.

CORDEX-Africa, launched by the World Climate Research Program (WCRP), offers the possibility of developing high-resolution (HR) regional climate predictions over Africa. It is used to predict the effects of climate change at the local and regional levels. The CORDEX-Africa regional climate model (RCM) provides HR historical and future climate predictions at the regional level by downscaling global climate models (GCMs) driven by representative concentration pathways (RCPs) based on the Coupled Intercomparison Project Phase 5 (CMIP5) (Nikulin et al. 2012). For future climate projections, RCP4.5 and RCP8.5 climate change scenarios with a resolution of 44° were employed from the CORDEX-Africa RCA4 RCM. Future precipitation data under RCP4.5 and RCP8.5 were downloaded from https://www.climate4impact.eu/c4ifrontend/.

In the Sululta catchment, four different types of soil were found: Chromic Luvisols, Eutric Cambisols, Eutric Leptosols, and Eutric Vertisols (Figure 3(c)). Their corresponding K values range from 0.15 to 0.25 (0.15, 0.2, 0.2, and 0.25 for Eutric Vertisols, Chromic Luvisols, Eutric Cambisols, and Eutric Leptosols, respectively) (Figure 3(d)). K values for different types of soil were obtained from various literatures (Hurni 1985; Bewket & Teferi 2009; Fenta et al. 2016; Gelagay & Minale 2016). The catchment's eastern part has a greater erodibility value, indicating lower resistance to soil erosion. The catchment's southern portion exhibits a moderate soil erodibility value.

The L and S parameters measure the effect of L and S on soil erosion. It raises the rate of runoff, resulting in particles of soil separation and, as a result, soil erosion. Soil erosion increases when the S and L sensitive to the erosion of soil increase (Wagari & Tamiru 2021). In this study, the LS factor generated from the DEM (Figure 3(e)) in the ArcGIS environment varies from 0 to 30.38, which is from low to moderate (Figure 3(f)).

The LULC in the Sululta catchment was divided into four main categories. Grassland, cultivated land, shrubland, and woodland are the dominant LULCs (Figure 3(g)). Grassland has the highest coverage (38.35%), followed by cultivated land (33.14%) and shrubland (19.65%), while woodland covers the least (8.86%). The catchment's central part and southern portion, with the lowest C factors, are covered by grassland and woodland, respectively. They contribute to less soil erosion by preventing the soil from the rainfall effects. Shrubland with a high C factor is prone to soil erosion because soil erosion increases with the C value. The catchment's C values range from 0.01 to 0.2 (0.01, 0.015, 0.15, and 0.2 for woodland, grassland, cultivated land, and shrubland, respectively) (Figure 3(h)).

In the Sululta catchment, the P factor varies from 0.1 to 1, which was obtained from slope class and LULC. Non-agricultural land received the same P factor regardless of slope class, whereas cultivated land was classified into different slopes with P values that vary between 0.1 and 0.33 (Figure 3(i) and 3(j)). The higher P factor 1, representing other land uses (woodland, shrubland, and grassland), covers a larger area. The P factor increases as the slope increases for cultivated land. For cultivated land, a lower P factor of 0.1–0.19 with a slope of 0–20% is observed in the western and some central regions. This indicates that the eastern, northern, southern, and some middle parts have a high contribution to soil erosion. However, the western and some central parts of the catchment contribute less to the soil loss. The lowest P factor indicates good conservation practices like terracing, contouring, or strip cropping, whereas the highest value implies little or poor conservation practice.

The soil loss map for the baseline period (1989–2018) was prepared from RUSLE factors by applying map algebra in ArcGIS. This study's C factor is similar to that of Hurni (1985), Tiruneh & Ayalew (2015), and Girmay et al. (2021) for similar LULC, and the K factor is similar to that of Hurni (1985), Fenta et al. (2016), and Gelagay & Minale (2016). The P value of the current study, which ranges from 0.1 to 1, agrees with that of Mengistu et al. (2015) at the Upper Blue Nile River Basin, Tufa & Feyissa (2019) at the upper Didessa watershed, and Gelagay & Minale (2016) at the Koga watershed.

The Sululta catchment projected rate of soil erosion is below the global soil erosion rate and the projected erosion rate in Ethiopia in the future climate change scenario like that of Moges et al. (2020) and Girmay et al. (2021), which could be related to the choice of RCM and the deriving model (GCM), which affect rainfall and rainfall erosivity.

As illustrated in Figure 8, a large soil loss rate was observed in the eastern part of the Sululta catchment. Climate change-induced soil erosion can have an impact on the catchment's ecosystem. Recognizing soil loss response therefore allows for determining how the frequency of erosion of soil varies with changing climates, and this serves in selecting critical regions for adopting strategies for soil management. Improved cover management (C) and conservation practices (P) factors may decrease increasing soil erosion. To determine the C factor, various remote sensing methods can be considered, enhancing the analysis of the soil loss process (Almagro et al. 2019). According to Tian et al. (2021), the P factor improved by considering different water and soil protection techniques with a slope gradient of 15–35°. Therefore, the projected erosion of soil may be reduced by recommending appropriate land use practices and adopting soil conservation practices. To validate the model estimation result, it is important to obtain observed soil loss data and conduct field-level experiments.

The RUSLE and GIS methods were used to evaluate the present and projected effects of climate change on soil loss in the Sululta catchment. RUSLE factors, baseline, and future rainfall were used to estimate the soil loss. The catchment's average annual soil loss in the current time period was estimated to be 5.03 tons/ha/year. The mean annual R factor for the catchment could decrease by 3.41 and 1.31% in 2021–2050 and 2051–2080, respectively, under RCP4.5 when compared to the baseline period. Under RCP8.5, the mean annual erosivity showed an increase of 7.18 and 2.48% in 2021–2050 and 2051–2080, respectively. Under RCP4.5, the average annual soil loss could decrease by 2.78% in 2021–2050 and by 0.80% in 2051–2080. In RCP8.5, the catchment's mean annual soil loss will increase by 7.75% during 2021–2050 and by 2.98% in the 2051–2080 future time period. Under both time periods, the study revealed that mean annual soil loss increases in future periods compared to the baseline period under RCP8.5. According to the study, the eastern part of the catchment has a high erosion rate, which needs intervention to manage the soil loss. Projection studies of climate change scenarios may therefore assist in determining when and where to prioritize soil erosion protection measures. Future researchers should consider the impacts of changes in LULC on soil erosion in addition to climate change in the Sululta catchment.

I would like to express my gratitude to the Ministry of Water, Irrigation, and Electricity, as well as the National Meteorological Service Agency, for their assistance in obtaining the essential data. I thank the WCRP's Working Group on Regional Climate and Coupled Modeling, as well as CORDEXE's previous coordinating body and responsible panel for CMIP5. I also recognize the Earth System Grid Federation infrastructure, which is a global effort led by the US Department of Energy's Program for Climate Model Diagnosis and Intercomparison the European Network for Earth System Modelling.

No funding was received.

K.T. designed the research, collected the necessary data, analyzed it, and drafted the paper.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.