Joint 70th Rocky Mountain Annual Section / 114th Cordilleran Annual Section Meeting - 2018

Paper No. 71-3
Presentation Time: 8:30 AM-4:30 PM

SPECTRAL ANALYSIS OF SEISMIC NOISE FROM THE 2017 OROVILLE DAM CRISIS


DORN, Kathryn Margo, BRAUNMILLER, Jochen and THOMPSON, Glenn, University of South Florida, 4202 E. Fowler Avenue, Tampa, FL 33620

In early 2017, winter rains in northern California caused the Oroville Dam reservoir to become dangerously full; to accommodate this, the dam operators released water at higher-than-usual flow rates, exceeding 10,000 cubic feet per second for the first time since 2011. Shortly after the flow down the service spillway spiked above 50,000 cubic feet per second on February 7, a piece of the concrete-paved spillway tore away from its foundation, exposing the nearby concrete and underlying hillside to severe erosion.

Heavy rain caused water to overtop the emergency spillway crest on February 11, despite ongoing releases down the damaged service spillway. Strong headcutting erosion quickly progressed towards the emergency spillway crest and led to an evacuation order for over 180,000 people. The dam's operators released more water down the service spillway, eventually ceasing the overflow down the emergency spillway and preserving its integrity at the cost of further damage to the service spillway.

The rushing water and its bedload generated significant seismic noise at the nearby broadband station ORV, operated by the University of California Berkeley Seismological Laboratory. We obtained 100-Hz vertical-component ORV data from the Northern California Earthquake Data Center, along with hourly dam outflow and precipitation data from the California Data Exchange Center, and compared the dam outflow and meteorological data to the seismic spectrum and the per-minute average 0.5-to-40-Hz bandpass-filtered seismic amplitude during the first two months of 2017.

After excluding meteorologically noisy times from the seismic data, we found that seismic noise correlates with dam outflow rates for frequencies above 0.5 Hz, but the relationship is neither linear nor consistent. We then investigated temporal changes in seismic amplitude as a function of water flow rate; this type of hysteresis can indicate changes in bedload transport, such as when the water began to carry spillway and hillside debris. We observe slight counterclockwise hysteresis at most frequencies between 0.5 Hz and 40 Hz, possibly due to the shift from a stream of relatively clean water flowing down an intact spillway to a cascade of debris-laden water.