Locked Megathrust Near a Coastal Hospital Site
A coastal city of 400,000 people sits above a subduction megathrust. GPS stations show the coast moving 38 mm/yr toward the trench, but the seaward 60 km of the plate boundary is locked from 15 to 40 km depth. The last great earthquake occurred in 1710; written records are sparse, but a coastal lake contains eight sandy tsunami layers in the past 4,000 years. Marine terraces at 6 m elevation record repeated coseismic uplift, and tide-gauge data show slow interseismic subsidence of about 1 mm/yr. A developer proposes a 120-bed hospital on a 4-m-high spit 300 m from the shoreline, citing the absence of a major earthquake in over 300 years as evidence that the fault is no longer dangerous. The city asks for a hazard assessment and siting advice before approving the permit.
1. Estimate the average recurrence interval from the geologic record and evaluate the developer’s “300 years without an event” argument.
Read the worked answer
Eight events in 4,000 years gives a rough average of 500 years between events, though the record may miss smaller events and the intervals are probably irregular. The last event was roughly 310 years ago, about 62% of the average, so the elapsed time does not indicate that the fault is dead; on a locked megathrust, strain continues to accumulate. The developer’s argument confuses “no recent event” with “no hazard”-it is the opposite: a long quiet period on a locked fault increases the accumulated strain. The average also carries large uncertainty; the next event could occur before or after 500 years. Planning should use the geologic and geodetic evidence, not a single elapsed-time comparison, and should consider a near-maximum rupture rather than the average.
2. Interpret the GPS, terrace, and tide-gauge data together. What do they tell you about the earthquake cycle and tsunami potential?
Read the worked answer
The 38 mm/yr trenchward GPS velocity shows the upper plate is being compressed and the megathrust is locked; elastic strain is accumulating. The 6-m marine terraces record abrupt coseismic uplift during past earthquakes, so the coast has experienced repeated large ruptures. The 1 mm/yr tide-gauge subsidence is interseismic: the coast sinks slowly between earthquakes as strain builds, then may jump upward during a rupture. Eight tsunami sand layers in 4,000 years show that at least some ruptures generated tsunamis that reached the coastal lake. A locked zone 15–40 km deep is offshore and shallow enough to generate significant seafloor displacement. The combined record supports a high probability of a future great earthquake and local tsunami, not a quiet margin. The 4-m spit is likely within the tsunami inundation zone.
3. Recommend siting and engineering measures for the hospital, and identify the data you would require before approval.
Read the worked answer
The hospital should not be sited on a 4-m spit 300 m from shore. It is a critical facility requiring post-disaster function, so it should be relocated inland and to higher ground, ideally above the mapped tsunami runup and outside liquefaction-prone beach sands. If no alternative site exists, the floor should be elevated above the design runup, with vertical-evacuation access, reinforced construction, and tsunami-resistant foundations, but these measures do not eliminate risk during a near-field event with minutes of warning. Require a probabilistic tsunami hazard assessment, a site-specific seismic hazard study including amplification and liquefaction, and a geotechnical investigation. Identify evacuation routes and early-warning coverage. The approval should also consider redundancy: a hospital that becomes inaccessible or inundated cannot serve the injured.