*3.4.2. West Coast of USA*

**Figure 6.** The location of the radars A087, A088, the tide gauge at Hakodate, and the offshore bathymetry.

**Figure 7.** Perpendicular band velocities from A088 and derived *q*‐factors. (a) Distance from radar: 0–2 km (blue), 2–4 km (red), and 4–6 km (black). (b) The *q*‐factor for range 0–6 km. (c) Distance from radar: 6–8 km (blue), 8–10 km (red),

and 10–12 km (black). (d) The *q*‐factor for range 6–12 km.

86 Tsunami

Radar spectra measured by 10 radars located along the US West Coast were analyzed to give band velocities and *q*‐factors. Arrival times were compared with those at local tide gauges. **Figure 8** shows radar and tide gauge locations and the offshore bathymetry. As the adjoin‐ ing continental shelf is narrow off California and Oregon, the tsunami is often detectable on‐ ly for close‐in ranges.

**Figure 8.** Locations of radars and tide gauges in California and Oregon. The offshore bathymetry indicates a narrow offshore continental shelf.

**Figure 9.** Perpendicular band velocities and derived *q*‐factors from YHS2. (a) Distance from radar: 2–4 km (blue), 4–6 km (red), and 6–8 km (black). (b) The *q*‐factor for range 2–8 km. (c) Distance from radar: 8–10 km (blue), 10–12 km (red), and 12–14 km (black). (d) The *q*‐factor for range 8–14 km.

We illustrate the tsunami detections with two examples of measured band velocities and derived *q*‐factors.

Our first example is the tsunami detection by the radar at YHS2, Oregon (transmit frequency 12 MHz).

**Figure 9** shows the band velocities and corresponding *q*‐factors.

**Figure 8.** Locations of radars and tide gauges in California and Oregon. The offshore bathymetry indicates a narrow

**Figure 9.** Perpendicular band velocities and derived *q*‐factors from YHS2. (a) Distance from radar: 2–4 km (blue), 4–6 km (red), and 6–8 km (black). (b) The *q*‐factor for range 2–8 km. (c) Distance from radar: 8–10 km (blue), 10–12 km

(red), and 12–14 km (black). (d) The *q*‐factor for range 8–14 km.

offshore continental shelf.

88 Tsunami

The correlation is evident between the velocities in different bands starting at about 3:45 pm Coordinated Universal Time (UTC), resulting in a sharp decrease in the *q*‐factor, which indicates the tsunami moving offshore, resulting in a decrease in water level. The neighboring South Beach tide gauge observed an initial water level increase due to the tsunami of just 0.3 m, which was inadequate to produce a radar detection. However, we note that in **Figure 9(a– c)** the band velocities show the typical correlation due to the tsunami just before the sharp decrease.

Our second example is the tsunami detection by the radar at ESTR in Southern California (transmit frequency 13 MHz). **Figure 10** shows the band velocities and *q*‐factors for ESTR.

**Figure 10.** Perpendicular band velocities and derived *q*‐factors from Radar ESTR. (a) Distance from radar: 2–4 km (blue), 4–6 km (red), and 6–8 km (black). (b) The *q*‐factor for range 2–8 km. (c) Distance from radar: 8–10 km (blue), 10– 12 km (red), and 12–14 km (black). (d) The *q*‐factor for range 8–14 km.

The observed background current velocities are quite variable for this site: it is the correlations between velocities in different bands that allow the tsunami to be detected by the patternrecognition algorithm described in Section 3.3.

The Japan tsunami took approximately 10 h to reach the US West Coast; it moved down the coast from north to south. Tsunami and tide gauge arrival times are compared in **Table 2**. Radar stations are listed in the order of the expected arrival of the tsunami.


**Table 2.** Comparison of tsunami arrival times from radars and neighboring tide/wave gauges.

**Table 2** shows that listed arrival times obtained from the radar *q*-factors reported are normally in the correct order; thus, it arrives in Southern California after it gets to Northern California and Oregon. Arrival times measured by the radars preceded those at neighboring tide gauges by an average of 15 min, due to both the "quadrature relation" between velocity and height (discussed in Section 4.1) and the tsunami propagation delay between the two observations. We expect to be able to quantify the propagation delay using simulated tsunami velocity patterns derived from the analytical model, as discussed later in Section 5.1.

We note from **Table 2** that the tsunami was detected even though water-level changes at neighboring tide gauges were not large, varying between 0.3 and 2 m.
