Estimating sediment yield from the Swift Creek landslide, Whatcom County, Washington State

The amount of suspended sediment carried by streams in mountainous watersheds is an important factor in environmental and engineering planning, especially when the material happens to be of toxic nature. The Swift Creek watershed contains a deep-seated landslide composed of weathered serpentinite, which includes chrysotile (capable of asbestiform morphology), chlorite, illite, and hydrotalcite. The United States Environmental Protection Agency has determined that the asbestiform material contains particles of sufficient size and quantity that could be hazardous to human health. The suspended sediment load from Swift Creek is primarily influence by the steep, disturbed and unvegetated toe of the landslide which provides a large surface for overland flow directly into the stream creating temperamental conditions, as well as an effectively endless supply of sediment. The remainder of the watershed is heavily forested and consequently supplies relatively little sediment to the stream. I attempted to develop a means to estimate sediment yield from the landslide, and provide a consistent method to monitor the stream in spite of the flashy conditions. I used the Distributed Hydrology-Soil-Vegetation Model to create a continuous discharge dataset from point discharge measurements. I also used the Turbidity Threshold Sampling method to collect physical water samples drawn during specific changes in turbidity and used the turbidity data as a proxy for suspended sediment. I developed linear models based on discharge and turbidity to estimate an annual yield. Pacific Surveying and Engineering conducted a similar study three years after my data collection period and estimated a sediment yield that did not support my sediment yield estimates. The methods were slightly different as necessitated by the difference in quality of the various data types. As a result, I evaluated the differences between the methods in an effort to determine if the disparity between my estimate and Pacific Surveying and Engineering's estimate was caused by procedural differences. I included an analysis of timing between peak turbidity and precipitation and between peak turbidity and discharge. I found that the time between storms was important to the suspended sediment magnitudes. Future modeling efforts will need to incorporate this discharge-sediment hysteresis over the linear models in this research and by Pacific Surveying and Engineering. The United States Geological Survey collected turbidity and discharge data on the Sumas River for several seasons. Monitoring here in the future will likely be more effective than monitoring Swift Creek directly because of the rivers discharge stability. If direct monitoring of the Swift Creek is to be continued, the relationship between discharge and suspended sediment should be further developed rather than turbidity and suspended sediment due to measurement stability.