Land use impact on water quality in two river systems in South Africa

Abstract

South Africa is a water-scarce country that is expected to face challenges with freshwater resources in the future. These concerns are not only related to the quantity of available water, but also to its quality. The country’s water is not always of an acceptable quality, and this is due to nutrient pollution caused by runoff from agricultural land, informal settlements, mines, industries and urban areas with inadequate water treatment systems. By analysing the land use of an area over time one can explore the link between different land use types and water quality within a region.

In order to explore change in land use over time, it is necessary to know what the land use was in history. In most cases, unless a study spans a large number of years, such data are not available. However, satellite imagery and remote sensing can assist us in ‘looking back in time’. Satellites such as Landsat have been recording imagery of the world since 1984, making it possible to select an image of the time and place of interest. Using remote sensing in a geographic information system (GIS) these images can be classified into land cover and land use maps. The area of each land use type can then be calculated. Long-term water quality data collected by the government was used to explore the relationship of water quality parameters such as electrical conductivity, phosphates, nitrates and ammonium nitrogen, with land use types.

Two study sites were selected. The Duiwenhoks catchment area is in the Western Cape, in a temperate, all-year-rainfall climate, which was an established farming region of the country. The Mzimkulu catchment area is in the KwaZulu-Natal Province, in a subtropical, summer-rainfall climate, which was more rural in nature with more potential for change. The study spanned from 1984 to 2009.

Results
It was expected that land use would not change significantly in the Duiwenhoks catchment due it being so established with little room for change, and this was also the finding of the study. In contrast, it was expected, and shown, that the Mzimkulu catchment would show a significant change in land use. In fact, all land use types changed significantly in Mzimkulu, with the exception of urban areas. Rural areas increased the most. Although the urban increase was not statistically significant, both the urban and rural areas increased between 1990 and 1997. This corresponds with the end of Apartheid in 1994, and the resultant movement of people. The increase in population brought with it a need for more cultivated crops, and the traditional clearing of ground around houses to keep it bare, resulting in an expected increase in cultivated and bare areas.

Some correlations between land use and water quality were found. Electrical conductivity was found to increase with an increase in fynbos, wooded/savannah, cultivated crops, sugar cane, plantations, urban areas and rural areas. This shows that these land use types increased the total of dissolved salts in the water. Phosphates increased with plantations, bare ground and cultivated crops. Phosphates are linked to effluent and fertilisers. Nitrates increased with an increase in cultivated crops, sugar cane, rural areas and bare ground. Fertilisers are the main reason for an increase in nitrates, which corresponds nicely with the increase in crop farming in both catchment areas. Ammonium nitrogen was found to increase with an increase in bare ground, cultivated crops and urban areas. This is normally linked to sewerage and fertilisers, which can be explained by the increase in urban areas and the likelihood that waste water is not being treated sufficiently, as well as the increase in crops.

Conclusion
Remote sensing is a useful method to classify and analyse land use. Land use was found to change as expected in the selected areas, and water quality parameters showed the expected correlations with land use type. Using this information one can predict how water quality is likely to change in the future by predicting how land use will change, for example using climate change models. Water is our lifeblood, and if we can predict potential problems in advance, we could help avert serious problems with water quality in the future.

Abstract

South Africa is a water-scarce country that is expected to face challenges regarding freshwater resources in the future. This study aimed to identify the land cover in two different catchments and to analyse the water quality in relation to land cover data. The study shows that GIS and remote sensing can be successfully used as a tool in classifying land cover over time in order to use the results in analyses that aim to answer questions about variables related to changes in land cover.

South Africa requires not only a certain quantity of water but also a certain quality. The water quality of the country is not always of an acceptable standard and nutrient pollution is one of the problems faced by water bodies. Nutrient pollution occurs when runoff rich in nutrients from agricultural land, informal settlements, mines, industries, and urban areas with inadequate water treatment systems, flows overland or via shallow groundwater into surface water systems. The water eventually flows into estuaries just before it reaches the sea, carrying its pollutants with it. Estuaries are ecologically and economically important to the country and need to be conserved and managed properly.

Temporal and spatial changes in land cover and water quality differed greatly between the temperate, well-developed Duiwenhoks catchment area and the subtropical, rural Mzimkulu catchment area. Land cover was not found to change significantly in the Duiwenhoks catchment between 1984 and 2009, but all land cover classes besides urban areas changed significantly over time in the Mzimkulu catchment, particularly a decrease in grassland and an increase in bare unvegetated ground, plantations and sugar cane.

A relationship was found to exist between land cover and water quality, where concentrations of phosphates and ammonium nitrogen increased with increases in land cover classes such as bare ground, cultivated grains, sugar cane, urban and rural areas and plantations. All of these classes are anthropogenic and suggest that with an increase in human activities water quality would worsen. The resulting correlations found between land cover and water quality can be used to infer changes in future scenarios.

• author: Ogden, Monika
• supervisor: Martin Berggren; Melanie Lück-Vogel
• organization: Dept of Physical Geography and Ecosystem Science
• alternative title: Land use impact on water quality in South Africa
• course: GISM01 20131
• year: 2013
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: remote sensing, physical geography and ecosystem analysis, GIS, land cover change, water quality.
• publication/series: LUMA-GIS Thesis
• report number: 23
• language: English
• additional info: External supervisor: Dr. Melanie Lück-Vogel, Coastal Systems Research Group, Natural Resources and the Environment at the Council for Scientific; and Industrial Research (CSIR), Stellenbosch, South Africa
• id: 3910312
• date added to LUP: 2013-06-27 10:23:44
• date last changed: 2013-06-27 10:29:43