In the United States, the annual cost of on-site soil erosion problems such as soil and nutrient losses, and off-site soil erosion problems such as sedimentation of lakes and river, loss of navigable waterways, flooding and water quality impairment, has been estimated at 44 billion USD (Pimentel, 1995; Telles, 2011). While eroding sediment sources can either be from land or from stream/river systems, the erosion from streambanks can be quite significant, reaching up to 80% of sediment leaving a watershed (Simon et al 2002; Simon and Rinaldi 2006). Despite many decades of research one the erosion of cohesive soils by flowing water (fluvial erosion), this significant aspect of environmental sustainability and engineering is still poorly understood. While past studies have given invaluable insight into fluvial erosion, this process is still poorly understood. Therefore, the objective of this dissertation was to examine the relationship between time and erosion resistance of remolded cohesive soils, and to quantify and model the effects soil and water temperature on the fluvial erosion of cohesive soils

First, erosion tests were performed to investigate how soil erosion resistance develops over time using three natural soils and testing in a laboratory water channel. Results showed that the erosion rate of the soils decreased significantly over the time since the soils were wetted. This study indicates researchers need to report their sample preparation methods in detail, including the time between sample wetting and sample testing.

Second, erosion tests were performed at multiple soil and water temperatures. Results showed that increases in water temperature led to increased erosion rates while increases in soil temperature resulted in decreased erosion rate. When soil and water temperatures were equal, erosion results were not significantly different. Results also showed a linear relationship between erosion rate and the difference between soil and water temperatures, indicating erosion resistance decreased as heat energy was added to the soil.

Lastly, two common erosion models (the excess shear stress and the Wilson models) were evaluated, and were modified to account for soil and water temperature effects. Results showed that, compared to the original models, the modified models were better in predicting erosion rates. However, significant error between model predictions and measured erosion rates still existed.

Overall, these results improve the current state of knowledge of how erosion resistance of remolded cohesive soils evolves with time, showing the importance of this factor in the design of cohesive erosion experiments. Also, the results show that by accounting for thermal effects on erosion rate, the usability of erosion models can be improved in their use for erosion predictions in soil and water conservation and engineering practice.