Decadal Oscillations Of The Pacific Kind

Guest Post by Willis Eschenbach

The recent post here on WUWT about the Pacific Decadal Oscillation (PDO) has a lot of folks claiming that the PDO is useful for predicting the future of the climate … I don’t think so myself, and this post is about why I don’t think the PDO predicts the climate in other than a general way. Let me talk a bit about what the PDO is, what it does, and how we measure it.

First, what is the PDO when it’s at home? It is a phenomenon which manifests itself as a swing between a “cold phase” and a “warm phase”. This swing seems to occur about every thirty or forty years. The changeover from one phase to the other was first noticed in 1976, when it was called the “Great Pacific Climate Shift”. The existence of the PDO itself, curiously, was first noticed in its effects on the salmon catches of the Pacific Northwest.

pdo warm and cold phases

Figure 1. The phases of the PDO, showing the typical winds and temperatures associated with its two phases. The color scale shows the temperature anomalies in degrees C.

Figure 1 is a clear physical depiction of the two opposite ends of the PDO swing, based on how it manifests itself in terms of surface temperatures and winds. But to me that’s not the valuable definition. The valuable definition is a functional definition, based on what the PDO does rather than on how it manifests itself. In other words, a definition based on the effect that the PDO has on the functioning of the climate as a whole.

A Functional Definition of the PDO

To understand what the PDO is doing, you first need to understand how the planet keeps from overheating. The tropics doesn’t radiate all the heat it receives. If it did the tropics would be much, much hotter than it is. Instead, the planet keeps cool by constantly moving huge, almost unimaginably large amounts of heat from the tropics to the poles. At the poles, that heat is radiated back to space.

The transportation of the heat from the equator to the poles is done by both the atmosphere and the ocean. The atmosphere can move and respond quickly, so it controls the shorter-term variations in the poleward transport. However, the ocean can carry much more heat than the atmosphere, so it is doing the slower heavy lifting.

The heat is transported by the ocean to the poles in a couple of ways. One is that because the surface waters of the tropical oceans are warm, they expand. As a result, there is a permanent gravitational gradient from the tropics to the poles, and a corresponding slow movement of water following that gradient.

The major movement of heat by the ocean, however, is not gravitationally driven. It is the millions of tonnes of warm tropical Pacific water pumped to the poles by the alternation of the El Nino and La Nina conditions. I described in “The Tao of El Nino” http://wattsupwiththat.com/2013/01/28/the-tao-of-el-nino/ how this pump works. Briefly, the Nino/Nina alteration periodically pushes a huge mass of warm water westwards. At the western edge of the Pacific Ocean, the warm water splits, and moves polewards along the Asian and Australian coasts. Finally, at the poles it radiates its heat to space. Figure 1a from my previous post shows the action of the pump.

nino nina tao triton temp and dynamic heightFigure 1a. 3D section of the Pacific Ocean looking westward alone the equator. Each 3D section covers the area eight degrees north and south of the equator, from 137° East (far end) to 95° West (near end), and down to 500 metres depth. Click on image for larger size.

Figure 1a shows a stretch of the top layer of the Pacific Ocean. It runs along the Equator all the way across the Pacific, from South America (near end of illustration) to Asia (far end of illustration). During the El Nino half of the pumping cycle, which corresponds to the input stroke of a pump, warm water builds up along the Equator as shown in the left 3D section. Then in the La Nina part of the cycle, the pressure stroke, that water is physically moved by the wind across the entire Pacific, where it splits and moves toward both poles.

Now, this El Nino/La Nina pumping action is not a simple feedback in any sense. It is a complex governing mechanism which kicks in periodically to remove excess heat from the tropical Pacific to the poles. As such it exerts control over the long-term energy content of the planet.

So here’s the first oddity about the PDO. The two alternate states of the PDO look very much like the two alternate states of El Nino/La Nina. In both, heat builds up in the eastern tropical Pacific, while the poles are cool. And in both, the alternate situation is where the heat is moved to the poles, residual warmth remains along the coasts of Asia and Australia, and the eastern tropical Pacific is cool.

This is an important observation because in addition to regulating the amount of incoming energy through the timing of the onset of the clouds and thunderstorms, the planet regulates its heat content by varying the rate of “throughput”. I am using “throughput” to mean the rate at which heat is moved from the equator to the poles. When the movement of heat to the poles slows, heat builds up. And when that pole-bound movement speeds up, the heat content of the planet is reduced through increased heat loss at the poles.

The rate of throughput of heat from the tropics to the poles is controlled at different time scales by different phenomena.

On an hourly/daily scale, the variations in the amount of heat moved are all in the atmospheric part of the system. The timing and amount of thunderstorms directly regulate the amount of heat leaving the surface to join the Hadley circulation to the poles.

On an inter annual basis, the throughput is regulated by the El Nino/La Nina pump.

And finally, on a decadal basis, the throughput is regulated by the PDO.

So as a functional definition, I would say that the PDO is a another part of the complex system which controls the planetary heat content. It is a rhythmic shift in the strength and location of the Pacific currents which alternately impedes or aids the flow of heat to the poles.

The Climate Effects of the PDO

As you might imagine, the state of the PDO has a huge effect on the climate, particularly in the nearby regions. The climate of Alaska, for example, is hugely influenced by the state of the PDO.

Nor is this the only effect. The PDO seems to move in some sense in phase with global temperatures. Since the Pacific covers about half the planet, this should come as no surprise.

How We Measure the PDO

The PDO was first measured in salmon catches. Historical records in British Columbia up in Canada showed a clear cyclical pattern … and since then, a number of other ways to measure the PDO have been created. Current usage seems to favor either the detrended North Pacific temperature, or alternately using the first “principle component” (PC) of that temperature. Since the first PC of a slowly trending time series is approximately the detrended series itself, these are quite similar.

To measure the PDO or the El Nino, I don’t like these types of temperature-based indices. For both theoretical and practical reasons, I prefer pressure-based indices.

The practical reason is that we don’t have much information about the North Pacific historical water temperatures. Sure, we have the output of the computer reanalysis models, but that’s computer model output based on very fragmentary input, and not data. As a result it’s hard to take a long-term look at the PDO using temperatures, which is important when a full cycle lasts sixty years or so.

The same issue doesn’t apply as much to pressure-based indices. The big difference is that the pressure field changes much more gradually than the temperature field at all spatial scales. If you move a thermometer a hundred metres you can get a very different temperature. That is not true about a barometer, you get the same pressure anywhere in town.  Indeed, they don’t suffer from many of the problems in temperature based indices, in part because the instruments used to measure pressure are not subject to the micro-climate issues that bedevil temperature records. This means that you can directly compare say the pressure in Darwin and the pressure in Tahiti. So those two datasets are used to construct the pressure-based Southern Ocean Index.

As a result, it is much easier to construct an accurate estimate of the entire pressure field from say a few hundred stations than it is to estimate the temperature field. Indeed, this kind of estimation has been used for many decades before computers to construct the weather maps showing the high and low-pressure areas. This is because the surface pressure field, unlike the surface temperature field, is smooth and relatively computable from scattered ground stations.

The theoretical reason I don’t like temperature based indices is that people always want to subtract them from the global temperature for various reasons. I see this done all the time with temperature-based El Nino indices. It all seems too incestuous to me, removing temperature of the part from temperature of the whole.

The final theoretical reason I prefer pressure-based indices is that they integrate the data from a large area. For example, the Southern Ocean Index (which measures pressures in the Southern Hemisphere) reflects conditions all the way from Australia to Tahiti.

In any case, Figure 2 shows a typical PDO index. This is the one maintained by the Japanese at JISAO. It is temperature based.

monthly values JISAO pdo indexFigure 2. The temperature-based JISAO Pacific Decadal Oscillation Index. It is calculated as the leading principal component of the North Pacific sea surface temperature. 

As I mentioned, for the PDO, I much prefer pressure based indices. Here is the record of one of the pressure-based indices, the “North Pacific Index”. The information page says:

The North Pacific (NP) Index is the area-weighted sea level pressure over the region 30°N-65°N, 160°E-140°W.

NPI per trenberth hurrell

Figure 3. The pressure-based North Pacific Index, calculated as detailed above.

As you can see, the sense of the NP Index is opposite to the sense of the JISAO PDO Index. They’ve indicated this in Figure 3 by putting the red (for warm) below the line and the blue (for cool) above the line, but this doesn’t matter, it’s just how the index is constructed. It moves roughly in parallel (after inversion) with the JISAO PDO Index shown in Figure 2.

Now, for me, both of those charts are totally uninteresting. Why? Because they don’t tell me when the regime changes. I mean, in Figure 3, was there some kind of reversal around 1990? 1950?  It’s all a jumble, with no clear switch from one regime to the other.

To answer these types of questions, I’ve become accustomed to using a procedure that other folks don’t seem to utilize much. I’ve taken some grief for using it here on WUWT, but to me it is an invaluable procedure.

This is to look at the cumulative total of the index in question. A “cumulative total” is what we get when we start with the first value, and then add each succeeding value to the previous total. Why use the cumulative total of an index? Figure 4 shows why:

cumulative monthly north pacific index

Figure 4. Cumulative North Pacific Index (inverted). The data have been normalized, so the units are standard deviations. The cumulative index is detrended, see Appendix for details.

I’ve inverted the cumulative NPI to make it run the same direction as the temperature. You can see the advantage of using the cumulative total of the index—it lays bare the timing of the fundamental shifts in the system.

Now, looking at the Pacific Decadal Oscillation in this way makes it a few things clear.

First, it establishes that there are two distinct states of the PDO. It’s either going up or going down.

In addition, it shows that the shift from one to the other is clearly threshold-based. Until a certain (unknown) threshold condition is reached, there is no sign of any change in the regime, and the motion up or down continues unabated.

But once that (unknown) threshold is passed, the entire direction of motion changes. Not only that, but the turnaround time is remarkably short. After only a few months in each case the other direction is established.

Finally, to me this shows the clear fingerprint of a governing mechanism. You can see the effects of the unknown “thermostat” switching the system from one state to the other.

RECAP

I’ve hypothesized that the Pacific Decadal Oscillation (PDO) is another one of the complex interlocking emergent mechanisms which regulate the temperature and the heat content of the climate system. They do this in part by regulating the “throughput”, the speed and volume of the movement of heat from the tropics to the poles via the atmosphere and the oceans.

These emergent mechanisms operate at a variety of spatial and temporal scales. At the small end, the scales are on the order of minutes and hundreds of metres for something like a dust devil (cooling the surface by moving heat skywards and eventually polewards).

On a daily scale, the tropical thunderstorms form the main driving force for the Hadley atmospheric circulation that moves heat polewards. Of course, the hotter the tropics get, the more thunderstorms form, and the more heat is moved polewards, keeping the tropical temperature relatively constant … quite convenient, no?

On an inter-annual scale, when heat builds up in the tropical Pacific, once it reaches a certain threshold the El Nino/La Nina alteration pumps a huge amount of warm water rapidly (months) to the poles.

Finally, on a decadal scale, the entire North Pacific Ocean reorganizes itself in some as-yet unknown fashion to either aid or impede the flow of heat from the tropics to the poles.

CONCLUSION

So … can the PDO help us to forecast the temperature? Hard to tell. It is sooo tempting to say yes … but the problem is, we simply don’t know. We don’t know what the threshold is which is passed at the warm end of the scale in Figure 4 to turn the PDO back downwards. We also don’t know what the other threshold is at the cool end that re-establishes the previous regime anew. Not only do we not know the threshold, we don’t know the domain of the threshold, although obviously it involves temperatures … but which temperatures where, and what else is involved?

And most importantly, we don’t know what the physical mechanisms involved in the shift might be. My speculation, and it is only that, is that there is some rapid and fundamental shift in the pattern of the currents carrying the heat polewards. The climate system is constantly evolving and reorganizing in response to changing conditions.

As a result, it makes perfect sense and is in accordance with the Constructal Law that when the sea temperature gradient from the tropics to the poles gets steep enough, the ocean currents will re-organize in a manner that increases the polewards heat flow. Conversely, when enough heat is moved polewards and the tropics-to-poles heat gradient decreases, the currents will return to their previous configuration.

But exactly what those reversal thresholds might be, and when we will strike the next one, remains unknown.

HOWEVER … all is not lost. The reversals in the state of the PDO can be definitively established in Figure 4. They occurred in 1923, 1945, 1976, and 2005. One thing that we do NOT see in the record is any reversal shorter than 22 years (except a two-year reversal 1988-1990) … and we’re about eight years into this one. So acting on way scanty information (only three intervals, with time between reversals of 22, 31, and 29 years), my educated guess would be that we will have this state of the PDO for another decade or two. I’ve sailed across the Pacific, it’s a huge place, things don’t change fast. So I find it hard to believe that the Pacific could gain or lose heat fast enough to turn the state of the PDO around in five or ten years, when we don’t see that kind of occurrence in a century of records.

Of course, nature is rarely that regular, so we may see a PDO reversal next month … which is why I say that tempting as it might be, I wouldn’t lay any big bets on the duration of the current phase of the PDO. History says it will continue for a decade or two … but in chaotic systems, history is notoriously unreliable.

w.

PS—This discussion of pressure-based indices makes me think that there should be some way to use pressure as a proxy for the temperature. This might aid in such quests as identifying jumps in the temperature record, or UHI in the cities, or the like. So many drummers … so little time.

MATH NOTE: The shape of the cumulative total is strongly dependent on the zero value used for the total. If all of the results are positive, for example, the cumulative total will look much like a straight line heading upwards to the right, and it will go downwards to the right if the values are all negative. As a result, it cannot be used to determine an underlying trend. The key to the puzzle is to detrend the cumulative total, because strangely, the detrended cumulative total is the same no matter what number is chosen for the zero value. Go figure.

So I just calculate the trend starting with the first point in whatever units I’m using, and then detrend the result.

Advertisements

  Subscribe  
newest oldest most voted
Notify of
jcspe

yeah, but will we catch more or less fish?

The key question: what tips the pendulum?

Janice Moore

Good questions asked, here, Mr. Eschenbach, and plausible conjecture on your part.
I can’t give you any answers (except that God is an amazing designer!). But, I don’t need to understand the PDO to realize that it is, indeed, wonderful.
Thank you for sharing and opening yourself up, once again, to both kindly, constructive, criticism by WUWT scientists of integrity and also, inevitably, I’m afraid, to the disingenuous, thoughtless, harsh, attack of less gracious (and, often, less intelligent) souls.
You are a brave man! A fine spirit.

Willis Eschenbach

Anthony Watts says:
June 8, 2013 at 9:06 pm

The key question: what tips the pendulum?

Indeed. As I said, it has to be temperature related, but what temperature, and where, and what else is involved? Gotta love settled science … only thing for sure is that CO2 isn’t directly involved.
w.

Werner Brozek

Anthony Watts says:
June 8, 2013 at 9:06 pm
The key question: what tips the pendulum?
Could it be plasma speed from the sun? Recent large El Ninos were 1987, 1998 and 2010. Check out the low plasma speeds each time at:
http://snag.gy/UtqpX.jpg

Gary Hagland

Question for Mr. Eschenbach; do you have a good source on PDO phases affecting salmon catch? Salmon return is a controversial issue here in the Pacific Northwest as the tribes are using it as a pretext in an attempt to deny landowners access to their well water. Instream flow rules and all that. Since the PDO has flipped to its cool phase, should we expect larger catch and return or is it the other way around. Noted that the coho return was much larger than predicted last year.

Manfred

The positive PDO warm water off North America up to Alaska is leftover warm water from previous El Ninos. Those water pools carry tropical fish up to Alaska.
http://www.elnino.noaa.gov/enso4.html
The bottom line:
PDO is the low frequency tail of ENSO.
http://www.esrl.noaa.gov/psd/people/gilbert.p.compo/CompoSardeshmukh2008b.pdf
“Because its [ENSO’s] spectrum has a long low frequency tail, fluctuations in the timing, number and amplitude of individual El Nino and La Nina events, within, say, 50-yr intervals can give rise to substantial 50-yr trends…”
“…It [The Pacific decadal oscillation or the interdecadal Pacific oscillation] is strongly reminiscent of the low-frequency tail of ENSO and has, indeed been argued to be such in previous studies (e.g. Alexander et al 2002, Newman et al 2003, Schneider and Cornuelle 2005, Alexander et al 2008)…”
“…it also accountd for an appreciable fraction of the total warming trend…” (see figure 9b )
The question is then, what tips ENSO ?

jorgekafkazar

Anthony Watts says: “The key question: what tips the pendulum?”
As with El Niño/La Niña, there’s a mouse running up and down the back of the pendulum at irregular intervals. The periodicity of the pendulum changes when the mouse moves, making it very difficult to associate the pendulum swinging the other way with any given potential cause, since we can’t see or identify the mouse.

Drave Robber

Is the allusion in the headline intentional? 😉

Girma

Excellent post Willis.
Thank you.
If you can, please do a similar one on the AMO.

davidq

The only thing I can remember is that the Winter 76/77 was fantastic. The best skiing snow. We could ski out of our kitchen down the farmers field infront of us until late March. Didn’t happen again. The years before that the winters were short and often “green”. After that it was icy and miserable.
However, the point is, that the switch could be anywhere in the system. But, for sure, we felt that switch in northern Europe, when I was young.

Girma

Willis
It all seems too incestuous to me, removing temperature of the part from temperature of the whole.
But you are forced to do that when you want to decompose the global mean temperature into secular and cyclic components.

Willis wrote:

“As a result, it makes perfect sense and is in accordance with the Constructal Law that when the sea temperature gradient from the tropics to the poles gets steep enough, the ocean currents will re-organize in a manner that increases the polewards heat flow. Conversely, when enough heat is moved polewards and the tropics-to-poles heat gradient decreases, the currents will return to their previous configuration.”

The Drinking Bird heat engine pendulum comes to mind…in slow motion over decades.

All the same elements in play: heat, evaporation, condensation, temperature differential, liquid flow, gas laws, Maxwell-Boltzman distribution.
http://en.wikipedia.org/wiki/Drinking_bird

jai mitchell

If the el nino and la nina don’t cause the planet to heat up or cool down, would it be safe to say that during an el nino, more ocean heat is transported to the air and during a la nina more air heat is transported to the oceans? It seems that if the seas are colder then they will receive more warming.

Fascinating! So it’s not just water, atmosphere, etc, that are essential for life-as-we-know-it on earth, but a working mechanism for moving heat from tropics to poles that keeps temperatures within narrow limits. Amazing.

In_A_Nut_Shell

[Snip – more Slayer junk science from Doug Cotton]

Janice Moore

Jai Mitchell! LOL. When do you sleep?
Here is some “transporting” information for you (just for fun):
(note: blue for La Nina and red for El Nino — “terrible way to travel, spreading a man’s molecules all over the universe” — Bwah, ha, ha, ha, haaa!)
http://www.bing.com/videos/search?q=Star+Trek+transporter&view=detail&mid=DD73ED6F768194D510AADD73ED6F768194D510AA&first=0&FORM=NVPFVR
Be well, O Jai. Live long and prosper.

Jon

The temperature gradient creates thermal wind and jet streams and pressure differences. And most of the sea currents are mostly wind driven are they not?

Anthony Watts says: June 8, 2013 at 9:06 pm
“The key question: what tips the pendulum?”
At least in mid 20th Century, at about 1943 a main contributor could have been the commencement of the naval war in the Pacific, discussed in Chapter H: “Pacific War, 1942-1945, contributing to Global Cooling?” (about 12 pages ) at: http://www.seaclimate.com/h/h.html .
Kindly pay particular attention to Fig. H-14 (based on Rundenov and Bond, 2004) showing that the PDO-shift in 1943 happened without any delay, while the subsequent shift about 40 years later, happened earlier in winter (ca. 1889), and years later in summer (ca. 1998), as shown in the image here: http://www.seaclimate.com/h/images/buch/big/h-14.jpg

Girma


I’ve inverted the cumulative NPI to make it run the same direction as the temperature. You can see the advantage of using the cumulative total of the index—it lays bare the timing of the fundamental shifts in the system.

Your result shown in Fig 3 can be arrived at as follows:
http://www.woodfortrees.org/plot/hadcrut4gl/mean:60/detrend:0.8/from:1880

Leonard Lane

Nice post Willis, thanks. In the case of sea breezes flowing from the sea to the land and then reversing to flow from the land to the sea, do you know how quickly these turn around? Or are the sea breezes and PDO’s so different that such comparisons cannot be made?

Girma


Not only do we not know the threshold, we don’t know the domain of the threshold, although obviously it involves temperatures … but which temperatures where, and what else is involved?

We do know.
Here is how:
http://www.woodfortrees.org/plot/hadcrut4gl/compress:12/from:1880/plot/hadcrut4gl/from:1880/to:2012/trend/plot/hadcrut4gl/from:1880/to:2012/trend/offset:0.25/plot/hadcrut4gl/from:1880/to:2012/trend/offset:-0.25/plot/hadcrut4gl/scale:0.00001/offset:2/from:1880
The global mean temperature can move from through to peak by not more than about 0.5 deg C as shown. The threshold is a warming of 0.5 deg C or a cooling of 0.5 deg C.
The ocean has enormous heat capacity and inertia. It is like a moving tanker. Once its trajectory is established, it changes little with time. Expect the pattern in my link above to continue for several decades.

CodeTech

Jon – a quote from wiki:

An ocean current is a continuous, directed movement of ocean water generated by the forces acting upon this mean flow, such as breaking waves, wind, Coriolis effect, cabbeling, temperature and salinity differences and tides caused by the gravitational pull of the Moon and the Sun.

also:

Surface ocean currents are generally wind-driven…..
Deep ocean currents are driven by density and temperature gradients.

Girma

The best study of the PDO ever for me. Using presure data is great science. Thanks Willis for not pretending to know the answer and stating the qustion so clearly. We will all be thinking and calculating and searching for the answer. I hope someone smarter than me will post that answer. Meanwhile, just knowing the trend with long term data to support the length of the moves will help long term outlooks. Thank you, sir.

Girma

They occurred in 1923, 1945, 1976, and 2005.
I belive it is instead:
1909, 1941, 1973, 2005 and hopefully 2037!
http://www.woodfortrees.org/plot/hadcrut4gl/mean:60/detrend:0.8/from:1880

James

I saw figure 4 and my mouth dropped open, because it looks exactly like this graph: http://www.climate4you.com/GlobalTemperatures.htm#Cyclic air temperature changes
just lagged by about 5 years. It was such a lightbulb moment. I see you guys are already on this, but the climate4you version seems so much clearer to me. I’m still reeling from how clear it is.
It does seem to breakdown prior to 1920 but that may be a data issue.

Willis Eschenbach

Girma says:
June 9, 2013 at 12:07 am

They occurred in 1923, 1945, 1976, and 2005.

I belive it is instead:
1909, 1941, 1973, 2005 and hopefully 2037!
http://www.woodfortrees.org/plot/hadcrut4gl/mean:60/detrend:0.8/from:1880

Thanks, Girma. Regarding your recent posts, perhaps you didn’t notice, but I’m talking about the PDO, and you are talking about the HadCRUT4 temperature record … your data is interesting, but it says absolutely nothing about the PDO.
w.

Girma

Willis
Are not the PDO and HadCRUT4 closely related?

Willis Eschenbach said @ June 9, 2013 at 12:28 am

Thanks, Girma. Regarding your recent posts, perhaps you didn’t notice, but I’m talking about the PDO, and you are talking about the HadCRUT4 temperature record … your data is interesting, but it says absolutely nothing about the PDO.

Perhaps not absolutely nothing. I was quite taken by the NASA JPL paper linking Nile floods with Aurora Borealis over several centuries. Rainfall in Africa is strongly linked to PDO and obviously Nile floods depend on that rainfall.

The N. Atlantic oscillation has some ‘resonance’ with geological events there; these are also plentiful in the Pacific, it appears that may be a similar link to the atmospheric pressure oscillation (southern oscillation index).
http://www.vukcevic.talktalk.net/SOI.htm

I wonder what the sunspot index would look like if cumulated? Presumably it would be necessary to choose the starting point carefully. Also it would be wise to use the corrections proposed by Lief Svalgaard.

Willis Eschenbach

Girma says:
June 9, 2013 at 12:53 am

Willis
Are not the PDO and HadCRUT4 closely related?

Related, yes. The same, no. Your claim that the reversal dates of the PDO were wrong and should be replaced by dates related to HadCRUT4 reveals a misunderstanding. They are not the same, and no, you can’t claim that dates relating to one should replace dates relating to the other.
The Pompous Git says:
June 9, 2013 at 1:07 am

Willis Eschenbach said @ June 9, 2013 at 12:28 am

Thanks, Girma. Regarding your recent posts, perhaps you didn’t notice, but I’m talking about the PDO, and you are talking about the HadCRUT4 temperature record … your data is interesting, but it says absolutely nothing about the PDO.

Perhaps not absolutely nothing. I was quite taken by the NASA JPL paper linking Nile floods with Aurora Borealis over several centuries. Rainfall in Africa is strongly linked to PDO and obviously Nile floods depend on that rainfall.

Note the response above. I was not speaking theoretically. I was speaking about Girma’s data regarding reversal dates, which he claimed should replace the actual reversal dates of the PDO …
vukcevic says:
June 9, 2013 at 1:14 am

The N. Atlantic oscillation has some ‘resonance’ with geological events there; these are also plentiful in the Pacific, it appears that may be a similar link to the atmospheric pressure oscillation (southern oscillation index).
http://www.vukcevic.talktalk.net/SOI.htm

Thanks, Vuk. The SOI is an entire story into itself. I had a couple of free hours last week, so I constructed an “NOI” based on the SOI. The SOI looks at the pressure difference Tahiti to Darwin, Australia. For the NOI, I’ve used the exact same technique to relate the pressures in Tahiti and Tokyo. Remember that the PDO affects both oceans. There are interesting differences in the timing of the reversal in the South Pacific as opposed to the North.
w.

lgl

Girma
“If you can, please do a similar one on the AMO.”
I’ve done that one too long time ago, http://virakkraft.com/NAO-AMO.png
(in a discussion with Vuk)

Girma

Willis
Is there any relationship between PDO and the great conveyor belt?
What is your current understanding on the roll of the conveyor belt on global mean temperature?

JJM Gommers

It ressemblence a Belousov-Zhabotinsky process and it’s thermodynamically driven. You would expect to find the anthropogenic fingerprint in it besides the natural variablility. Time lapse , deflection and treshold value should be effected. Or is the human impact to small to find.

It comes down to the state of the polar cells: the colder they are, the denser their high pressure meanders. Like there are rivers of low pressure travelling polewards, so there are meanders of pressure travelling equatorwards. The colder, the denser, the drier the air that reahes the Hadle and the higher the SST the more the downwelling from aloft, the stronger the trades. Indeed it is an expression of the meridional thermocline, but can only be understood when you understand that the PDO and the AMO moves in phase with the AO. Solarcycle length seems to determine some of the rhythm, probably by regulating low level clouds in polar and subpolar atmospheres.
WIllis: you state that heat is radiated at the poles, which is obviously true, but you seem to neglect, that the Ferrell cell is convecting tremoundous amounts of energy far aloft, just as the thunderstorms under the Hadley regime. I suspect that this renders co2 neglegtible, but probably not ozone, which may be a real driver for polar climates under changing UV regimes.
Just my 2 cents, which are definately not expert. 🙂
Per

Hi Willis
I enjoyed greatly your stories from Solomons, a feel of Hemingway, if I may say so.
Solomons and SOI ? there could be a lot more to it.

Girma
If you can, please do a similar one on the AMO.
North Atlantic (atmospheric) Oscillation –NAO (or some of its components) and the AMO are closely related. I did a detailed analysis (personally encouraged by Dr. J.. Curry), if you follow this link you will find lot of info in there

Girma

Willis
Could we say the following:
Increase in the surface ocean current speed from the equator to the poles results in global cooling. Decrease in the surface ocean current speed from the equator to the poles results in global warming.

Girma

vukcevic
Thank you for the link.

Greg Goodman

Another excellent insight Willis. The cumulative integral or cumulative distribution function (CDF hereafter ) is indeed quite revealing. IIRC this was used by Hirst as a means of detecting ‘regime changes’ in Nile flood data
I used it in the volcano stack plots where I also removed a linear function. Some explanation of what it is why it is legit to remove the linear slope was given below the plot.
http://climategrog.wordpress.com/?attachment_id=285
Some of that is applicable here, so I’ll adapt it to your pressure-PDO .
Firstly, why this works is that integrals in general are low pass filters, so they take out the fast changes and leave the long term behaviour. Now the trouble with cumulative integral is that it is not well behaved, constant filter. It filters more and more heavily as it goes along. It’s a variable length filter not one that does the same thing to all the data like kernel based convolution filters. It is more comparable to an iteratively defined filter which requires some ‘spin up’ period before it stabilises. It’s crude but it works. This needs to be born in mind when looking at the output.
For example the much larger swing at the beginning is not (necessarily) because climate was change faster of with a larger swing, it’s because there’s not much ‘ballast’ in the accumulating kitty, so changes make a bigger difference.
Now what is not always obvious is that a straight line slope in such a plot represents a constant . Clearly much of this record is dominated by essentially constant values of the pressure_PDO . Much of the record seems dominated by one of two values which on this representation are roughly equal in magnitude.
Since the “detrending” operation, which represents removal of a constant value from the index, is arbitrary it may well be useful to chose a detrending value that makes the two slopes equal, thus using it to _define_ the neutral point of the pressure_PDO index.
So I would say hat’s off to Willis, I think you have defined a useful index and shown that PDO is not in fact an oscillation but another bipolar state in climate.
I see two notable features in this plot straight away. Firstly, the drop around 1990 was well under way before Mt. Pinatubo eruption as I pointed out in the volcano stack analysis.
Secondly, the steep jump around 1940 corresponds to the steep jump in SST that Hadley Centre decided was a sampling error and removed 0.5 K from the remainder of the climate record.
Thirdly, the early 20th c. rise is almost identical to the later rise. Not much evidence of a planet threatening AGW effect in this index.

johnmarshall

Climate is cyclic with many drivers all of which are cyclic but with different cycle lengths. Sometimes these cycles are in phase, sometimes out of phase so the end product, the climate cycle, can occasionally look anything but cyclic carrying temperature piggyback so this has great variations giving some the impression of a tipping point where none exists. Just the cycle operating as it does, with wide variations.

Willis writes: “So here’s the first oddity about the PDO. The two alternate states of the PDO look very much like the two alternate states of El Nino/La Nina.”
There are a number of reasons for this.
First, the maps you’ve used from JISAO presents most of the Pacific, but the PDO is not calculated from the sea surface temperature anomalies of the entire Pacific. The PDO is determined from (and represents only the spatial pattern of) the sea surface temperature anomalies of the North Pacific, north of 20N—basically from Hawaii north. I marked up the JISAO maps in the following illustration:
http://i53.tinypic.com/opzpqh.jpg
It’s from this post:
http://bobtisdale.wordpress.com/2011/06/30/yet-even-more-discussions-about-the-pacific-decadal-oscillation-pdo/
Second, it has to be kept in mind that the PDO represents the spatial pattern of the sea surface temperature anomalies of the North Pacific north of 20N—not the sea surface temperature anomalies themselves. The sea surface temperature anomalies of the North Pacific north of 20N are actually inversely related to the PDO. I presented that here:
http://bobtisdale.wordpress.com/2010/09/14/an-inverse-relationship-between-the-pdo-and-north-pacific-sst-anomaly-residuals/
That inverse relationship impacts the statement in your post: “Since the first PC of a slowly trending time series is approximately the detrended series itself, these are quite similar.”
Third, because the PDO represents the spatial pattern of the sea surface temperature anomalies of the North Pacific north of 20N, it is dependent on ENSO, which is the dominant process in the Pacific. In other words, El Niño and La Niña events are the primary causes of the spatial patterns in the North Pacific. El Niño events create the pattern where it’s warm in the eastern North Pacific but cool in the west and central portions of the North Pacific—and the opposite pattern is created in response to La Niña events. But also keep in mind that, while the PDO represents the dominant spatial pattern in the North Pacific, there are other spatial patterns; the PDO pattern simply occurs most often. Also, the spatial pattern in the North Pacific is different during El Niño Modoki than it is during a full-blown east Pacific El Niño event.
Fourth, the reason the PDO has a different pattern in time than ENSO is because the spatial pattern of the sea surface temperature anomalies in the North Pacific is also impacted by the sea level pressure in the North Pacific. The sea level pressure of the North Pacific, and the wind patterns associated with it, can resist or enhance the poleward migration of warm water poleward from the tropics.
Fifth, the other thing to keep in mind about the PDO: it has been standardized—divided by its standard deviation. That is, the values of the leading principal components of North Pacific sea surface temperature residuals are much smaller than the values presented by JISAO:
http://i53.tinypic.com/2yjxydk.jpg
Or to phrase it another way, the JISAO PDO index exaggerates the variations in the North Pacific by about 5.6 times. Refer to the discussion in the following post under the heading of DOES THE PDO DATA EXAGGERATE ITS RELATIVE SIGNIFICANCE?:
http://bobtisdale.wordpress.com/2011/06/30/yet-even-more-discussions-about-the-pacific-decadal-oscillation-pdo/
Willis, you wrote: “Since the Pacific covers about half the planet, this should come as no surprise.”
You’ve exaggerated a little here. The Pacific Ocean covers about one-third. Surface area of the Pacific = 165.2 million km^2. Surface area of Earth = 510 million km^2.
Regards.

Thanks for taking so much time to clarify this issue Willis.
My takeaway is that since the PDO is chaotic, it doesn’t have predictive value – but based on the short series of observations to date, it seems likely the current trend will continue for at least another decade or two – but how likely, we haven’t got enough data to say.
Regards,
Eric

lgl

“So I would say hat’s off to Willis, I think you have defined a useful index”
Except I defined that “index” years ago:
http://wattsupwiththat.com/2011/12/17/frank-lansner-on-foster-and-rahmstorf-2011/#comment-836884

Rhys Jaggar

A few questions arise:
1. Can you do the same with pressure for the Atlantic Multidecadal Oscillation?
2. Can you explain the lag between the Pacific and Atlantic oscillation phases through a heat transfer process from the Western Pacific through the Southern Atlantic to the North Atlantic or not?
3. What role does the global deep ocean circulation pattern have to play in all of this??

Neville.

But what about NOAAs reconstruction of the PDO for the last 1000 years? Why did the the two different phases remain locked in for such long periods during the early part of the record? http://en.wikipedia.org/wiki/File:PDO1000yr.svg
I think I’ve asked this before but never seem to get an answer. I just wish Willis or Bob or anyone could have a go.
We know about the mega droughts that affected the west coast of USA and into Canada at that earlier period and there seems to be evidence of very wet periods over eastern Australia at the same time.

jai mitchell says: “If the el nino and la nina don’t cause the planet to heat up or cool down…”
They do cause the planet to heat up and cool down. The sea surface temperatures for the entire East Pacific ocean mimics the variations in the tropical Pacific.
http://oi47.tinypic.com/hv8lcx.jpg
All of the left over warm water from El Niño events, on the other hand, cause the sea surface temperatures of the Atlantic, Indian and West Pacific, to effectively shift upwards in response to strong El Niños:
http://oi49.tinypic.com/29le06e.jpg
jai mitchell says: “…would it be safe to say that during an el nino, more ocean heat is transported to the air and during a la nina more air heat is transported to the oceans?”
Yes and no. An El Niño releases more heat than normal from the tropical Pacific to the atmosphere. That occurs primarily through evaporation. During an El Niño, there is more warm water covering the surface of the tropical Pacific, which causes more water to evaporate from its surface. When the evaporated water condenses again and comes out as rain, it heats the atmosphere. A La Niña, on the other hand, releases less heat than normal from the tropical Pacific ocean to the atmosphere.
The recharge of ocean heat in the tropical Pacific during a La Niña is a function of cloud cover and sunlight. Because there is less evaporation during a La Niña, there is less cloud cover. Less cloud cover means more sunlight can enter and warm the tropical Pacific. This recharges (or replenishes) the heat released during the El Niño.
Regards

jai Mitchell: Oops, forgot. For a further explanation, see here (54mb):
http://bobtisdale.files.wordpress.com/2013/01/the-manmade-global-warming-challenge.pdf
Regards