Saturday, February 28, 2015

A short pause: Bob Tisdale thinks climate models are weather forecasts of the Northern Hemisphere

Sou | 3:53 AM Go to the first of 3 comments. Add a comment

There's a new paper out in Science mag, which is another one looking at the so-called hiatus. This time it's from a different angle. The researchers, Byron A. Steinman, Michael E. Mann and Sonya K. Miller, were exploring climate models and observations in relation to natural variation. They studied surface temperature variations in the northern hemisphere over the past 150 years.

A temporary respite before more heat kicks in

The abstract sums up the research. The study suggests the supposed pause is merely a coincidence of two features of natural oceanic fluctuations - a peak in the Atlantic Multidecadal Oscillation (AMO) and a strongly negative-trending Pacific Multidecadal Oscillation (PMO). That it is the combined effect of these (in other words, natural variability) that partly offset the ongoing greenhouse warming.

The paper suggests that it may not be long before we get a lot hotter.

Here is the abstract:
The recent slowdown in global warming has brought into question the reliability of climate model projections of future temperature change and has led to a vigorous debate over whether this slowdown is the result of naturally occurring, internal variability or forcing external to Earth’s climate system. To address these issues, we applied a semi-empirical approach that combines climate observations and model simulations to estimate Atlantic- and Pacific-based internal multidecadal variability (termed “AMO” and “PMO,” respectively). Using this method, the AMO and PMO are found to explain a large proportion of internal variability in Northern Hemisphere mean temperatures. Competition between a modest positive peak in the AMO and a substantially negative-trending PMO are seen to produce a slowdown or “false pause” in warming of the past decade.

What's left after forcing is removed

The scientists defined the AMO, PMO and Northern Hemisphere Multidecadal Oscillation (NMO) as the difference between observations and the estimated forced temperatures for each of the repective regions of the ocean - using a low-pass filtered frequency of forty years. The purpose of the low-pass filter is to isolate the multidecadal frequency.

Some of you will understand that preceding paragraph much better than I do. What the research was focused on was the northern hemisphere, the North Atlantic and the North Pacific oceans in particular. The authors describe how these two oceans are important sources of natural variability in the climates of the northern hemisphere over timescales of multiple decades. However, "there is substantial uncertainty in their relative contributions to the observed variability".

To tease out their contributions, the scientists used what they describe as a new semi-empirical method, which "uses a combination of observational data and a large ensemble of coupled climate model simulations". The paper refers to this as a "target region regression" approach. Their aim was to distinguish forced from internal variability in the Northern Hemisphere.

One of the paragraphs I found quite interesting. It was about how the AMO in the North Atlantic operates on a time scale of 50 to 70 years. I think that they mean a full cycle lasts around 50 to 70 years, although the point is made that these are not necessarily oscillations (ie not like a swinging pendulum). The authors described the PDO as having "at least two signals":
  • one has a 16 to 20 year period
  • the other has something like a 50 to 70 year period. This latter, they indicate, could be partly related to the AMO.

The authors refer to this latter multi-decadal component of the Pacific variability as the PMO (presumably the Pacific Multidecadal Oscillation).

It seems to me that analysing patterns on multi-decadal time scales must be extremely difficult because there aren't really good records going back over multiple periods of multiple decades. Land surface temperatures can, at a pinch, go back to the mid 1800s. Sea surface temperatures not nearly as far. That's around 160 years at best, which only encapsulates two multidecadal periods of 70 years or so. Not a lot to hang your hat on, is it.

Defining the AMO

The paper describes how other people have defined the AMO (my paras): 
Prior methods used to define these internal variability components and their influence on Northern Hemisphere temperature include
  1. a simple linear detrending of the mean North Atlantic SST time series (17–21), 
  2. estimating the forced trend based on regression of North Atlantic SST against global mean SST and removing the forced trend to yield an estimate of the internal variability (16, 22, 23), and 
  3. defining the forced component as the mean of North Atlantic SST in an ensemble of climate model simulations and defining the internal variability component as the difference between the observed SST series and the multimodel mean (24, 25). 
These methods, as shown below, do not in general yield correct results. We estimated the Atlantic and Pacific-basin multidecadal internal variability components and their contribution to Northern Hemisphere temperature change on the basis of a new target region regression approach.

Target region regression

I had to do some hard thinking to get my head around what the authors meant by "target region regression" and how it differs from other definitions, like the simple linear detrending. Michael Mann has helped me understand this, in his article at realclimate.org. He wrote (my emphasis):
We propose and test an alternative method for identifying these oscillations, which makes use of the climate simulations used in the most recent IPCC report (the so-called “CMIP5” simulations). These simulations are used to estimate the component of temperature changes due to increasing greenhouse gas concentrations and other human impacts plus the effects of volcanic eruptions and observed changes in solar output. When all those influences are removed, the only thing remaining should be internal oscillations. We show that our method gives the correct answer when tested with climate model simulations.

There is additional information on this in the supplementary material, if you want to find out more. For example:
Definition of the AMO, PMO, and NMO
The observed low-frequency internal variability may represent the sum of true oscillatory (i.e., 50–70 year) signals and simple low-frequency red noise. For our purposes the distinction is unimportant, and we use AMO, PMO, and NMO in a loose sense to denote the multidecadal (>40 year timescale) internal variability in the respective series, regardless of whether it constitutes a true oscillation or not.

The realclimate.org article included Figure 3 (C) from the paper, which illustrates the contributions from the AMO, the PMO and the NMO to northern hemisphere temperatures. Below are all panels from Figure 3 (animated):

Fig. 3 Semi-empirical estimate of AMO, PMO, and NMO based on target region regression using historical climate model realizations. (A) CMIP5-GISS. (B) CMIP5-AIE. (C) CMIP-All. In (A) to (C), blue, AMO; green, PMO; and black, NMO. Bivariate regression-based approximation of NMO (red) strongly correlates (R2 = 0.86/0.88/0.91 for CMIP5-All/CMIP5-GISS, CMIP5-AIE, respectively) with semi-empirical NMO estimate (black). 95% confidence limits of the AMO, PMO, and NMO CMIP5-All means were determined by using the ensemble of target region mean series resulting from bootstrap resampling (Fig. 1) and are shown as colored shading.
Source: Steinman15

Since the dip in the mid 1970s, the AMO (blue line) has risen some to a peak in around 2005. It's the PMO (green line) that took a big dive from a peak around 1990. That's what's dragged down the NMO (black and red lines), which as you'll recall, the authors define as the multidecadal component of internal Northern Hemisphere mean temperature variability.

Get ready for hotter

Therefore, the authors argue that, rather than contributing to recent warming, the multidecadal variability has offset anthropogenic warming. They warn that, going by what's happened in the past, this won't last and will probably reverse. In other words, internal variability will add to human-caused warming in coming decades. We could well be in for some hot shocks.

Bob Tisdale, WUWT pseudo-scientist, takes his best shot - and misses

Bob Tisdale doesn't like the approach taken by Steinman15 (archived here). He wrote in his pseudo-science fashion:
It appears their grand assumption is that the outputs of the climate models stored in the CMIP5 archive can be used as a reference for how surface temperatures should actually have warmed…when, as shown as an example in Figure 1, climate models show no skill at being able to simulate the multidecadal variability of North Atlantic.

Well what it appears to Bob is nothing like what the scientists assumed. Bob gets it back to front and upside down as usual. For one thing he is confusing a deduction and an assumption, and he'd be wrong even so. Bob is correct in that the scientists found that models can be used to estimate the effect of climate forcings. But he also seems to think that the scientists assume something they didn't. The point is that climate models will not necessarily reflect internal variability at the same time as it occurs. They do reflect internal variability, but it is not necessarily in phase with what actually happens. In fact, in the paper itself is this statement (my emphasis):
Our method is based on the principle that internal variability is uncorrelated among distinct realizations of a large ensemble. We therefore used the mean of the Coupled Model Intercomparison Project Phase 5 (CMIP5) ensemble (26) as an initial estimate of the forced component of surface temperature for the North Atlantic, North Pacific, and the entire (land + ocean) Northern Hemisphere region 

What the scientists did was use the models to determine what of the observed changes were from anthropogenic and other external forcings. That is using the models to estimate the changes that can be attributed to increasing greenhouse gas concentrations, volcanic eruptions and changes in solar output. They could then distinguish those changes from the actual observations. They could deduct from observations the changes from forcings as estimated by models, and what was left was attributed to internal variability.

Bob then writes this:
Steinman et al., according to Mann’s blog post, have subtracted the models from the data. This assumes that all of the warming since the mid-1970s is caused by the forcings used to drive the climate models. That’s a monumental assumption when the data have indicated the surfaces of the North Atlantic are capable of warming at rates that are much higher than the forced component on the models. In other words, they’re assuming that the North Atlantic since the mid-1970s has not once again warmed at a rate that is much higher than forced by manmade greenhouse gases.

Bob is incorrect when he writes "This assumes that all of the warming since the mid-1970s is caused by the forcings used to drive the climate models". It doesn't assume any such thing. It's no monumental assumption. It's not an assumption at all. Look again at Figure 3 above.  From around 1970 to around 1990, the PMO curve was rising as was that of the NMO, the AMO curve dipped and rose again. None were flat, though they didn't contribute a great deal (check the Y axis). In more recent years the PMO takes a dive. The scientists have concluded that internal variability suppressed temperatures in the last few years. As Michael Mann wrote at realclimate.org:
Applying our method to the actual climate observations (see figure above) we find that the NMO is currently trending downward. In other words, the internal oscillatory component is currently offsetting some of the Northern Hemisphere warming that we would otherwise be experiencing. This finding expands upon our previous work coming to a similar conclusion, but in the current study we better pinpoint the source of the downturn.
The much-vaunted AMO appears to have made relatively little contribution to large-scale temperature changes over the past couple decades. Its amplitude has been small, and it is currently relatively flat, approaching the crest of a very shallow upward peak. That contrasts with the PMO, which is trending sharply downward. It is that decline in the PMO (which is tied to the predominance of cold La Niña-like conditions in the tropical Pacific over the past decade) that appears responsible for the declining NMO, i.e. the slowdown in warming or “faux pause” as some have termed it.

Contrast what Michael Mann wrote with what Bob Tisdale wrote. Bob doesn't have the excuse that he didn't bother reading the paper itself, he referred more than once to the realclimate article, so he could hardly have missed it.

Bob went further into crankery, writing:
What Steiman et al. have done is similar to subtracting an exponential curve from a sine wave…where the upswing in the exponential curve aligns with the last minimum to maximum of the sine wave…without first establishing a relationship between the two totally different curves.

That's gobbledegook masquarading as pseudo-science! Bob's "establishing a relationship" between forced temperature changes and internal variability was at the heart of the paper. The scientists determined that there is no correlation. That's the essence of the work. To separate out internal variability from forced changes. The aim was to get two totally different curves.

If Bob had bothered himself to read the supplementary material he would have found this discussion of how linear detrending to estimate the AMO, mixes up forced temperature changes with internal variability, for example:
Mann and Emanuel (47) showed that what is commonly termed the “AMO” (e.g., 17–19) may substantially be an artifact of the misidentification of forced trends as internal variability. Focusing on tropical Atlantic SST, they showed that a commonly employed method of estimating the AMO—simple linear detrending followed by low-frequency filtering of the residual series—results in an artificial, apparent low-frequency “oscillation”, since the true forced signal is not linear in time. The cooling from the 1950s–1970s AD associated with a substantial increase in anthropogenic aerosols in the Northern Hemisphere and a subsequent warming due to the decrease in aerosols commencing in the late 1970s, masquerades as a low-frequency “oscillation”. Other more recent work has supported that finding (54–58).

Bob demonstrates further how he doesn't understand the first thing about climate models (or this paper). He wrote, referring to Figure 3 (C) I showed above:
I had to laugh when I saw the following illustration presented in Michael Mann’s blog post at RealClimate. I assume it’s from Steinman et al. In it, the simulations of the surface temperatures (represented by the multi-model mean of CMIP5-archived models) of the North Atlantic, North Pacific and Northern Hemisphere surface temperatures have been subtracted from the data. That illustration clearly shows that the climate models in the CMIP5 archive are not capable of simulating the multidecadal variations in the sea surface temperatures of the North Atlantic and the North Pacific or the surface temperatures of the Northern Hemisphere.

We're laughing too. But we're laughing at Bob not with him. Bob doesn't seem to have twigged to the fact that subtracting forcings from observations leaves the internal variability. He still thinks that climate models are weather forecasts. He thinks that models should be an exact replica of observations, including the precise timing not just of long term multidecadal internal variability, but short term random weather events like ENSO. I don't think he'll ever understand climate models. He's not very smart.

Another perspective by Ben B Booth 

There's a rather nice graphic in a perspective piece on the paper, written by Ben B. Booth. It shows the contribution from internal variability to northern hemisphere temperatures, as estimated by Steinman15.

Natural Variability: On the basis of current climate simulations and observed temperature records, Steinman et al. estimate the role that natural ocean variability has played in modulating temperature trends over the past 130 years. Source:  Ben B. Booth, Science

Contrast the above figure with what Bob Tisdale claimed about assuming "that all of the warming since the mid-1970s is caused by the forcings used to drive the climate models". He was wrong, wasn't he, particularly in regard to the impact of the internal variability in the North Pacific.

One issue that Steinman15 does raise is how scientists view the variability of sea surface temperatures. As Ben Booth wrote:
Fifteen years ago, Richard Kerr coined the phrase “Atlantic multidecadal oscillation” (AMO) to describe the apparent 40- to 80-year oscillation in observed North Atlantic SSTs (9). At the time, the observed variability and natural ocean variability were taken to be synonymous. Since then, research has identified industrial and volcanic aerosols and solar changes as drivers of observed SST variations (7, 8, 10). This has led to confusion, with papers either referring to the total SST variability (both natural and externally driven) as the AMO (8) or only the natural ocean component as the AMO (as done in Steinman et al.). There is a real need in the community to agree on common terms to distinguish the two. One possibility would be to use Atlantic multidecadal variability (AMV) to refer to the total temperature variability and AMO for the natural component.

Booth also urged caution in regard to this paper, particularly in regard to the reliance on climate models to accurately reflecting forcings (eg cloud/aerosol interactions). He does agree, however, that Steinman15 "provides much-needed longer-term context for the role that natural ocean-driven variations have played in past climate change".

From the WUWT comments

I can hardly be bothered reading the WUWT comments. They are pure illiterati - as bad as WUWT is every day. I'll add a few to show you what I mean.

Richard M, who wouldn't know straight logic if he tripped over it, wrote:
February 26, 2015 at 5:54 pm
Looks like more circular logic from the AGW propagandists.

Tim's an obedient trained monkey, if rather stupid (for a monkey), and behaves in the manner he's been coached at WUWT.
February 26, 2015 at 5:55 pm
More garbage by Mann and his associated clowns. 

bones seems to understand Bob Tisdale's pseudo-science, and like Bob, hasn't got the first clue about climate models.
February 26, 2015 at 5:58 pm
Subtracting the models from the data only changes the sign of the result. Either way it shows that the models have completely failed to capture the oscillations. Steinman et al. must have applied some serious smoothing to the differences in Figure 3.

Michael Jankowski seems to think that surface temperature trends in recent years have been dramatic. His notion of what constitutes drama will probably get a shake-up in the not too distant future.
February 26, 2015 at 6:00 pm
Mann and others are providing the “it’s worse than we thought” excuse about how it’s REALLY going to warm when the “pause” ends.
Forget that they poo-poo’d any idea that natural variability could be so dramatic. It somehow can cancel all the warming from the current high levels of CO2 in the atmosphere but was claimed to be basically negligible back when CO2 levels were lower.
It’s amazing that they are given any credibility at all anymore.

Bob Tisdale is feeling awfully pleased with his demonstration of just how much of a crank he is. He has become quite excited, and delighted with the depth of sciency knowledge displayed in the comments like the ones above. He really showed them sciency types what for, didn't he.
February 26, 2015 at 6:08 pm
We could have a field day with Mann’s post. Lets see how the warmists respond around the blogosphere.

Oh, here's a decent comment. It's from Robert Way (of Cowtan and Way fame), who's pretty well saying pointing out the same sort of things that I keep criticising Bob Tisdale for.
February 26, 2015 at 6:15 pm
This post shows a fundamental misunderstanding about how climate models work. You would never expect for climate models to all simulate multidecadal variability in the same phase and magnitude simply because these are process-based models that have their own synoptics and various forcings as input. This is why taking a large ensemble should average out all the natural variability leaving only the forced response. However, the forced response in the models is underestimated because the lack of updated forcings for the historical runs – e.g. many (almost all) models assume no volcanic forcing past 2005 whereas there’s strong evidence of a moderate forcing. Unless you have the forcings correct the forced response the approach of detrending using the models will cause some attenuation of the actual signal.

Robert may have a point re the post-2005 forcings, if you remember the two recent papers on volcanoes.  On the other hand, as Michael Mann says in an in-line comment, you can only work with what you've got. What he and his team did in regard to models is indicated in the supplementary materials as:
We extended the CMIP5 mean series (which spans AD 1850–2005) to 2012 using the slope of the 30 year trend (AD 1975–2005)
So the last few years of their analysis may have overestimated the forcing - or not. Remember this is a regional analysis as much as hemispherical. And as Michael Mann points out, his downturned PMO is consistent with the largely La Nina-like state of the Pacific Ocean in recent years (until last year).

Bob Weber is either having a shot at WUWT-ers' idiotic wishful thinking, or else blissfully unaware of his own.
February 26, 2015 at 7:30 pm
The answer is they believe in “consensus reality”, not “objective reality”.
In other words ‘make believe’ is more real to them than truth, call it wishful thinking.

A lot of people at WUWT don't realise that it was scientists, like Michael Mann and others, who first described all the various types of internal variability, like the AMO, the PDO, now the PMO and the NMO, and ENSO and the AO and whatever other acronyms you can think of. Richard M doesn't know this irrefutable and most obvious fact. Where does he think the understanding of natural variability came from?
February 26, 2015 at 7:26 pm
You mean that “natural variability” that was ignored by climate scientists in the past? Where was that in their previous model work? Oh wait, they obviously didn’t have a clue how the climate actually works but you now want us to believe they’ve had a revelation and have been blessed with divine knowledge. Yeah, what flavor kool-aid are you imbibing? 

That should be more than sufficient to demonstrate the incompetence and ignorance of readers at WUWT. I couldn't be bothered reading any further. If you want more you can get it here.

Byron A. Steinman, Michael E. Mann, Sonya K. Miller. "Atlantic and Pacific multidecadal oscillations and Northern Hemisphere temperatures". Science, 2015  DOI: 10.1126/science.1257856

BBen B. Booth. "Why the Pacific is cool". Science, 2015 DOI: 10.1126/science.aaa4840

Additional reading:

Press release at ScienceDaily.com: Interaction of Atlantic and Pacific oscillations caused 'false pause' in warming

Michael E. Mann. "Climate Oscillations and the Global Warming Faux Pause", RealClimate.org Feb 2015 


  1. Perhaps slightly ambiguous Sou

    >It's from Robert Way (of Cowtan and Way fame), who's pretty well saying the sort of things that I keep criticising Bob Tisdale for.

    Maybe "pointing out the things that.."

    1. Cowtan is an x-ray crystallographer. He is used to dealing with data sets that have a lot of noise at the high resolution end. We have one thing in our favour the Universe follows the laws of Physics. We expect internal consistency in our theories. Teasing out the pathetic signal to noise is so basic because we can use independent information to tie down the phase and magnitude of our noisy measurements. Bert


Instead of commenting as "Anonymous", please comment using "Name/URL" and your name, initials or pseudonym or whatever. You can leave the "URL" box blank. This isn't mandatory. You can also sign in using your Google ID, Wordpress ID etc as indicated. NOTE: Some Wordpress users are having trouble signing in. If that's you, try signing in using Name/URL. Details here.

Click here to read the HotWhopper comment policy.