La présente recherche répond au besoin actuel d'évaluer l'impact des activités humaines sur les ressources naturelles de manière à encourager une exploitation durable de ces ressources. On s'intéresse précisément ici à une centaine de parcelles situées autour du terroir d'Umarbugu au Sud du Mali, une région particulièrement sensible au phénomène de dégradation des sols. L'application du modèle SLEMSA aux données caractérisant les parcelles agricoles mentionnées précédemment a permis d'obtenir des estimations annuelles des taux de pertes en sols. Celles-ci reflètent la grande variabilité et l'intensité du climat avec des années de sécheresse et très peu de pertes en sols mais aussi des années dans la moyenne, 850 mm de précipitations répartis sur environ quatre mois, ou davantage, avec des épisodes de pluies beaucoup plus intenses et de fait des pertes beaucoup plus alarmantes pouvant dépasser les 35 t/ha/année. L'analyse des estimations obtenues a également permis d'identifier certains facteurs aggravant tels que la culture du coton ou les champs ayant une grande superficie. Le modèle choisi permet de tenir compte d'éventuelles pratiques de conservation et l'on sait que certains propriétaires des champs d'Umarbugu commencent à utiliser la culture en courbes de niveau, ce qui réduit l'écoulement et favorise la rétention de l'eau. On a donc effectué quelques simulations afin d'évaluer l'impact de telles pratiques et l'on a abouti à une diminution d'environ 60 % des taux de pertes de sols.

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

MENVSC (Geography) / Department of Geography and Geo-Information Sciences / Environmental degradation is a quite familiar factor of the mining industry that has been associated with South African mining industry from the beginning. The decommissioning of abandoned mines before the environment legislation, The National Environmental Management Act 107 of 1998 and the Minerals and Petroleum Resources Development Act 23 of 2002, was introduced is of great concern as the abandonment of mines without appropriate remediation and pollution monitoring was the result. Soil erosion has been recognised as an environmental hazard that emanates from abandoned mines. This study seeks to assess the soil erosion hazard around Nyala abandoned mine. The modified method of Soil Loss Estimation Model for Southern Africa (SLEMSA), for assessing soil erosion hazard, was used to estimate the spatial variation of erosion to achieve the goal of the study. Parameters that were considered for the model include relief (Slope steepness, S & slope length, L), soil erodibility (Fb), vegetation cover (C) and rainfall erosivity (E). Soil samples were collected from the field and; sieve and hydrometer analysis was conducted to determine the erodibility factor value of the study area. The model was run in a GIS environment (ArcGIS) and the parameters were multiplied to generate a soil erosion hazard map for the abandoned Nyala mine area. Results from the study indicated that 74.3 % of the watershed experiences low to moderate erosion hazard, with an estimated annual soil loss of 2.76 tons/ha/yr. The low rates of soil erosion in most parts of the watershed are associated with the low topographic ratio and low rainfall erosivity. The research demonstrated that the modified SLEMSA model used within GIS is a very useful tool as it enhances the capacity to assess and model the spatial variation of soil erosion hazard in a timeously and affordable manner. / NRF