The overwhelming focus of documentary sources indicate that traditional approaches to land degradation and soil erosion in South Africa have focussed on the physical dimension of the problem and the development of practical solutions to its reduction. This study was undertaken from the viewpoint that this emphasis has resulted in the neglect of other (for example, socio-political) aspects of soil erosion and that such neglect has exacerbated the soil erosion problem manifest in South Africa. An examination of the 'human dimension' of soil erosion in South Africa was therefore undertaken through an analysis of soil conservation policy and legislation promulgated to effect policy objectives between 1910 and 1992. Acknowledging that the policy environment is influenced by factors within the economic, political, historical and perceptual (all human) environments, as well as the natural environment, this study attempts to integrate information relating to each of these parameters within the overall framework of South African soil conservation policy. Particular emphasis is placed on the role of environmental perception in the decision-making process, together with the critical influence of intervening variables found to be operative within the perceptual environment as represented in the South African context. In the absence of substantive empirical data, this study posits a number of a priori assumptions regarding the extent and causes of soil erosion, support for which was initially derived from the extensive literature sources reviewed for the study. The basic premise of this study is that soil erosion persists in South Africa, and despite considerable government and public inputs and participation, and the existence of a legislative machinery created specifically to address the problem, progress in promoting soil conservation through implementation of specific measures has been slow. Following a review of soil conservation policy and legislation up to 1992, it is further submitted that factors other than legislative inadequacies could account for this problem and therefore warrant particular and thorough investigation. To this end, this work firstly describes the physical context within which South African soil erosion occurs, followed by a brief appraisal of socio-economic and political variables which together have shaped contemporary perceptions regarding the nature, extent and causes of soil erosion in the country. A comprehensive review of relevant policy and legislation in the period 1910-1992 was then undertaken by reference to published and unpublished sources. The evolution of soil conservation policy in the country was charted through reference to relevant legislation and parliamentary debate. The temporal variation in the relative success of the policy and legislative enactments was measured by reference to relevant indicators. The adequacy of South African soil conservation (as indicated by four key legislative enactments formulated to specifically address soil erosion) was evaluated using key elements of the World Soils Policy as a baseline. Analysis of the Forest and Veld Conservation Act of 1941, the Soil Conservation Act of 1946, the Soil Conservation Act of 1969 and the Conservation of Agricultural Resources Act of 1983, revealed a clear evolutionary progression in which successive acts sought to build upon successes and minimise the weaknesses of previous efforts. This analysis reveals marked temporal variability in the extent to which each element is addressed and explores the multi-environmental (political, social, historical, economic and perceptual) constraints on attainment of all goals. This assessment permitted the compilation of, what the author has termed the South African Policy Environment Model, which takes the form of a working hypothesis. This hypothesis was subsequently tested utilising the 103 responses obtained from a postal questionnaire survey directed at 242 scientists, policy developers and extensionists (representing a 43 per cent return), that is, those persons who either currently or during the study period, were actively involved in the development of soil erosion research and/or the implementation of conservation policy objectives. In spite of the apparent effort by the South African government to address soil conservation, contemporary opinion (according to documentary evidence) suggests that the policies formulated have failed to attain soil conservation goals and reduce the manifest extent and rate of land degradation in the country. This study broadly concludes that such inefficacy of policy may be ascribed to: 1. lack of importance ascribed to soil 2. national level control 3. non-uniformity in application of law 4. inadequacies in the implementation of policy 5. paucity of information on real nature and extent of problem 6. perceptions in an uninformed environment. The study furthermore submits that images concerning the realities of soil erosion are shaped by perceptual filters and the value systems of individuals active in the soil conservation arena, or more specifically, key players' perceptions regarding the causes, extent and nature of the soil erosion problem, are what underpin and ultimately give rise to the relative effectiveness of soil conservation strategies. This study identifies a multiplicity of factors which operate within five dynamically interative environments (the political, economic, historical, perceptual and natural environments) considered influential in shaping the temporal (and spatial) variation in the policy environment represented in this study. This examination of the multidimensionality of soil erosion has led to the conclusion that in addition to the problems broadly outlined above, soil erosion is also a problem of: 1. accountability; 2. focus; 3. priorities and government commitment; 4. situational incompatibility; 5. misinformed perceptions; and 6. timing. It is submitted that lack of recognition of these inter- and intra-environment dynamics could account for the relative inefficacy of soil conservation policy to promote the sustained adoption of conservation practices. Such factors will in the past have been overlooked due to the neglect of the 'human dimension' of the problem in South Africa in the period under review. It is believed that the measure of consensus derived from the results of this study, reflects contemporary realities concerning the status of soil conservation in the country, at least amongst those individuals most intimately involved in the development, formulation and administration of soil conservation policy. As such it provides an appropriate foundation upon which to base future policy decisions and more importantly, to derive optimum compliance with conservation norms and standards of practice amongst land users. Only by recognising the multidimensionality of the soil conservation policy environment and its components, can the past inefficacies be overcome. It is submitted therefore that for South Africa to meet its challenges of the 21st century concerning the conservation and sustainable utilisation of soil, the priority of policy developers must be - the expedient adoption of a multi- and interdisciplinary approach to agricultural resource management, with particular emphasis on its 'human dimension' . / Thesis (Ph.D.)-University of Natal, Durban, 1996.

Land and water degradations are serious environmental concerns facing South Africa. One of the major causes is human-induced soil erosion due to intensified land uses and environmental degradation caused by bad agricultural practices and inappropriate land uses. Soil erosion is a typical and important example of land degradation that the LandCare program intends to address. In order to evaluate the success of LandCare project in addressing soil erosion evaluation tools which can be applied during the project monitoring and evaluation process. Several soil loss models and field assessment methods were theoretically evaluated on criteria such as the scientific principles, availability and the impact of data requirements of the models. Soil loss models, SLEMSA (Soil Loss Estimation for Southern Africa) and RUSLE (Revised Universal Soil Loss Estimation) and the ACED (Assessment of Current Erosion Damage) method were selected to be tested in a study area that is naturally susceptible to erosion. The soil loss as predicted by RUSLE and SLEMSA and that resulting from visible damage as accounted by ACED method is not comparable. The SLEMSA and RUSLE models vary considerably in extent and approach compared with the field assessment method, ACED. ACED can be used as a participatory learning erosion tool and to identify critical areas on hillslopes. SLEMSA and RUSLE had predictive advantage over ACED and could predict soil loss before and after the LandCare project. Therefore, the models were considered valuable tools to guide decision-making based on the management and use of the natural resources on farmland or by the community. Soil loss models that require readily available input data, such as RUSLE and SLEMSA, are suitable evaluation tools for monitoring and evaluation of soil erosion in LandCare project. Based on the results of the scenario prediction study, it was found that RUSLE could simulate the impact of different agricultural practices much better than SLEMSA. However, the reliability of SLEMSA and RUSLE should be verified with measured data from erosion plots as RUSLE also underestimate the erodibility of the Escourt soil. / Dissertation (MSc (Agriculture))--University of Pretoria, 2006. / Plant Production and Soil Science / unrestricted

Thesis (PhD) -- Stellenbosch University, 2003. / Please refer to full text to view abstract. / ENGLISH ABSTRACT: The effects of different land use practices on soil quality in South Africa were investigated in five contrasting biomes, with a particular emphasis on the tendency of soils to crust and soil C content. Soil quality is a nebulous concept and its applicability in the South African landscape is scrutinised. A wide range of chemical and physical soil properties were examined. The tendency of soils to crust was assessed using modulus of rupture, water dispersible clay and a new method of laboratory infiltration which was verified with rainfall simulation. Crusting was greater in bare, exposed soils than soils under vegetation and varied with soil parent material. Differences in crusting are explained by factors relating to clay dispersion such as clay mineralogy, soil C, labile or readily oxidisable C, concentration of soluble salts, soil texture and exchangeable Na percentage (ESP). Results from longterm bum plots in savanna and grassland revealed that annual burning can increase the tendency of soils to crust. Greater crusting in burnt plots is ascribed to greater dispersion of clay, which in tum is attributed to a decline in soil C, a decline in EC and an increase in ESP. The loss of nutrients from burnt plots over time is ascribed to removal of ash in surface runoff. Calcium, Mg, and K were lost more readily than Na probably because plants take up these nutrients in greater concentration than Na. The net effect was an increase in ESP. Crusting on burnt plots may be self-perpetuating, because increased runoff is likely to incsease the loss of soluble salts. Removal of vegetation due to cultivation, grazing or burning reduced soil C at all sites. Mean soil C in the 0-1 cm layer of unburnt plots in the Kruger National Park was more than three times greater than in burnt plots (2.7 vs 0.8%). The difference in soil C between treatments decreased with depth and illustrated that sampling to depths greater than a few centimetres can obscure effects of land use. The top few centimetres of soil have a disproportionate effect on soil infiltrability and nutrient cycling. This layer was named the pedoderm. Tree cover on burnt plots in the southern Kruger Nationa--l- Park is highly variable, and was hypothesised to be a function of herbivory pressure. Herbivores tended to congregate on plots with the greatest clay, Zn and Mn content and the lowest tree cover. It is suggested that soil properties determine the abundance of herbivores after fire, which in tum affects tree cover. In the Eastern Cape,intensive stocking with goats transforms dense thicket to an open savanna. Soils from goat-transformed sites had a greater tendency to crust than soils from intact thicket, probably due to aggregate weakening associated with a decline in soil C. Mean soil C content of intact thicket was almost double that of goat-transformed thicket (5.6 vs. 3% to a depth of 10 cm) and is exceptionally high for a semi-arid region. The potential to sequestrate carbon in degraded thicket landscapes is thus considerable. Managing the land for greater sequestration of C will have the added benefit of increasing soil aggregate stability, reducing the tendency of soils to crust and therefore increasing the rate of water infiltration through the pedoderm. The benefits of such an approach have been recognised by specialists in soil conservation and rural land use for many decades, based largely on empirical observation. The results of this thesis provide a more quantitative basis for appreciating the effects of soil C across a broad spectrum of South African biomes. / AFRIKAANSE OPSOMMING: Die invloed van verskillende landgebruikspraktyke op grondkwaliteit in Suid-Afrika is in vyf kontrasterende biome ondersoek met spesifieke klem op die neiging van gronde om korste te vorm en die grond koolstofinhoud. Grondkwaliteit is 'n vae konsep en die toepassing daarvan in die Suid-Afrikaanse grondlandskap is noukeurig ondersoek. 'n Wye reeks van chemiese en fisiese grondeienskappe is ondersoek. Die neiging van korsvorming by gronde is beraam deur die gebruik van breukmodulus, waterdispergeerbare klei en 'n nuwe metode van laboratorium-infiltrasie wat met behulp van reënvalsimulasie gekontroleer is. Korsvorming was groter in kaal, blootgestelde gronde as in gronde met 'n plantbedekking en het gewissel volgens moedermateriaal. Verskille in korsvorming word verklaar deur faktore wat verband hou met kleidispergering soos byvoorbeeld kleimineralogie, grondkoolstof, labiele of maklik oksideerbare koolstof, konsentrasie oplosbare soute, grondtekstuur en uitruilbare natriumpersentasie (UNP). Resultate van langtermyn brandpersele in savanna en grasland het getoon dat jaarlikse brand die neiging tot korsvorming kan verhoog. Meer korsvorming in brandpersele word toegeskryf aan groter kleidispergering, wat waarskynlik verband hou met 'n afname in grondkoolstof, 'n afname in elektriese geleiding (Be) en 'n toename in UNP. Plantvoedingstowwe gaan oor tyd verlore uit brandpersele, waarskynlik deur die verwydering van as in oppervlak afloop. Kalsium, Mg en K gaan meer geredelik verlore as Na, waarskynlik omdat plante hierdie voedingstowwe in groter hoeveelhede opneem as Na. Die netto effek is 'n toename in UNP. Korsvorming op brandpersele kan self instand gehou word omdat verhoogde afloop die moontlike verlies van oplosbare soute kan verhoog. Verwydering van plantegroei deur bewerking, beweiding of brand het grondkoolstof op alle plekke verlaag. Die gemiddelde grondkoolstof in die 0-1 cm laag van ongebrande persele in die Kruger Nasionale Park was meer as drie maal groter as in --.. brandpersele (2.7 vs 0.8 %). Die verskil in grondkoolstof tussen behandelings neem af met diepte wat daarop dui dat monsterneming tot dieptes groter as 'n paar sentimeters die effek van landgebruik kan verberg. Die boonste paar sentimeters van 'n grond het 'n oneweredige invloed op infiltrasie en voedingstofsirkulasie. Hierdie laag word die pedoderm genoem. Boombedekking op brandperseie in die suidelikeKruger Nasionale Park is hoogs variërend. Die hipotese was dat dit 'n funksie van druk deur planteters is. Planteters neig om op persele met die hoogste klei, Zn- en Mn-inhoud en die laagste boom bedekking te versamel. Daar word voorgestel dat grondeienskappe die hoeveelheid planteters na 'n brand bepaal. Dit beïnvloed op sy beurt weer die boombedekking. In die Oos-Kaap het intensiewe bokboerdery digte bosruigtes verander na oop savannas. Gronde van bok-veranderde lokaliteite het 'n groter neiging tot korsvorming as gronde van onveranderde bosruigtes, vermoedelik as gevolg van verswakking van aggregate met 'n. afname in grondkoolstof. Die gemiddelde grondkoolstof van onveranderde bosruigtes was byna dubbel soveel as die koolstof van bok-veranderde bosruigtes (5.6 vs 3 % tot 'n diepte van 10 cm) en buitengewoon hoog vir 'n semi-ariede streek. Die potensiaal vir koolstof sekwestrasie in degradeerde bosruigte landskappe is dus aansienlik. Bestuur van land vir groter sekwestrasie van koolstof het die bykomende voordeel van verhoogde grond aggregaatstabilteit, verlaging van die neiging tot korsvorming en daardeur 'n verhoging in die tempo van waterinfiltrasie deur die pedoderm. Die voordele van so 'n benadering is vir baie dekades deur spesialiste in grondbewaring en landelike landgebruik herken. Dit was grootliks gebaseer op empiriese waarneming. Die resultate van hierdie tesis bied 'n meer kwantitatiewe basis tot die verstaan van die invloed van grondkoolstof oor 'n breë spektrum van Suid-Afrikaanse biorne.

M.Ing. (Civil Engineering) / The purpose of this study was to determine the effectiveness of a sisal fibre natural geotextile in controlling the rate and extent of soil erosion. The effectiveness of this locally produced material was compared to other natural fibre geotextiles and other methods commonly used for the control of soil erosion. The different materials and methods were tested under the same conditions to compare the relative effectiveness of these products. The various erosion control products were subjected to laboratory tests. A rainfall simulator and an adjustable slope simulator were used to simulate the process of soil erosion. During these tests the effectiveness of the different products in controlling soil erosion was measured and conclusions were made about the effectiveness of the products. Mechanical sieve analysis tests were used to determine the erodibility of the different soil particles. Field trials that made use of the sisal fibre geotextile are also discussed.

Thesis (MTech (Civil Engineering))--Cape Technikon, 1992. / In the Western Cape, gully ("donga") erosion is characteristic of the Swartland area where the effects of the problem have reached spectacular dimensions. In 1942 the Social and Economic Planning Council commissioned an investigation into the distribution and prevention of gully erosion. This investigation was led by Professor W J Talbot of the University of Cape Town, with the collaboration of researchers from the University of Stellenbosch. This led to the "Swartland and Sandveld" publication in August 1945. Little progress has since been made. The gullies identified by Prof Talbot's team remain to this day. In most cases they have grown even longer and deeper. The present investigation aims to study a typical network in one catchment area and to propose measures for the reclamation or stabilization of gullies. Such measures must not only be effective but also economically justifiable. In reclaiming or stabilizing a donga, a key factor is the establishment of a foundation structure. Such a structure is essential for supporting semi-permanent structures and promoting the growth of grass and bushes. At present, however, no satisfactory methods are known for the design and construction of supporting walls In narrow, deep dongas, other than those methods that require a rock foundation. Furthermore, conventional methods would require a buttress of length 17 m for a retaining wall of height 4,4 m. The amount of concrete needed, would clearly be excessive. It is therefore one of the aims of this study to develop a retaining wall with a relatively short buttress. If finances permit, the retaining structure, will be built and the stabilization process carefully monitored for possible applications in the future.

Commercial forestry is recognized for both its economic contribution as well as its environmental impact. Of particular concern, is the soil erosion and sedimentation of watercourses associated with forestry plantations. Environmental laws regulate many of the activities of the forestry sector. It is critical that the forestry sector ensure that its operations are compliant with the legal requirements that govern its use of natural resources. In pursuing legal compliance it is necessary to ensure that erosion control strategies are developed so as to ensure the positive effects of any interventions are optimised. The identification of areas that are particularly at risk to erosion or contribute to sediment delivery is an essential component in prioritising areas for management interventions. Establishing the erosion potential for commercial forestry areas is readily accomplished through the application of existing models. Process based erosion models generally have greater data requirements than the empirically derived USLE-based models. Given the paucity of data available, the latter approach was adopted. Two methods of topographic sub-factor derivation were investigated, those associated with the RUSLE (Renard, Foster, Weesies & McCool1991) and the Unit Stream Power method presented by Moore and Burch (1986). Since no existing methods identifying delivery risk areas existed, a method was developed based on principles and factors identified in the literature. Additionally, methods for identifying topographic assets, in terms of sediment attenuation, were developed. From these models three indices were derived; sediment supply, delivery risk and sediment attenuation. Thereafter, the mean Sediment Supply Index was divided by stream length for small catchments defined within the landscape to derive an index of sediment loading to streams. This index is used to identify priorities for management intervention across the landscape. The mean slope and sediment supply is used to develop buffer width recommendations for the streams draining the catchments, using a method developed by Karssies and Prosser (2001). Using the three indices in conjunction it is possible to make on-site and off-site erosion control recommendations as well as identify and exploit any natural features that can be utilized in erosion control. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2002

The Matlabas mire, an actively peat accumulating wetland, is located in the headwaters of the Matlabas River, Marakele National Park, Limpopo Province, South Africa. Various seepage zones and artesian peat domes are contained in this peatland that consists of two tributaries of which the western one is partially channelled. The occurrence of decaying peat domes and desiccated areas with terrestrial vegetation, as well as the apparent erosion on the western tributary, have raised concerns on the health of this wetland. A network of piezometers was installed in the mire and results confirm that the system is fed primarily from seepage from the slopes of the catchment. Chemical analysis and temperature recorded indicate an isolated groundwater source of which the water does not mix with surface water. This is linked with isotope analysis of the age of peat in various sections of the mire. Erosion was attributed to anthropogenic changes in the catchment. Management recommendations include rehabilitation and reinstating the driving forces that support the mire. / Environmental Sciences / M. Sc. (Environmental Management)

Soil erosion is a persistent problem that requires continued control efforts as agricultural land loses productivity and communities dependent on the land become increasingly vulnerable to decreased food security. The negative effects of soil erosion in Khamopele River catchment, in the Mount Fletcher southern Drakensberg region of South Africa, are manifest in extensive gullying and wetland loss. Soil erosion has resulted in siltation in a recently constructed dam and the alteration of aquatic habitats. This research was undertaken to identify the sources of eroded sediment in the small upper catchments of the Mzimvubu River catchment to inform broader catchment management strategies. The scale of erosion was quantified using field surveys of gully extent and form. Environmental magnetic tracing techniques were used to determine the sources of eroded sediment in Khamopele River and upper Tina River catchments. The radionuclide ¹³⁷Cs was used to determine soil loss over a 55 year period in Khamopele River catchment. The Landscape Connectivity framework was used to describe the sediment source, pathway and sink interactions at sample area level. Results indicated that historical and contemporary land management practices such as uncontrolled grazing, grassland burning and furrows promoted soil erosion in the catchment. Soil erosion was most pronounced in the Taung sample area where there was extensive gullying, tunnelling and subsurface erosion. Environmental magnetic tracing results indicated that there were clear differences in source areas. Despite its prevalence in the area, gully erosion was not shown to be a major source of sediment to downstream sinks. Topsoil and hillslope derived sediment were shown to be mobile in the catchment, suggesting that sheet erosion processes were dominant in the catchment. Radionuclide tracing studies showed that at least 20 cm of soil had been eroded from the Khamopele River catchment surface since 1956. This research has shown that it is possible to distinguish source areas of erosion in the catchment by matching catchment mineral magnetic signatures to those in sink areas. This means that rehabilitation projects can use resources efficiently as the areas needing the most attention can be identified.

[Portion of abstract]: Sediment dynamics are influenced by transformed landscape connectivity in catchments worldwide. The upper Thina catchment, an important high rainfall resource in the northern Eastern Cape, South Africa, is an example of where ongoing subsistence farming on communal land has led to overgrazing and trampling that has initiated large erosive features (e.g. gullies) and river incision. The formation of gullies led to increased hillslope-channel connectivity and the resultant river incision decreased the channel-valley fill connectivity. These two changes in connectivity led to increased sediment export from the catchment that has various down-stream ecological and socio-economic impacts. This study investigates how the change in hillslope-channel and channel-valley fill connectivity has altered the sediment dynamics in the Vuvu catchment, a headwater tributary of the Thina River. A combination of methods were used to assess the changes in hillslope-channel and channel-valley fill connectivity. High resolution aerial images were used to map source features, such as fields, gullies, sheet erosion, landslides, roads and livestock tracks. Topographic and geological characteristics of the source features were extracted using a Geographic Information System. Furthermore, hillslope-channel pathways, such as the natural drainage network, continuous gullies, discontinuous gullies, roads and livestock tracks were mapped and analysed in terms of topographic and geological characteristics. Historic aerial images were assessed to calculate the date the larger gullies began forming. Recent aerial photos and cross sectional surveys of the valley fill were combined to map the various sediment sinks. Particle size and organic content were analysed for flood bench cores and terrace samples. The chronology of the flood benches was determined using unsupported Pb-210 and Cs-137 dating, and determined for the terraces using Optically Stimulated Luminescence dating. Quantitative and qualitative sediment tracing approaches, using mineral magnetic properties, were used to trace the origin of suspended sediment (collected during flood events), sediment stored in the flood benches and sediment stored in the terraces. Hydrological monitoring was used to assess the potential to store sediment on flood benches along the valley fill through flood bench inundation frequency. Hydrological and hydraulic modelling extended the measured inundation frequencies to a 73 year period and other cross sections along the valley fill. Furthermore, a future scenario of an increased vegetation cover and reduced hillslope-channel connectivity was assessed in terms of channel-valley fill inundation frequency.