Observed and modeled tsunami current velocities in Humboldt Bay and Crescent City Harbor, northern California

A pilot project was initiated in 2009 in Humboldt Bay, about 370 kilometers (km) north of San Francisco, California, to measure the currents produced by tsunamis. Northern California is susceptible to both near- and far-field tsunamis and has a historic record of damaging events. Crescent City Harbor, located approximately 100 km north of Humboldt Bay, suffered US 20 million in damages from strong currents produced by the 2006 Kuril Islands tsunami and an additional US 20 million from the 2011 Japan tsunami. In order to better evaluate these currents in northern California, we deployed a Nortek Aquadopp 600kHz 2D Acoustic Doppler Current Profiler (ADCP) with a one-minute sampling interval in Humboldt Bay, near the existing National Oceanic and Atmospheric Administration (NOAA) National Ocean Service (NOS) tide gauge station. The instrument recorded the tsunamis produced by the Mw 8.8 Chile earthquake on February 27, 2010 and the Mw 9.0 Japan earthquake on March 11, 2011. Currents from the 2010 tsunami persisted in Humboldt Bay for at least 30 hours with peak amplitudes of about 0.3 meters per second (m/s). The 2011 tsunami signal lasted for over 86 hours with peak amplitude of 0.95 m/s. Strongest currents corresponded to the maximum change in water level as recorded on the NOAA NOS tide gauge, and occurred 90 minutes after the initial wave arrival. No damage was observed in Humboldt Bay for either event. In Crescent City, currents for the first three and a half hours of the 2011 Japan tsunami were estimated using security camera video footage from the Harbor Master building across from the entrance to the small boat basin, approximately 70 meters away from the NOAA NOS tide gauge station. The largest amplitude tide gauge water-level oscillations and most of the damage occurred within this time window. The currents reached a velocity of approximately 4.5 m/s and six cycles exceeded 3 m/s during this period. Measured current velocities both in Humboldt Bay and in Crescent City were compared to calculated velocities from the Method of Splitting Tsunamis (MOST) numerical model. For Humboldt Bay, the 2010 model tsunami frequencies matched the actual values for the first two hours after the initial arrival however the amplitudes were underestimated by approximately 65%. MOST replicated the first four hours of the 2011 tsunami signal in Humboldt Bay quite well although the peak flood currents were underestimated by about 50%. MOST predicted attenuation of the signal after four hours but the actual signal persisted at a nearly constant level for more than 48 hours. In Crescent City, the model prediction of the 2011 frequency agreed quite well with the observed signal for the first two and a half hours after the initial arrival with a 50% underestimation of the peak amplitude. The results from this project demonstrate that ADCPs can effectively record tsunami currents for small to moderate events and can be used to calibrate and validate models (i.e. MOST) in order to better predict hazardous tsunami conditions and improve planned responses to protect lives and property, especially within harbors. An ADCP will be installed in Crescent City Harbor and four additional ADCPs are being deployed in Humboldt Bay during the fall of 2012.