Sediment entering streams and rivers is a global concern for water quality. Suspended sediment can reduce visual clarity, light penetration, and amenity values of water. Transported sediment can also smother benthic fauna and flora and increase the infilling of receiving environments such as lakes and estuaries. In New Zealand catchments, there are three primary sources of erosion i) hillslopes through sheetwash and rill erosion ii) stream bank erosion and iii) mass wasting of hillslopes. There has been little research conducted on the relative contribution of sediment sources to fine sediment within New Zealand. However, some limited research in New Zealand, and research from other countries has identified stream bank erosion as a major source of fine sediment. The aim of this thesis was to identify the contribution of various erosion sources to sediment in streams. This study tested the hypothesis that stream bank erosion would be dominant in a pasture catchment, and hillslope erosion would be dominant in a native catchment. This study used paired, adjacent headwater catchments with similar sizes, slopes, soils, and underlying geology but contrasting land uses. The Mangaotama catchment was predominantly pastoral land use, with some areas of pine plantation, while the Whakakai catchment had regenerated and pristine indigenous forest. Qualitative field surveys were conducted in the Mangaotama catchment (total length 1.3 km with 0.9 km under pastoral land use and 0.4 km under pine plantation) and Whakakai catchment (0.9 km) to assess stream and riparian conditions including stream bank erosion. Stream banks in the Mangaotama catchment were primarily stable (pasture section (left: 64%, right: 65%); pine section (left: 91%, right: 85%)) and the Whakakai catchment was generally stable (left: 84%, right: 86%). Stock damage dominated stream bank erosion in the pastoral section of the Mangaotama catchment (left: 23%, right: 19%), while stream bank undercutting dominated the pine section of the Mangaotama (left: 9%, right: 13%) and the Whakakai catchment (left: 15% and right: 13%). Sediment fingerprinting used radionuclides (¹³⁷Cs, ²¹⁰Pbₑₓ, ²²⁶Ra and ²²⁸Ra), to determine the relative contribution of stream bank erosion and hillslope erosion in suspended sediment samples collected during high flow events. The concentrations of all radionuclides were higher in the hillslopes (P <0.001) ((mean ¹³⁷Cs: 5.8 Bq kg⁻¹, std: 2.4 Bq kg⁻¹); (mean ²¹⁰Pbex: 32.5 Bq kg⁻¹, std: 16.0 Bq kg⁻¹); (mean ²²⁶Ra: 24.9 Bq kg⁻¹, std: 2.1 Bq kg⁻¹); (mean ²²⁸Ra: 35.3 Bq kg⁻¹, std: 4.0 Bq kg⁻¹)) of both the Mangaotama and Whakakai catchments than in the streambanks ((mean ¹³⁷Cs: 0.9 Bq kg⁻¹, std: 0.3 Bq kg⁻¹); (mean ²¹⁰Pbex: 8.4 Bq kg-1, std: 4.6 Bq kg⁻¹); (mean ²²⁶Ra: 28.8 Bq kg⁻¹, std: 1.7 Bq kg⁻¹); (mean ²²⁸Ra: 47.2 Bq kg⁻¹, std: 2.2 Bq kg⁻¹)), clearly distinguishing the sources of sediment. A numerical mixing model determined that stream bank erosion was the dominant source of sediment (90%) within both catchments, with the remaining 10% originating from hillslope erosion. Simultaneously recorded continuous turbidity and discharge (Q) data were analysed for the previous year (2017). Six flood events were analysed in the Mangaotama and Whakakai catchments (three large events Q > 1,000 l s⁻¹ and three smaller events (Mangaotama: Q between 127 l s⁻¹ to 606 l s⁻¹, Whakakai: Q between 219 l s⁻¹ and 1398 l s⁻¹). Turbidity was greater in the Mangaotama catchment for four events (ranging from 237 NTU to 3,402 NTU) and greater in the Whakakai catchment for two events (1,293 NTU and 1,435 NTU). Strong clockwise hysteresis dominated (5 out of 6) of the flood events in the Mangaotama catchment, suggesting that stream bank erosion was the dominant source of sediment. Weak anti-clockwise hysteresis dominated the Whakakai catchment (5 out of 6) flood events, indicating sources further away from the sample point. Anti-clockwise hysteresis is primarily associated with hillslope erosion, however, in this study, it was most likely attributed to stream bank erosion sources in the upper reaches of the Whakakai catchment. Terrestrial photogrammetry was tested, as it is potentially an effective, simple and cost-effective tool to monitor stream bank erosion over time. Further, a baseline was established which could be re-surveyed though it is recommended that further baseline sites be established to provide improved rigour. Stream bank erosion was the greatest contributor (90%) to stream sediment loads in both the Mangaotama and Whakakai catchments, with the remaining 10% the stream sediment in both catchments being derived from hillslope erosion. Thus, the hypothesis for this study was rejected. This study demonstrates that stream bank erosion is a natural process and can dominate the supply of sediment to river systems even in relatively unimpacted catchments. Catchment managers need to take this into account when considering catchment rehabilitation projects aimed at reducing the delivery of sediment to streams.