Show simple item record

dc.contributor.advisorJewitt, Graham Paul Wyndham.
dc.creatorNamugize, Jean Nepomuscene.
dc.date.accessioned2019-05-07T09:03:40Z
dc.date.available2019-05-07T09:03:40Z
dc.date.created2017
dc.date.issued2017
dc.identifier.urihttps://researchspace.ukzn.ac.za/handle/10413/16271
dc.descriptionDoctor of Philosophy in Hydrology University of KwaZulu-Natal. Pietermaritzburg, 2017.en_US
dc.description.abstractChanges of land use and land cover are important drivers of the quality of water reaching a waterbody. These changes affect the catchment and modify the chemical composition of the atmosphere, and thus altering the cycle of nutrients and the flux of energy. With current developments in Geographic Information Systems (GIS) techniques, hydrological modelling and statistical analyses, one or a combination of many methods can be used to assess the relationships between land use and land cover (LULC) classes and water quality variables. However, all these approaches are reliant on the collection of field measurements, LULC data and water sampling. Typically funding for such long-term information is not generally available in Africa. A three-year study involving analysis of historical data, field work and desktop investigations was conducted in the upper reaches of the uMngeni Catchment (1653 km2), South Africa, to assess the spatial and temporal variation of land use and land cover and its influence on the flux of water, nutrients (nitrogen and phosphorus) and Escherichia coli (E. coli) in the catchment. This involved the analysis of historical land use and land cover information (1994, 2000, 2008 and 2011), analysis and processing of historical datasets of E. coli, electrical conductivity, ammonium, nitrate, soluble reactive phosphorus (SRP), total phosphorus (TP), total suspended solids (TSS), temperature and turbidity. A water quality index based on a long-term data base of water quality emanating from existing monitoring programmes was assessed. In addition, stations were established for river sampling (14) and collection of bulk atmospheric deposition (3) of ammonium, nitrates, SRP and TP, in the Midmar Dam catchment (927 km2). These were consolidated with the application and testing of the Hydrological Predictions for the Environment (HYPE) model in the catchment, in simulating streamflow, transport and dynamic of inorganic nitrogen and total phosphorus, resulting from LULC changes. Results showed that the natural vegetation declined by 17% between 1994 and 2011, coinciding with an increase in cultivated, urban/built-up and degraded lands by 6%, 4.5% and 3%, respectively. This resulted in high variability in the concentrations of water quality parameters, but Midmar and Albert Falls Dams retain over 20% of nutrients and sediment and approximately 85% of E. coli. It was concluded that these dramatic changes in LULC directly affect the chemical composition of water in the catchment. However, these linkages are complex, site-specific and vary from one sub-catchment to another and decision-making regarding water resources management in the catchment must recognise this. The level of E. coli in water is a major issue for human contact during recreational activities in the entire study area. Higher concentrations of E. coli, ammonium, nitrates, SRP and TP were attributed to the poor or lack of sanitation facilities in the informal settlements, dysfunctional sewage systems, effluent discharged from wastewater works, expansion of agricultural activities, as well as a runoff from livestock farming and urban areas. Moreover, water quality in the catchment ranged between “marginal” and “fair”, predominantly “marginal” in 90% of the sites and completely poorer in the Mthinzima Stream, an important tributary of Midmar Dam. A declining monitoring frequency and resultant poorly reporting of water quality in the catchment, led to a recommendation for the establishment of automatic or event-based samplers, which should provide the optimum information on nutrient loadings to the waterbodies. Bulk atmospheric deposition and river inflows into the Midmar Dam studies were conducted under severe drought conditions. Higher concentrations of NH4, NO3 and TP in precipitation samples than those of rivers were found because of the high retention of nutrients in the landscape. In terms of loading, the bulk atmospheric deposition provided significant quantities of NH4, while TP, SRP and nitrates were predominantly from river flows. Specific loads of DIN (nitrate + ammonium) and TP in the catchment were slightly higher that the previously reported values for the catchment and are comparable to the other human-disturbed catchments of the world. HYPE model has successfully simulated streamflow (1961-1999), DIN and TP (1989-1999). For simulations of streamflow NSE values = 0.7 in four out of the nine sites (at a monthly time-step) and NSE > 0 in eight out of nine sites (at a daily time-step). Major floods and drought events were represented very well in the model, with a general over-simulation of baseflow events. The water balance was captured well at calibration sites with over-simulation of streamflow on the Lions River (PBIAS=28%) and their under-simulation in outlet sub-catchments (PBIAS < 0). This is ascribed to the simplification of some processes in the model i.e. evapotranspiration, water release, water abstraction and inter-basin transfer. There has been good fit between the simulations and observations of TP and streamflow with a lagging of the observed values. However, mismatches were noted for DIN. Evaluation of seasonal distribution of DIN suggested that denitrification, crop uptake of DIN and dilution were intensive during the period of rainfall and high temperatures in the catchment, while TP was highly mobilised during rainfall events, due to its strong binding with the soil. The information from this study highlighted the current state of LULC changes, the sub-catchments with the potentiality to export high levels of DIN and TP, the complexity of the relationship between LULC-water quality, the gaps in existing data collection programmes, the catchment responses to LULC changes and the usefulness of hydrological models which may apply beyond the upper reaches of the uMngeni Catchment.en_US
dc.language.isoenen_US
dc.subject.otherLand use land cover.en_US
dc.subject.otherWater quality index.en_US
dc.subject.otherBulk atmospheric deposition.en_US
dc.subject.otherUmngemi catchment.en_US
dc.subject.otherHYPE model.en_US
dc.titleEffects of land use and land cover changes on water quality of the upper Umngeni River, KwaZulu-Natal province, South Africa.en_US
dc.typeThesisen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record