CSZ History

When will the next major Cascadia Subduction Zone earthquake will strike? Scientists can’t know for sure, but there are some scientific clues which provide glimpses to the answer.

What’s commonly known
1) The Pacific Northwest has a ‘37% chance of a major Cascadia Subduction Zone earthquake in the next 50 years’.
The 37% figure comes from research headed by OSU Professor Chris Goldfinger, who helped me understand what data was used for the calculations and why the Log-Normal probability model was used.

2) Three hundred and twenty-one years have passed since the Pacific Northwest saw its last major Cascadia Subduction Zone earthquake.

What’s less well known
Take a look at this ten thousand year historical chart1. Each red and green bar represents an interval, or the number of years between major Pacific Northwest earthquakes on the Cascadia Subduction Zone. The horizontal blue line runs at 321 years. Far fewer red lines stretch above it, than the green lines hovering below

The intervals which stretched longer than 321 years are shown as red bars in the chart above. Only eight out of forty-six lines are red… meaning 83% of the time, the fault has not had to wait 321 years for the strain to break it.

Unlike probability modeling, this is not a 50-year outlook. This is a comparison of our fault’s 10,000 year historical behavior to where the fault currently stands (as of 1/26/2021).

Let’s play a game
If you knew that 83% of the time, your toddler was asleep by 8:00 pm, would you expect him to be up until 9:00 pm tonight? Probably not.

Knowing that 83% of the time, our fault has broken prior to reaching its 321st year of quiet… should we expect the fault to hold for another 50 years? 20 years?
Now, One Step Farther
Placing a yellow border around the longest intervals illustrates how they are grouped toward the oldest chunk of the 10,000-year history (intervals 34-46).

Intervals one through thirty-three make up the most-recent 6,030 years of history. During that time only 2 lines stretch beyond the 321-year horizontal blue line. Those two lines don’t reach far above. To be clear, this means 94% of the time in the last 6,000 years, the Cascadia Subduction Zone did not have to wait 321 years between ruptures.

Let’s play that game again
If you knew 94% of the time your toddler was asleep by 8:00 pm, would you expect her to be up until 9:00 pm tonight? Nope. Just nope.

Why do the oldest 4,000 years and the most recent 6,000 years look so different on the chart?

EOS’s image below is one possible explanation. Graphic A shows earthquake behavior which more closely matches the CSZ’s oldest 4,000 years (in the 10,000-year timeframe). There is a magnitude 9.0 followed by a single, smaller 8.0 range, then back to the 9.0.

Graphic B is an excellent illustration of what the CSZ has experienced over the last 6,000 years (more magnitude 8 earthquakes coming between the 9s).

Another angle
The Excel-generated charts from above show interval timing. Below is another chronological chart; however, each bar represents an earthquake rather than an interval. Again, the most recent history on the right side.

Like EOS’s figure B, notice how the CSZ has experienced more 8.0s (short lines) between the 9.0s (long lines). This isn’t always the case, as the last two CSZ earthquakes were 9.0s. Still, those earthquakes came fairly close together with only 232 years between. (Curious how big the next one be? Click here for some related research/thoughts.)

The short bars above represent magnitude 8.0 – 8.9 earthquakes (smaller events), while the tall bars represent 9.0s & above.

Here is what a couple of professionals in the field have to say about the matter.

website 4.JPG
website 3.JPG

Final thoughts
The ‘37% chance in 50 years’ probability is based on the entire 10,000 years, rather than just the most recent 6,000. If our current fault behavior matches that of graphic B in the EOS image above, then in my humble opinion we should not be looking 50 years out when we think about risk. We need a different narrative.

If you look at the comparison images below, you’ll see that the Log-Normal probability model changes from the 37% (36.88% rounded) chance in 50 years to 53% if only the most-recent 6,000 years are used in the calculations. The Gaussian model, equally accepted in the field, goes from a 47% to a 79%. Clearly the probability jumps in both models when looking at the current fault behavior. More, the percentages vary greatly based on which model is used.

This worksheet used by Goldfinger and his team at OSU originates from Northwestern University

It is my hope that you will consider this information, as I did, when you take steps toward getting prepared for the earthquake. If you need extra incentive, here is a list of very reputable organizations asking you to take it seriously & get two-weeks-ready. (There are many others!)


website 6

*The age of the 5th earthquake in this series is currently unknown. Due to the ambiguity, intervals 5 & 6 are unknown, as well. What we do know is that the interval between the 4th & 6th earthquakes equals only 115 years. Therefore, intervals 5 & 6 are each far less than 318 years. The same is true for the 25th & 31st undated earthquakes, with 235 and 37-year intervals surrounding them, respectively.