Guest Post by Willis Eschenbach
A while back, folks noticed that a couple of months after the El Nino kicked in across the Pacific, the earth would warm up a bit. Since then, people have engaged in what they describe as “removing the El Nino signal” from the global temperature record. A while back I wrote a post called “Why El Nino and not the AMO“. If you have not read that post, it lays out some of my objections to the procedure of “removing the El Nino signal”. This post carries those ideas forward.
I got to thinking about just how well the El Nino 3.4 area does or doesn’t forecast the evolution of the global temperature, and I realized that I could look at that question using the CERES satellite data. Figure 1 shows the correlation of local temperatures with the global average temperature two months later:
Figure 1. Correlation of local temperatures with the global temperature two months later. All observations have had the monthly variations removed. Blue rectangle outlines the El Nino 3.4 area from 120° West to 170° West and extending 5° north and south of the Equator. Data from CERES Mar 2000 – Feb 2014.
As you can see, while the El Nino 3.4 Index will give us an idea of future global temperatures, it is by no means the only area which can do so. In fact, the red area in the tropical Atlantic is better correlated with future global temperatures than is the El Nino 3.4 area.
Having seen that, I thought, well, what areas are the best at predicting the temperature of just the Northern Hemisphere two months ahead? Figure 2 shows that result.
Figure 2. Correlation of local temperatures with the northern hemisphere temperature two months later. Blue rectangle outlines the El Nino 3.4 area. Data from CERES Mar 2000 – Feb 2014.
This is most interesting. While the El Nino 3.4 area is barely better than random at predicting the NH temperature, the area in the Atlantic is quite strongly correlated with the future NH temperature.
And the Southern Hemisphere? Figure 3 below has those results:
Figure 3. Correlation of local temperatures with the southern hemisphere temperature two months later. Blue rectangle outlines the El Nino 3.4 area. Data from CERES Mar 2000 – Feb 2014.
I do enjoy surprises. I’d expected somewhere in the Southern Hemisphere to have the greatest correlation with . Instead, it turns out to be exactly centered on the El Nino 3.4 area outlined by the blue rectangle.
Now, it does strike me that we could do a passable job at predicting the global temperatures a couple of months ahead using nothing more than those last two maps for the northern and southern hemispheres. Might make an interesting and possibly profitable project for someone with more time on their hands than I have.
But I’m still leery of subtracting those predictions from the actual data and calling it “removing the El Nino effect” for a couple of reasons. First, “removing the El Nino” doesn’t help us with the Northern Hemisphere temperature. To the contrary, it just adds noise.
But more importantly, the El Nino/La Nina pumping action is not a true cause of changing temperatures in the sense that a volcanic eruption or a change in the amount of top-of-atmosphere total solar insolation is a cause of changing temperatures. Instead, the fluctuations in oceanic temperatures are in turn a result of some previous condition. I term this a “chain of effects”. For example:
Figure 4. Correlation of local temperatures with the El Nino 3.4 area two months later. Blue rectangle outlines the El Nino 3.4 area. Data from CERES Mar 2000 – Feb 2014.
In Figure 4 we can see that high temperatures in the El Nino 3.4 region are in turn presaged by low temperatures off of Australia and the Philippines (blue areas). As I said, the El Nino/La Nina is not a cause. It’s just part of a chain of effects.
And as a result, yes, you can pick out the red area in the tropical Atlantic and “remove its effects” from the temperature … but what is it that you end up with? Or you could “remove” both the tropical Atlantic and the El Nino variation … and “remove” the PDO, and the QBO as well … but what is it that remains once you’ve done that?
As always, my best wishes for everyone on what is a lovely rainy day here …
w.
My Usual Request: If you disagree with me or anyone, please quote the exact words you disagree with. I can defend my own words. I cannot defend someone’s interpretation of my words.
My Other Request: If you think that e.g. I’m using the wrong method on the wrong dataset, please educate me and others by demonstrating the proper use of the right method on the right dataset. Simply claiming I’m wrong doesn’t advance the discussion.
“But I’m still leering of subtracting those predictions”
leery?
Thanks, clipe, fixed.
w.
Hahaha…and people say that climate study isn’t ‘sexy’.
My guess is that your graphs change significantly if you replace “correlations between temperatures” with “correlations between temperatures after subtracting the global trend”. If that’s the case, if might be interesting to see the alternative graphs, which should be easy to compute.
For the curious, my concern relates to the fact that “mean effects” can be important. Take two timeseries of interest and compute their correlation. Depending on the autocorrelation, when you correlate the timeseries after subtracting their respective mean innovations, you might find substantially different answers. A point to watch out for is that the variance of a timeseries can be dominated by an explanatory factor with a long autocorrelation length (perhaps several years). It might be impossible to establish statistical significance for the long-autocorrelation component of the variation, because only a few decades (say 5 to 10 effective independent datapoints) are available. So instead you should ask about the explanatory power of the remaining variables for predicting the residual variance. But in most climatology papers I see little or no analysis of this issue. What exactly is being predicted, how many independent data points are available after accounting for autocorrelations, and after accounting for these issues, what is the statistical signficance of the claimed relationship? That is all statistics 101, so any paper that doesn’t directly address such concerns should probably be discarded as just another ignorant, uninformative contribution to the literature.
Sam December 17, 2015 at 7:20 pm
Thanks, Sam. Seems doubtful, as there is little trend over the period (May 2000 – Feb 2014).
w.
Sorry, to clarify, by “trend” on date D I mean the trend estimated only using the data prior to date D. Does that change your conclusions?
The maps show correlation with what happened 2 months later. The global trend, even according to the linear trend in the latest version of GISS for 1979-onward, is about .165 degree C per decade, which means .00275 degree C in 2 months. These maps are concerned with changes a couple orders of magnitude larger.
I wonder how El Nino correlates with 12(?) months previous.
IMO, that would be a very difficult guess as no one knows when the trade winds will slacken, stop, or reverse direction.
My how the climate of the earth is so sensitive to just a little warm water in the South Pacific. It makes one wonder as to why this same earthly climate has been so insensitive to an increase of 33% of the master control knob of climate – CO2.
El Nino and La Nina increase and decrease coverage of the ITCZ with convection to the tropopause from the tropics, and shift where in the world that happens.
LOL very well put.
“how the climate of the earth is so sensitive to just a little warm water…has been so insensitive to an increase of 33% of the master control knob of climate – CO2”
FJ, It’s my understanding that El Nino is redistribution of heat, where CO2 is proposed to increased heat.
And I’m pretty sure that it’s far from settled that CO2 is the “master control knob of climate”. BTW, that “little warm water” brought noticeable changes to our temperatures a lot faster than CO2 has ever been blamed for.
I think the Nino 3.4 region does not tell everything about El Nino. Most of the Nino 3 region that is not in the 3.4 region has almost as much correlation with lagging SH temperature and leading opposite temperature east of the Philipines and northern Australia as the Nino 3.4 region has. This comes to my mind because in the 1997-1998 El Nino, the part of the Nino 3 region that is not in the Nino 3.4 region warmed more than it did in the current El Nino.
Agreed but fortunately we have other, partially independent measures to look at like SOI and MEI. How do such indicators affect your own analysis?
Sadly, currently NOAA uses Nino 3.4 SST anomaly as its official ENSO index. Meanwhile, the MEI shows the 1982-1983 El Nino as a very strong one, and as for its effects on USA weather were as major as MEI indicates. I attempted web research of what regions of the Pacific are considered for SST anomaly, and they don’t make it easy for me to find this quickly. But it seems to me that the Nino 1+2, 3 and 4 regions are all considered for MEI. I found one paper saying that the Nino 3 SST correlates better with MEI than every other individual component of MEI, IIRC with 90% correlation.
Once cite: http://onlinelibrary.wiley.com/doi/10.1002/j.1477-8696.1998.tb06407.x/pdf (which for me loads incompletely for some reason, but loads enough for me to see some useful stuff)
As for parts of the western tropical pacific having negative correlation with Nino 3.4 region temperatures 2 months later: I think that wind variations in the western tropical Pacific are a major contributor to what happens farther east in the tropical Pacific 2 months later.
I just noticed that the color coded maps only cover a 14 year period starting in March 2000. The great El Nino of 1997-1998 and the very strong one of 1982-1983 and the very strong strong La Nina of 1984-1985 are not included – because there was not much use of CERES instruments before 2000. What happened so far with the current strong El Nino is also excluded.
The Atlantic / NH connection was a good idea, I don’t recall having seen that anywhere before in term of correlation.
Willis:
Indeed. The whole idea of wanting to subtract it is based on the unfounded *assumption* that it is some kind of net zero “internal oscillation” : just ‘noise’ on top of the AGW signal.
I may be informative to do a similar exercise with negative lag, ie is there any areas which are precursors to El Nino or the tropical Atlantic changes. In fact correlations should be inspected in both directions for at least a year. An animation of that would be a quick way to eyeball what is happening.
TOA energy budget drives tropical SST with a lag of about 13mo. The feedback of SST to TOA peaks at about 3mo in the other direction IIRC.
That last bit was the case following Mt P , I think Dr Spencer produced similar lag relation using the CERES data.
Spencer & Baswell 2011:
http://www.mdpi.com/2072-4292/3/8/1603/pdf
Willis, you non-reader! If you had read my book you would know that the reason Nino3.4 has any predictive powers is that it sits smack dab in the middle of the equatorial counter-current and watches all the El Nino waves go by. They are on the way from the Indo-Pacific Warm Pool to the South American coast. The reason there is a lag time between Nino3.4 and atmospheric warming is that Nino3.4 records the passage of the El Nino wave in the middle of the Pacific but there is still half an ocean for it to go before it hits South America. Once it gets there it spreads out north and south along the coast and warms the air above it. This warm air then rises, joins the westerlies, and the world notices the arrival of an El Nino peak. But this is not the end of it. Any wave that runs ashore must also retreat. As an El Nino wave retreats water level behind it drops by as much as half a meter, cold water from below rises up to fill the vacuum, and a La Nina has started. As much as the El Nino warmed the atmosphere the La Nina that follows will now cool it and the mean SST will remain the same. An El Nino wave just causes a temperature oscillation about the mean but cannot change the mean itself. If you have a full record of an ENSO wave train you can follow the water temperature change too. Just mark the midpoints of an El Nino peak and its neighboring La Nina valleys and connect the dots. It is is stupid to get rid of the El Nino factor because you are destroying this information. Knowing the locations of the El Nino peaks also helps to interpret the so-called volcanic cooling phenomena which actually don’t cool anything as the book explains.
Arnoarrak, Book / you just sold one
Thanks for the concise clarification, the best I came across (imho); but what is the cause initiating the Indo-Pacific Warm Pool’s wave?
I suspect it may be something to do with the tectonics of the region (need to update the graph
arnoarrak December 17, 2015 at 10:53 pm
Thanks, Arno. Mmmm … well, perhaps, but if that is the case then why is the Nino3.4 area so ineffective in predicting the NH temperatures?
And by the same token, if that is the case then why does the northern tropical Atlantic have any predictive power at all? It certainly doesn’t “sit smack dab in the middle of the equatorial counter-current“, it doesn’t have any two-month lagged waves going on … and despite that the tropical Atlantic is more correlated with future temperatures, both global and NH alone, than is the Nino3.4 area. How does your detailed theory explain that?
Nature is rarely simple …
Regards,
w.
PS—If I read every book I’d never have time to do my own investigations …
North Atlantic SST variability effect on the N. Hemisphere (and consequently global) temperatures depends on the polar jet stream orientation, which (I think) is determined by two semi-permanent atmospheric low pressure systems (Icelandic and Aleutian) both ‘correlating’ to the tectonics of the corresponding regions. The Enso (I think) is basically neutral, and has very little if any effect on the polar jet stream.
What concentrates heat in the tropical East of Indonesia so it can roll across the Pacific as a warm blob that seems to get larger and hotter underwater as it travels East?
I think that the Solomon Sea is the key to the ENSO, it is relatively shallow water, one of the tectonically most complex and very active regions I have come across in my very limited research.
http://volcano.oregonstate.edu/vwdocs/volc_images/southeast_asia/papua_new_guinea/PNG1.jpg
Ocean currents here (in contrast to the relatively stable 10Sv in the Arctic-North Atlantic outflow) are highly variable (normally + – ~2-3 Sv, but occasionally grow strongly, e.g. up to 10 Sv during 1997/8 ENSO).
Are these changes in the flow ‘forced’ by local tectonics?
Degree of correlation suggest ( see graph that that may be the case.
Geoff , this is a common misreading of what is shown in equatorial vertical section graphs and animations.
What is usually plotted is a temperature anomaly . This means that that a particular red blob is warmer at any particular spot is for that time of year. So a red blob that travels eastwards in an animation does not necessarily represent a movement of a body of water in that direction.
In a similar way a surface wave may propagate and we perceive it as moving but the physical movement of water is a local oscillation and does not involve continued lateral displacement.
Also is should be remembered tha oceans are 3D not 2D. There is a N-S displacement of water that contributes to the “anomaly” we see in the vertical cross-section or the SST maps.
El Ninos and La Ninas often don’t balance each other out in the short term, and the imbalance between the two can vary over decades. Have a look at the graph of the Multivariate ENSO Index at: http://www.esrl.noaa.gov/psd/enso/mei/
Good description of the processes, but the implied assumption of symmetry in the phases is faulty.
Nice Willis. Presents a lovely fact about the climate that I didn’t know, and does it in a simple and elegant fashion with pretty pictures. Thanks.
In usefulness, it ranks right alongside the claims the deep ocean is warming and so…global warming in spite of the hiatus. Even if it were true, then so what? Surface temperature is what matters to life on earth, not 0.01C change in deep ocean temperature.
Last time I looked several of the patterns appears to have peaked suggesting that we are about to start seeing cooling. However, since we’ve seen very little warming – it may be that the “warming” has been offset by an underlying global cooling trend. If so when we start seeing the cooling El Nino it will add to the underlying cooling so we may see such significant cooling that even NOAA and NASA’s continued fraudulent upjusting will be unable to mask it.
Speculation about why the Tropical Atlantic correlates well with temperatures in the Northern Hemisphere.
The area is cooled by dust particles from the Sahara. When the wind is blowing to the east the dust particles don’t cool and the Tropical Atlantic is warm. This is not a cause of the northern hemisphere becoming warmer – it is a separate effect from the same cause.
The cause being that the jet stream is kinked allowing warm air from the tropics to reach the cooler north quicker. As the heat radiates back into space quicker from hotter things, spreading the heat out reduces the heat loss and warms the planet.
Ill thought out speculation but it makes some sense.
Now that’s interesting. Can you give a pointer to the data and how you worked with it to get those images? R by any chance?
The correlations are interesting.
By interesting I mean that you may have uncovered something that I wasn’t previously aware of.
Where does the energy in the Kelvin wave go? Arno describes the observed fact of what happens in the Pacific with regards to the wave.
But it would appear that, like any wave traveling towards a beach, it may slop over onto the shore. In this case the Atlantic.
Hence the correlation you see. Possibly.
RichardLH December 18, 2015 at 1:45 am
[Willis: The blockquote divisions above (between your quoted sentences, and your own sentences) are not clear. Did you mean to blockquote Richard only in the first paragraph? .mod]
OK. So I’m fairly new to R. Wrote some stuff a while back to get some better filtering on UAH which was published here on WUWT. Decided that any real ‘proofs’ would be outside my statistic lifetime and moved on.
Then a coupe of things happened. I made a comment on one your other threads which started a train of thought I am busy pursuing in Academia (at first anyway).
May I just leave it as a ‘possible thing to bear in mind’ for your work?
As an aside, I think I may have come up with a way to scientifically dismiss all those straight line ‘trend’ suggestions being put forward.
A dry observation that the capture widow available does not support the bandwidth required to get to that frequency.
“Indeed. As I mentioned, inter alia I was quite surprised that for the Northern Hemisphere, the Nino3.4 area had so little predictive power.”
That possibly because the energy is not now (2 months on) in the Pacific. Its in the Atlantic (in deltas anyway).
Willis
We live on a water world. Water is the key, it is the control.
So how does El Nino conditions impact upon evaporation? Note the Pacific is warmer than the Atlantic and hence different evaporation and with it sensible energy transfer.
El Nino impacts upon rainfall patterns, again sensible energy transfer. Does the Atlantic have the same impact?
Then of course there are the prevailing winds, and these winds are necessary to drive El Nino. Again, this acts to distribute heat and energy in the system in 3D
“removing the El Nino signal” betrays a fundamental flaw in the thinking of those who try it. It assumes that the El Nino – La Nina quasi-periodic seasaw is some kind of perturbation on an underlaying climate system, which therefore can be “corrected” for. But the seasaw is an essential ingredient of it.
Last week I was trying to understand the baking of bread by my lady. I played with the idea that I could gain an insight by removing each component of the recipe in turn. Removing the salt worked ok, Removing the yeast (such a small ingredient whatever difference could that make?) gave me quite something else.
The modelers, and for that matter the climate science, doesn’t know what is the root cause of the El Nino – La Nina cycle; they only know its phenomenology. The idea to “correct” for it betrays a profound ignorance of their own ignorance.
Strange not to see McLeans 2008 paper being mentioned here. It was discussed at WUWT before. The correlation to global temperatures is quite large, about 70%. Why don’t we try to remove offects of PMD? Do we have enough data to do that? The PMD isn’t that frequent and there isn’t data to paramtrize its effect on the global temperature, right?
Thanks, Londo … Link to the McLeans paper? I have no memory of this … which doesn’t mean much these days …
And what is the “PMD” when it is at home? Do you mean the PMO? If so, no, the data would be pretty scanty for that question.
w.
http://onlinelibrary.wiley.com/doi/10.1029/2008JD011637/full
Sorry. I ment PDO but now I see you wrote AMO but that is in the spirit of my answer the same.
richard verney December 18, 2015 at 2:00 am
Thanks, richard. As I showed in “Tropical Evaporative Cooling“, in the deep tropics the evaporation goes up by about 25 W/m2 for each degree of increased temperature …
w.
Just noting that correlations of 0.1 or 0.2 are actually high numbers with respect to the ENSO.
Nino 3.4 can vary by +/- 3.0C (maybe 2.5C is better) while if another region on the planet changes by 0.2 times that, it is still 0.6C change in local temperatures. The 0.5 in the southern north atlantic sounds too high to me.
And the lag is better at 3 months, rather than 2 months.
What is going on here if anything? It looks like the previous extent shows this year has made a breakout rise in ice, but the new ice extent graph introduced this year shows no significant change, why? El Nino?
http://ocean.dmi.dk/arctic/old_icecover.uk.php
http://ocean.dmi.dk/arctic/icecover.uk.php
From the UK’s Met Office: Forecast expects 2016 to be among the warmest years
“The forecast is based on the key drivers of global climate, but it doesn’t include random events, such as large volcanic eruptions – which can cause a temporary cooling effect.
Man-made global warming, combined with a smaller effect from El Niño from unusually warm waters in the tropical Pacific Ocean feature within our forecast.
The global mean temperature for 2016 is expected to be between 0.72 °C and 0.96 °C above the long-term (1961-1990) average of 14.0 °C, with a central estimate of 0.84 °C, according to the Met Office annual global temperature forecast….”
http://www.metoffice.gov.uk/news/releases/archive/2015/global-temperature
The Met noting the impact of the volcanoes on these correlations is important because El Chichon happened right when the 1982-83 super-el nino was happening and Pinatubo also occurred right when the large 1991-92 El Nino was starting. These volcanoes impact the numbers .
So if one is testing correlations of the ENSO, one needs to also adjust for the 0.3C to 0.5C impact of these volcanoes if the time-periods overlap.
http://s21.postimg.org/gbr75yjfb/Nino_3_4_lag3_UAH_RSS_temps_Nov2015.png
Clearly runaway temperature with no connection to natural phenomena.
So maybe these El Nino events were triggered by the eruptions?! Climate feedback dumps ocean ocean heat to atmosphere to compensate for reduced in coming.
Furthermore, if each volcanic event is happening concurrent with an El Nino , how do you assess the magnitude of the “0.3 or 0.5” ? Maybe much of the later dips attributed to volcanic blocking is , inf act , La Nina. You also need to add in solar peaks. Both supposed volcanic cooling event s happened on steepest section of drop off of the solar cycle.
This is how it all ends up as rather arbitrary multivariate linear regression or something similar.
This is highly prone to problems of confounding variables and false attribution.
IMO the cooling due to volcanoes is severely over-estimated and was compensated by exaggerated CO2 warming. This works while both are present and falls apart when on is missing: eg. post Y2K.
The AMO is not a runaway temperature.
http://i772.photobucket.com/albums/yy8/SciMattG/RSS%20Global_v_RemovedAMO2_zpsssrgab0r.png
Well guys, its official. Next year was the hottest on record.
http://www.bbc.co.uk/news/science-environment-35121340
Why stop with next year? We all know the trend is linearly increasing so there will be years following next year that will be hotter – and so on – ad infinitum. /sarc
BBC says:
As usual, most of the “global warming ” is due the hot air spouted by the BBC.
Wow, If this warming continues, I may actually notice it in my lifetime.
Maybe for satellites, but probably not for atmosphere at surface.
With regional circulation patterns destroying the “typical” El Niño footprint(and embarrassing most long range forecasters) this winter…this type of study is perhaps one
of the most important you have ever written.
The “climate establishment” attempted this same thing 25-years ago.
And I kept wondering. Why???
I’ve wondered before why only a little rectangle was used. What would you get if you took the whole strip +5 to -5 around the circumference of the globe, cool and warm and used changes in that? I’m encouraged to make this suggestion from your finding of the effect of the Tropical Atlantic. It wasn’t a surprise to me that it had a commensurate effect, although just for the NH was a surprise (currents, winds??).
I think I probably agree, Willis. If I subtracted the effects of Winter, it would be Summer. Which is nice.
Very cool. Extremely interesting the nino connection to SH temperature. Maybe just because it’s even more a water world. Maybe because the southern vortex hinders mixing with antarctic air.
…yes, you can pick out the red area in the tropical Atlantic and “remove its effects” from the temperature … but what is it that you end up with? Or you could “remove” both the tropical Atlantic and the El Nino variation … and “remove” the PDO, and the QBO as well … but what is it that remains once you’ve done that?…
Presumably, if you do it correctly, you end up by seeing smaller effects from other phenomena which have been hidden by the big El Nino impact.
And if you ‘remove’ those as well, you would end up with a single flat temperature line which should match the theoretical temperature for a sphere spinning at 1 rev/24 hours at an inclination of 23 deg warmed by a nuclear reactor 94m miles away…
The assumption is that you will be left with an alarming upward curve due to AGW. But this is the semantic trick of calling them “oscillations” .
Though it is NEVER said explicitly this introduces an implication and the totally unfounded *assumption* that all these “oscillations” are a net zero effect. Like a swinging pendulum.
It’s all carefully crafted terminology like temperature “anomalies” implying something abnormal is happening. Word games masquerading as science.
A fascinating idea if it could be done. You should be able then to obtain a rough bundle of feedbacks ( individuals maybe not so much) that at least would show the magnitude of what’s to be dealt with.
If the gross net feedback is small – no worries. Something different needs to be done to cure the infestation of political science in the wrong domain.
I’m wondering what’s missing here: “I do enjoy surprises. I’d expected somewhere in the Southern Hemisphere to have the greatest correlation with .”
Sorry for the confusion, Meridan. Not sure anything is missing, but what I meant was that I was surprised that the area with the greatest correlation with the Southern Hemisphere two months ahead was not actually in the Southern Hemisphere, but was on the Equator.
w.
Remember that a preposition is not something to end a sentence with.
Hi willis/everyone,
Help needed. Is there an article somewhere which explains exactly how El Nino warms the planet? Is the warming due to an increase in tropical deep convection? Is it due to a disruption of the Walker circulation, which moves tropical surface heat to the upper troposphere where it radiates into space? If it is a disruption in the Walker circulation, what is that disruption? Is it something else?
For years I assumed that there was an increase in tropical precipitation but I haven’t spotted such a pattern in rainfall data. Reanalysis data, which has inherent limitations, doesn’t seem to provide clear fingerprints of the mechanisms. Most articles I’ve read seem to handwave their explanations of the warming mechanism. Perhaps I’m dense and am missing the obvious.
Thanks in advance for any help,
David
I’d say all you need to do is search for “bob tisdale enso” in your favorite search engine.
davidsmith651 December 18, 2015 at 11:47 am
Good question, David. Actually, the El Nino/La Nina function together as a pump which moves huge volumes of warm surface water from the equatorial Pacific to the poles, where it is more free to radiate to space. In addition, this pumping action exposes the atmosphere to the cooler subsurface Pacific waters.
Overall, both of these processes cool the planet. I’ve put up a couple of posts about this, hang on … OK, here they are:
My regards to you,
w.
How about looking at a correlation between global temp and ENSO only when the NINO 3.4 region is either +- 0.5? By removing the neutral conditions then I bet you’d see a very convincing correlation.
But I certainly agree with the premise of your essay, why remove it at all? Shouldn’t they just look at a 3-5 year running average?
Can anyone tell me what if anything is increasingly warm body of water off the east coast of N.America? Or it’s due completely to the cold blob that seems to have blocked the way of the Gulf currents? I’ve been eyeballing this setting from REN’s links for a yr or more and it also seems to be increasing in size. http://earth.nullschool.net/#current/ocean/surface/currents/overlay=sea_surface_temp_anomaly/orthographic=-71.81,21.55,376
We’ve had a nice a balmy fall and now almost winter compared to the last few we’ve had. Nice difference that I love, what’s so wrong with a few degrees of warmth? Flora and fauna are loving it too.
Great post again Willis,
I’m sure you have already done this, as you are usually pretty thorough, but have you looked at an array of time-lags to see if the predictive power is better using a different lag value?
In other words, maybe the anecdotal “2 months” is just off… and a different lag works better?
Regards.
-a
Thanks, Anton. I did look at longer lags. Since the results at a lag of 2 months was about the same as the results at 3 months, I used two months.
Part of the problem is that the equatorial areas are on an ~ 6 month cycle for things like sun position and temperature. As a result, when you get to 3 month lags that’s half a cycle, which can give you anomalous results not due to an actual lag. So I tend to shy away from three month lags.
w.
Cheers to Willis Eschenbach and the WUWT commenters here for a delightful and informative discussion.
An atmosphere sadly lacking in mainstream climate exploration (I can’t bring myself to call it science anymore)
Thanks, Willis.
A very interesting article.
Yes, I think ENSO is the best correlated phenomena to global temperatures.
Heat from the Sun, modulated by shading clouds and winds warming the Pacific Ocean. Then oceanic currents moving it. Atmospheric CO2 just points to an ever-warming Earth, with little modulation apart from seasonal fluctuations.
Bob Tisdale has been showing this.
Willis,
You remove the obvious anual cycle from the CERES data, right? That makes perfect sense, because it is clear the anual cycle doesn’t contain information about a long term trend. Removing the influence of ENSO has exactly the same logic. ENSO is not a clean a cycle as the annual cycle, of course, but that is a liitle beside the point: there is a recognized internal variation which in the long term averages to zero in influence. Makes perfect sence to take it into account, just as with the annual cycle.
The lack of strong correlation of ENSO with temperatures outside the tropics has been recognized for a long time. The leading correlation of the region you identified in the Atlantic with northern hemisphere temperatures is something new, and potentially very useful in understanding short term variation.
Thanks for the interesting write up Willis. I’m not quite sure where it leads yet, but at least your ‘teleconnections’ make a lot more sense than the tree ring ones proposed by the hockey team a few years back.
Having some of “simple country boy” left in me, I ask how it can be that databases accused of inadequate oceanic coverage can be corrected for an oceanic condition.