A Japanese Puzzle

Guest Post by Willis Eschenbach

A reader who posts under the name “tokyoboy” sent a link to a very interesting sea level record from the Japanese Meteorological Agency. It covers the period 1906–2010, and when I first saw it I thought they’d made some mistake.

So I got their data, and plotted it up. I also got the satellite records for the area. Finally, I got records of one of the sites that the Japanese used, but I obtained it from the PSMSL records. All of them agree very well, so I am forced to assume that there are no obvious errors in the Japanese records. Figure 1 shows the results:

Figure 1. Japanese sea level records. Two records marked “Japan” are from the citation above. They are averages of long-term records since 1906 (4 sites, blue line), and shorter-term records since 1960 (16 sites, red line). Satellite records (green, 1993-2010) are from the University of Colorado interactive wizard. Wajima records (purple, 1930-2010) are from the PSMSL

You can see why I thought there was a mistake. Sea level around Japan rose steadily from 1906 to 1950. Then it dropped for fifteen years and bounced around until 1980. Since then it has risen again, but it is about 20 mm lower than it was in 1950.

Now, I can’t find anything at all wrong with the data. The satellite record agrees with the Japanese averages, as does the PSMSL record. So we have to assume it is accurate.

But it is unlike any record I’ve seen of the global average change in sea level. That global record climbs steadily over the century.

Image added by Anthony via Wikipedia

So I fear I have no great insight into what is going on. Over the last 80 years, the sea level in Japan was level, went up, went down, went back up, and now is not far from where it started.

I’m happy for suggestions and comments, as I’m in mystery over this one. It’s one of the great things about the climate, always more puzzles to solve.



newest oldest most voted
Notify of

Is that data corrected for the land rising or subsiding? With the high level of earthquakes in Japan I would expect that earth movement has more effect on sea level than actual changes in sea level.

The Japanese land moves a lot and obviously not only when quakes strike…how’s that for an answer?

Sea level varies with atmospheric pressure as well as the volume of water in the oceans. Japan is in a seismically active zone so the earth’s crust will be moving up and down by some amount. Determining an average sea level is not an easy task. Remember the arguments between John Daly and John Hunter over the mean sea level mark on Dead Man’s Isle made by Lemprier in the 19thC. The mark is still 300mm or so above current MSL.


Yes, my initial thought is tectonic activity.

Anthony, as you are aware, Japan is plainly one of the most seismic regions on the planet…
This has implications: correlating sea-level rise and fall with local large seismic events, most notably 1923 and 1995? Most certainly, the Sendai region subsided more than a meter after the March 2011 earthquake, but there are almost certainly other ‘aseismic’ swellings and subsidences in the Japanese archipelago which would have a direct influence on sea levels recorded there.

And I should have addressed “Willis” above^^^

Japan sits on the edge of tectonic plates – perhaps that makes the land rise and sink?
Here in Scandinavia the sea level steadily sinks contrary to global trends, but that’s easily explained by the land rise still active since the last ice age.


Anthony, How do the lows & highs in the Japanese data compare to where GISS, etc. have “adjusted” the data over time–compare with your “blink charts” & maybe the source of disagreement will flash into view!


I would think earthquakes could be the cause. It is a very geologically active area.


Agree with above two posters, it will be the changes in coastal formations. Japan has constant tremors and quakes. Last time I was there a twenty second shake rocked the hotel in Tokyo I was staying in, various bits flew around the room, don’t recall off hand how many in that area but loads a year.
Had a strange experience many years before that when living there for a while on the edge of Kyoto. Taking a break in the garden sitting on the ground watching the paddy fields a little below me I felt and heard a tremor coming from a distance away to the mountains around me, and it was coming in a straight thin line directly under where I was sitting, happened very quickly. Saw the line of it shaking the trees in the distance and watched as it came under me and carried on.


Well, it looks like global temperature (without UHI).


My initial thought was Godzilla.




And they’re also deep, this the latest one:
January 1, 2012 10:14 AM EST
Less than a year after Japan was devastated by a massive earthquake and tsunami in 2011, the first strong earthquake of 2012 has struck Japan.
The strong earthquake with a magnitude of 6.8 shook Japan off the country’s coast on Sunday, the first day of the new year, according to the U.S. Geological Survey.
The USGS said the earthquake struck 495 kilometers (307 miles) south-southwest of Tokyo at a depth of 348.5 kilometers (216.6 miles). Because the jolt was so deep, it was less likely to cause damage at the surface than quake last March 11, which caused a tsunami and, ultimately, a nuclear crisis in the country.
However, there were no reports of damage this time around, and the Pacific Tsunami Warning Center did not issue a tsunami warning.

Ryan Simpson

Average sea level is always a function of tectonic subsidence and rise. It is generally accepted that mean sea level has been rising at a constant rate of 2mm/yr over the last century. Local variation will be the result of tectonic activity which can have a rate as high as 9cm/yr. (Usually much lower.) Bangledesh and the Mississippi delta are excellent examples of a subsiding landmass. Conversely parts of Canada, Europe and the US are still isostatically rebounding from the last glaciation at a rate that exceeds sea level rise.

Simple. Overfishing and 10-year fallout from Hiroshima and Nagasaki led to a serious depletion of ocean fish stocks in the 1950s. Because of the lack of fish, the volume of the Ocean around Japan was reduced. As the fish stocks recovered, so did the ocean volume, affecting its level.

Ian W

I see that no-one has come up with the obvious answer.
‘Global Average Sea Level’ is a useless metric. Like an average telephone number it is meaningless.



Willis Eschenbach

Several people have proposed geophysical movements. However, the changes in the record seem too abrupt for such a cause. There’s a discussion of the issue at the PSMSL, but it doesn’t really resolve anything.


If tectonic in origin do all the 4/16 sites move up or down at the same time, or do some go up and some go down. I am doubtful if they (the sites) move in the same direction at the same time. How does the data compare with Iceland, New Zealand and the Bay of Naples three areas of volcanism I can think of, or the Eastern Mediterranean which I think is quite seismically active. I am sure that others can suggest better areas for comparison. If there is similar deviation from the norm then perhaps it is down to the earth moving.


“Here in Scandinavia the sea level steadily sinks”
Well.. another place where the sea level sinks…
“That global record climbs steadily over the century.”
No it doesn’t.

An Inquirer

Look at other specific localities. i suspect Japan is not the only locality whose trend differs from the Wikipedia graph.


I have put together a video of the entire set of Permanent Service for Median Sea Level (PSMSL) data set to music. (I admit it – I’m a geek). The idea it to visually look for an acceleration of rise rate in the graphs. The first two minutes of text may be a little boring, but they explain the context.
If interested, have a look here…

Alexander K

As so many others have said, sea level is usually measured relative to land and if the land frequently bounces up and down…


I think Willis addresses landmass movement by pointing out that Japanese measurements match satellite measurements, even through the recent massive quake.


mkelly says:
January 31, 2012 at 11:05 am

That has my vote :). May I propose an addition? Global Sloshing. I just need a grant now…

Apropos my comment above regarding atmospheric pressure, Kolker & Hameed’s paper, Meteorologically driven trends in sea level rise makes for interesting reading.

Determining the rate of global sea level rise (GSLR) during the past century is critical to understanding recent changes to the global climate system. However, this is
complicated by non-tidal, short-term, local sea-level variability that is orders of magnitude greater than the trend. While the non-dimensional North Atlantic Oscillation (NAO) index can explain some of this variability in the Atlantic, significant results have been largely restricted to
Europe.We show that dimensional indices of the position and intensity of the atmospheric centers of action (COAs) comprising the NAO are correlated with a major fraction of the variability and trend at 5 Atlantic Ocean tide gauges since 1900. COA fluctuations are shown to influence winds, pressure and sea-surface temperatures, thereby influencing sea level via a suite of coastal oceanographic processes. These findings reduce variability in regional sea level rise estimates and indicate a meteorological driver of sea-level trends.

Russ in Houston

Ian W nails it. Global sea level is almost as useless as an Average Global Temperature.


Sea floor volcanoes hold water on top of them….higher water levels
Every tsunami and eruptions shows on your plot……especially Wajima records


Looks like El Niños piling up water in the western Pacific.
PDO peaked briefly around 1940 was high from 1980 to 2000.


Wise math calls it, and isn’t it so obvious? Good call wise math – hey – would’ja like to apply with me for a grant to study it because I’m tellin ya now: the science is settled, buddy.
The sintz is settlt. Thim’s sum POWWWWWW’eRFUL PROGNOSTIFUKASHUNS Wise Math, Powwwww’ERFUL!!!
All up for a study on Wise Math (and now mine – we great minds you know – well maybe you guys DON’T know, caws YEW AIN’T CLIMATOLUJISTS!) and his FANTASTIC THEORY- it’s practically a LAW already – just sign some petition I’ll throw up when I sober up, and we’ll all be rich,
we’ll be in W.I.K.I.P.E.D.I.A. as sum uv thuh smartist minn in thuh WERLD! *
*that was a paraphrase from al (the goracle) gore
Tony – P.E.A.C.E. on GOOD GUYS like you and Willis and everyone.


Your 17-year Gaussian average occults the data in the late 1950’s, making it hard to judge how well the splices correlate. Is it possible that the big downswings are a product of the transition from the 4 old sites to the 16 new sites?
The two big upswings together look like around 120mm, which isn’t too far off the 180mm in the chart of global sea rise Anthony added.
I see you chose to truncate your 17-year average line by about 4 years from either end. I love that you didn’t try to run the line to the ends of the data, but would you have a few minutes to detail the thinking behind the 4-year cutoff you chose?

Major oscillation appears to be due to the tectonic plate movements.


To mis-quote a Sarah Palin parody, “I can see the PDO from my house”.

Ian W said @ January 31, 2012 at 11:05 am

I see that no-one has come up with the obvious answer.
‘Global Average Sea Level’ is a useless metric. Like an average telephone number it is meaningless.

OTOH local sea level, like local temperature is important.


Ugh, it is of course La Niñas that pile up water in the western Pacific, not El Niños as I stated above. 😉

FergalR says:
January 31, 2012 at 11:35 am
Looks like El Niños piling up water in the western Pacific
I was going to guess wind and barometric pressure.

Hi willis
I wrote about sea level changes from the Holocene to the Romans in part 1 of my article here;
My own study-backed up by several graphs by current researchers cited in my study-demonstrate that sea levels continually rise and fall by around 30cm either side of a mean average. Great care must be exercised to take into account land rise and fall-there are several links in my study to some interesting papers demonstrating the impact. All in all, we are still currently around 20cm below the sea levels experienced at several points in our recent history.
The chart you cite is one I wrote about also, it is based on a tiny numer of tide gauges mostly in the NH on to which is spliced the sarellite record a la the hockey stick. If you could zoom out to see it in historical context it reached its last peak around 1600 and has had a couple of oscillations since then, with a slow but very undramatic rise in recent decades. The current rate of sea level rise is virtually zero despite all the models and statistically the rise in the second half of the 20th Century was marginally less than in the first half according to Simon Holgate

It looks like AMO.

Mike Pickett

Regardless of various opinions of what the mechanisms are or aren’t, it would be very interesting to run a Fourier analysis of the curve, because I think I see at least one periodic function in it. Maybe the period could be matched to some other known period in our planet or planet-moon coupling.

“The Japanese land moves a lot and obviously not only when quakes strike…how’s that for an answer?”
Does the satelite measure relative to the land? To itself?

doug s

If I were a cynical gambling man, I’d say the “global data” was cooked, and the Japanese data is as it was recorded.


I believe studies of Tuvalu also showed that sea level had gone up and down over the years and was presently lower than in 1950. The islanders did not like the report and suppressed it from being made public. But, I think they forgot to ban writing a research paper about it.
Perhaps the steady sea level rise graph has been a tad corrupted in the same way the temperature data has been.
I tend to believe the real world data more than anything from the large organizations.

Hi Willis: The SST anomalies around Japan show some interesting variability.
If I had the time, I’d look into the impacts of multidecadal variability of ENSO and North Pacific Sea Level Pressure (NPI) to see how closely they relate to your Sea Level curves. Unfortunately, I’m trying to finish up a long-term project, If you were to check, invert the NINO3.4 SST anomalies first.


Is the sea level record from JMA adjusted for thermal expansion consistently with the global record? The 1950s were of course warm; in the 1970s it was cold (hence all the concerns about end of the interglacial), then in the 1980s it got warm and Hansen started strutting his stuff.
Or, referring back to your recent article on “Decimals of Precision” (URL as below) on standard error of the mean, the global plot is taken from only 23 tide gauge records worldwide – but the individual traces on the global chart appear to swing regularly at least +/- 5cms from the average. Might the Japan plot (4 sites from 1906, 16 from 1960 albeit for a much smaller area), showing the approximate 7 cm decline from 1950 to 1985, be a just statistical artefact?
URL: http://wattsupwiththat.com/2012/01/26/decimals-of-precision-trenberths-missing-heat/


An interesting way to study the rising or sinking of the land is to check the current level of the coastline from the previous interglacial ca 120 000 years ago. In Italy which is a seismically active area this varies from 185 meters above sea-level near Messina to 120 meters below sea level near Ravenna. So, on average, the Ravenna area has sunk about 1 mm/year during the last 100,000 years while Messina has risen about 1.5 mm/year during the same time
If we instead study Australa, famously the most tectonically stable part of the world, the variation is indeed only from + 40 meters in Tasmania, to several meters below sea-level in northern Queensland, so there the movement has probably everywhere been well below 0.5 mm/yr.
So are there any areas which haven’t moved during this time period? Perhaps. There are some largish tectonically stable blocks with an interglacial coastline about 2-3 meters above sea level, including the Gawler Block in Australia and the Tyrrhenian block in the western Mediterranean. Since such large areas never move “as a piece” as far as we know, it seems plausible that instead they haven’t moved at all.

Tom Ragsdale

I haven’t read all of the previous comments but the thing that comes to my mind that could be at least part of the cause is a variation in the paths or strengths of pacific ocean currents.
Just a thought.

Pete in Cumbria UK

Howzabout something similar to what’s described here yesterday….
was it a huge undersea heat-source/volcano/magma chamber expanding/lifting the water – were there there any geologic events in the 20s, 30s or 40s that correspond to the high point?