Introduction

Sustainable land management is an important key to the increasing land productivity, better livelihoods and improved ecosystem health (Liniger et al. 2011). Factors that lead to land degradation are population pressure, overgrazing, deforestation, crop cultivation expansion on steep slope and severe soil loss in Ethiopia (Bishaw 2001; Taddese 2001; Tamene and Vlek 2008; Hurni et al. 2010; Gashaw et al. 2014). Land management is about exploring existing and possible land use/land cover (LULC) and making decision on choosing to implement the one that ensure sustainable production (UNEP 2014). Such decisions could directly affect ecosystem functions and services and alter condition of ecosystem resources such as soil, water, flora and fauna (Lemenih 2004; Maitima et al. 2009). Decision makers, therefore, need to carefully weigh the trade-off between increasing productivity on the one hand and loss of other ecosystem functions and services on the other.

In Ethiopia, land degradation had begun with the emergence of cultivation thousand years ago (Hurni 1990; Yirdaw 1996). Ethiopia is recognized for its land resource degradation, food insecurity, which has deforestation and forest degradation as its root causes (Bishaw 2001; Hurni 1990; Tamene and Vlek 2008). Various studies have shown changes in LULC, with most of the changes being the expansion of cultivated land, the increment of bare land, decline in forest areas and reduction in grazing land (Dwivedi et al. 2005; Haile et al. 2010; Kidane et al. 2012). The expansion of Eucalyptus woodlots and plantation are also observed in the highlands of Ethiopia (Jenbere et al. 2012; Chanie et al. 2013; Jaleta et al. 2016a). These dynamics in LULC have direct and indirect impact on soil and water resources of the country.

Studies have been done to assess the impact of different LULC changes on runoff and sediment yield in different parts of Ethiopia since late 1970s (Bayabil et al. 2010; Taye et al. 2013; Tebebu et al. 2015). These studies focused on effects of cultivated land, conservation areas or grassland. Results of the studies recorded the highest runoff from cultivated land (Hurni et al. 2005; Descheemaeker et al. 2006; Girmay et al. 2009; Adimassu et al. 2014).

The recent uncontrolled expansion of Eucalyptus could have significant effects on various ecosystem processes (Kebebew and Ayele 2010; Jenbere et al. 2012; Chanie et al. 2013; Tadele and Teketay 2014; Jaleta et al. 2016b). Eucalyptus expansion has been a contentious matter due to its argued ecological effects (Dessie and Erkossa 2011; Tadele and Teketay 2014; Yitaferu et al. 2013; Jaleta et al. 2016a). Various studies have assessed its effects on soil (Jenbere et al. 2012; Chanie et al. 2013; Yitaferu et al. 2013), water efficiency, allelophatic effect (Nigatu and Michelsen 1993; Fikreyesus et al. 2011) and socio-economy (Mekonnen et al. 2007; Adimassu et al. 2010; Kebebew and Ayele 2010). However, there are few studies that assessed its effect on surface runoff. Moreover, runoff-rainfall effects are site specific, due to various local effects such as climate and biophysical characteristics (Critchley et al. 1991; Girmay et al. 2009). Therefore, it is important to understand how Eucalyptus alters surface runoff patterns compared to other land use system so as to make decision on ecosystem management. The objective of this study was to evaluate surface runoff from three LULC in Meja River watershed, Oromia Regional State, central Ethiopia.

Study area

The study was carried out at Meja River watershed, Jeldu District in west Shewa, central Ethiopia. The watershed is located 114 km west of the capital, Addis Ababa, Ethiopia (Fig. 1). The watershed is experiencing rapid expansion of Eucalyptus. The altitude of the watershed ranges from 2400 to 3200 m above sea level. The two subcatchments in Meja River watershed called Sochoa and Tiki were selected to carry out the study. The mean annual temperature ranges from 17 to 22 °C. The rainfall is bi-modal with recent fluctuations with the short rainy season from February to May and long rainy season from June to September. The mean annual rainfall is 1400 mm. The agro-ecology of the area belongs to cool highland with sufficient rainfall.

Fig. 1
figure 1

Map of the study sites in Meja River watershed, Ethiopia

Full size image

Land use is dominated by a mixed crop-livestock system. Main crops grown in the watershed are barley (Hordium vulgare), wheat (Triticum vulgare) and potato (Solanum tuberosum). The major sources of cash for the community of the area are potato and Eucalyptus products. Average family size is six people, and land holding ranges from 0 to 4 ha. Eucalyptus globulus woodlots are abundant in the study area and established mostly by replacing cultivated land and grazing lands. The soil of the area is Pellic Vertisol.

Methods

Experimental design

The rainfall amount at each study catchment was collected using the rain gauge installed to record daily rainfall amount. The rainfall depth was measured every morning at 6:00 am and in the evening at 6:00 pm. Three LULC types, namely, Eucalyptus woodlot, cultivated land and grassland in each study catchment were selected for comparison as treatments. Four replications of each LULC type were used forming a total of 12 runoff plots. Each runoff plot consisted of an area of 10 m × 4 m with triangular funnelling plot of 4 m × 2.5 m × 2.5 m and cylindrical collecting metal trench of size 0.6 m (depth) × 0.6 m (diameter) (Fig. 2). Thus, the runoff plot has an area of 43.3 m2 (i.e. the entire area of harvesting runoff). External runoff flow into and out flow from the plots were protected by a plastic structure constructed around the plots.

Fig. 2
figure 2

Experimental setup of the runoff plots

Full size image

Runoff depth was measured using volume-depth relationships using the water depth of the trench. Measurement was done every morning at 6:00 am and evening at 6:00 pm. Total harvested runoff was collected from the collecting trenches and measured using a graduated cylinder for further check-up of the volume at every morning and evening. The harvested runoff was removed from the collecting trenches. Runoff coefficient was calculated as the ratio of total runoff depth harvested in each plot by total rainfall depth. From each runoff plots, the slope gradient, soil moisture, soil temperature, electric conductivity and ground cover by above-ground biomass were recorded. The data is summarized in Table 1. Slope gradient was measured by clinometer. Soil moisture, soil temperature and electric conductivity were measured at three places along the slope by time domain reflectrometry (TDR) device at the end of the study period, and the average was taken. Stone cover was determined by taking a quadrant of 50 × 50 cm made of wood stick placed at three locations in each plot. Then, the length of the stone surface that comes in contact with ruler which was laid down on five locations was registered, and average was calculated into percentage of the total coverage. The ground cover with stubble, weeds and organic residues was measured in a 1 m × 1 m quadrant laid in three locations. The counting was done following a similar procedure stated above for stone cover. The tree crown cover measurement was done taking the average value of north-south and east-west measurement with tape measure and calculated into percentage (Girmay et al. 2009).

Table 1 Mean of biophysical conditions of the runoff plots
Full size table

Data analysis

Runoff coefficient was calculated using the following formula:

$$ \mathrm{R}\mathrm{C} = \kern0.75em \frac{{\displaystyle \sum }\ \mathrm{R}\mathrm{o}\mathrm{F}}{{\displaystyle \sum }\ \mathrm{R}\mathrm{F}} $$
(1)

where RC is the runoff coefficient, ∑ RoF is the total runoff depth harvested in each plot and ∑ RF is the total rainfall depth over the entire rainy season.

The significance of variance of biophysical conditions of the runoff plots, runoff coefficient and runoff volume due to the effect of land use was evaluated using analysis of the significance of variation. Genstat1 15th edition was used to analyse significance of variation. Least significant difference (LSD) test was used to compare mean value at p < 0.05. The correlation analysis was done to observe the relationship of rainfall and runoff volume for each LULC.

Results

The total number of days with rainfall was 75 out of 91 total study days (July–September), which are typical rainy season for the study area. There were two extreme rainfall events registered in this period: one on July 10, 2015 and the other on July 13, 2015. The recorded rainfall amounts were 49 and 48 mm (average of the two rain gauge values), respectively (Fig. 3). The experimental year was a year of overall low rainfall registered in the area, and across the country, it was the most severe drought year registered in 50 years.

Fig. 3
figure 3

Daily rainfall for the study period from July to September 2015 at study area

Full size image

There is significant difference at p < 0.05 in soil moisture content among LULC types (Table 1). The moisture content in cultivated land was significantly less than the grassland and Eucalyptus woodlot. The soil electric conductivity of the grassland was significantly different at p < 0.05 with the cultivated land and Eucalyptus woodlot. The soil electric conductivity of the grassland was higher than Eucalyptus woodlot and cultivated land. There is significant difference at p < 0.05 in organic residues coverage among LULC types where Eucalyptus woodlot has significantly higher organic residues over the cultivated land and grassland.

The result indicated that LULC have significant influence on runoff volume and runoff coefficient (percentage) (Table 2). The highest significant mean runoff volume was found on cultivated land (191.9 mm). The lowest runoff volume was recorded from the grassland, but it was not statistically significantly different (p < 0.05) from the runoff volume recorded in Eucalyptus woodlots. The runoff coefficient was in the order of cultivated land > Eucalyptus stand > grassland (Table 2). The runoff coefficient recorded under cultivated land was significantly different with grassland and Eucalyptus (p < 0.05).

Table 2 Mean of rainfall, runoff and runoff coefficient from three LULC
Full size table

The runoff coefficient for the Eucalyptus woodlots and grassland was not significantly different at p < 0.05. The mean runoff generated from each LULC in the study period was 191.9, 147.8 and 154.0 mm (LSD 5.57) for cultivated land, grassland and Eucalyptus stand, respectively. The relative effect of Eucalyptus on runoff was calculated against that of cultivated land by considering cultivated land as 100%. The result indicated that Eucalyptus reduced the runoff generated from cultivated land by 21%. The rainfall and runoff volume have high correlation coefficient in the cultivated land (0.8022), grassland (0.8018) and Eucalyptus (0.8349) as shown in Figs. 4, 5 and 6.

Fig. 4
figure 4

Rainfall and runoff volume correlation graph in Eucalyptus

Full size image
Fig. 5
figure 5

Rainfall and runoff volume correlation graph in grassland

Full size image
Fig. 6
figure 6

Rainfall and runoff volume correlation graph cultivated land

Full size image

Discussion

Continuous ploughing of land may result in poor soil aggregate and soil crusting that reduces infiltration of rainfall. This is why cultivated land generated high runoff (Girmay et al. 2009). In addition, this study has found less soil moisture content in the cultivated land (Table 1), which directly strengthens the findings of high surface runoff from cultivated land, that means there was less rainfall infiltrate to the soil. Studies, such as Descheemaeker et al. (2006) and Girmay et al. (2009) also reported a similar finding; higher runoff from cultivated land than other land uses. However, Defersha and Melesse (2012) reported lower runoff generated from field with grown-up maize than in grassland and bare land in Kenya, which might be due to the effect of the maize crop. Conversely, Hurni et al. (2005) reported higher runoff on cultivated land than grassland and forest land in the northern highlands of Ethiopia. According to Adimassu et al. (2014), cultivated land with soil bunds generated less runoff than fallow and non-conserved cultivated land in central Ethiopia. The above listed findings could be related to the biophysical conditions of the plots together with the LULC of the plots as this study found. Applying soil and water conservation measures, therefore, reduces runoff generated especially on steep slope (Nyssen et al. 2010; Adimassu et al. 2014; Dagnew et al. 2015).

Our study found that grassland has generated least surface runoff as compared to other land uses. It is due to the dense ground coverage with grass that intercepts raindrops and reduces surface runoff to give it a time for infiltration. The moisture content in grassland was higher than other LULC, which could be one of the reasons for least surface runoff generation from the grassland. Similarly, Hurni et al. (2005) found less runoff coefficient in grassland than degraded area and cultivated land, which was similar also to the study of Girmay et al. (2009). Another study by Bayabil et al. (2010) also found a lower runoff from grassland than cultivated land with maize in Maybar watershed.

On the other hand, Eucalyptus stand generated less surface runoff compared to cultivated land. This is also due to the interception of raindrops by the stand canopy. The ground was also covered by litter fall that reduces speed of runoff and allows relatively better infiltration. This finding conforms to the findings of other studies. For instance, a study by Girmay et al. (2009) reported that in Eucalyptus-dominated plantation with limited understorey vegetation, there was no significant difference in runoff with grassland. Zhou et al. (2002) also stressed that runoff from Eucalyptus plantation decreased with accumulation of litter. However, some other studies reported result contrary to our findings. For instance, Descheemaeker et al. (2006) has found higher runoff under old Eucalyptus plantations (greater than 20 years), which was attributed to limited understorey vegetation cover.

Given the current result and other similar studies, Eucalyptus plantation could be used for catchment protection to reduce surface runoff. Its role can be enhanced with better litter accumulation and managing undergrowth. Canopy interception of Eucalyptus has made runoff generation to be less compared to the cultivated land. The intercepted water loss from Eucalyptus field is lower than other tree plantation and forests (Lima 1993). Tree planting spacing can also influence the amount of runoff generated from the field (FAO 2009). Generally, comparing runoff under Eucalyptus of different places is not advisable as other influencing factors such as soil, slope, precipitation regimes, climate, the growth stage of the forest, the use of ground vegetation and litter by local people often vary (Descheemaeker et al. 2006; FAO 2009). According to Hurni et al. (2005), surface runoff is expected to increase with land use expansion and intensification without soil and water conservation. Similar to this study, Girmay et al. (2009) and Adimassu et al. (2014) have found high correlation coefficient in the rainfall and runoff volume.

Conclusions

LULC can meaningfully influence runoff generation and runoff coefficient. There was significant difference in surface runoff among the compared LULC types. Cultivated land has generated higher surface runoff volume in the study period. However, there was no significant difference between grassland and Eucalyptus woodlots. This could be related with canopy cover, ground cover, litter availability or soil infiltration capacity during study period. According to the finding of this study, shifting the land use from cultivated land to Eucalyptus could reduce 21% of the surface runoff volume generated from the area. Where there is an ample amount of precipitation, using Eucalyptus as soil conservation tree could be one option. This is because it can reduce surface runoff generated from the area as compared to cultivated land.

The main reason for reduced surface runoff in Eucalyptus plantations is believed to be canopy interception that leads storage and slowly movement of water in order to percolate to the ground. As the runoff study spatially limited to the local conditions, further multiple studies should be done. Otherwise, it could not be used to compare the results from different areas. In general, the expansion of Eucalyptus has no significant impact on surface runoff generation if it is expanded on previously grassland. However, it could also significantly reduce the surface runoff generated if it is planted on previously cultivated land. In addition, this study has also observed higher soil moisture under Eucalyptus woodlots than cultivated land. Depending the above observations, Eucalyptus can be used as area conservation tree, especially to reduce soil erosion by water, where high runoff recorded fields with consideration of tree planting spacing. The effect of proper Eucalyptus planting spacing and litter accumulation level on the runoff generation needs further studies in the country.