Sediment load and distribution in the lower Skagit River, Skagit County, Washington


2016: Sediment load and distribution in the lower Skagit River, Skagit County, Washington

Christopher A. Curran, Eric E. Grossman, Mark C. Mastin, and Raegan L. Huffman

The Skagit River delivers about 40 percent of all fluvial sediment that enters Puget Sound, influencing flood hazards in the Skagit lowlands, critically important estuarine habitat in the delta, and some of the most diverse and productive agriculture in western Washington. A total of 175 measurements of suspended-sediment load, made routinely from 1974 to 1993, and sporadically from 2006 to 2009, were used to develop and evaluate regression models of sediment transport (also known as “sediment-rating curves”) for estimating suspended-sediment load as a function of river discharge. Using a flow-range model and 75 years of daily discharge record (acquired from 1941 to 2015), the mean annual suspended-sediment load for the Skagit River near Mount Vernon, Washington, was estimated to be 2.5 teragrams (Tg, where 1 Tg = 1 million metric tons). The seasonal model indicates that 74 percent of the total annual suspended‑sediment load is delivered to Puget Sound during the winter storm season (from October through March), but also indicates that discharge is a poor surrogate for suspended‑sediment concentration (SSC) during the summer low-flow season. Sediment-rating curves developed for different time periods revealed that the regression model slope of the SSC-discharge relation increased 66 percent between the periods of 1974–76 and 2006-09 when suspended-sediment samples were collected, implying that changes in sediment supply, channel hydraulics, and (or) basin hydrology occurred between the two time intervals. In the relatively wet water year 2007 (October 1, 2006, through September 30, 2007), an automated sampler was used to collect daily samples of suspended sediment from which an annual load of 4.5 Tg was calculated, dominated by a single large flood event that contributed 1.8 Tg, or 40 percent of the total. In comparison, the annual load calculated for water year 2007 using the preferred flow-range model was 4.8 Tg (+6.7 percent), in close agreement with the measured value.

Particle size affects sediment transport, fate and distribution across watersheds, and therefore is important for predicting how coastal environments, particularly deltas and beaches, will respond to changes in climate and sea-level. Particle-size analysis of winter storm samples indicated that about one-half of the suspended-sediment load consisted of fines (that is, silt– and clay-sized particles smaller than 0.0625 mm in diameter), and the remainder consisted of mostly fine– to medium-sized sand (0.0625–0.5 mm), whereas bedload during winter storm flows (about 1–3 percent of total sediment load) was predominantly composed of medium to coarse sand (0.25–1 mm). A continuous turbidity record from the Anacortes Water Treatment Plant (water years 1999–2013), used as a surrogate for the concentration of fines (R2 = 0.93, p = 4.2E-10, n = 17), confirms that about one-half of the mean annual suspended-sediment load is composed of fines.

The distribution of flow through the delta distributaries (that is, the channels into which the main stem splits as it approaches the delta) is dynamic, with twice as much flow through the North Fork of the Skagit River relative to the South Fork during low-flow conditions, and close to equal flows in the two channels during high-flow conditions. Turbidity, monitored at several locations in the lower river in spring 2009, was essentially uniform among sites, indicating that fines are well mixed in the lower Skagit River system (defined as the Skagit River and all its distributaries downstream of the Mount Vernon streamgage). A strong relation (R2 = 0.95, p = 3.2E-14, n = 21; linear regression) between the concentration of fines and turbidity measured at various locations in summer 2009 indicates that turbidity is an effective surrogate for the concentration of fines, independent of location in the river, under naturally well-mixed fluvial conditions. This relation is especially useful for monitoring suspended sediment in western Washington rivers that are seasonally dominated by glacier meltwater because glacial melting typically produces suspended-sediment concentrations that are not well correlated with discharge. These results provide a comprehensive set of tools to estimate sediment delivery and delta responses of interest to scientists and resource managers including decision-makers examining options for flood hazard mitigation, estuary restoration, and climate change adaptation.