Wetland degradation, which includes deterioration in functional performance and erosion, is a problem around the world. This has engendered a quest for causes and attempts to prevent the problem or to rehabilitate wetlands already degraded or undergoing degradation. The Craigieburn wetland system in Mpumalanga Province, South Africa has undergone erosion due to two downstream discontinuous gullies that have drained and considerably reduced the size of the wetland system. Measurements from 1954 to 1997 aerial photographs showed that over 40 years, the upper gully migrated headward over a distance of 30 m, while the lower gully eroded 522 m headward raising the question as to what caused their erosion? Prior to this study, the predominant view was that human activities, namely poor land use management within the wetland system, increased human occupation, and overgrazing on the adjacent catchment that caused a reduction in vegetation cover, were responsible for this wetland erosion. Detailed field observation, aerial photograph interpretation, soil analyses for mineralogy, chemistry and particle size distribution, landscape mapping, dumpy level survey of the wetland valley and statistical analysis were undertaken to establish the relationships between gullying and possible contributing factors. Human impacts on wetland gully development between 1954 and 1997 were estimated using the number of individual homes, and total lengths of footpaths, animal tracks and dirt roads. Agricultural activities and the stocking rate of livestock were excluded due to the poor quality of aerial photographs and lack of historical records. Results of multiple regression correlating lengths of the two gullies (upper and lower gullies) and the sum of these human factors gave a high correlation (adjusted R² = 0.92 and 0.90, respectively) but a low significance (p = 0.18 and 0.21, respectively). However, time has played a significant role in the erosion of both the upper gully (R² = 0.82, p = 0.02) and the lower gully (R² = 0.98, p = 0.02) at Craigieburn. X-ray diffraction and X-ray fluorescence spectrometry of weathered parent materials showed that the area has undergone deep weathering, supplying sediments to the wetland valley through surface run-off. The accumulation of these sediments resulted in localized over-steepening of certain sections of the valley floor with raised gradients of 0.0336 and 0.0337 at the two headcuts relative to the upper and lower non-eroding sections with lower gradients. These localized steep sections increased flow velocity and stream power and therefore stream erosivity thus triggering gully erosion. In addition to localized areas of raised valley floor, results from multiple regression showed a significant relationship (p = 0.002) between areas of earthflow scars and gully length, especially at the lower gully, thus further suggesting that physical factors are largely responsible for gully erosion at Craigieburn. Long-term climate change has resulted in the formation of two terraces, an older, D1 (USU-760, 1.67 ± 0.89 ka) and a younger, D2 (USU 761, 0.32 ± 0.08 ka). The former probably eroded during the medieval warming around 1230 AD while the younger terrace, which likely formed during the last half of the Little Ice Age, has been eroding since the renewed warming thereafter. This erosion has been exacerbated by short-term periodic or seasonal climatic changes, especially episodic summer rainfall events, which have likely played a key role in the headward migration of the two gullies. The result has been shrinkage of the wetland system by about 15 m on both sides of the valley, leaving behind a greyish soil colour indicating wet and reducing conditions in the past. These, together with dried relict mottles left behind in the soil matrix at the margin of the shrunken wetland system suggest past seasonal fluctuation of the water table engendering the belief that the wetland system once extended beyond its present limit. The overwhelming contribution of these physical factors, in addition to the fact that the two gullies predate human occupation of the study area catchment and environs, strongly argues for their responsibility in gully initiation and development at Craigieburn. Human presence and activities, which only became evident in the catchment from the 1950s onwards, may be secondary contributory factors. This conclusion encourages a rethink of previous views that human occupation and activities are solely responsible for this wetland gully erosion at Craigieburn and provides a rationale for including physical processes and climate change as factors when investigating causes of wetland erosion elsewhere. Such an understanding should be used to inform any rehabilitation or conservation efforts that are related to wetland ecosystems. / Thesis (Ph.D.)-University of KwaZulu-Natal, Westville, 2011.