Noxon Rapids Dam Sudden Seepage and Erratic Piezometer Behavior

Noxon Rapids Hydroelectric Development, owned by Avista Corporation, was completed in 1960. It is a 260 foot high concrete gravity dam with two earthfill embankments. The southwest embankment and reservoir overlie a lacustrian clayey silt layer underlain by a clean, pervious gravel (Noxon Gravel) and bedrock. Upon first filling, the partial blockage by bedrock of seepage through the Noxon Gravel to the tailwater below the southwest embankment created a relatively high hydraulic gradient and flow velocity. Also, the Noxon Gravel did not meet reverse filter criteria for the lacustrian layer. Consequently, sink holes developed upstream of the dam, causing the phreatic surface in the Noxon Gravel to rise substantially. The result was serious slope movements and seepage downstream of the southwest embankment. In response, from 1960 through 1969, Avista conducted extensive clay blanketing of the reservoir bottom to seal the leakage into the Noxon Gravel. This blanketing proved to be successful and the slope movement and seepage were eliminated.
In 2008, a non-turbid seep was discovered approximately 100 yards downstream of the southwest embankment. A weir was established at the seepage area and monitoring was initiated. In 2009, the quarterly water level measurement for well 701, located on the southwest embankment, dropped 10.9 feet. Historical changes between quarterly measurements are normally less than one foot.
The new seepage and piezometer changes raised concerns that the historic seepage issues at this location may be reinitiating. In response, Avista compiled historic information, including the locations and logs of original design borings, and performed a diving inspection of the area upstream of the southwest embankment. No indications of failure of the upstream clay blanket were found. A hydrogeological review was also conducted, which further refined the understanding of subsurface conditions at the dam.
Review of the available information and understanding of the subsurface strata indicated the seepage point elevation did not correlate with water levels observed in the nearby wells. Rather, the seepage was postulated to be related to the presence of shallow ground water in a surficial sand and gravel fill which is separated from the Noxon Gravel by an aquitard formed by the lacustrian clay. But why did this seepage suddenly appear? This paper will discuss Avista’s response to the seepage and erratic piezometer behavior to understand and address what was occurring in a logical progression that may be useful to other dam owners. 16 pp., various diagrams.