In the United States, the annual cost of on-site soil erosion problems such as soil and nutrient losses, and off-site soil erosion problems such as sedimentation of lakes and river, loss of navigable waterways, flooding and water quality impairment, has been estimated at 44 billion USD (Pimentel, 1995; Telles, 2011). While eroding sediment sources can either be from land or from stream/river systems, the erosion from streambanks can be quite significant, reaching up to 80% of sediment leaving a watershed (Simon et al 2002; Simon and Rinaldi 2006). Despite many decades of research one the erosion of cohesive soils by flowing water (fluvial erosion), this significant aspect of environmental sustainability and engineering is still poorly understood. While past studies have given invaluable insight into fluvial erosion, this process is still poorly understood. Therefore, the objective of this dissertation was to examine the relationship between time and erosion resistance of remolded cohesive soils, and to quantify and model the effects soil and water temperature on the fluvial erosion of cohesive soils First, erosion tests were performed to investigate how soil erosion resistance develops over time using three natural soils and testing in a laboratory water channel. Results showed that the erosion rate of the soils decreased significantly over the time since the soils were wetted. This study indicates researchers need to report their sample preparation methods in detail, including the time between sample wetting and sample testing. Second, erosion tests were performed at multiple soil and water temperatures. Results showed that increases in water temperature led to increased erosion rates while increases in soil temperature resulted in decreased erosion rate. When soil and water temperatures were equal, erosion results were not significantly different. Results also showed a linear relationship between erosion rate and the difference between soil and water temperatures, indicating erosion resistance decreased as heat energy was added to the soil. Lastly, two common erosion models (the excess shear stress and the Wilson models) were evaluated, and were modified to account for soil and water temperature effects. Results showed that, compared to the original models, the modified models were better in predicting erosion rates. However, significant error between model predictions and measured erosion rates still existed. Overall, these results improve the current state of knowledge of how erosion resistance of remolded cohesive soils evolves with time, showing the importance of this factor in the design of cohesive erosion experiments. Also, the results show that by accounting for thermal effects on erosion rate, the usability of erosion models can be improved in their use for erosion predictions in soil and water conservation and engineering practice. / PHD / In the United States, the annual cost of on-site soil erosion problems such as soil and nutrient losses, and off-site soil erosion problems such as sedimentation of lakes and river, loss of navigable waterways, flooding and water quality impairment, has been estimated at 44 billion USD (Pimentel, 1995; Telles, 2011). While eroding sediment sources can either be from land or from stream/river systems, the erosion from streambanks can be quite significant, reaching up to 80% of sediment leaving a watershed (Simon et al 2002; Simon and Rinaldi 2006). Despite many decades of research one the erosion of cohesive soils by flowing water (fluvial erosion), this significant aspect of environmental sustainability and engineering is still poorly understood. While past studies have given invaluable insight into fluvial erosion, this process is still poorly understood. Therefore, the objective of this dissertation was to examine the relationship between time and erosion resistance of remolded cohesive soils, and to quantify and model the effects soil and water temperature on the fluvial erosion of cohesive soils First, erosion tests were performed to investigate how soil erosion resistance develops over time using three natural soils and testing in a laboratory water channel. Results showed that the erosion rate of the soils decreased significantly over the time since the soils were wetted. This study indicates researchers need to report their sample preparation methods in detail, including the time between sample wetting and sample testing. Second, erosion tests were performed at multiple soil and water temperatures. Results showed that increases in water temperature led to increased erosion rates while increases in soil vi temperature resulted in decreased erosion rate. When soil and water temperatures were equal, erosion results were not significantly different. Results also showed a linear relationship between erosion rate and the difference between soil and water temperatures, indicating erosion resistance decreased as heat energy was added to the soil. Lastly, two common erosion models (the excess shear stress and the Wilson models) were evaluated, and were modified to account for soil and water temperature effects. Results showed that, compared to the original models, the modified models were better in predicting erosion rates. However, significant error between model predictions and measured erosion rates still existed. Overall, these results improve the current state of knowledge of how erosion resistance of remolded cohesive soils evolves with time, showing the importance of this factor in the design of cohesive erosion experiments. Also, the results show that by accounting for thermal effects on erosion rate, the usability of erosion models can be improved in their use for erosion predictions in soil and water conservation and engineering practice.

Sediment is one of the leading non-point source pollutants in the U.S and has detrimental effects on biological communities such as aquatic communities; human use such as recreation; and natural processes such as flood water storage. For silvicultural operations, the majority of sediment is produced from erosion on highly disturbed areas, such as skid trails, haul roads, and log landings. Erosion from silvicultural activities not only has the potential to introduce sediment into waterways but can also decrease site productivity through the removal of topsoil. In order to minimize erosion from silvicultural operations, forestry Best Management Practices (BMPs) have been developed, but efficacies of various BMP options are not well documented. This study evaluated five closure and cover BMPs for the control of erosion on bladed skid trails through both field based measurements with sediment traps and soil erosion modeling. The erosion models used were the Universal Soil Loss Equation for Forestry (USLE – Forest), the Revised Universal Soil Loss Equation version 2 (RUSLE2), and the Water Erosion Prediction Project for Forest Roads (WEPP – Forest Roads). Erosion model predictions were also regressed against field based results to determine accuracy. The bladed skid trail BMP treatments evaluated were: 1) water bar only (Control); 2) water bar and grass seed (Seed); 3) water bar, grass seed, and straw mulch (Mulch); 4) water bar and piled hardwood slash (Hardwood Slash); and 5) water bar and piled pine slash (Pine Slash). Field based results show that the Control treatment was the most erosive (137.7 tonnes/ha/yr), followed by the Seed treatment (31.5 tonnes/ha/yr), Hardwood Slash treatment (8.9 tonnes/ha/yr), Pine Slash treatment (5.9 tonnes/ha/yr), and finally the Mulch treatment was the most effective erosion control technique (3.0 tonnes/ha/yr). Model accuracy results show that RUSLE2 performed the best overall. Both USLE – Forest and WEPP – Forest Roads under predicted values on the Control treatment, where erosion rates were very high. WEPP – Forest Roads under predicted these values the most. All models generally show that the Control was the most erosive followed by the Seed, Hardwood Slash, Pine Slash, and Mulch treatments. / Master of Science

Accelerated soil erosion is one of the major threats to agricultural production in Ethiopia and the Harerge region is not exceptional. It is estimated that about 1.5 billion tones of soil is being eroded every year in Ethiopia. In the extreme cases, especially for the highlands, the rate of soil loss is estimated to reach up to 300 t ha-1yr-1 with an average of about 70 t ha -1yr-1 which is beyond any tolerable level. The government have made different attempts to avert the situation since 1975 through initiation of a massive program of soil conservation and rehabilitation of severely degraded lands. Despite considerable efforts, the achievements were far bellow expectations. This study was aimed at assessing the effect of some soil properties, rainfall intensity and slope gradients on surface sealing, soil erodibility, runoff and soil loss from selected sites in the Harerge region, eastern Ethiopia, using simulated rainfall. Soil loss was also estimated for the sites using Soil Loss Estimation Model for Southern Africa (SLEMSA) and the Universal soil Loss Equation (USLE). Moreover, the effectiveness of various rates and patterns of wheat residue mulching in controlling soil loss was also evaluated for one of the study sites, (i.e. Regosol of Alemaya University), under both rainfall simulation and field natural rainfall conditions. For most of the erosion parameters, the interaction among soil texture, slope gradient and rainfall intensity was significant. In general however, high rainfall intensity induced high runoff, sediment yield and splash. The effect of slope gradients on most of the erosion parameters was not significant as the slope length was too small to bring about a concentrated flow. The effect of soils dominated by any one of the three soil separates on the erosion parameters was largely dependent on rainfall intensity and slope gradient. The soils form the 15 different sites in Harerge showed different degrees of vulnerability to surface sealing, runoff and sediment yield. These differences were associated with various soil properties. Correlation of soil properties to the erosion parameters revealed that aggregate stability was the main factor that determined the susceptibility of soils to sealing, runoff and soil loss. This was in turn affected by organic carbon content, percent clay and exchangeable sodium percentage (ESP). Soils with relatively high ESP such as those at Babile (13.85) and Gelemso (7.18) were among the lowest in their aggregate stability (percent water stable aggregates of 0.25 –2.0mm diameter); and have highest runoff and sediment yield as compared to other soils in the study. Similarly, most of those soils with relatively low ESP, high organic carbon content (OC%) and high water stable aggregates such as Hamaressa, AU (Alemaya University) vertisol and AU regosol were among the least susceptible to sealing and interrill erosion. Nevertheless, some exceptions include soils like those of Hirna where high runoff was recorded whilst having relatively high OC%, low ESP and high water stable aggregates. Both the SLEMSA and USLE models were able to identify the erosion hazards for the study sites. Despite the differences in the procedures of the two models, significant correlation (r = 0.87) was observed between the values estimated by the two methods. Both models estimated higher soil loss for Gelemso, Babile, Karamara and Hamaressa. Soil loss was lower for Diredawa, AU-vertisol and AU-Alluvial all of which occur on a relatively low slope gradients. The high soil loss for Babile and Gelemso conforms with the relative soil erodibility values obtained under rainfall simulation suggesting that soil erodibility, among others, is the main factor contributing to high soil loss for these soils. The difference in the estimated soil losses for the different sites was a function of the interaction of the various factors involved. Though the laboratory soil erodibility values were low to medium for Hamaressa and Karamara, the estimated soil loss was higher owing to the field topographic situations such as high slope gradient. SLEMSA and USLE showed different degrees of sensitivities to their input variables for the conditions of the study sites. SLEMSA was highly sensitive to changes in rainfall kinetic energy (E) and soil erodibility (F) and less sensitive to the cover and slope length factors. The sensitivity of SLEMSA to changes in the cover factor was higher for areas having initially smaller percentage rainfall interception values. On the other hand, USLE was highly sensitive to slope gradient and less so to slope length as compared to the other input factors. The study on the various rates and application patterns of wheat residue on runoff and soil loss both in the laboratory rainfall simulation and under field natural rainfall conditions revealed that surface application of crop residue is more effective in reducing soil loss and runoff than incorporating the same amount of the residue into the soil. Likewise, for a particular residue application method, runoff and soil loss decreased with increasing application rate of the mulch. However, the difference was not significant between 4 Mg ha-1 and 8 Mg ha-1 wheat straw rates suggesting that the former can effectively control soil loss and can be used in areas where there is limitation of crop residues provided that other conditions are similar to that of the study site (AU Regosols). The effectiveness of lower rates of straw (i.e. less than 4 Mg ha-1 ) should also be studied. It should however be noted that the effectiveness of mulching in controlling soils loss and runoff could be different under various slope gradients, rainfall characteristics and cover types that were not covered in this study. Integrated soil and water conservation research is required to develop a comprehensive database for modelling various soil erosion parameters. Further research is therefore required on the effect of soil properties (with special emphasis to aggregate stability, clay mineralogy, exchangeable cations, soil texture and organic matter), types and rates of crop residues, cropping and tillage systems, mechanical and biological soil conservation measures on soil erosion and its conservation for a better estimation of the actual soil loss in the study sites. Copyright 2004, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. Please cite as follows: Bobe, BW 2004, Evaluation of soil erosion in the Harerge region of Ethiopia using soil loss models, rainfall simulation and field trials, PhD thesis, University of Pretoria, Pretoria, viewed yymmdd < http://upetd.up.ac.za/thesis/available/etd-08022004-141533 / > / Thesis (PhD (Soil Science))--University of Pretoria, 2004. / Plant Production and Soil Science / unrestricted

L’objectif de cette thèse est de proposer un modèle d’érosion diffuse qui intègre les principaux processus de ce phénomène (détachement, transport, sédimentation) et qui prend en compte l’interaction des gouttes de pluie avec ces processus. Dans un premier temps, nous avons établi une loi de détachement par la pluie qui inclut l’effet des gouttes et celui de l’épaisseur de la lame d’eau qui couvre la surface du sol. Pour obtenir cette loi, une étude numérique avec le logiciel Gerris a permis de modéliser les cisaillements créés par l’impact des gouttes sur des épaisseurs de lame d’eau variables. Ces cisaillements estiment la quantité de sol détaché par chaque goutte. Nous avons montré, à travers une étude probabiliste, que les gouttes sont quasiment indépendantes lors du détachement. Les détachements de l’ensemble des gouttes sont donc sommés pour établir la loi de détachement pour la pluie. Par ailleurs, l’étude probabiliste a montré la possibilité d’une forte interaction entre les gouttes de pluie et les particules en sédimentation. Par conséquent, pour le processus de transport-sédimentation, nous avons privilégié une approche expérimentale. Cette étude a révélé que l’effet des gouttes de pluie est d’augmenter la vitesse de sédimentation des particules. Enfin, nous avons proposé un nouveau modèle d’érosion qui généralise plusieurs modèles d’érosion de la littérature et décrit l’évolution des concentrations en sédiments avec des effets linéaires et non-linéaires. Selon le choix des paramètres du modèle, celui-ci peut représenter l’érosion diffuse et concentrée à l’échelle du bassin versant, le transport par charriage dans les rivières ou encore le transport chimique. L’intégration du modèle dans le logiciel de ruissellement FullSWOF est aussi réalisée. / The aim of this work is to formulate an interrill erosion model. This model should take into account the main erosion processes (detachment, transport and sedimentation) and the interaction of raindrops during these processes. First we develop a law for rainfall detachment that includes the effects of the raindrops and the water layer thickness at the soil surface. We use the Gerris software to simulate the shear stresses created by the impacts of raindrops at the soil surface. These shear stresses allow to evaluate the quantity of soil detached by each raindrop. We have shown with a probabilistic approach that raindrops are almost independent during soil detachment. Then by summing all the raindrops detachments we obtain the rainfall detachment law. Futhermore the probabilistic study has revealed the possibility of a strong interaction between raindrops and settling particles. So, we used specific laboratory experiments to investigate the particles transport and sedimentation processes. These experiments show that the effect of raindrops is to increase the particles settling velocity. Finally, we propose a new erosion model which encompasses previous literature erosion models and that can describe the behavior of sediments concentrations with linear and non-linear behaviors. The model is able to simulate interrill and rill erosions at the watershed scale, bedload transport in rivers and chemical transfer. The integration of the model in the FullSWOF runoff software is also carried out.