Effect of soil and water conservation measures on hydrological processes and sediment yield in the highlands of North-Western Ethiopia

Samenvatting

Prolonged soil drying and severe moisture stress during the dry season, and soil saturation, erosion, flooding and sedimentation during the short rainy season are critical problems in the North-Western Ethiopian highlands. Despite massive investment in SWC strategies to reduce these problems, soil erosion, moisture stress, flooding and sedimentation are still severe problems throughout Ethiopia and specifically in fragile regions of humid/sub-humid highlands of North-Western Ethiopia. The success to prevent erosion, moisture stress, flooding and sedimentation and effective management and utilisation of water and soil resources have been very limited (Bewket, 2007; Nyssen et al., 2007; Adimassu et al., 2012; Jemberu et al., 2017). The prevention of soil and water loss, flooding and sedimentation requires a thorough understanding of hydrological and sedimentary processes and selecting suitable SWC strategies to increase land productivity for effective management and utilisation of soil and water resources. However, very few studies are available about the impacts of SWC measures on hydrological processes and sediment yield in Ethiopia (Herweg and Ludi, 1999; Dagnew et al., 2015; Jemberu et al., 2017). Further, in order to foster SLM systems in Ethiopia, scientifically validated appropriate SWC strategies are needed.

After the general introduction (chapter 1), chapter 2 describes participatory assessment of soil erosion severity and performances of mitigation measures with local community. This research aimed to assess soil erosion and performance of different SWC measures from the farmers’ perspective by documenting their perceptions and experiences in Koga catchment, highlands of Ethiopia. To this aim, workshops were organised in three sub-catchments differing in slopes and SWC measures. Workshops included group discussions and field monitoring of sheet erosion indicators, and rills and systematically describing the status of soil erosion, soil fertility and yield to assess the performance of SWC measures. Results show that farmers are aware of not only the harmful effects of ongoing soil erosion but also the impacts of mitigation measures on their farms. Sheet erosion was found to be the most widespread form of erosion while rill damage was critical on plots cultivated to cereals on steep slopes.

Farmers perceived that the desired impacts of SWC practices were attained in general: runoff and soil loss rates decreased, while soil fertility and production increased. However, they also believed that SWC measures did not prevent soil erosion completely. Comparatively, graded stone-faced soil bunds revealed maximum desired impacts and were most appreciated by farmers whereas all level bunds caused water logging. Most traditional ditches were highly graded and begun incising and affected production of cereals.

Chapter 3 deals with the effect of bund structures dimensions on soil selected properties and hydrology. Field investigations and laboratory analysis were carried at three sub-watersheds of Koga catchment in the highlands of NW Ethiopia to evaluate the variation in bund structure design and its impact on soil-water retention, bulk density, porosity and soil depth. Results show that the dimensions and stability of bund structures, and consequently, the impacts of bunds varied widely. The widths and heights of bunds, and slope gradients were significantly correlated with infiltration rate and available soil-water content. Water holding capacity was 24% higher compared to that on untreated farms. Bunds with larger dimensions of top width (>85 cm) and height (>75 cm) showed 17 and 18% higher water holding capacity compared to fields with bunds having lower dimensions. In addition, results exhibited that the porosity and soil depth were significantly affected by the dimensions of bund structures and increased with increasing widths and heights of bunds. A 14% increase in porosity; a 41% increase in infiltration rate; and a 18% decrease in bulk density was found in soils of treated farms compared to those of untreated farm plots. Differences in particle size distribution and soil erodibility among widths of the bunds were low.

Chapter 4 discuss the quantitative relationships for rainfall features, runoff characteristics and sediment yield patterns for treated sub-watersheds areas in Koga catchment, highlands of North-Western Ethiopia. Multiple regression models were formulated to predict sediment discharge using daily: rainfall, 10-minute maximum rain intensity, rain intensity, SSC and runoff discharge. FDC, hydrographs, sedigraphs and statistical analysis were analysed to assess the hydrological and sedimentary processes and the response of treated sub-watersheds to extreme events of rainfall and runoff. Results show that very strong seasonal and event variability in rain intensity, runoff and sediment yield was observed in all study sites. The rain intensity was the strongest predictor of sediment discharge. This implies that the hydrological response of the study sub-watersheds mainly depended on rain intensity, as they have minimal variation in slope, topography and soils. Area-specific sediment yield was different for the study sub-watersheds and did not show scale dependence. There was a striking difference between seasonal sediment and runoff dynamics, with suspended sediment concentration being very high at the start of the rainy season and gradually decreasing over time, while this did not follow the temporal dynamics of the runoff, which was low at the start of the rainy season and peaked halfway in the rainy season. FDC illustrates that discharge variability is high and storage capacity very low in Koga catchment specifically in Asanat.

The importance of individual events was very high and more than 50% of runoff and 52% of total annual sediment load was produced by the five largest events. This dominance of the largest events in terms of runoff and suspended sediment transport clearly emphasises the high temporal variability of runoff and sediment production. Runoff and SSC event analysis during different seasons showed three different types of hysteretic loops: clockwise and linear (mainly for early rain season), complex and negative for middle and late rainy season. This indicates the complex and heterogeneous nature of runoff and sediment yield response of study sub-watersheds.

Chapter 5 considers the spatially distributed WATEM/SEDEM modelling approach to quantify the effects of SWC strategies on soil erosion and sedimentary processes. WATEM/SEDEM was calibrated for three sub-watersheds in Koga catchment (Asanat, Debre Yakob and Rim). Five scenarios of SWC measures were evaluated: scenario I: a baseline condition (present-day situation) treated with bund structures; scenario II: bund structures and contour cultivation; Scenario III: combination of bunds, contour cultivation and strip cropping; scenario IV: integrated use of bunds, contour cultivation, strip cropping and grass strips, and Scenario V: a scenario without SWC practices. The model simulation results showed that all SWC scenarios reduced soil erosion and sediment yield in all the sub-watersheds. Bund structures reduced soil erosion by more than 57, 65 and 53% in Asanat, Debre Yakob and Rim sub-watersheds, respectively.

The integrated use of bund structures, contour cultivation, strip cropping and grass strips (scenario IV) showed the largest effect on erosion and sediment yield. In scenario IV, compared to a scenario without SWC measures, sediment yield was reduced from 66.9 to 17.0 t ha-1 y-1 in Asanat, 41.1 to 13.9 t ha-1 y-1 in Debre Yakob and 59.0 to 12.5 t ha-1 y-1 in Rim sub watersheds. Bund structures and grass strips showed the largest contribution in controlling soil erosion and sediment yield in all study sub-watersheds. For the present-day situation, the measured and simulated erosion and sediment yield values were relatively lower for Debre Yakob which might be due to the lower transport capacity and sediment connectivity, as a result of larger coverage of bunds and other conservation measures such as traditional ditches and diversion channels. This emphasises the importance of integrated use of conservation strategies to reduce soil erosion and sediment delivery. This study illustrates that the measurement of sediment yield at the outlets of the catchment areas can be poor indicator of upland soil erosion processes.

The calibration of WATEM/SEDEM at sub-watershed level has provided good model performance for simulated erosion and sediment yields. Despite the semi-quantitative nature of the methodology, using farmers’ perceptions and experiences to document land degradation and the impacts of SWC measures is crucial as they are the daily users of the land and therefore directly affecting the success or failure of SWC measures. Field measurements and analysis of the local biophysical settings at catchment level are essential to enhance on- and off-site impacts of SWC measures. Therefore, WATEM/SEDEM representing the underlying soil erosion and sedimentary processes, documenting the local community perceptions and experiences, field measurements and analysis are crucial to evaluate the effect of SWC measures on hydrological processes and sediment yield on existing land use and SWC measures or new land use and SWC scenarios.