DOGAMI SPECIAL Paper Series

[Go back] | Publications Center | DOGAMI Home

Publication Preview

Special Paper 51, Columbia River tsunami modeling: toward improved maritime planning response, by Jonathan C. Allan, Joseph Zhang, Fletcher E. O'Brien, and Laura L. Gabel.

PUBLICATION CONTENTS

Report (77 p., 24.7 MB PDF)

Animations:

• RUN06a_animation_200m.avi (1 GB)
  Current vectors showing the progression of an XXL1 tsunami occurring on a flood tide and with average river flows (Run06) into the Columbia River estuary (example image below from animation shows progression at t = 1 hour)
  
  
  
  
  
• RUN06d-pmel01_animation_200m.avi (1.35 GB)
  Current vectors showing the progression of an AKMax tsunami occurring on a flood tide and with average river flows (Run06-pmel01) into the Columbia River estuary (example image below from animation shows progression at t = 1 hour)
  
  

Related publication:
DOGAMI Maritime Tsunami Response Guidance (MTRG) - Port of Astoria, Clatsop County, Oregon and Lower Columbia River Estuary
—Guidance for operators of small watercraft and commercial fishing vessels in the event of a distant tsunami.

EXECUTIVE SUMMARY

Recent tsunamis affecting the West Coast of the United States have resulted in significant damage to ports and harbors, as well as to recreational and commercial vessels attempting to escape the tsunami. This study evaluates new tsunami modeling results completed for both distant and local tsunamis in the Columbia River system. The overall goal is to examine the interaction of tsunamis with dynamic tides (as opposed to modeling using a fixed tidal elevation such as mean higher high water), different riverine flow regimes, and friction to provide an improved understanding of tsunami effects on maritime traffic operating offshore the mouth of the Columbia River and within the estuary, as well as upriver toward the ports of Longview, Vancouver, and Portland. This was accomplished by evaluating a suite of tsunami simulations (35 in total) for the Columbia River focused on two distant earthquake scenarios: the 1964 Anchorage, Alaska (AK64) earthquake and a maximum considered eastern Aleutian Island (AKMax) earthquake, and two local Cascadia subduction zone (CSZ) scenarios: Large1 (L1) and Extra-extra-large1 (XXL1).
The Alaska 1964 scenario provides an excellent reference point for the potential maritime effects of the most extreme distant event to strike the Oregon coast in the past century. Our model simulations reveal the following:

• The simulation indicates that the tsunami arrived on the Oregon coast ~ 4 hours after the earthquake and occurred during a spring flood tide;
• Modeled maximum tsunami water levels reached 1 to 2 m (3–7 ft) along the open coast, while water levels at the mouth of the Columbia River (MCR), in Baker Bay, Warrenton, and Astoria ranged from 0.6 to 1.2 m (2–4 ft) high.
• Within the estuary, the simulated tsunami waves were found to rapidly decrease in height becoming negligible east of Tongue Point.
• Moderately strong currents of ~ 1.5–3 m/s (3–6 knots) are concentrated within the MCR, particularly the navigation channel, before dropping below 1.5 m/s (3 knots), the threshold below which damage within the ports and harbors is unlikely to occur.
• Damage for this event was minor and was limited to the community of Ilwaco, Washington, and included flooded streets and damage to pilings at a cannery.

For the maximum considered AKMax distant tsunami, our analyses indicate:

• The initial wave arrival at the MCR occurs ~3 hours 38 minutes after the start of the earthquake.
• The tsunami takes an additional 19 minutes to travel into Baker Bay, where it eventually inundates the lower areas of Ilwaco; the AKMax tsunami takes an additional 27 minutes to reach Warrenton, 39 minutes to reach Tongue Point, and arrives ~91 minutes later at Wauna; total time to Wauna is 5 hours 9 minutes.
• The tsunami is detectable as far up stream as St. Helens, although the amount of energy in the tsunami is effectively negligible such that the event can largely be ignored upriver of Longview.
• The simulations demonstrated significant along-coast and in-water variability in the maximum tsunami water levels and currents. The most dangerous conditions will occur at the MCR, within Baker Bay and Young's Bay, and by Tongue Point. Strong currents exceeding 4.5 m/s (9 knots) will dominate the estuary west of Chinook and Hammond, and some 8 km (5 mi) seaward of the MCR.

With respect to maritime evacuation for a maximum considered distant tsunami, we note the following:

• We recommend vessels seaward of the MCR evacuate to depths greater than 46 m (25 fathoms/150 ft). Dangerous currents (> 2.6 m/s [5 knots]) are expected to occur at depths shallower than 27 m (15 fathoms/90 ft) in this scenario.
• Offshore maritime evacuation may be feasible for some vessels operating out of the ports of Ilwaco, Chinook, Hammond, and Warrenton.
• Vessels east of Warrenton may choose to evacuate upriver; seaward evacuation for vessels upriver of Warrenton is not advised because those vessels might be transiting the mouth at around the time at which the tsunami waves arrive.
• Large ships moored between Tongue Point and Rice Island, should probably consider deploying additional drag anchors to further safeguard their vessels; evacuation upriver is probably not warranted for these ships. No additional measures are required for vessels operating upstream of Fitzpatrick Island other than to ensure the presence of sufficient mooring lines.

Results from modeling a maximum considered locally generated CSZ tsunami revealed the following:

• The tsunami reaches the MCR in as little as 7 minutes. The tsunami takes an additional 18 minutes to travel northward up into Baker Bay, where tsunami waves inundate Ilwaco 25 minutes after the earthquake; the XXL1 local tsunami arrives at 31 minutes at Warrenton, 46 minutes at Tongue Point, and at ~99 minutes at Wauna. By the time the tsunami reaches Tongue Point, the tsunami is travelling at a speed of ~37.8 km/hr (23.5 mi/hr).
• The CSZ tsunami is detectable upriver as far as the confluence of the Columbia and Willamette Rivers and arrives 4 hours and 40 minutes after the start of earthquake shaking.
• The CSZ tsunami contains little energy upriver of St. Helens and can be effectively ignored. Accordingly, a maximum-considered CSZ tsunami will not impact the ports of Portland or Vancouver and no additional measures will be needed to safeguard vessels at these locations, aside from having to deal with the earthquake shaking. The latter is likely to cause navigation hazards where water-saturated sediment liquefies and causes navigation channel sides, levees, and port foundations to fail.
• At the open coast, maximum water levels exceeding 18 m (60 ft) will be observed at Seaside, decreasing to 10–12 m (~33 to 40 ft) along the Clatsop Plains; water levels in the MCR will range from 6 to 12 m (20–40 ft) (relative to mean higher high water [MHHW]), decreasing to ~ 5 m (16 ft) by Sand Island.
• Extreme currents exceeding 6.1 m/s (12 knots) will be observed across much of the lower estuary, west of Astoria. These currents will be enhanced during ebb tide conditions, which could contribute to localized amplification of tsunami and wind waves at the MCR. In this scenario damage will probably be devastating for all ports and harbors in the lower estuary.
• Tsunami current velocities fall significantly upriver of ~Cathlamet and fall below the 1.5 m/s (3 knot) port damage threshold by the time the tsunami reaches Wauna. A transition area between Fitzpatrick Island and Cathlamet is characterized by current velocities in the 1.5 to 2 m/s range (3–4 knots).

With respect to maritime evacuation for a maximum considered CSZ tsunami, we note the following:

• There is insufficient time for mariners in ports to respond to this event other than to evacuate by foot to high ground.
• For vessels operating west of the MCR, the most effective strategy is to immediately evacuate toward deeper water and, accordingly, toward decreasing tsunami-generated currents. We recommend that vessels evacuate directly toward Astoria Canyon to depths greater than 146 m (80 fathoms).
• Mariners should prepare to remain offshore for potentially days as the MCR is unlikely to be navigable following a CSZ earthquake and tsunami. As a result, plans to evacuate to potentially safe ports located south of Cape Mendocino on the California coast should be developed.
• Limited options are available for vessels caught out in the estuary west of Astoria.
• For vessels located east of Astoria, evacuation upriver toward Fitzpatrick Island may be an option. However, this is likely to be limited to smaller, faster boats.
• For vessels located upriver of Wauna, we recommend that ship operators attach additional mooring ropes to secure their boats or move vessels into the navigation channel.
• Because the tsunami is negligible by the time it reaches St Helens, no additional maritime safety requirements are needed upriver of this point.

Modeling undertaken using dynamic tides and varying river flows has yielded some useful insights, when compared with static models undertaken at MHHW and with no flows. These include:

• The predicted maximum velocities exhibit more local extrema along the coast and within the lower estuary, especially near the mouth where the interaction is found to be strongest due to powerful currents and shoaling of tsunami waves.
• The conventional wisdom is that tsunamis arriving with a flood spring tide are usually more damaging. This is largely true in the deep channel around Tongue Point. However, the situation becomes very complex in the shallow waters of the upper estuary, where tsunamis arriving at ebb and flood slack were found to be considerably more energetic.
• Tsunamis arriving at flood slack resulted in the greatest upriver penetration of the tsunami, with strong currents in the shallows east of Tongue Point and as far upriver as Wauna.
• The violent collision between tidal and tsunami currents at the MCR makes the ebb scenarios especially dangerous for ships of all sizes. Our modeling confirms that tsunamis arriving during an ebb phase produce considerably stronger currents when compared with the flood scenario.
• Model results that included varying the river discharge indicate generally little difference in the current velocities associated with low, average, or high flow scenarios in the Columbia River.
• Effects associated with low river flows are largely confined to areas offshore the MCR and along the open coast, where slightly stronger tsunami currents predominate. In this scenario, we find that the tsunami water levels in the estuary are generally reduced by ~ −0.25 m (−0.8 ft).
• The high river flow scenario yielded even stronger currents at the MCR and contributes ~0.2 m (0.7 ft) to the maximum tsunami water levels in the estuary.
• Changes caused by varying river flows on the modeled tsunami currents were found to be negligible in the lower estuary (< ±0.2 m/s [< ±0.4 knots]).
• The above observations apply to both the distant and local tsunamis.