Having spent some time yesterday going over ocean buoy data and observing tsunamis, I was surprised to hear about a "27 foot" tsunami seen in the open ocean. The source of this amazing rumor is, of all places, CNN:
JERAS: OK, so these are the detection which are out there in the Pacific Ocean. And you can see the flashing ones. These are active. These are the ones that we're watching. And there's Hawaii right from there. About 140 miles away from the Hawaiian island, we have a Bouie out there and this is what it is showing here. There you can see the line and notice this big drop down here. We have this big drop. This is about a nine-meter drop.
SANCHEZ: But what we can say is, tell me if I'm wrong, there is a tsunami there and it was just detected that it caused a 27-foot drop.
FRANKEL: Yeah, we recoded the tsunami passing that buoy, yes.
That's the essential part of the "reporting", the gist of which is that CNN claims a nine meter tsunami was detected in deep water. There's just so much wrong with this I don't know where to begin:
- The wave is two orders of magnitude larger than an open-water tsunami
- The wave is an order of magnitude shorter than a tsunami
- The wave precedes the expected arrival time of the tsunami by two hours
(The tsunami arrival time was estimated no earlier than 11:05 HST (for Hilo), which is 21:05 UTC -- two hours later than this "tsunami". CNN didn't catch on.)
Now there is something very important about this data, the reason why I added the data "dots" to the graph (on top of the lines). Right when the "tsunami" occured, the reporting frequency increased from 15 minutes to 1 minute, as you see by the density of the data points. Clearly, this "tsunami" is an artifact of measurement. To speculate, I think it is most likely they are catching ordinary ocean swells (or the integral over several of them, over 1 minute). There must be some sort of averaging algorithm that removes this short-scale variation. Most of the data, such as the 1-minute data featured in yesterday's blog, does not have such spikes. And in fact this dataset, over the last hour I've graphed, is perfectly flat as well.
To conclude, here is the rest of this buoy's data for the evening (postdating the CNN report), and again with the very obvious outliers excised. In the final, adjusted version, you can see the actual tsunami arriving right on schedule, around 21:10 UTC (11:10 HST).
By the way, take a look at this amateur video of the tsunami at Hilo (where it is amplified by the bay). It causes a small river to suddenly reverse direction.