The coastline of the Canterbury Bight between Timaru and the southern end of Kaitorete Barrier is in long term erosion. Paradoxically, alluvial cliffs and several large, braided rivers drain from the rapidly eroding Southern Alps into the Canterbury Bight, transporting indurated sandstone of the Torlesse Supergroup, or greywacke, to the coast. Catchment specific sediment yields average 1856 ± 261 t km-2yr-1 compared with the world average of 182 t km-2yr-1 and are among the highest known specific sediment yields in the world. The large sediment loads from these Canterbury rivers are comprised predominantly of fine sediments as suspended load (more than 90%), with only a small proportion of coarse bedload (less than 10%), which is reflected in the nature of the mixed sand and gravel beaches. The Canterbury coastline is apparently abundantly supplied with sediment, which is transported northward forming Kaitorete Barrier. However, since the 1950's there has been very little accumulation of sediment against Banks Peninsula at the downdrift end. Previous research suggested that sediment is lost from the beaches due to abrasion.

Abrasion has been used as both a generic term and to describe a specific process of mechanical reduction. However, it is not the only process that causes sediment particle change. For this reason, the term reduction is preferred to describe the overall change to sediment particles, and abrasion is simply one process of reduction. Reduction includes both physical and chemical processes operating concurrently.

This thesis examines reduction and its impact on the mixed sand and gravel beaches of the Canterbury Bight. To do this, sediments were collected from 55 field sites along the Canterbury Bight and three series of tumbler experiments were carried out in the laboratory using a concrete mixer. Series One, identified the significant variables for the study of reduction. Other studies have considered sediment particle size a significant indicator of the rate of reduction, however, results showed that larger sizes did not always have the higher reduction rates. These experiments identified other areas that required further investigation, including sediment lithology due to the variations in reduction rates between sites, shape and variations in the quality of sediment at the coast and also suggested chemical weathering was important.

Series Two investigated the significance of textural mix to reduction, and results showed that textural mix was a more important indicator of loss than sediment size alone. Other influences identified as significant to the reduction of sediments included lithology, where variations of sediment particles provided some explanation for variations in reduction rates between sites. Shape was investigated, but results were inconclusive as to the significance of this variable to reduction. The dominance of shapes on Canterbury Bight beaches also showed that, unlike other mixed sediment beaches, there was no zonation across the beach profile.

Series Three investigated the significance of chemical weathering. Weathered and unweathered material was compared to establish the significance of the quality of sediments to reduction. The depth of weathering rinds was found to be a significant indicator of the quality of sediments, where weathered material showed greater losses than unweathered material when tumbled.

Having identified the significant variables for reduction, it was possible to develop numerical models that allowed for the application of the laboratory results in the field. A textural mix model was developed based on the textural mix of sediments. Results confirmed a high variability in reduction rates along the Canterbury Bight.

A sediment transport model, calibrated to 20 years of hindcast data, was developed to establish the distance travelled by sediment within the swash zone of these mixed sand and gravel beaches. A tracer experiment made it possible to establish the actual rate of sediment transport based on the movement of the faster pebbles in a sample.

By combining both the textural mix model and the sediment transport model it was possible to develop a sediment displacement model. This is a predictive model and provides an estimate of the amount of time it would take for the displacement of sediment from one section of beach to another. It works on the premise that what is not lost to reduction is moved by longshore sediment transport.

Finally, the thesis suggested that there was a paradox, with large braided rivers and sediments from the alluvial cliffs providing sufficient sediment to renourish the coast, yet there is very little accumulation of sediment against Banks Peninsula. So why is the coast from Timaru to Taumutu eroding? The answer is sediment displacement, where both the processes of reduction and longshore sediment transport are moving and removing the sediments from the coastal system.