EcoRestore : a decision support system for the restoration of degraded rangelands in southern Africa

Thesis (M.Sc. (Botany))--North-West University, Potchefstroom Campus, 2004.

Land degradation is a process that causes the reduction in resource potential of natural rangelands and occurs widespread throughout southern Africa. This process is mainly characterized by the loss in vegetation cover, which leads to the occurrence of bare and denuded patches, increased soil erosion, changes in species composition as well as bush encroachment by indigenous and alien invasive plant species in savannah areas. Degradation of rangelands has drastically extended at an alarming rate during the last few decades with the main causes being overstocking, extended periods of drought, global climate change, overgrazing and general mismanagement of the land. Many researchers, however, feel that rangeland degradation is mainly caused by a combination of changes in land use practices and climate variability. Land users have, however, been applying a variety of technologies over the years in order to restore affected rangelands and mitigate the effect of degradation. These technologies include passive and active intervention methods, aimed at restoring bare and denuded areas and controlling indigenous bush and alien plant species encroachment. Bush control can be carried out by applying different technologies, involving chemical, mechanical, manual or biological control. The focal point of this study is on bush encroachment, the factors causing the problem, the possible ways of controlling this phenomenon and lastly the incorporation of such information into a user-friendly Decision Support System (DSS). The Decision Support System comprises of two databases as well as a related expert system. Bush encroachment is a matter of great concern in most southern African countries. This study therefore mainly included data from Namibia and to a lesser extent, South Africa, as the main study areas, seeing that this form of degradation greatly influences the biodiversity of rangelands in both these countries. The Namibia Agricultural Union (NLU) identified the need for the development of a user-friendly Decision Support System, in which case studies concerning the different bush control technologies could be stored in a database. Restoration technologies, regarding the occurrence of bare and denuded areas, that have been applied by the land users over a period of time and in a specific environment in the past, have been captured in a computerized database and expert system, serving as a Decision Support System (DSS) and user-friendly consulting tool in a similar study, carried out by Mr. Van der Merwe (1997). This DSS was based on CBR (Case Based Reasoning) methodologies by which a number of case studies, that have previously been stored in the database, can be searched by means of an expert system approach to advise the land user concerning the most appropriate solution (action) to similar degradation problems. The DSS developed by Mr. Van der Merwe was never published or made accessible to the land user in a format that could be consulted by either CD-ROM or the internet. Seeing that the NLU identified the need for a similar DSS containing bush control technologies, it was decided to incorporate both these databases into a single DSS, concerning bush control as well as the restoration of bare and denuded patches. The newly converted DSS is currently known as Eco Restore and consists on two databases: Grass Expert, which focuses on technologies to reclaim degraded rangelands, and Bush Expert, which is more focused on the control of bush encroachment and combating of alien invasives. As mentioned, this study focussed on the development of the Bush Expert database and will therefore only include results, discussions and conclusions of these case studies. The case studies in the Bush Expert database consist of results obtained by means of a questionnaire completed by the land user, in collaboration with the agricultural extension officer, as well as a quantitative vegetation assessment, to determine the success rate of the applied technology. The Bush Expert questionnaire, comprises of questions concerning personal information of the land user (e.g. location of the farm), the situation on the farm before bush control was applied (e.g. information on the environmental factors, such as density of problem trees), as well as the type of control technology applied and the situation of the rangeland after control (e.g. establishment of the herbaceous species). The quantitative vegetation assessments involved the sampling of the woody and herbaceous components in the area where a specific control technology was applied. The density and height classes of the woody component were determined by means of the belt-transect method. By using the descending-point method, the herbaceous component was surveyed to determine the abundance/frequency of the annual and perennial grass species. In order to increase the success of any restoration project, it is important to take the existing indigenous knowledge of local land users, concerning the problem of degradation and mitigation thereof, into consideration. By doing so, the local people and communities have greater control and responsibility over their resources and are able to command a greater range and level of resourcefulness. Taking indigenous knowledge into consideration finally enables the local land users to actively participate in and influence higher-level decision-making processes by which they are affected. A total of 175 case studies in Namibia and nine case studies in South Africa were surveyed. The Namibian case studies were surveyed in the central and northern arid and semi-arid regions, and South African case studies in a limited location within the Limpopo Province. Only 100 of the Namibian case studies have thus far been incorporated into the Bush Expert database. Multivariate data analyses techniques, analysis of variance and correlation analyses were used to analyse the data obtained from the questionnaires and quantitative vegetation surveys. Results were represented in the form of histogrammes, tables and multivariate analysis ordinations. From the results obtained for the Bush Expert database, it was clear that chemical control technologies were most often applied in Namibian and South African case studies (61%). The herbicides most commonly applied as chemical control technology in Namibia included Grazer (20%) and Savana (15%), whilst in South Africa these included Access (33.3%) and Tordon Super (33.3%). Herbicides were mostly applied by means of aerial application (46%) methods in Namibia and as cut-stump treatment (55.5%) by means of knapsack spraying or with a brush in South Africa. The dominant woody species causing bush encroachment problems in Namibia were found to be Acacia mellifera, Acacia reficiens and Dichrostachys cinerea, whereas in South Africa these species included Dichrostachys cinerea, Acacia erubescens and Acacia karroo. The wood of the controlled problem species (dead woody material) is mostly not utilized after control, but rather left on the land to disintegrate and thus contribute to the organic material content in the soil. Dead branches are also used for brush packing, which forms and ideal micro-climate for the germination and establishment of grass seeds, which serves as an erosion control medium and protects grass seedlings against grazing impacts. Some land users do however produce charcoal from certain controlled woody species, in order to recover some of the input costs of bush control. The majority of the case study sites (68%) in Namibia occurred within the 300-450 mm short- and long-term rainfall zones and in South Africa the majority of case study sites occurred within the short-term rainfall zone of 550-600 mm (66.6%) and 400-500 mm long-term rainfall zone (55.5%). Case studies where chemical and manual bush control technologies were applied indicated the highest success rates after control (81.7% and 75.2% respectively). Success rate as an entity was greatly influenced by the type of control technology applied, the density of the problem woody species after bush control as well as environmental variables such as rainfall and soil clay percentage. No definite trend could be determined concerning the application of a specific bush control technology and a certain problem species. Land users tend to apply a chosen control technology, according to the resources available, such as labour, mechanical implements and finances. The only positive correlation between control technologies and the type of problem species could be found regarding Dichrostachys cinerea. This species was mainly chemically controlled by means of the application of certain herbicides. The most important lesson to be learnt from the surveys completed in the two countries is that it is an absolute necessity to apply a proper after-care programme as a management practice following the initial control of problem woody species. The implementation of after-care determines the final success rate of any applied bush control technology as a restoration practice within a rangeland. Only 11% of the case studies surveyed for Namibia and South Africa indicated the implementation of an after-care programme, which usually involved biological control (e.g. browsing by boer goats or the use of controlled or accidental natural veld fires). The EcoRestore Decision Support System is currently available as an online webversion (www.puk.ac.za/EcoRestore), as well as a CD-ROM version. The CD-ROM version is available in a package containing the CD and user's manual. An example of the package is included in this dissertation. In consulting the databases through question-and-answer procedures, the best action will be proposed to the land user for future rangeland restoration, either the reclamation of denuded areas or the control of bush encroachment. Since the case studies are based on past and existing experiences and research, the land user will have an indication of the expected outcome, should the same advised technology be applied. The EcoRestore DSS does not only offer a consulting tool for extension workers and technicians, but also creates networking and participation between land users and researchers, both locally and between neighbouring countries. The DSS is linked to other national and international websites and databases, to offer users a wider range of information and technologies with regard to agricultural and conservation practices. Better awareness is created amongst land users concerning the problem of rangeland degradation, which might encourage closer monitoring of the degradation and mitigation processes. The EcoRestore DSS was developed in such a way for it to be as user-friendly as possible, in order to reach as many parties involved in current or future restoration programmes. This study involved the development of the first version of the DSS (Version 1.0) and is thus only the prototype system. It is proposed that the Bush Expert database of the EcoRestore DSS, will be expanded in future and additional bush control case studies from other Southern African countries will be included. The addition of such case studies will ultimately increase the effectivity of this DSS.

Masters