Effects of Land Use Land Cover and Slope Gradients on Soil Fertility at Kori Sub-Watershed, East Wollega, Ethiopia


 BackgroundThe complex nature of the relationship and interaction between LULC and slope gradients resulted in the decline of soil fertility parameters, which aggravate the reduction of sustainable productivity in Ethiopia in general and the study area in particular. This study was aimed to determine the effects of land use land cover and slope gradients on the physicochemical properties of soil in study area A total of 27 composite soil samples were collected from 0-20cm depth under three dominant adjacent LULC across three slope with three replications. The collected soil samples were analyzed for selected soil physicochemical properties. Two-way ANOVA was used to test the mean differences of the soil fertility parameters. ResultThe mean values of soil physicochemical parameters showed that, SOC, TN, AvP, CEC, exchangeable bases (Ca2+ Mg2+, K+, and Na+), PBS, and percentage of clay contents of cultivated land and steep slope gradient (15-30%) were low and significantly different at (P≤0.05) than forest and grad grassland of the same slope gradient.. The gentle slope (3-8%) gradients of the forest lands had the lowest BD and high TP as compared to the others.Conclusion﻿The overall soil fertility status of the steep slope gradient (15-30%) of cultivated lands is lower than others and cultivating the steep slope is the cause for productivity loss in the study sub-watershed. Therefore, proper land-use planning and the use of integrated soil fertility management strategy give better production and keep the soil fertility status to a better level.


Methods
Soil sampling design and sample size determination Soil-sampling sites were selected by stratifying the study sub-watershed on the dominant LULC and difference in slope gradient. From each slope gradients and LULC, a plot with 25mx25m size was marked as a sample plot following a method applied by The soil samples were then taken by using a zigzag sampling technique from the surface 0-20cm depth.
Twenty-seven (27) composite soil samples, from 3 slope gradients,3 LULC, and 3 replications were collected to determine selected physicochemical properties of soil and analyzed in Nekemte Soil Research Center. In the same way, twenty-seven (27) undisturbed soil samples were taken from the surface 0-20cm depth through a 100cm3 volume steel core sampler. All sampling coordinate points were taken by using the Global Positioning System (GPS) to locate the study area (Fig. 2).
A preliminary eld survey was carried out to have a general view of LULC and slope gradients in the sub-watershed. Throughout the visual observation of the study area, its geographic coordination (latitudes and longitudes) and elevation were recorded by the global positioning system (GPS). Finally, the study area, Kori sub-watershed was strati ed into three major LULC and slope gradients.

Soil Laboratory Analysis
The disturbed soil samples were air dried, thoroughly mixed, homogenized into a single composite sample and pass through 0.5 and 2mm sieve for SOM and other physicochemical analysis respectively. Soil texture was estimated using Hydrometer method (Day, 1965) after destroying organic matter by adding hydrogen peroxide (H2O2) and separating the soil through adding sodium hexametaphosphate (NaPo3)6. The BD of undisturbed soil samples were determined by using metal sampler). The mass of solids and the water content of the core were determined, by weighing the core, drying it to constant weight in an oven at 105 o C for 24 hours and calculated as oven dried mass per core volume.
The TP was calculated using BD and particle density (Pd) as described in Hao et al. (2008) by the following formula.
The pH and EC of soil samples were measured from a soil suspension solution prepared with 1:2.5 soils water ratios. SOM, SOC and TN content were determined following the Walkey and Black titration (1934) procedure. AvP of soil samples was determined following the (Olsen et.al., 1945). The CEC and exchangeable bases (Ca2+, Mg2+, K+ and Na+) were determined after extracting the soil samples by ammonium acetate (1N NH4OAc) at pH 7.0.
Exchangeable Ca2+ and Mg2+ in the extracts were analyzed using titration method while Na+ and K+ were analyzed by ame photometer (Chapman, 1965). The CEC was estimated titrimetrically by distillation of ammonium that was displaced by sodium from NaCl solution (Chapman, 1965). The PBS was calculated by dividing the sum of the base forming cations (Ca2+, Mg2+, K+ and Na+) by the CEC of the soil and multiplied by 100. Similarly, ESP was calculated by dividing exchangeable Na+ to CEC of the soil and multiplying by 100.

Data Analysis
The physicochemical properties of the soil generated from soil laboratory test were subjected to two-way analysis of variance (ANOVA) at p≤0.05 & 0.01 to nd out the in uence of LULC and slope gradients variation on soil fertility parameters and their interaction effects using the general linear model (GLM) procedure using S PSS version 24 and XLSTAT 2015 software. The least signi cant difference (LSD) (p≤0.05 & 0.01) test was used to separate statistically signi cant means of soil parameters. The Pearson correlation coe cient analysis was applied to determine the relationship between some soil physicochemical variables.

Results And Discussion
Effect of LULC and slope gradient on soil physical properties Soil texture The analysis of variance (ANOVA) results showed that the particle size distribution of the soil in the study sub-watershed was signi cantly affected by LULC and slope gradient (P≤0.01). Accordingly, the highest percentage of sand content (62.78%) was observed under the cultivated land and the lowest (44.89%) in the forest land. The percentage of the silt fractions was 23.25, 27.67, and 31.33% for cultivated, grazing, and forest lands, respectively (Table 3) The combined effect result (Table 4) at all LULC also showed that the percentage of the clay increased as slope gradient decreases while the percentage of the sand decreased down the slope gradients.

Soil bulk density and total porosity
The result showed that soil bulk density (BD) was signi cantly affected by LULC and slope gradient (P ≤0.01).
As speci ed table 4, the mean value of BD was lowest (0.81gcm-3) under the forest and highest (1.30 gcm-3) in grazing land. The BD of soil is also affected by the texture of the soil. Sandy soils have higher BD when OM content and total pore space are less than clay or silt soils that can limit root penetration. On the other hand, the mean values of lowest (0.73gmcm-3) and highest (1.29gmcm-3) soil BD were recorded from the gentle slope (3-8%) and moderately steep (15-30%) slope gradient of the forest and grazing lands of the study subwatershed, respectively (Table 4).
The TP analyses result showed that the soil TP in all LULC and slope gradients were increased as the BD decreased and vice versa (Tables 3 and 4). The mean percentage values of TP were (51.29), (59.47 and (69.4) for grazing, cultivated, and forest lands, respectively (Table 3). The mean percentage values of the TP of forest land is higher than both cultivated and grazing land and statistically, forest land had a signi cant difference (P ≤0.01) than cultivated and grazing land. The mean percentage values from the slope and LULC interaction showed a decrease in TP from the gentle slope (3-8%) to moderately steep slope (15-30%) gradients for all LULC ( Table 4).

Effect of LULC and slope gradients on selected soil chemical Properties
Soil reaction (pH) and electrical conductivity (EC) The mean value of soil pH did not show a signi cant difference between all LULC and slope gradients. Relatively the lowest (5.34) and highest (5.57) mean value of soil pH was observed under cultivated and forest lands respectively (Table 5).
Likewise, the lowest pH value (5.27) was observed in soils of moderately steep (15-30%) slope gradient of the cultivated land, whereas the highest pH (5.69) was observed on gentle (3-8%) slope gradient of the forest land (Table 6).
Soil organic carbon, total nitrogen, and available phosphorous The analysis of variance results showed that soil organic carbon (SOC) contents were signi cantly affected by LULC and the interaction of LULC with slope gradients (P ≤ 0.05) ( Table 6). Contents of SOC in the study area ranged between 2.53% and 3.72%. Relatively, SOC content in forest land was signi cantly higher (3.72%) as compared to the SOC content in cultivated land (2.53%) (Fig. 3).
The interaction of cultivated land with all slope gradients produced a lower soil SOC than other LULC. The highest mean value of SOC was recorded from the interaction of forest land with 3-8% slope (4.22%) gradient and had a signi cant difference at (P ≤ 0.05) than other lands of the same slope gradient, that is, cultivated land (2.86%) and grazing land (3.17%). The lowest mean value of SOC (2.61%) was recorded from cultivated land at the slope gradient of 15-30% (Table 6).
The SOC in the moderately steep slope (15-30%) was signi cantly different (P ≤ 0.01) as compared to the undulating slope (8-15%) and gentle slope (3-8%) gradients of each LULC. The total nitrogen (TN) content of soils of the study sub-watershed was signi cantly affected by both LULC and slope gradients (P ≤ 0.01). The mean value of TN was highest (0.32%) in soils of the forest land while the lowest (0.22%) in the cultivated land of the study sub-watershed (Fig. 3).
The result of the multiple comparisons showed that there is a signi cant variation in TN among all LULC in all slope gradients of the study sub-watershed (P ≤ 0.01). With the interaction of LULC by slope gradients, the highest (0.36%) mean value of TN was observed at the gentle slope (3-8%) gradients of the forest land. On the other hand, the lowest (0.23%) interaction mean value of TN was observed in the moderately steep slope (15-30%) gradients of the cultivated land (Table 6).
According to the analysis of variance, results showed that the available phosphorus (AvP) of the study sub-watershed was signi cantly affected by LULC and the interaction of LULC with slope gradients (P ≤ 0.01) ( Table 6 and Fig. 3). The mean values of AvP in the study sub-watershed were 4.16, 4.58, and 7.96ppm for grazing, cultivated, and forest lands respectively (Fig. 3).
Concerning the interaction effect of LULC with slope gradients, the highest (8.70ppm) and the lowest (3.04pmm) of AvP contents were recorded under the gentle slope (3-8%) and moderately steep slope (15-30%) gradients of the forest and grazing lands, respectively ( Table 6).
The summary of the SOC, TN, and AvP contents as affected by LULC in the study sub-watershed was indicated below (Fig. 3).

Exchangeable basic cations, CEC and PBS
The analysis of variance results showed that the mean values of basic cations on the exchange site of soils of all LULC and slope gradients of the study subwatershed were in the order of Ca2+ > Mg2+ > K+ > Na+ (Table 5 and 7). The nding also showed that the lowest and highest values of exchangeable bases were observed for moderately steep (15-30%) and gentle (3-8%) slope gradients at all LULC, respectively (Table 7).
The cation exchangeable capacity (CEC) mean values of the soils in the study sub-watershed were signi cantly affected by LULC and the interaction of LULC by slope gradients (P ≤ 0.05). The result of the mean values of CEC is ranged from 22.28 to 27.29 Cmol (+) /kg) in the study subwatershed. Relatively, the mean CEC values are higher for forest (27.29 Cmol (+) /kg) and lower in cultivated (22.28Cmol (+) /kg) land (Table 5).
In another way, the lowest and highest CEC values were recorded for moderately steep slope (17.43cmol (+) kg-1) and gentle slope (31.61cmol (+) kg-1) gradients of cultivated land and forest lands of the study sub-watershed respectively (Fig. 4).
Similar to the exchangeable basic cations the percent base saturation (PBS) of the study area was signi cantly affected by LULC and slope gradients (P ≤ 0.05) (Tables 5 and 7).

Discussion
The cultivated land had signi cantly lowest (13.97%) clay content compared to the forest (23.78%) and grazing (17.10 %) land uses (Table 3), because most of the cultivated lands in the study sub-watershed lack soil erosion control practices. This nding is in agreement with that of Teshome et al. (2013) who reported that the reason for low clay content on cultivated lands might be due to selective removal of clay from the surface by erosion.
At all LULC the percentage of the sand decreased down the slope gradients. This might be due to the removal of the clay particles by erosion is greater on the moderately steep (15-30%) slope gradient of mainly under the cultivated land while deposition of these particles occurs higher on the gentle (3-8%) slope gradients of the forest land (Table 4). This nding is in line with Mohammed et al. (2005), who reported that ner soil materials are deposited at the lower slope gradients, where they are coming from the upper sites.
The higher BD values (Table 3) in grazing land and cultivated land as compared to forest land might be due to the result over grazing and continuous cultivation and low OM input. The ndings of the present study are in line with the ndings of Nega (2006) and Solomon et al. (2002).
The variation of soil BD between the slope gradients and LULC might be recognized to the variation of soil particle size distribution and disturbance of soil particles with erosion. This result was also in line with the works of Yihenew and Getachew (2013).
The highest soil means percentage values of TP under cultivated and forest land might be due to lower animal trampling. This means that soils having higher SOM had lower BD which indicates higher total porosity. The result of Pearson's correlation coe cient (r=0.62) indicated TP is positively and signi cantly correlated (P ≤0.01) with SOC (Table 8). The results were in agreement with that of Achalu et al. (2012).
The observed comparatively higher pH value in forest land soils could be related with higher SOM content. In agreement with this, Abreha et al. (2012) reported that the high pH of soils from the forest land might be due to the high accumulation of organic matter at the surface. In contrast, the lower pH in soils of the cultivated land might be due to the removal of basic cations by surface runoff and deep percolation in cultivated land because of less plant cover in cultivated land as compared to other land uses.
Person's correlation matrix also showed that CEC, Ca2+, K+, Mg2+, and pH have had a positive relationship with each other (Table 8). This, in line with Brady and Weil (2002) who stated as, in acid soils, Al3+ becomes soluble and increases soil acidity while in alkaline soils; exchangeable basic cations tend to occupy the exchange site of the soils by replacing exchangeable H1+ and Al3+.
Generally, the similar low pH in all LULC and slope gradients of the study sub-watershed might be related to the high rainfall in western Ethiopia that causing the leaching of basic Cations (K+, Mg2+, and Ca2+) and replaced by acidic cations like Al3+, Fe3+, Cu2+, Mn2+. In line with this nding, Nega and Heluf (2013) reported that loss of base-forming cations through leaching and runoff produced from accelerated erosion reduces soil pH and thus increases soil acidity. Based on Foth and Ellis (1997) pH, between 4.6 and 5.5 is classi ed as strongly acidic. Accordingly, the pH of the study sub-watershed falls under strongly acidic soil.
The consistently low electrical conductivity (EC) value under different LULC and slope gradients (Tables 5 and 6) might be due to the low exchangeable Na+ content observed in the study sub-watershed. The result from Pearson's correlation matrix revealed that EC is positively and signi cantly correlated with exchangeable Na+ (r=0.75**) (Table 8). According to FAO (2006) and Landon (1991) rating, EC values of the study subwatershed were considered to be very low (0-2ds/cm) for all LULC and slope gradients which show the study sub-watershed as non-saline soil.
The higher SOC observed in the forest land of the study sub-watershed might be due to the addition of more plant residues on the surface of forest soils and their reduced rate of disturbance as compared to the other land-use types. In contrast, the decline in SOC contents in the cultivated land might be due to intensive cultivation and soil erosion. This result is in agreement with the nding of Eylachew (2001) and Yihenew (2002) who reported that most cultivated soils of Ethiopia are poor in organic matter contents due to the low amount of organic materials applied to the soil, intensive cultivation and high erosion problems.
The SOC in the moderately steep slope (15-30%) was signi cantly different (P ≤ 0.01) as compared to the undulating slope (8-15%) and gentle slope (3-8%) gradients of each LULC, which might be due to the higher soil erosion problem and a comparatively high percentage of sand in the steeper slope while the ner soil materials are deposited at the lower slope gradients of the land use types. This result is similar to Bovine et al. (2000) that indicated SOC shows signi cant variation in LULC and slope gradients. According to Tekalign (1991) general guidelines on the interpretation of soil SOC test results, the study sub-watershed is considered as moderate SOC content.
The higher TN in forest land could be due to an addition of a comparatively higher plant residue and slight rate of decomposition and low soil erosion impact. However, the lower TN in cultivated land might be linked to the rapid mineralization of the organic substrates following intensive cultivation. In addition, as the area receives high rainfall, the nitrogen leaching problem can be another reason for the decline of TN in soils of cultivated land. The results of the present study agree with the ndings of Solomon et al. (2002); and Yihenew and Getachew (2013).
On the other hand, intensive and continuous cultivation and lower activity of N-xing bacteria due to strong acidity pH < 5.34 might have resulted in a decrease of TN in soils of cultivated land as compared to natural forest land. The results were following the ndings of Wakene and Heluf (2004) who stated that intensive and continuous cultivation forced oxidation of SOC and thus resulted in a decrease of TN.
The TN contents in the soil taken from the gentle slope (3-8%) gradients of each LULC were comparatively higher because it is not susceptible to soil erosion thereby leading to accumulation of nitrogen in soil. On the other hand, the moderately steep slope (15-30%) gradients are highly susceptible to runoff and erosion that caused the removal of nitrogen and SOC from the area ( Table 6). The result of Pearson correlation analysis has also shown the contents of TN in the study sub-watershed had a strong positive and signi cant correlation with SOC (r = 1.00**) as shown in (Table 8). As per the rating of TN indicated by Tekalign and Landon (1991), the study watershed quali es for the moderate status of TN.
Relatively the high content of AvP in the forest land might be due to the high content of SOM resulting in the release of organic phosphorus thus increases AvP under forest land. Similarly, this result is in agreement with the ndings of Abad et al. (2014) who reported that the AvP was high in forest land compared to the adjacent grazing and cultivated land at 0-30 cm soil depth.
The low AvP status in the cultivated and grazing land of the study sub-watershed might be linked with the low pH and high exchangeable acidity. In agreement with this, the Pearson correlation analysis has shown a positive correlation (r = 0.41*) of AvP with soil pH. Similarly, there exists a positive correlation between AvP and SOC (r = 0.61**) ( Table 8). The low contents of AvP in the soil agree with the results reported by Murphy (1968), andDawit et al. (2002) that the AvP under most soils of Ethiopia decline by the impacts of xation, abundant crop harvest, and erosion. According to the rating of AvP stated by Barber 1984, the study subwatershed quali es for the low status of AvP.
In general, the variation of AvP content between the slope gradients of each LULC might also be the variation of SOC. This shows that SOC could contribute to the presence of more AvP in the soil. In agreement with this, Fisseha et al. (2014) found low AvP in soils having a low content of organic matter.
Similarly, exchangeable cations were higher under forest land than cultivated and grazing land and had a signi cant difference at (p< 0.05) between them (Table 5). This is because of the presence of different woody species and perennial plants in forest land which can add SOM and reduce the rate of soil erosion. This result is in agreement with the work of Yitbarek et al. (2013). The variation of exchangeable basic cations from upper to lower slope gradients under each LULC might also be due to their loss through runoff and erosion in the high slope gradients and accumulation in areas having lower slope gradient. In general, as per the ratings of FAO (2006), all exchangeable cations (Mg2+, Ca2+, and K+) of the study sub-watershed were found to be moderate except exchangeable Na+ which was low.
The CEC values in the cultivated land decreased mostly due to the decrease in SOC content. In agreement with this, Pearson correlation analysis has shown a strong positive correlation (r = 0.64**) of CEC with SOC (Table 8). This result is in agreement with the ndings of Yitbarek et al. (2013) who suggested that the CEC of soil was higher in forest land compared to that of the adjacent grazing and cultivated lands.
The highest CEC value in the gentle slope (3-8%) of the forest land is linked to the availability of a high concentration of SOC and clay content. This is an indication that SOC and clay are major sites for exchangeable cations. This is in agreement with the nding of Teshome et al. (2016), Western Ethiopia.
The Pearson correlation result also revealed that CEC was positively and signi cantly correlated with clay (r = 0.61**) and SOC (r = 0.64**), while it was inversely and signi cantly correlated with sand (r =−0.83**) (Table 8). Based on the rating established by Landon (1991), the CEC of the study area is considered as a medium under both LULC and slope gradients.
The variances in PBS and exchangeable sodium percentage (ESP) in both LULC and slope gradients are mainly associated with differences in CEC. This indicates that high exchangeable bases and high CEC may give low PBS and ESP, while the higher exchangeable bases and lower CEC may give higher PBS and ESP. The result of Pearson correlation also showed that the PBS of the studied soils increased with increase in clay content (r = 0.56**), CEC (r = 0.61**) and basic cations mainly Ca (r = 0.87**) and Mg (r =0.89**) ( Table 8).
The result revealed that the effects of LULC types with slope gradients in uenced signi cantly the soil fertility status in the study area. The ndings indicated that, the steepest slope gradients of cultivated land had the lower as compared to the same slope gradient of forest and grazing land, this is due to a continuous cultivation of the steep land for long. Therefore, it is recommended to avoid the cultivation of steep slope for crop growing. Implementing appropriate soil and water conservation techniques should be undertaken by local communities particularly on the moderate to steep slope gradients of the cultivated lands. This study result should be complemented by further research with plant tissue analysis that grow in the study sub-watershed, and the nutrient ratings should also be done by considering the local situations of the study area. The data used in this paper is with the authors and can be available upon demand.

Ethics approval and consent to participate
Not applicable

Competing interests
The authors declare that they have no competing interests.

Funding This Research was funded by Bureau of Oromia Public Service and Human resources
Availability of data and material The data used in this paper is with the authors and can be available upon demand.