Introduction

Land degradation is a leading factor of low crop productivity in Pakistan (Shah and Arshad 2006). Soil erosion is considered as one of the most damaging type of land degradation across the globe (Ashraf et al. 2000). Soil erosion is the loss or removal of surface soil by the moving water, wind or ice and adversely disturbs agricultural land by affecting soil fertility, landscape beauty, water ecosystem, environmental management and crop production (Ashraf et al. 2002; Jie et al. 2002; Ananda and Herath 2003; Lal 2003). Rainfed tracts are severely suffering from erosion due to inappropriate land uses, uncontrolled grazing and illegal cutting of trees and associated vegetation (Khan et al. 2012). Severe soil erosion in rainfed areas and the possible factors responsible for this menace include inappropriate land uses for growing unsuitable crops, livestock grazing without considering the carrying capacity of rangelands and illegitimate deforestation (Irshad et al. 2007). Approximately one billion tons of fertile soils are eroded annually and are deposited in dams and carried into the Arabian Sea (Anjum et al. 2010).

High rainfall areas with steep slopes are mostly prone to water erosion, particularly in the Potohar region (Shah and Arshad 2006). The main areas affected by water erosion are Potohar, Bannu, Peshawar valley, Piedmont plains and the Himalayan Mountains. A large area of the region has been affected by gullies, is steeply dissected, and pinnacle erosion, piping and slumping are also quite prominent (Baig et al. 1999; Farooq et al. 2007). Potohar region is a geographically important rainfed region of Pakistan with depleted underground water resources, undulating terrain and higher drainage concentration that makes it high soil erosion-prone area. If the soil erosion is not checked, it may lead to the soil degradation and thereby to desertification, rendering the cultivable land unfit for cultivation. Potohar region lies at the top of the Indus Basin, the world’s largest irrigation networks of canals and barrages. The soil erosion in the region can produce an enormous amount of sediments load in the downstream and barrages, which will adversely affect the already loaded barrages. This will also add to the miseries caused by floods.

Despite the severity of the problem and associated damages, the spatio-temporal information about soil erosion measurement and evaluation in Pakistan is scarce. Few studies reported the erosion problem but these were carried out at a watershed scale. Similarly, the spatio-temporal information about soil erosion is also missing (Ahmad et al. 1990; Nasir et al. 2006; Renard et al. 1997). Larger scale studies of soil erosion assessment are vital for improving land conservation practices and the sustainable land management in rainfed areas of Pakistan. Such kind of information is also indispensable for sustainable land use planning and policy making. Inclusion of natural hazard maps, geomorphological and geographical parameters in suitability mapping for urban growth and light industry development is a prerequisite for sustainable land use evaluation (Bathrellos et al. 2012, 2017; Andriyanto et al. 2015).

Geospatial technology especially Satellite Remote Sensing and GIS is considered as the most useful tools for monitoring of natural resources and their sustainable management. Digital elevation model (DEM) combined with remote sensing data and GIS can be highly beneficial in rapid assessment of land hazards like erosion. Soil loss is commonly predicted using empirical models. Revised Universal Soil Loss Equation (RUSLE) is a commonly used model (Wischmeier and Smith 1978). RUSLE got wide acceptance and use because of being simple and easily operational, and needs less data and time (Jones et al. 1996). The integration of Geographical Information System and Remote Sensing (GIS & RS) with RUSLE model makes soil erosion assessment more accurate with precise spatial distribution (Millward and Mersey 1999; Wang et al. 2003; Naipal et al. 2015). Other models such as the Limburg Soil Erosion (LISEM) Model (Jetten 2002), the WEPP (Water Erosion Prediction) model (Flanagan and Nearing 1995) and the Soil and Water Assessment Tool (SWAT) (Gassman et al. 2010) have limitations regarding large data requirements, complex laboratory operations and complex field data collection information (Beskow et al. 2009). RUSLE model presents an easier implementation; the required data are easily available for most of the countries. The information produced through such studies is useful to start taking the maintenance and support procedures with spatial implementations in targeting the areas requiring priority action to combat erosion and to protect the natural environment (Abdo and Salloum 2017; Djoukbala et al. 2018).

This study aims to make soil erosion profile of the rainfed Potohar region of Pakistan in a spatially explicit way. The work involves estimation of soil loss and sedimentation yield using the modern geospatial approaches along with globally known RUSLE model.

Materials and methods

Study area

Potohar region is a geographically important rainfed region of Pakistan. It is a plateau that lies between River Indus and River Jhelum, and is comprised of four districts including Jhelum, Attock, Chakwal and Rawalpindi (Fig. 1). The total geographical area of the region is around 2.2 million hectares. The terrain of Potohar is bumpy and the slope range is 8–40% that is inclined between 32.5°N and 34.0°N latitude, and between 72°E and 74.0°E longitude. The annual rainfall in the area varies from 450 to 1750 mm and almost 70% precipitation occurs during monsoon season (Cheema and Bastiaanssen 2012). There are many seasonal streams in the region including River Haro and River Soan as the main streams (Ghani et al. 2013). The region is much susceptible to erosion because of its deep underground water, undulating terrain and complex drainage system. The soil erosion can cause silting in the linked barrages/canals, which cannot only reduce their carrying capacity for water management but also for flood hazard in the country.

Fig. 1
figure 1

Study area map of Potohar region

Full size image

Revised universal soil loss equation (RUSLE) model

The model was originally developed for soil erosion estimation as USLE (Wischmeier and Smith 1978). It has been used in different scenarios such as forests, rangelands and disturbed areas in modified form as RUSLE (Renard et al. 1997). The modified form (RUSLE) included soil type, climatic variables, land use effect, topography, raindrop impact and surface runoff (Renard et al. 1997). It intended to calculate long-term average annual soil loss. The equation of RUSLE model is given as follows:

$$A=R~ \times K~ \times {\text{LS}} \times C \times P$$
(1)

where A is the average soil loss in tons/ha/year, R is rainfall–runoff erosivity factor, K is the soil erodibility factor, LS is the slope length and steepness factor, C is the land cover-management factor and P is the Erosion Control Practice factor.

Factors of RUSLE model

  1. (a)

    Topographic factor (LS)

The ASTER Global Digital Elevation Model (ASTER GDEM V2: 30 m Spatial Resolution) was used to calculate combined topographic factor (LS) required for the model. The steps involved in the model for creating the slope angle, slope length and combined LS factor are given below:

  • Identification and filling of sinks in the DEM.

  • Determination of Flow Direction using filled DEM as input.

  • Derivation of Flow Accumulation using the Flow Direction as input grid.

  • Stream Network Extraction using the Flow Accumulation as input grid.

  • Stream Network ordering using the Extracted Drainage/Stream Network.

  • Derivation of watershed Flow Length using the stream network-ordering map.

  • Derivation of the slope from DEM

  • Derivation of LS factor map from the flow length and the slope maps.

The following Eq. (2) was used to determine LS factor as recommended by Morgan and Davidson (1991).

$${\text{LS}}=~\sqrt {\frac{L}{{22}}\left( {0.065+{{0.0450.045}^*}s+{{0.0065}^*}~{S^2}} \right)}$$
(2)

where L = slope length in meter and S = percent slope (%)

  1. (b)

    Cover factor (C)

C factor was calculated from the land use/land cover map of the region. The land cover assessment is necessary to monitor the influence of various types of land cover on soil erosion. To produce C factor map, the satellite-based land cover/use map was developed by maximum likelihood classification of Landsat 8.0 data. Ground truthing surveys were carried out in the region to validate the classification results and get accuracy. In current study, the C factor values were allotted to different land cover/use types based on literature and from already available experimental value as given in previous studies (USDA 1978; Rao 1981).

  1. (c)

    Soil erodibility factor (K)

The K factor (soil erodibility) explains the susceptibility of diverse types of soils to erosion and their contribution to rate of runoff. Soils of the study area being poor in organic matter, for the current assessment the K factor was estimated keeping in view the soil texture type, the major factor which affects soil erodibility as assumed by Stewart et al. (1975). Soil aggregation, being dependent on many soil properties such as organic matter and texture, is one of the major factor affecting soil erodibility (Stanchi et al. 2015). The soil type map of Potohar region was extracted from the Soil Survey of Pakistan maps. K factor values were allocated to each soil type by editing the soil map attribute table. Using ArcGIS, the shape file was transformed to raster of 30 m cell size. The K factor values explain the influence of soil properties on soil erosion (Renard et al. 1997).

  1. (d)

    Rainfall erosivity factor (R)

The Rainfall–Runoff erosivity (R) explains the extents and intensities of individual rainfall events across a year (Hudson 1981; Weng and Mokhtar 2004). The yearly precipitation data for the meteorological stations situated in the study area were obtained from Pakistan Meteorological Department. Spatial interpolation of data was conducted using Ordinary Kriging method (Isaaks and Srivatava 1989) through the ArcGIS software. The need for very specific data mostly makes the estimation of R difficult to a certain degree. This has led to derivation of simplified methods. In this study, the model proposed by Morgan and Davidson (1991) was used to calculate factor R from rainfall interpolated map using following equation:

$$R=P \times 0.05~~~~$$
(3)

where P is average yearly rainfall (mm).

  1. (e)

    Erosion control practice factor (P)

The erosion control practice factor (P) is generally calculated according to the practices that minimize the effects of different influential factors on erosion. As no conservation practices are in place in the region, therefore, a P-factor value of ‘1’ was assigned to the whole region. In previous studies conducted for erosion assessment at watershed level in the region, P factor value of ‘1’ was also assigned (Ahmad et al. 1990; Renard et al. 1997; Nasir et al. 2006).

  1. (f)

    Sediment yield (Y)

Sediment yield explains the total sediments that are carried to the outlet of the watershed (Julien 2010). It is the quantitative sediment outflow out of a basin across a specific time duration and is measured in tons per year. The sediment delivery ratio (SDR) was calculated through distributing sediment yield “Y” by the drainage area “A” of the watershed. Therefore,

$$SDR={\raise0.7ex\hbox{$Y$} \!\mathord{\left/ {\vphantom {Y A}}\right.\kern-0pt}\!\lower0.7ex\hbox{$A$}}$$
(4)

or

$$Y={\text{SDR}}~ \times A$$
(5)

where “Y” is sediment yield, “SDR” is sediment delivery ratio and “A” is the drainage area.

Results

C-factor

The C-factor values assigned to different land cover classes for the study area are given in Table 1 while Fig. 2 presents the map of C-factor values and its spatial distribution. In the current study, the values range from 0.004 in forest area and 1.0 in barren land. These values are in accordance with the generally used value of C-factor for bare soils, i.e. 1.0, for root/tuber crops values of 1.0–0.9 are used, while for grasslands and cover plants values of 0.01 and for forests c-factor values of 0.001 are used (Roose 1996).

Table 1 C factor values in the Potohar Region
Full size table
Fig. 2
figure 2

C-factor map of Potohar region

Full size image

K-factor

Table 2 and Fig. 3 present the resultant K values of the studied area. These values range from 0 for the bare rock land to 0.48 for the silt loam area. Soils rich in clay possess low K values, about 0.1–0.25, because these are generally well structured and resist dispersion. Low K values of 0.05–0.2 are assigned to coarse textured soils, like sandy soils, because these soils are less prone to runoff despite their ability of easy detachment. Moderate K values of about 0.25–0.4 are given to the medium textured soils like silt loam soils, because of their susceptibility for detachment and role in production of moderate runoff. Soils with higher silt contents are the soils most susceptible to erosion. They could be easily dispersed, lead to surface crusts, decreasing water infiltration and producing high runoff rates. K-factor values higher than 0.4 are taken for these soils. Soil calcareousness is another important factor to be considered for K factor estimation. Calcareousness influences soil structure and disturbs water-related soil characteristics. Soil aggregation is higher in calcareous soils and water diffusivity is greater that makes soils less erodible in comparison with non-calcareous soils of the similar texture. Urban soils are considered very poorly erodible because of the built-up cover in urban areas. Salt-affected soils are highly prone to erosion because of their poor structure and less aggregation while patchy, broken and gullied soils are included among soils with high K factor.

Table 2 K factor values in the Potohar Region
Full size table
Fig. 3
figure 3

K-factor map of Potohar region

Full size image

R-factor

The R factor (rainfall erosivity) is given in Fig. 4. The map shows that precipitation intensity is higher in the area with high altitude and its adjoining areas that can greatly enhance the chances of soil erosion.

Fig. 4
figure 4

R factor map of Potohar region

Full size image

LS factor

The LS factor map (Fig. 5) depicts the situation of both the important factors of length and slope in soil erosion. In northeast, the hilly areas including Murree hills and other upland areas in the region have high LS value to contribute in final soil loss. The maps show the complexity of the topography, which make it vulnerable for soil erosion.

Fig. 5
figure 5

LS map of Potohar region

Full size image

Soil loss

Figure 6 provides the status of annual average soil loss. The values of annual soil loss in the region range from 0 to 208 tons ha−1 year−1 with mean of 19 tons ha−1 year−1 for the whole region. The total mean yearly soil loss rate of the whole region is approximately 47 million tons.

Fig. 6
figure 6

Soil loss map of Potohar region

Full size image

Figure 6 further implies that most of the barren lands and sloppy highlands are prone to severe erosion. Dense natural forest cover once covered northeastern part of the region but it is going under extensive deforestation that make it exposed to severe erosion as pointed out in the map (Jamal et al. 2018). Similarly, all the regions showing varying amount of soil loss that can contribute enormously to soil deposition in the precious Indus water distribution system. Barren lands by producing about 38% were the second largest contributor of the overall soil loss rate in the basin.

Using the ArcGIS classification function, the soil erosion estimation map was further classified and a soil erosion severity map was generated. The map presented in Fig. 7 shows the soil erosion severity in the Potohar region classified into four classes of erosion severity, i.e. low, moderate, high and very high erosion. This classification of these classes is highly localized and is based on the expert opinions of local experts involved in soil erosion assessment and land degradation work. It was found that 69.25% of the area lies in the low (tolerable), 15.35% in moderate and 13.34% in high-erosion zones, while the severe erosion zone occupies 2.06% of the total area. Table 3 presents the erosion severity classes, the range of annual soil loss in each class and the area falling under each class of erosion severity.

Fig. 7
figure 7

Soil Erosion severity classes of Potohar region

Full size image
Table 3 Soil erosion severity classes and area under these classes in Potohar Region
Full size table

Table 4 presents the annual soil loss in different districts of Potohar Region. It was observed that most important annual soil loss occurs in Rawalpindi district. As evident from soil loss map (Fig. 6) the northern part of the district is majorly contributing to this soil loss. It is because of the hilly areas of Murree that is part of this district. This area also receives much higher rainfall than the remaining areas, thus more prone to erosion. Attock district having most of the plain areas is least in annual soil loss.

Table 4 District-wise annual soil loss in the Potohar Region
Full size table

Sediment yield

Sediment yield is the eroded soil material that is carried out of the basin along runoff. Figure 8 shows the map that depicts that soil sediment yield is highest in area having steep slope, sandy soil and scrubland. These areas have a sediment yield as high as 148 tons ha−1.

Fig. 8
figure 8

Sediment yield map of Potohar region

Full size image

Discussion

Soil erosion is a serious issue, particularly in Potohar region where various contributing factors such as climate and terrain conditions speed up the runoff and sedimentation. The area is located at the top of the world’s largest irrigation system. Ignoring the phenomena can lead to reduced water distribution and ultimately affect the food security of the area that is highly dependent on irrigation water. The yearly mean soil loss rate of the region was around 19 tons ha−1 year−1, and the total yearly soil loss from the region was approximately 47 million tons/year. Barren lands by producing about 38% were the second largest contributor of the overall soil loss rate in the basin. The results of soil erosion severity classification showed that 69.25% of the area in Potohar region lies in the low (tolerable limit i.e. 0–2 tons ha−1 year−1). This value corresponds to an annual soil loss of about 0.2 mm, which can also be considered as the annual soil formation rate. Therefore, this value is recommended by the OECD (2001) as the soil loss tolerance level (Oweis and Ashraf 2012). It was found that 15.35% area of the region falls under moderate and 13.34% under high-erosion zones, while the severe erosion zone occupies 2.06% of the total area. The areas under high erosion include most of the sloppy areas with less vegetation. Deforestation, urbanization, and other land-use change activities can significantly contribute to erosion and make it severe in sloppy areas (Oweis and Ashraf 2012; Farhan and Nawaiseh 2015).

RUSLE has been used and adopted by many research organizations for erosion prediction because of its high reliability in erosion prediction studies. Due to practical and financial difficulties, the validation analysis was not performed. The reason behind is unavailability of the required data and difficulties involved in estimation of erosion soil loads at a regional scale (Anton et al. 2003). However, the model results were also considered credible and accepted by research community. The upshots of the current study are also supported by the outcomes of the previous studies conducted for this area at watershed level. During a study conducted in the Fateh Jang watershed, it was revealed that for the area having slopes of about 1–10%, soil loss rate was 17–41 tons ha−1 year−1 when the land was fallow and at a rate of 9–26 tons ha−1 year−1 for the land under vegetative (Ahmad et al. 1990). In another soil erosion assessment study conducted by Nasir et al. (2006) for a small hilly watershed through using RUSLE (Revised Universal Soil Loss Equation), the soil loss rate was 0.1–8 tons ha−1 year−1 for plain area soils. The average rate of soil loss from the entire 13 ha watershed was 19.1 tons ha−1 year−1. Steep slopes generated 74% of the total soil loss. Similarly, Bashir et al. (2013) indicated a potential soil loss of 28 tons ha−1 year−1 for Rawal watershed situated within our study area.

The study results are valuable to understand the soil erosion risk in the region and the dominant contributing factors including land cover, land use, soil type and topography of the region. The information produced is highly beneficial in the natural resource management and policy making/planning to reduce land degradation and transportation of sediments in water reservoirs. In the international context, such studies should be conducted in the areas with truncated topography and high rainfall to assess the potential risk. This will help identifying priority areas for implementing strategies to combat erosion and other land degradation issues (Bathrellos et al. 2017). For the Potohar region, this assessment is very useful especially in the scenario of increased deforestation and conversion of land use patterns. The quantification of erosional losses in the Potohar region would help to make suitable water harvesting structures that would provide water in the lean season as well as prevent the soil erosion in this region. Conservation practices such as tree plantation and strip cropping should be promoted to control the soil erosion in the region.

Conclusion

The study indicated that Potohar region is highly susceptible to soil erosion and if the soil erosion at this rate continues, it may lead to the soil degradation and thereby to desertification rendering the cultivable land unfit for cultivation. The average annual soil loss was predicted up to 19 tons ha−1 year−1 of which the maximum erosion (70–208 tons ha−1 year−1) was near the river channels and hilly areas. The sediment yield due to the erosion is as high as 148 tons ha−1 year−1 with an average of 4.3 tons ha− 1 year− 1. Chakwal and Jhelum districts of the region are found to be more affected by erosion owing to their topography and soil properties. The enormous amount of sediments produced in the streams and barrages could adversely affect the already loaded canals and barrages. The information generated in this study regarding the erosion intensity, sediment yield and erosion severity classes will help the policy makers, planners, and soil conservation-related institutes to focus on the highly susceptible zones. It is, hence, recommended that suitable water harvesting structures and other conservation practices be made to prevent the soil erosion in this region.