# JPL: Japan Quake May Have Shortened Earth Days, Moved Axis

From NASA’s Jet Propulsion Lab: The March 11, magnitude 9.0 earthquake in Japan may have shortened the length of each Earth day and shifted its axis. But don’t worry-you won’t notice the difference.

Using a United States Geological Survey estimate for how the fault responsible for the earthquake slipped, research scientist Richard Gross of NASA’s Jet Propulsion Laboratory, Pasadena, Calif., applied a complex model to perform a preliminary theoretical calculation of how the Japan earthquake-the fifth largest since 1900-affected Earth’s rotation. His calculations indicate that by changing the distribution of Earth’s mass, the Japanese earthquake should have caused Earth to rotate a bit faster, shortening the length of the day by about 1.8 microseconds (a microsecond is one millionth of a second).

The calculations also show the Japan quake should have shifted the position of Earth’s figure axis (the axis about which Earth’s mass is balanced) by about 17 centimeters (6.5 inches), towards 133 degrees east longitude. Earth’s figure axis should not be confused with its north-south axis; they are offset by about 10 meters (about 33 feet). This shift in Earth’s figure axis will cause Earth to wobble a bit differently as it rotates, but it will not cause a shift of Earth’s axis in space-only external forces such as the gravitational attraction of the sun, moon and planets can do that.

Both calculations will likely change as data on the quake are further refined.

In comparison, following last year’s magnitude 8.8 earthquake in Chile, Gross estimated the Chile quake should have shortened the length of day by about 1.26 microseconds and shifted Earth’s figure axis by about 8 centimeters (3 inches). A similar calculation performed after the 2004 magnitude 9.1 Sumatran earthquake revealed it should have shortened the length of day by 6.8 microseconds and shifted Earth’s figure axis by about 7 centimeters, or 2.76 inches. How an individual earthquake affects Earth’s rotation depends on its size (magnitude), location and the details of how the fault slipped.

Gross said that, in theory, anything that redistributes Earth’s mass will change Earth’s rotation.

“Earth’s rotation changes all the time as a result of not only earthquakes, but also the much larger effects of changes in atmospheric winds and oceanic currents,” he said. “Over the course of a year, the length of the day increases and decreases by about a millisecond, or about 550 times larger than the change caused by the Japanese earthquake. The position of Earth’s figure axis also changes all the time, by about 1 meter (3.3 feet) over the course of a year, or about six times more than the change that should have been caused by the Japan quake.”

Gross said that while we can measure the effects of the atmosphere and ocean on Earth’s rotation, the effects of earthquakes, at least up until now, have been too small to measure. The computed change in the length of day caused by earthquakes is much smaller than the accuracy with which scientists can currently measure changes in the length of the day. However, since the position of the figure axis can be measured to an accuracy of about 5 centimeters (2 inches), the estimated 17-centimeter shift in the figure axis from the Japan quake may actually be large enough to observe if scientists can adequately remove the larger effects of the atmosphere and ocean from the Earth rotation measurements. He and other scientists will be investigating this as more data become available.

Gross said the changes in Earth’s rotation and figure axis caused by earthquakes should not have any impacts on our daily lives. “These changes in Earth’s rotation are perfectly natural and happen all the time,” he said. “People shouldn’t worry about them.”

Article Rating
Inline Feedbacks
jb
March 17, 2011 12:59 am

Hooray for the NASA scientitst not using fear for funding.

MangoChutney
March 17, 2011 1:54 am

Does this mean the UK is now in the tropics?
🙂
/Mango

Bulldust
March 17, 2011 2:09 am

Interestingly a program I watched on Tuesday (How Earth Made Us – Water episode) in which it was stated that the huge volumes of stored water in dams in the northern hemisphere has altered the earth’s rotation as well (I forget when they said the bit about the dams):
http://www.abc.net.au/iview/?gclid=CNSZtrGd1acCFc2DpAodfTiq8w#/view/728666
The first show is also worth watching (Deep Earth) for the amazing gypsum crsytals in a mine in northern Mexico:
http://www.abc.net.au/iview/?gclid=CNSZtrGd1acCFc2DpAodfTiq8w#/view/728662
See from 2 minute mark onwards, or skip to the magical part at 5 minutes and on. I wish there were more shows like this to get kids interested in scientific fields like geology.

TFN Johnson
March 17, 2011 2:16 am

Slowing the Earth’s rotation causes the moon to move away a little bit, in order to preserve angular momentum of the whole Earth/Moon system.

Alan the Brit
March 17, 2011 2:45 am

So, he is telling us that the measured shift amounts to diddly-squat when all is taken into account – why did he bother?

Patrick Davis
March 17, 2011 2:46 am

Reports now are turning from the death and destruction of the quake, tsunami and threat from radiation to extreme cold. No power, many at risk from hypothermia.
Talking of the moon, now that it is closest to earth, it does appear brighter, and I can see more detail than usual. Amazing really, can’t say I have ever noticed before.

UK Sceptic
March 17, 2011 3:08 am

I’m with Alan the Brit. Why bother?

David L
March 17, 2011 3:12 am

Based on a Model.

JDN
March 17, 2011 3:15 am

So is GPS off by about 6′ in Japan, or, is everything else off by 6′?

Myrrh
March 17, 2011 3:21 am

Do these changes in wobbles affect the wind patterns?

John Marshall
March 17, 2011 3:28 am

And increased ice mass at the South Pole will increase rotation speed a tad. Or at least reduce the rate of rotational degradation due to ocean and atmospheric tidal friction.

steveta_uk
March 17, 2011 3:32 am

Every morning, lots of people commute into cities, and every evening they go home again. As a result, city centres will gain the mass of many people, cars, and other stuff, and so get heavier, and flex downwards, then revert in the evening. Since this is cyclic, the cities may start to oscillate up and down in a 24-hour rhythm. If some city somewhere happens to have an 11.574 uHz resonant frequency (that’s 24 hrs), this could be catastrophic!
Has anyone at NASA looked into this yet?

March 17, 2011 4:07 am

Alan the Brit and UK Sceptic;
Because it is interesting.
Isn’t that what science and research are SUPPOSED to be about?

amabo
March 17, 2011 4:43 am

Not necessarily, but it’s certainly a bonus.

martin brumby
March 17, 2011 4:49 am

Eh?
The hypothermalists won’t like that.
What would happen if every scientist went around talking about natural cycles and saying there was nothing to worry about?
This guy had better watch out.
Joe Romm and Bob Ward will be on his case in a heartbeat.

Joe Lalonde
March 17, 2011 5:11 am

Anthony,
I always wondered where they kept getting these figures.
Off of a model and not actually measurements from the planet.
How do they check if they are correct to the model?

Mark Wagner
March 17, 2011 5:21 am

why did all the reported shifts work to shorten the days? coincidence? would not some seismic events work to lengthen days as well?
@ TPN: the moon/earth system is moving the moon moves a couple of inches farther away every year. the moon pulls tides, which are pushed ahead of the moon by the rotation of the earth. the slightly off-axis gravitational pull works to pull the moon along faster, and therefore a higher orbit, every year. This also works to slow the earth day. In some billions of years the earth/moon will become locked in position with the moon directly over some position on earth, and a ~40 hour day.
or something like that.

Bill Illis
March 17, 2011 5:28 am

As noted, even the ENSO affects the length of day by about 1,000 times more than this earthquake. A large El Nino will shorten the rotation by up to 1 millisecond (slower equatorial winds and stronger mid-latitude winds in an El Nino slow the rotation).
There is a pretty strong correlation as shown in the numbers up to February 2011.
http://img38.imageshack.us/img38/6569/ensovsaamfeb11.png

ew-3
March 17, 2011 5:41 am

There is an upside to shorter days – we get to live to an older age 😉

TFN Johnson
March 17, 2011 5:48 am

Thanx, Mark Wagner.
I’ve been wondering how the Moon gets to know the Earth is rotating more slowly.

March 17, 2011 6:00 am

To Alan the Brit:
I find this very interesting. This is true science!

March 17, 2011 6:15 am

TFN Johnson says:
March 17, 2011 at 2:16 am
Slowing the Earth’s rotation causes the moon to move away a little bit, in order to preserve angular momentum of the whole Earth/Moon system.
No, as the angular momentum didn’t change.

March 17, 2011 6:21 am

My daughter is currently in Japan and she sent me the following link, which lists all the earthquakes they’re having–they post new temblors just a few minutes after they happen. They’ve had nine in the last hour and 17 minutes alone–four of which registered 5.8 on the Richter scale. (Occurred at: 20:41 JST 17 Mar 2011 to 21:55 JST 17 Mar 2011)
http://www.jma.go.jp/en/quake/quake_local_index.html

Rod
March 17, 2011 6:29 am

Changes in the length of day mean changes in earth’s rotation and thus its angular momentum. Over the years of looking at global weather satellite imagery, I observed what appear to be periods of time when cyclonic motion across the globe (i.e. storminess) is more prevalent than at other times. There is likely to be some correlation here between these change events and the degree or prevalence of cyclonic motion in the atmosphere. Perhaps also on anticyclonic motion and the degree of blocking patterns in the atmosphere. This would seem to be a fruitful area for research.

Pamela Gray
March 17, 2011 6:38 am

It is this atmospheric/oceanic phenomenon that prompted that other paper-reviewer to suggest that AGW causes rotational change. Yet in the same paper, ignored by the reviewer, it was clearly stated that the modeled AGW portion of atmospheric change is not powerful enough to show up in measurable rotational changes.
However, I trust climate scientists to eventually splice together a graph that shows it does, theoretically.

Coach Springer
March 17, 2011 6:41 am

So, plate tectonics can change the climate? And melting ice causes plates to move, causing earthquakes – meaning plate movement? I think I’ve solved all of our problems: Move something of weight equal to that of melting ice to Greenland. So, Al Gore really can save the planet!

March 17, 2011 7:10 am

Interesting—and not a word about ‘climate change’. Bravo!
/Mr Lynn

2klbofun
March 17, 2011 7:14 am

Well, the silver lining is that Obama’s term will be about 4 ms shorter.

A G Foster
March 17, 2011 7:54 am

Gross is as sharp as a tack on LOD problems, but for any who are wondering about the mechanism, I hope this helps. Hot mantle slowly rises, and cools, then sinks, but not smoothly. Over the long haul the lithosphere’s moment of inertia doesn’t change, but when a cooled mass drops suddenly, and is replaced by warmer mass of less density, rotational inertia decreases and the earth speeds up, until the sunken mass warms and the replacement mass cools. So it has to do with uneven cooling and flow of the mantle. Think of big kids and little kids on the merry-go-round switching places from the inside to the outside.
True, this is so far an academic excercise, but it has its raison d’etre in the new science of satellite measured earth rotation, which has thrown lots of light on earth processes. And the time may come when even microseconds can be measured, which would correspond to half a millimeter at the equator, measured instantaneously.

March 17, 2011 7:56 am

“These changes in Earth’s climate are perfectly natural and happen all the time. People shouldn’t worry about them.”

March 17, 2011 8:04 am

Rod says:
March 17, 2011 at 6:29 am
Changes in the length of day mean changes in earth’s rotation and thus its angular momentum.
No, in its moment of inertia. Think of a spinning ice skater. She has a certain [fixed] angular momentum. If she stretches out her arms, she increases her moment of inertia, and since the angular momentum is not changed her spin must slow down.

dp
March 17, 2011 8:05 am

I would think that if the center of mass of the Earth has shifted, so too have the orbits of the satellites that encircle us, including the GPS birdcage. The entire reference frame has shifted along with everything in it. If so this will have interesting effects in chaotic systems. At least our projections from our models will be impacted because we now have new parameters injected into the system. The butterfly’s wing beats in Paraguay and something happens somewhere, but what? Where?

truthsword
March 17, 2011 8:05 am

MAY? Heh

CRS, Dr.P.H.
March 17, 2011 8:51 am

@steveta_uk says:
March 17, 2011 at 3:32 am
Every morning, lots of people commute into cities, and every evening they go home again. As a result, city centres will gain the mass of many people, cars, and other stuff, and so get heavier, and flex downwards, then revert in the evening. Since this is cyclic, the cities may start to oscillate up and down in a 24-hour rhythm. If some city somewhere happens to have an 11.574 uHz resonant frequency (that’s 24 hrs), this could be catastrophic!
Has anyone at NASA looked into this yet?
————-
Interestingly, the World Wildlife Federation’s latest report on sustainable energy mentions that the vibrations of city centers might be harnessed for electrical generation!
http://wwf.panda.org/what_we_do/footprint/climate_carbon_energy/energy_solutions/renewable_energy/sustainable_energy_report/
I humbly offer this great work of fiction for Anthony, readers & mods of WUWT!

Rod
March 17, 2011 9:00 am

Leif Svalgaard says:
March 17, 2011 at 8:04 am
No, in its moment of inertia. Think of a spinning ice skater. She has a certain [fixed] angular momentum. If she stretches out her arms, she increases her moment of inertia, and since the angular momentum is not changed her spin must slow down.

Rod
March 17, 2011 9:07 am

Leif Svalgaard says:
March 17, 2011 at 8:04 am
No, in its moment of inertia. Think of a spinning ice skater. She has a certain [fixed] angular momentum. If she stretches out her arms, she increases her moment of inertia, and since the angular momentum is not changed her spin must slow down.
Thanks. You are right. What I am seeking to understand are changes in vorticity within the atmosphere when earth’s rotation speed changes.

TonyG
March 17, 2011 9:32 am

Mark Wagner says:
why did all the reported shifts work to shorten the days? coincidence? would not some seismic events work to lengthen days as well?
Exactly what I was wondering. Anyone have an answer?

DAV
March 17, 2011 9:43 am

“The March 11, magnitude 9.0 earthquake in Japan may have shortened the length of each Earth day … ”
Swell! As if there was too much time in the day already. Add that to the hour lost last Sunday. We need to start saving time if only for our children.

A G Foster
March 17, 2011 10:16 am

Mark Wagner at 521, Tony G at 932: Volcanic activity has the opposite and steadier effect of slowing the earth down. Hot magma arrives at the surface allowing the older and cooler surface to sink. This is a microcosm of large scale mantle convection. Lateral fault movements have no effect on LOD, and uplift of cold rock, while common on the continents, runs against the grain in the mantle. Slow rising of hot mantle slows the earth down. I suppose the interface between tectonic plates is smoother than that between the top plate and the continents, leading to upward movement of hot lower plates being smoother than downward moving upper plates. So bigger, deeper quakes typically involve a loss of inertia, which loss is balanced out by the steadier rise of the lower plates. If you want a better answer you can contact Gross at NASA.

A G Foster
March 17, 2011 10:58 am

I’ll have to modify that. Whereas the rise and fall of masses is simultaneous, the cooling is steady, and the process of cooling leads to contraction, a T inversion, and the ensuing convection leading to quakes. Forget about plate roughness.

March 17, 2011 11:15 am

I call bull-crap!!!

Agile Aspect
March 17, 2011 12:19 pm

by Alan the Brit
So, he is telling us that the measured shift amounts to diddly-squat when all is taken into account – why did he bother?
—————————————-;
It was a model prediction – it was not a measurement. Until it’s confirmed by measurement it’s fiction.
The season change in the LOD is on the order of millisecs when compared to UTC.

Agile Aspect
March 17, 2011 12:51 pm

by TFN Johnson says:
March 17, 2011 at 2:16 am
Slowing the Earth’s rotation causes the moon to move away a little bit, in order to preserve angular momentum of the whole Earth/Moon system.
——————-;
To first order, it’s actually a 3 body graviation problem since you can’t ignore the Sun.
The LOD of the Earth will continue to decrease and the Moon’s LOD will continue to continue to increase as the Moon continues to move away from the Earth since the Earth and the Moon are gravitionally phased locked.
Interestingly, Poincare demostrated over 100 years ago the 3 body gravitation problem was chaotic (which until recently – with the exception of Einstein, has been largely ignored.)

Agile Aspect
March 17, 2011 12:56 pm

“The LOD of the Earth will continue to decrease and the Moon’s LOD will continue to continue to increase as the Moon continues to move away from the Earth since the Earth and the Moon are gravitionally phased locked.”
Ahh, the rotation of the Moon will speed up as the rotation of the Earth slows down. Sometime after the Sun becomes a Red Drawf, the speeds will be the same.

Philip T. Downman
March 17, 2011 1:08 pm

The Earth’s axis has long had a considerable tilt. Now I fear it is bent too.

Sun Spot
March 17, 2011 3:54 pm

Is this just more media science ? Media science is good for funding.

Dave Worley
March 17, 2011 4:23 pm

It’s ok because the melting of the last ice age surely lowered lots of formerly elevated water, increasing the speed of rotation. We’re just making up for that a little.

JudyW
March 17, 2011 6:00 pm

The magnetic pole is shifting. The tectonic plate will adjust accordingly. Expect more big earthquakes as the magnetic pole move toward Siberia.

A G Foster
March 17, 2011 6:33 pm

When grounded polar ice melts the earth slows down, leaving the ground depressed. As the ground rebounds, the earth speeds back up, but at present not enough to make up for tidal acceleration (2.3ms/century less .6ms/century as I recall). As the relative masses of the earth, moon, and sun are easily obtained through Newtonian physics, and as the absolute distances may be measured by triangulation or calculated by measurement of the gravitational constant, so may the relative tidal contributions of the moon and sun be calculated precisely. And since the growing distance to the moon can now be measured by laser ranging, its growing angular momentum can be calculated, hence the earth’s loss of angular momentum due to lunar, then solar tides. This calculated loss is then compared to observed loss as obtained through ancient astronomical observations and modern telescopic sighting and satellite telemetry, and the long term difference is attributed to glacial isostatic adjustment. Then the considerable fortnightly contribution of zonal tides may be modeled and deducted from measured LOD, leaving seasonal variation to be accounted for, most of which may be explained by atmospheric coupling. Other components include groundwater depletion, terminal lake depletion (mainly the Caspian and Dead Sea) opposite artificial reservoirs, ocean basin adjustment to changing sea levels and other activity of the lithosphere, etc.
After all this we are left to argue whether remaining trends are due to core mantle coupling, ice growth in Antarctica, less sea level rise than claimed, and so on. At least LOD places limits on estimations of current rate of sea level rise.

March 17, 2011 6:58 pm

JudyW says:
March 17, 2011 at 6:00 pm
The magnetic pole is shifting. The tectonic plate will adjust accordingly. Expect more big earthquakes as the magnetic pole move toward Siberia.
No, the magnetic field is generated deep in the core and has nothing to do with the plates.

Paul Vaughan
March 17, 2011 8:13 pm

Re: Bill Illis
…And compare with global seismicity (Centennial Earthquake Catalog) – and rate of change of solar wind speed and rate of change of neutron count rates (proxy for cosmic ray flux). Nonrandom relations — but those who ignore the spatial dimension (and just run temporal spectral analyses) will draw fundamentally flawed conclusions. Spatial phase reversals are being misinterpreted as temporal features. The phase of interannual NPI, for example, can be observed to fit the phase of rate of change of the above-mentioned solar indices if one simply realizes the effect of spatial windowing parameters on summaries. The mainstream has fallen victim to the spatiotemporal version of Simpson’s Paradox.

Paul Vaughan
March 17, 2011 8:26 pm

Pamela Gray wrote, “However, I trust climate scientists to eventually splice together a graph that shows it does, theoretically.”
There’s a roughly monotonic trend in decadal extent AAM & LOD power at both semi-annual & annual grain (but since the vast majority of the mainstream ignores extent, this will likely remain unnoticed).
I’ve been wondering how long it might take some enterprising AGW alarmist to attempt to capitalize empirically upon the preceding.

March 17, 2011 9:22 pm

Paul Vaughan says:
March 17, 2011 at 8:13 pm
…And compare with global seismicity (Centennial Earthquake Catalog) – and rate of change of solar wind speed and rate of change of neutron count rates (proxy for cosmic ray flux). Nonrandom relations
My analysis http://www.leif.org/research/Earthquake-Activity.png is conclusive. There is no hint of any increase in earthquake activity following geomagnetic storms, because if they did, there would be a larger number of earthquakes on the day of the storm [or the next day], and there isn’t. The solar wind speed is organized by sector boundaries, but passage of these also have no signal in global seismicity.

Paul Vaughan
March 17, 2011 9:32 pm

Re: Leif Svalgaard
You have misunderstood what I wrote (or perhaps simply seized what appeared as an opportunity to denounce the claims of others?)

March 18, 2011 7:04 am

Paul Vaughan says:
March 17, 2011 at 9:32 pm
You have misunderstood what I wrote
Perhaps be less opaque, then.

JudyW
March 18, 2011 7:17 am

One articale about the impact of the pole shift on tectonic plates:
http://www.earth-issues.com/2011/03/geomagnetic-pole-shift-triggering-massive-superquakes-2/
This is not a technical paper . I’m not sure who the author is.
On the external influences to earth’s magnetic field, an earthquake due to a magnetic storm would not be immediate. The lag factor would at least depend on the existing pressure built up of the plates and the size of the CME.

March 18, 2011 8:34 am

JudyW says:
March 18, 2011 at 7:17 am
On the external influences to earth’s magnetic field, an earthquake due to a magnetic storm would not be immediate. The lag factor would at least depend on the existing pressure built up of the plates and the size of the CME.
Examining ALL such storms over the last 100 years shows that there is no such increase in the number of earthquakes on the day of the storm [or on any other day up to a hundred days after the storm for that matter]. As simple as that: http://www.leif.org/research/Earthquake-Activity.png which shows the number of earthquakes on the day of the storm at 0, one day after the storm at 1, etc.
The blue and the red points are from two different Earthquake Catalogs [the red going to slightly lower energy 5.5 Richter than the blue 6, hence the higher number].

Paul Vaughan
March 18, 2011 7:21 pm

Leif, you’re missing some fundamentals about the effect of integration across spatiotemporal harmonics.

March 18, 2011 10:57 pm

Paul Vaughan says:
March 18, 2011 at 7:21 pm
Leif, you’re missing some fundamentals about the effect of integration across spatiotemporal harmonics.
No less opaque.

March 19, 2011 1:47 pm

Paul Vaughan says:
March 18, 2011 at 7:21 pm
Leif, you’re missing some fundamentals about the effect of integration across spatiotemporal harmonics.
the quantities you were mentioning are hardly ‘spatiotemporal’ except in the trivial sense that everything exists in space and time [or even spacetime]. The ‘harmonics’ bit is extraneous as these quantities are not oscillators. But in solar physics we do have genuine spatiotemporal variables and we do know how to deal with them. No ‘fundamentals’ are missing. Here is a good example: http://soi.stanford.edu/results/td.html
Note how [last figure] data is plotted against both space [abscissa] and time [ordinate].

Paul Vaughan
March 20, 2011 9:10 am

Leif, the issue is with sampling intervals (and the effect on statistical summaries of pattern). In layman’s terms, I am raising issues with apples and you are commenting about oranges. My commentary is deliberately brief to save time, so I accept responsibility for misunderstandings arising due to lack of elaboration. To clarify, the “apples” include not only NPI & global seismicity, but also SOI, SAM, AAO, AO, NAO, PDO’, AMO’, HadSST’, PWP, VEI, IVI2, SAOT, MSI, DVI, CO2′, LOD’, AAM, QBO, & others. Many analysts are misinterpreting spatial phase reversals as temporal features [due to a fundamental lack of awareness of (a) the effect of windowing parameters on statistical summaries of pattern and (b) the effect on pattern of integration across spatiotemporal harmonics]. Since related misunderstandings arising in the mainstream are so fundamental, I estimate that it could take decades to correct the widespread misinterpretations. The required teachings are challenging even for talented educators and the pool of citizens capable of grasping the concepts is limited.

March 20, 2011 10:18 am

Paul Vaughan says:
March 20, 2011 at 9:10 am
Leif, the issue is with sampling intervals (and the effect on statistical summaries of pattern). In layman’s terms, I am raising issues with apples and you are commenting about oranges.
The sampling intervals are irrelevant. When comparing, once always uses the same sampling interval for both [or all] quantities. Understanding of what a ‘sampling’ interval is may be sorely lacking. For example, the solar wind speed may be in reality sampled every few seconds. From such fine-grained data it is customary to form, say, hourly of daily averages. This is legitimate because there is a very high degree of autocorrelation [or to use the older word: positive conservation] in the data, i.e. the solar wind speed from one second to the next is almost the same.
Many analysts are misinterpreting spatial phase reversals as temporal features
Most of these quantities don’t have spatial phase reversals, e.g. the global temperature or TSI. Explain what spatial phase reversals are appropriate for those.

Paul Vaughan
March 21, 2011 10:28 pm

Leif, I apologize as I’ve limited time, so I’ll just list some of the things on my mind as I’ve commented:
EXTENT (not just grain),
SPATIAL sampling interval (not just temporal!) …and shape matters too, not just area,
multiscale spatial heterogeneity,
depth & altitude (“global” isn’t just a skin),
leverage (on statistical summaries by outliers, such as high amplitude regions…) – e.g. weighting (for example by specific heat – e.g. continental vs. maritime – e.g. why AMO’ relates more closely than SOI to HadSST’…)
The preceding’s not exhaustive – just suggestive.
I’ll elaborate briefly:
Plant ecologists throw down quadrats to measure pattern. If they change the quadrat size they measure a different pattern (particularly if there is clustering at some scale or another, etc.) IT’S NOT ENOUGH TO JUST PICK ONE SIZE AND STICK WITH IT; plenty of scolding articles, lectures, & presentations have been painstakingly made to emphasize (like thunder) this ABSOLUTELY CRUCIAL point. One EASILY finds paradoxes, scale discontinuites, etc. Take the word ‘paradox’ literally – (a good example to which lay-people can easily relate is the redrawing of electoral boundaries to favor a party). [Note: Some physical geographers refer to this as the “modifiable areal unit problem”, sometimes abbreviated MAUP.]
I trust that you get the essence without need for protracted exchange. My comments are most strongly influenced by the following subset of variables: NPI, AO, NAO, NAM, SAM, AAO, COWL, & SOI. Perhaps there will be time to elaborate further some less busy day (for example as to why simple linear methods canNOT handle this job). In the meantime I encourage you to have a look at some (not necessarily all) of what Tomas Milanovic has been saying over at Climate Etc. While I disagree fundamentally with him on some issues, I appreciate that he is aware of many of the windowing issues which I have described here; Milanovic is one of very few engaged in the online climate discussion possessing (& expressing explicitly) this awareness.
Climate science has itself tied in a knot by the spatiotemporal version of Simpson’s Paradox. With patience, freedom from the tangled bind is achievable.
Best Regards to All.

March 22, 2011 6:21 am

Paul Vaughan says:
March 21, 2011 at 10:28 pm
EXTENT (not just grain), grain????
SPATIAL sampling interval (not just temporal!) …and shape matters too, not just area,
Many analysts are misinterpreting spatial phase reversals as temporal features
Most of these quantities don’t have spatial phase reversals, e.g. the global temperature or TSI. Explain what spatial phase reversals are appropriate for those.