Two earthquake scenarios

The nearer earthquake does more to Vancouver itself. Cascadia affects the whole coast at once, which also delays outside help. [1] [2]

Two runs in the National Earthquake Scenario Catalogue underpin these scenarios. One models a full magnitude 9.0 rupture of the Cascadia fault affecting most of southwestern British Columbia. The other scales a real magnitude 4.6 earthquake beneath the Strait of Georgia in 1997 up to magnitude 7.0. [3]

Both runs use the same method, so they can be compared directly. Results are resolved to Census dissemination areas, roughly a few city blocks. They do not describe an individual address. [3]

The modelling counts shaking damage to buildings and the people inside them. It leaves out landslides, liquefaction, fire following and aftershocks. [3] That limit applies to almost every number below.

British Columbia’s damage and casualty figures for both scenarios use this federal modelling. [4] [3] [2] Emergency Management British Columbia and the Government Operations Centre in Ottawa commissioned the scenarios. [5] Two governments publishing the same figures is not independent agreement.

The province’s primary planning scenario is a shallow magnitude 7.0 beneath the Georgia Basin, modelled on the 1997 event. An earthquake of that size or larger happens in this broad region roughly once every 1,500 years. The rupture is under water, but a significant tsunami is not expected. [2]

The modelled fault plane dips at 47 degrees from the ground surface to a bottom edge about 13 km down, and the earthquake starts 3 km down. Its footprint runs from near Nanaimo to downtown Vancouver. That describes the fault itself, not where the damage falls. [6] Shaking across the Lower Mainland and Nanaimo is modelled at 10 to 60 per cent of gravity, strongest in downtown Vancouver, West Vancouver, Nanaimo, Gibsons, and on Gabriola and Bowen Islands. [5]

The scenario is set on a January afternoon, immediately after an atmospheric river has dropped 180 to 300 mm of rain over three days. The saturated ground makes landslides, liquefaction and dike failures more prominent. Warmth, dry shelter and unstable slopes are the immediate concerns. [2]

The earthquake may be heard before it is felt. A low rumble like a freight train is followed by 10 to 20 seconds of violent shaking. Near the epicentre, people can be knocked off their feet. Tall buildings sway, unsecured objects fall and roads crack. On soft, saturated soil, liquefaction removes support from buildings. [2]

Rain-soaked slopes fail and block transportation routes. Flooding worsens and some dikes fail. Damaged electrical equipment and broken gas lines start fires. Some buildings collapse; others shift, crack or burn. Glass and masonry fall into streets already blocked by debris. Running outside exposes people to those falling objects. Many who do are severely injured or killed, while thousands are trapped or injured inside buildings. [2]

Its modelled impacts, counting only shaking damage to buildings and to the people inside them:

Every figure in that table belongs to the daytime hours between 9am and 5pm, when most people are at school or work. They are minimum estimates because secondary hazards and damage to infrastructure are left out. [5]

The magnitude 9.0 scenario ruptures about 1,000 km of the fault from mid Vancouver Island to northern California. Its mainshock strikes at 10am on an August day during a 30 to 40 °C heatwave with wildfire smoke, making water, shade and clean air the urgent needs. Shaking lasts three minutes. [4]

Heaviest damage falls on Vancouver Island and in a band about 20 km deep along the mainland coast, from the border to the Sunshine Coast. The most severely affected area is roughly 45,000 square kilometres. Cascading events include tsunami, urban fires, liquefaction, landslides, floods, seiches and aftershocks. The west coast of Vancouver Island subsides by about 1 m. [4]

Mainshock shaking alone completely damages about 18,000 buildings, sends more than 10,000 people to hospital and kills 3,400. Those figures exclude landslides, liquefaction, tsunami, fires and other secondary hazards. [4]

It also models an aftershock: magnitude 7.1, 60 km beneath Sidney, about a month later, at 11pm, during an atmospheric river, lasting 20 seconds. [4]

The last Cascadia earthquake began at 9pm on 26 January 1700. The fault ruptured for about 1,000 km and produced a tsunami that crossed the Pacific. On Vancouver Island, shaking collapsed Cowichan houses and triggered landslides. People could not stand and became sick from the prolonged motion. The tsunami destroyed the Pachena Bay winter village, leaving no survivors. First Nations oral traditions record these events. [1]

Japanese records of the arriving tsunami establish the date and time. Those records and the First Nations accounts describe the same night. [1]

The crustal earthquake produces high-frequency shaking that is hardest on short buildings. Unreinforced masonry, brick or concrete block built without steel reinforcing, and unreinforced concrete buildings may collapse or become uninhabitable. [2]

The megathrust produces long-period waves that move tall buildings on deep, soft sediment, including parts of Richmond. [4]

A Lower Mainland study of deep bridge foundations shows why duration matters. During a 30-second design event, the soil liquefied while the structure was still shaking, so the moving ground and swaying structure loaded the foundations at the same time. During a 10-second event, those effects were largely separate. [7] These are design events for one structure, not descriptions of either scenario.

The province’s magnitude 7.0 Greater Vancouver planning scenario gives 10 to 20 seconds of violent shaking, while its magnitude 9.0 Cascadia scenario gives three minutes. [2] [4] A federal exercise scenario for a magnitude 6.8 near Tsawwassen lasts more than one minute. [8] Duration belongs to a named scenario, not to magnitude alone.

Official estimates put the average interval between Cascadia ruptures at roughly 400 to 600 years. The ranges around those estimates are much wider, from about a century to more than 1,000 years. An average cannot tell us when the next rupture will happen. [9] [2]

Current estimates put the chance of a rupture at the northern end of the fault, or a full-fault rupture, at roughly 5 to 20 per cent over the next 50 years. [10] [4]

A higher figure, about 37 per cent in the next 50 years, circulates widely. It applies to the southern end of the Cascadia fault, off southern Oregon and northern California, where ruptures are roughly twice as frequent. It is not a figure for this coast. [11]

For most systems, the published work assesses one design earthquake, so what is known about the weeks afterwards is the same for both. One thing differs: where help comes from. For a megathrust, the province states that the United States will be unable to deliver mutual aid if it is also overwhelmed; for a local crustal earthquake, its plan assumes agencies outside the impact area are available. [2] And one thing is known for the megathrust alone: the one study of the port, airport and ferry terminals modelled it and nothing else. [12]

Life afterwards covers every system.