Mann’s new AMO paper: ‘Had I been a reviewer, I would have pointed this out and recommended rejection. ‘

Mann’s new paper recharacterizing the Atlantic Multidecadal Oscillation

A guest post by Nic Lewis

Introduction

Michael Mann has had a paper on the Atlantic Multidecadal Oscillation (AMO) accepted by Geophysical Research Letters: “On forced temperature changes, internal variability, and the AMO”. The paper seeks to overturn the current understanding of the AMO, and provides what on the surface appears to be impressive evidence. But on my reading of the paper Mann’s case is built on results that do not support his contentions. Had I been a reviewer, I would have pointed this out and recommended rejection.

In this article, I first set out the background to the debate about the AMO and present Mann’s claims. I then examine Mann’s evidence for his claims in detail, and demonstrate that it is illusory. I end with a discussion of the AMO. All the links I give provide access to the full text of the papers cited, not just to their abstracts.

The abstract and access to Supplementary Information is here . Mann has made a preprint of the paper available, here . More importantly, and very commendably, he has made full data and Matlab code available.

The conventional view of the AMO

NOAA, which provides an AMO index, has a helpful FAQ on the AMO that says:

The AMO is an ongoing series of long-duration changes in the sea surface temperature of the North Atlantic Ocean, with cool and warm phases that may last for 20-40 years at a time and a difference of about 1°F between extremes. These changes are natural and have been occurring for at least the last 1,000 years… Since the mid-1990s we have been in a warm phase. The AMO has affected air temperatures and rainfall over much of the Northern Hemisphere… It alternately obscures and exaggerates the global increase in temperatures due to human-induced global warming.

The AMO is thought to be quasi-periodic with a typical cycle length of 60–70 years. It reached its nadir in the mid 1970s and, after reaching positive ground in 1995, may have peaked in the mid 2000s. NOAA’s AMO index[i] is a detrended average of mean North Atlantic (0°–70°N) sea surface temperature (SST) from the Kaplan dataset. Figure 1 shows that AMO index on both annual and centred 5-year mean bases.

Although the NOAA AMO index is based only on North Atlantic SST, both northern hemisphere (NH) temperature and global mean surface temperature (GMST) are quite strongly correlated with it. Something of the order of 0.2°C of the 0.5–0.6°C increase in GMST since the mid 1970s might be due to the strengthening AMO rather than to increasing anthropogenic radiative forcing. Consistent with this suggestion, a recent paper by Chylek et al concluded, using regression analysis, that about one-third of the post-1975 increase in GMST was likely due to the AMO. A 2013 paper by Zhou & Tung found an even stronger influence of the AMO on the post-1980 GMST trend.

clip_image002

Fig 1. NOAA AMO Index, annual (thin cyan line) and 5-year mean (thick green line), based on detrended North Atlantic (0°–70°N) SST from the Kaplan dataset.

Assuming that the AMO is natural, and it has had a positive influence on the increase in GMST over the last few decades, it follows that estimation of anthropogenic warming rates and the transient climate response (TCR) from post-1975 temperature changes will be biased upwards, showing high sensitivity, fast-warming climate models (CMIP5 GCMs, in particular) in an artificially favourable light, unless the AMO’s influence is adjusted for. A paper under discussion at Earth System Dynamics, here, makes just that point. It concludes that, adjusting for the influence of the AMO, global warming over the last 30 years indicates a best estimate for TCR of ~1.3°C. That is in line with the results of several studies based on warming since the second half of the 19th century – the trend of which will have been much less affected by the AMO – but well below the 1.8°C average TCR of current generation CMIP5 GCMs.

Did aerosols rather than the AMO drive 20th century Atlantic SST variations?

In 2012 a team of scientists at the UK Met Office published a paper claiming that anthropogenic aerosol indirect forcing, rather than natural variability, drove much of the 20th century variability in North Atlantic SST attributed to the AMO. This claim was based on simulations using the HadGEM2-ES climate model. However, in 2013 a team of scientists from GFDL and elsewhere published a counter-paper entitled “Have Aerosols Caused the Observed Atlantic Multidecadal Variability?“, which showed major discrepancies between the HadGEM2-ES simulations and observations in the North Atlantic.

Mann himself had argued that anthropogenic aerosols rather than the AMO drove variability in tropical Atlantic SST in a short 2006 paper, here . However, he accepted therein that his analysis relied upon the AMO having no influence on GMST, and he also used what is arguably a questionable statistical model. AR5 didn’t mention this paper when discussing the AMO (in Section 10.3.1.1.3).

Now, however, Mann has returned to this issue, making the extraordinary claim that trends forced by anthropogenic greenhouse gas and sulphate etc. emissions masqueraded as an apparent oscillation, and that, rather than warming the NH:

“The true AMO signal, instead, appears likely to have been in a cooling phase in recent decades, offsetting some of the anthropogenic warming temporarily.”

Mann’s other claims

The press release for the paper also says:

According to Mann, the problem with the earlier estimates stems from having defined the AMO as the low frequency component that is left after statistically accounting for the long-term temperature trends, referred to as detrending.

Mann and his colleagues took a different approach in defining the AMO…  They compared observed temperature variation with a variety of historic model simulations to create a model for internal variability of the AMO that minimizes the influence of external forcing — including greenhouse gases and aerosols. They call this the differenced-AMO because the internal variability comes from the difference between observations and the models’ estimates of the forced component of North Atlantic temperature change.

They also constructed plausible synthetic Northern Hemispheric mean temperature histories against which to test the differenced-AMO approaches.  Because the researchers know the true AMO signal for their synthetic data from the beginning, they could demonstrate that the differenced-AMO approach yielded the correct signal.  They also tested the detrended-AMO approach and found that it did not come up with the known internal variability.

While the detrended-AMO approach produces a spurious temperature increase in recent decades, the differenced approach instead shows a warm peak in the 1990s and a steady cooling since.

That is certainly a novel approach. By defining the AMO as the part of the smoothed temperature change simulated by the models that is not observed, the problem of models warming far too fast since ~2000 largely disappears. So does the inconvenient possibility that the fast model-simulated warming in the 1980s and 1990s might have only been matched in the real world due to a significant contribution from the AMO. Mann’s differenced-AMO is a high-sensitivity climate modeller’s dream. If climate models were perfect apart from not simulating the AMO, then the differenced-AMO approach would make sense. But models are by no means perfect – and if they were then they would simulate the AMO.

Mann’s differenced-AMO merely reflects, on a smoothed basis, the extent to which the observed NH temperature outpaces climate model simulated NH temperature, going negative when models simulate an unrealistically high temperature rise. It seems likely that it will represent model failings and unrealistic forcings to a greater extent than unforced multidecadal internal climate system variability. The CMIP5 models typically have very high aerosol forcing, and as aerosol forcing grew fast from 1950 to the mid/late 1970s it seems that their high aerosol forcing typically more than compensated for their high transient sensitivity, so that they partially emulated the effects of the AMO downswing.

After defining the differenced-AMO, Mann purports to show – using synthetic temperature histories containing a known AMO signal – that his differenced-AMO approach yields the correct signal, whereas the detrended-AMO approach does not. So how does Mann achieve this impressive feat?

The graphs in Mann’s paper are largely based on a simulation by a simple energy balance model (EBM). He obtains broadly similar, but less impressive, results using instead the GISS-E2-R GCM and the average of the full 40-model CMIP5 GCM ensemble. I’ll concentrate on his EBM simulation here, as it best illustrates how he achieves his surprising results.

Mann deals entirely with Northern Hemisphere, not global, temperatures. Figure 2 shows the evolution of NH surface temperature simulated by his EBM (blue line) when his code is run, compared to the HadCRUT4 observational record (black line). It also shows an alternative simulation by a low sensitivity EBM of my own specification (red line). The lines are aligned to all have the same overall mean. For HadCRUT4 the zero line is intended to represent preindustrial temperatures.

clip_image004

Fig 2. NH temperature anomalies from 1850–2012 per HadCRUT4 (black) and as simulated by Mann’s EBM (blue) and the alternative low sensitivity EBM (red)

 

Mann’s high sensitivity EBM and my alternative EBM

Mann’s EBM has an equilibrium/effective climate sensitivity (ECS) of 3.0°C and, unusually, no allowance for heat uptake by the ocean apart from in a 70 m deep mixed layer. As a result, its TCR – the simulated temperature rise from CO₂ concentrations doubling over 70 years as a result of 1% p.a. growth – is 2.8°C, very little lower than its ECS. His EBM uses a modest aerosol forcing that becomes only 0.3 W/m² more negative over 1950–1975. So why does its simulated temperature rise from 1920–1950, during the AMO’s upswing, but fall from 1950 to 1975, over which period the AMO was in a downswing but anthropogenic forcing excluding aerosols rose by over 0.7 W/m² (per AR5)?

The explanation is that Mann makes only a very modest allowance for the increase in non-CO₂ greenhouse gases and other non-aerosol anthropogenic forcings over 1950–1975, so his increase in total anthropogenic forcing over that period is only 0.4 W/m², 0.32 W/m² below AR5’s best estimate. During that period solar and volcanic forcings both had negative influences – totalling -0.2 W/m² per Mann’s data and -0.3 W/m² per AR5, taking trailing 5-year means to allow for the time constant of the ocean mixed layer. There was also sizeable negative volcanic forcing in 1963-64, again larger per AR5 than in Mann’s data. Therefore, Mann’s EBM had a negative forcing trend over 1950-1975 and shows cooling during that period.

On the other hand, during 1920–1950 the increasing trend in negative aerosol forcing was more than offset by trends in solar and volcanic forcings, and the very high TCR of Mann’s EBM made up for the shortfall in non-CO₂ greenhouse gas forcing and other non-aerosol anthropogenic forcings. After 1975, during the AMO upswing, much the same occurred, but by then the rise in CO₂ forcing was faster and more dominant, so Mann’s EBM simulated temperature rose fast. After 2000, since when the rise in CO₂ has strongly dominated changes in other forcings, Mann’s sensitive EBM outpaces HadCRUT4.

As a result of the particular forcing history used, Mann’s EBM, despite its very high TCR, is able to match – very closely on a smoothed basis – not only the overall HadCRUT4 20th century NH record but also its AMO-influenced ups and downs. But Mann chose the scaling for aerosol and solar forcing to optimise the fit, so it is not very surprising that it is good. The result is that his differenced-AMO smoothed time series is fairly flat over the 20th century, and declining post 2000.

My alternative, low-sensitivity, EBM is driven by the AR5 forcing best estimate time series. As is common when using a simple global model, volcanic forcing is scaled down, here by a factor of 0.5. It remains higher than the volcanic forcing series Mann uses for his EBM. The low-sensitivity EBM has the same ocean mixed layer depth as Mann’s EBM, but it is a 2-box model with the rest of the ocean’s heat capacity represented as well. The EBM has an ECS of only 1.65°C, in line with my best estimate using AR5 forcing and heat uptake data. One might expect a higher sensitivity (or a scaling of the simulated temperature) to be needed to match the warming in the NH, which is faster than the global rate, but different ocean parameters from those used for global temperature simulations suffice to allow for this.

The low-sensitivity EBM’s deep-ocean heat uptake coefficient is chosen to produce a TCR of 1.37°C which, on adding to the simulated temperatures a suitably scaled version of the 5-year mean NOAA AMO index, gives the best fit to the HadCRUT4 NH surface temperature record. That AMO index increases only modestly between the start and end of the simulation. The NH temperature simulated by my low-sensitivity EBM matches the overall NH temperature rise exhibited by Mann’s EBM up to the late 1990s, and matches the overall 1850–2012 observed (HadCRUT4) rise more closely. However, without the addition of the scaled 5-year mean NOAA AMO index the low-sensitivity EBM simulation’s fit to NH observations is a little worse than Mann’s in terms of mean square error, as it does not emulate the AMO’s fluctuations.

Mann’s differenced-AMO vs detrended-AMO

To recap, Mann’s differenced-AMO just represents actual minus model-simulated forced NH (not, as stated in the press release, North Atlantic) surface temperature. And whilst NOAA’s AMO Index is a detrended average of mean North Atlantic SST, for some unexplained reason Mann instead defines his detrended-AMO as the detrended average of mean NH temperature. In both cases, the AMO signal is smoothed by a 50-year low-pass optimising filter of Mann’s design – using slightly different variants in the two cases. Figure 3 shows the differenced-AMO (black) and the detrended-AMO (red), along with the unsmoothed annual time series that they are derived from.

Notice how Mann’s differenced-AMO, based on his EBM simulation, has a gentle peak just before 1990 before declining noticeably, so that it falls slightly from the mid-1970s to 2012. By contrast, Mann’s detrended-AMO rises strongly throughout that period. The smooth thick blue line shows the results of applying the detrended-AMO approach to the NH temperature evolution as simulated by Mann’s EBM. Its near coincidence with the smooth thick red actual detrended-AMO line shows how successful Mann has been in fitting his EBM to match the multidecadal fluctuations in NH surface temperature.

clip_image006

Fig 3. Version of Mann’s Figure 2.a). Estimated NH temperature anomaly variability: thin and thick lines are respectively annual, and 50-year low-passed smoothed, time series. The red lines show detrended observed (HadCRUT4) NH anomalies, the thick line being Mann’s detrended-AMO. The black lines show the observed NH temperature anomaly minus Mann’s EBM simulation, the thick line being the differenced-AMO. The blue lines show detrended anomalies from Mann’s EBM simulation, the thick line being what the detrended-AMO would be if based on the EBM-simulated rather than observed temperatures.

In Figure 2.a) of Mann’s actual paper (reproduced here as Figure 4), the smooth differenced-AMO line (grey dashed line in his figure) has a somewhat different shape, starting at a high level and ending at a lower level, with a peak around 1945 and minimum around 1965 that are missing when I run his code. The smooth blue line (in this case dashed rather than thickened) showing the results of applying the detrended-AMO approach to Mann’s EBM simulation is also marginally different. The EBM and smoothing code is deterministic so there should be no discrepancies.

clip_image008

Fig 4. Reproduction of Figure 2.a) from Mann’s GRL paper. This shows the same time series as Figure 3 and should be identical to it, but with grey lines in place of black lines and black lines in place of red lines. Dashed rather than thick lines are used to show the smoothed AMO-like signal versions of the annual time series.

As the jagged blue lines (the detrended EBM simulation anomalies) do not differ visually between my and Mann’s paper’s version of his Figure 2a, it seems possible that the difference lies in the smoothing used. Figure 5 shows the effect of changing the cut-off frequency of Mann’s low-pass filtering from the “freq0=0.02; % low-freq cutoff in cycles/year” in his archived code to freq0=0.025. That changes it from 50-year to 40-year low-pass filtering, which is in line with what the code comment says:

% determine multidecadal compoments via 40 year lowpassed versions of the residual series

The results, shown in Figure 5, are indeed much closer to those shown in Mann’s paper, although not identical. However, with 40-year low-pass filtering Mann’s results, as reflected in his subsequent figures, differ noticeably from those in his paper (and are slightly less impressive). I will leave this mystery for the future and continue for the rest of my present investigation making use of the 50-year low-pass filtering specified in Mann’s paper (resetting freq0 to 0.02). That filtering has broadly similar effects, leaving aside endpoints, to smoothing by a 15 or 20 year moving average, but it suppresses shorter-term fluctuations much more strongly.

clip_image010

Fig 5. Same as Figure 3 but using 40-year rather than 50-year low-pass filtering

 

Mann’s case against the detrended-AMO

Mann’s key claim is that, where the signal is known a priori, the detrended-AMO approach to estimating AMO-related variability fails to isolate the true internal variability, and yields an excessive and out-of-phase estimate of the true AMO signal. His Figure 3a, a version of which based on running his code is reproduced as Figure 6a, shows the differenced-AMO signal from five noisy variants of his EBM-simulated temperature time series with random realisations of red noise added – the noisy series being treated as surrogate NH temperature observations – (coloured lines), and the differenced-AMO based on actual NH temperature observations (black line). In all cases the differenced-AMO calculation deducts the noise-free EBM-simulated temperature time series from the noisy series (leaving just the red noise) and then applies low-pass filtering. Mann points out that the differenced-AMO signals represent independent realisations of multidecadal noise and are therefore uncorrelated, with random relative phases and a small amplitude. That is obviously so.

Mann’s Figure 3.b), a version of which based on running his code is presented as Figure 6b, shows detrended-AMO signal estimates from the same five noisy EBM simulations (thin coloured lines) and based on observed temperatures (red line of Figure 3) (black).

Mann writes in his paper:

The random surrogates are qualitatively similar in their attributes to the differenced-AMO estimate of the real-world AMO series. By contrast, the detrended-AMO signals (Figure 3b [here 6b]) show amplitudes ~0.25°C that are inflated by more than a factor of two. Further, they are largely all in phase with the detrended-AMO signal diagnosed from observations (Figure 2 [here 3]), an artifact of the common forced signal masquerading as coherent low-frequency noise.

a)

clip_image012

b)

clip_image014

Fig 6. Version of Mann’s Figure 3. Comparison of (a) “true” pseudo-NH AMO signal (as a priori defined using the differenced-AMO approach, not the real AMO) and (b) NH AMO signal as estimated by detrended-AMO procedure, applied to surrogate observational time series consisting of five noisy variants of Mann’s EBM simulation. (a) The differenced-AMO signal estimates relative to the noise-free EBM simulation: from the noisy simulations (thin coloured lines) and from the actual observed NH temperature time series (black; same as Figure 3 black line, but different scale). (b) The detrended-AMO signal estimates: from the noisy simulations (thin coloured lines); based on observed temperatures (red line of Figure 3) (black); and from the noise-free EBM simulation (blue dashed; same as Figure 3 blue line: omitted in Mann’s Figure 3).

 

The flaws in Mann’s case

The first part of what Mann writes is obviously true, but are his conclusions warranted? It is true that the detrended-AMO signals diagnosed from the noisy EBM simulations are indeed largely all in phase with, and very similar in amplitude to, the detrended-AMO signal diagnosed from observations (the black line). But their real such relationship is to the detrended-AMO signal diagnosed from the noise-free EBM simulation. Although that signal is not shown in Mann’s published Figure 3b, it is actually plotted by his code, and is shown by the blue dashed line in Figure 6b (same as Figure 3 thick blue line). As can be seen both there and in Figure 3 above, the low-passed detrended-AMO signal diagnosed from observations and the low-passed detrended-AMO signal diagnosed from the noise-free EBM simulation are almost identical, reflecting the success of Mann’s fitting of his EBM simulation to the smoothed observations. Therefore, the detrended-AMO signals diagnosed from the noisy EBM simulations appear also to be related to the detrended-AMO signal diagnosed from observations. But that apparent relationship is purely an artefact of the similarity of the detrended-AMO signal diagnosed from observations and the detrended-AMO signal diagnosed from the noise-free EBM simulation.

Mann’s random red-noise series have low-passed components typically only a quarter as large as the smoothed signal from applying his detrended-AMO approach to the EBM forced simulation (compare coloured lines in Fig 6a with blue dashed line in Fig 6b, noting different scales). So it is unsurprising that one recovers something close to that signal (as in Fig 6b) – and hence close to the nearly identical detrended-AMO from observed temperatures (black line in Fig 6b) – when applying the detrended-AMO approach to the EBM forced simulation with the random red-noise added, whatever realisation of noise is used.

So Figure 6 does not prove Mann’s claim. The detrended-AMO signals are in reality largely all in phase with, and of similar amplitude to, the detrended-AMO signal diagnosed from the noise-free EBM simulation, not (as Mann claims) with that signal derived from observations. One would expect to end up with something close to a smoothed version of the signal when adding a noise component with a small low-frequency amplitude to a signal with a ~4 times larger low-frequency amplitude and low-pass smoothing their sum, where there are only two cycles of signal in the pass band.

Figure S7.b3 in Mann’s Supplementary Information, reproduced as Figure 7, very much supports my conclusion. It shows the results when an alternative volcanic forcing series (Crowley) is used. When that is done, the application of the detrended-AMO approach to Mann’s EBM simulation gives a significantly different signal (blue line – present, but not identified, in Mann’s SI graphs) from when it is applied to the actual temperature observations (black line), and the coloured lines cluster closely around the blue line rather than the black line.

clip_image016

Fig 7. Reproduction of Mann’s Figure S7.b3: as Figure 3.b in his main article but using the Crowley volcanic forcing series. The detrended-AMO signal estimates: from the noisy EBM simulations (red, green, cyan, yellow and magenta lines); from observed temperatures (black line); and from the noise-free EBM simulation (blue line)

 

Mann’s attack on Stadium Waves

Essentially the same arguments apply to Mann’s critique of the “stadium wave” theory (Wyatt et al, 2012; Wyatt and Curry, 2013), about which the press release says:

Mann and his team also looked at supposed “stadium waves” suggested by some researchers to explain recent climate trends. The climate stadium wave supposedly occurs when the AMO and other related climate indicators synchronize, peaking and waning together.  Mann and his team show that this apparent synchronicity is likely a statistical artefact of using the problematic detrended-AMO approach.

Mann applies a similar procedure to what he terms synthetic AMO-related indices, which are pretty well the same as the noisy EBM simulations used already but with noise of a larger amplitude added. Figure 8, a version of Figure 4 in Mann’s paper produced by running his code, shows the outcome. Mann writes in his paper:

Indeed, the detrended-AMO approach (Figure 4b [here 8b]) yields an apparent multidecadal AMO oscillation that is coherent across the indices, an artifact of the residual forced signal masquerading as an apparent low frequency oscillation. The apparent AMO signal is most coherent across indices during the most recent half century, when the forcing is largest. Another important feature apparent in this comparison is that the low-frequency noise leads to substantial perturbations in the overall “phase” of the apparent AMO signal (Figure 4b [here 8b]) giving the appearance of a propagating wave or stadium wave in the parlance of Wyatt et al. [2012].

However, the thick blue line in Figure 8b, plotted by Mann’s code but missing from his published figure, shows the AMO signal as estimated by Mann’s detrended-AMO procedure applied to the noise-free EBM simulation, gives the lie to this claim. Rather than being “an artifact of the residual forced signal masquerading as an apparent low frequency oscillation”, the thin coloured lines are seen as modified, phase-shifted versions of the signal from applying the detrended-AMO approach to the noise-free EBM simulation.

a)

clip_image018

b)

clip_image020

Fig 8. Version of Mann’s Figure 4. Comparison of (a) true pseudo-NH AMO signal (as a priori defined using the differenced-AMO approach) and (b) NH AMO signal as estimated by detrended-AMO procedure. In both cases, no actual observational data is used: results are shown for the five synthetic standardized climate indices as described in Mann’s text, derived using his EBM simulation with added noise realisations (thin coloured lines). In (b) the result of applying the detrended-AMO approach to the noise-free EBM simulation is also shown (thick blue line; scaled version of that in Figure 3: omitted in Mann’s Figure 4). Series are standardized to have unit variance.

The results shown in Figure 8b are what one would expect to arise: a noise amplitude that is greater relative to the signal than before causes more modification and phase-shifting of the clean signal: (compare Figure 8b with Figure 6b). The extent of the differences between the coloured lines in Figure 8b derived from the noisy synthetic AMO-related indices and the blue line derived from the noise-free EBM simulation varies with the random realisations of noise, and can be much greater. The corresponding graphs (Figures S8.b9 and S8.b10) based on the GISS-E2-R and CMIP5 Ensemble simulations show a gradual loss of coherency between the five noisy versions and the detrended-AMO based on the forced simulations (the blue lines, which do appear in the graphs in Mann’s Supplementary Information). Before ~1960, the GISS-E2-R and CMIP5 Ensemble simulations do not follow the real-world detrended-AMO signal as well as Mann’s EBM simulation does.

Results using the low-sensitivity EBM

So far, I’ve been repeating Mann’s analysis using his EBM simulation. Now I’ll look at what happens when my low-sensitivity EBM simulation is used instead. Figure 9 shows the same as Figure 3 (my version of Mann’s Figure 2a) save for my low-sensitivity EBM simulation being used instead of Mann’s EBM simulation. Unlike the situation with Mann’s EBM simulation, the thick blue line (detrended-AMO based on EBM simulation) is not almost identical to the thick red observational detrended-AMO line, and the thick black line – the differenced-AMO – does bear a resemblance to the observational detrended-AMO.

clip_image022

Fig 9.Estimated NH temperature anomaly variability 1900-2012: thin and thick lines are respectively annual, and 50-year low-passed smoothed, time series. The red lines show detrended observed (HadCRUT4) NH anomalies, the thick line being the detrended-AMO. The black lines show the observed NH temperature anomaly minus the low-sensitivity EBM simulated temperature, the thick line being the differenced-AMO. The blue lines show detrended anomalies from the low-sensitivity EBM simulation, the thick line being what the detrended-AMO would be if based on the EBM-simulated rather than observed temperatures.

Figure 10 shows the same as Figure 6b (Mann’s Figure 3b) but using the low-sensitivity EBM simulation instead of that from his EBM. It is now fairly obvious visually that the coloured lines resemble the blue dashed line that represents an application of the detrended-AMO approach to the EBM-simulated temperatures, rather than resembling the black line representing the detrended-AMO derived from observed NH temperatures. That is confirmatory evidence that my analysis of what is going on is correct.

clip_image024

Fig 10. Version of Figure 6.b) based on the low-sensitivity EBM simulation

Is the detrended-AMO nevertheless questionable?

The detrended-AMO approach is not perfect, even when applied – as is standard – to SST temperatures in the North Atlantic, not to the full NH land and ocean surface temperature as Mann does. When the rate of increase in forcing secularly increases, as it has over the last hundred years, it is possible that the detrended-AMO may be biased towards high strengthening in recent decades. A comparison of the post mid-1970s segments of the red line (detrended-AMO from observations) and the black line (differenced-AMO from low-sensitivity EBM simulation) in Figure 9 illustrates this point. However, the basic shapes of the two lines are similar, and the differenced-AMO still accounts for about a 0.2°C rise in NH temperature over the last thirty or so years.

It would be preferable to find a way of estimating the AMO that was more independent of forced temperature trends. That is, in effect, what Delsole et al (2011) did in estimating their internal multidecadal pattern (IMP) in global SST. They employed a sophisticated statistical method based on maximising average predictability time, using simulations by a number of CMIP3 coupled GCMs as well as observed SST, to separate forced and unforced variability in SST. Although their method applies globally, the IMP they detect is remarkably similar to the standard NOAA detrended-AMO index. Figure 11, a reproduction of Figure 4 from Delsole et al’s paper, compares NOAA’s AMO index, suitably rescaled, (red line) with the ±1 standard deviation uncertainty range of their estimated IMP (shaded grey). The fit is remarkably close.

Using a different sophisticated statistical approach, Swanson et al (2009) also found an AMO-like pattern of multidecadal unforced variability, here in GMST rather than global SST, although with a somewhat lower recent level.

clip_image026

Fig. 11. Reproduction of Figure 4 from Delsole et al (2011). ±1 standard deviation uncertainty range of their estimated IMP (shaded grey) and scaled AMO index from NOAA based on detrended North Atlantic SST (red line). The vertical scale is arbitrary.

Finding a physical explanation for the AMO is of course desirable, and likely to lead to better estimation of its influence on temperatures and other climate phenomena both globally and regionally. That is a major attraction of the stadium wave theory. If it holds up under further examination, it promises a better understanding and estimation of multidecadal internal climate variability. Other papers, such as Dima & Lohmann (2007), have also put forward possible natural physical mechanisms for the AMO.

 

Conclusions

I have shown that the evidence Mann claims disproves the detrended-AMO, and supports his differenced-AMO, is illusory. I have also shown that his code produces different results from those shown in his accepted paper. I have pointed out that graph lines produced by his code that would have made it much easier to spot the flaws in Mann’s evidence, although appearing in the figures in his Supplementary Information, were omitted from the figures in his main paper.

A differenced-AMO approach has attractions in principle, but only makes sense if climate models are near-perfect, which is far from the case. The ease with which a simple EBM model can have its parameters adjusted to produce a nearly flat differenced-AMO shows the very low number of degrees of freedom involved, with only two full AMO cycles during the instrumental period. The very heavy, 50-year low-pass, smoothing applied by Mann arguably exacerbates this problem.

The detrended-AMO approach is not perfect, but the pattern exhibited by NOAA’s standard detrended AMO index based on North Atlantic SST appears to be supported by much more sophisticated approaches. The stadium wave theory, if it holds up, offers physical insight into the mechanisms underlying the AMO and may lead to more reliable estimation of its state and influence on surface temperatures and other climate variables.

Nicholas Lewis

A pdf version of this article is available here.


[i] Enfield, D.B., A.M. Mestas-Nunez, and P.J. Trimble, 2001: The Atlantic Multidecadal Oscillation and its relationship to rainfall and river flows in the continental U.S., Geophys. Res. Lett., 28: 2077-2080

About these ads

100 thoughts on “Mann’s new AMO paper: ‘Had I been a reviewer, I would have pointed this out and recommended rejection. ‘

  1. “They compared observed temperature variation with a variety of historic model simulations to create a model for internal variability of the AMO”. Hey wait, they modelled models in a model? Move on reality, nothing to see here.

  2. Statistically torturing the data until it shows what you wish to see is not science. That’s the history of AGW in a nutshell.

  3. If the supplied code does not give the results that Mann shows in the paper, how did it pass review?

  4. “A differenced-AMO approach has attractions in principle, but only makes sense if climate models are near-perfect, which is far from the case. The ease with which a simple EBM model can have its parameters adjusted to produce a nearly flat differenced-AMO shows the very low number of degrees of freedom involved, with only two full AMO cycles during the instrumental period.”

    Yes. If you assume that there is a perfect Tinker Toy reproduction of the Eiffel Tower then you can argue that your Tinker Toy reproductions approach that perfection and that your competitors’ work moves in the other direction. Surely, by now, I do not have to explain that the necessary assumption cannot belong to science. Whether computer models or statistical models, the number of moving parts in these methods prohibits testing of individual parts on their own; hence, the Tinker Toy metaphor. Theorists following these lines are limited to creatively mixing and matching “ad hoc” hypotheses.

    “The stadium wave theory, if it holds up, offers physical insight into the mechanisms underlying the AMO and may lead to more reliable estimation of its state and influence on surface temperatures and other climate variables.”

    Right, the stadium wave hypothesis posits a physical sequence, the wave moving around the stadium, which stands as a falsifiable physical hypothesis. That part is not analogous to Tinker Toys and cannot be replaced by an never ending sequence of “ad hoc” hypotheses.

  5. Thank you for this analysis. Rather than hiding the decline, here Mann hides the fail. But only in the main paper! Leaving it exposed in the SI and in the code. More evidence of incompetence.
    Given his own goal, you should request either a correction or a retraction from the journal.

  6. Nic,
    You should expect to be served regarding Mann’s suit against you any moment now.

  7. Closing the barn door after the AGW horse has left (17.5 yrs ago)… never works. GK

  8. MattN, you ask how it passed review – simples, the reviewers didn’t bother checking the paper in depth. The reviewers will have known who the authors are and would have automatically OK’d the paper without thought – why would they bother questioning “God”.

  9. Nic Lewis writes:
    “I have shown that the evidence Mann claims disproves the detrended-AMO, and supports his differenced-AMO, is illusory. I have also shown that his code produces different results from those shown in his accepted paper. I have pointed out that graph lines produced by his code that would have made it much easier to spot the flaws in Mann’s evidence, although appearing in the figures in his Supplementary Information, were omitted from the figures in his main paper.”
    _______________________
    This brief paragraph is about as damning as a peer review can get. Is anyone at Penn State, or elsewhere within the “Climate Science” community paying attention?

  10. “The true AMO signal, instead, appears likely to have been in a cooling phase in recent decades, offsetting some of the anthropogenic warming temporarily.”

    Making the claim that natural changes in the climate are strong enough to “offset” anthropogenic warming is an admission that they can also be strong enough to have caused much of the warming in the first place.

  11. This article if far too long for the time I have available today, however, the flaw in Mann’s work is much simpler, it seems. (Notwithstanding what looks like a credible criticism of redefining what AMO actually is!).

    The key flaw is idea that you can run a 50 filter up to the end of the data, you can’t (if you correctly align the phase and don’t let the filtered result lag 25 behind the data).

    To do this you are either padding that data, a la “Mike’s Nature Trick (TM), or you use and “adaptive” filter that changes from being a 50 y filter to a 25 year filter by the end of the data, or in some other undefined way that is determined by the data itself, and thus has not objective filter characteristics independent of that data.

    In sum what happens at the of the data is to an extent fictional and is not directly comparable to rest of the graph. Just how it varies depends upon which “trick” was used and whether the data was rising , falling or flat near the end.

    Now since the whole point of this paper seems to be to redefine the AMO and draw conclusions about how this affects the end of the data the whole effort is fundamentally flawed and compromised before the first graph is produced.

    I see little point in going into the extreme detail that Nic has done here. The last 25 years of the graphs are bunkum and can be chopped off. What is then remaining may or may not merit discussion.

  12. Nic – You have just won a trip to the Mann-Steyn trial. I am guessing there is a good reason why you were not selected as a referee.

    Engaging in academic debate over science is healthy. Good for you. The openness of the internet is superior to the closed door peer review process. Still hope some day we can launch an internet academic journal on climate science with open reviews such as this one. Perhaps the University of Phoenix could sponsor one.

  13. Climate science is really all about Simon Says pronouncements. Next up, the sun is not a factor at all.

  14. From lowpass.m :
    % (0) pad series with long-term mean value beyond x boundary
    % (1) pad series with values over last 1/2 filter width reflected w.r.t. x
    % [imposes a local maximum/minimum at x boundary]
    % (2) pad series with values over last 1/2 filter width reflected w.r.t. x and y (w.r.t. final value)
    % [imposes a point of inflection at x boundary]

    Like I said, padding.

    Figure 4 , reporting Mann’s 2a, shows the absurd way the “smoothed” result bends downwards at the end when driven by data that is strongly rising. That is pattently a gross distortion, and artefact of the filter.

    Any conclusion or discussion based on how these lines run in the last two decades is a pointless discussion of the artefacts of naively flawed data processing.

  15. “… the absurd way the “smoothed” result bends downwards at the end when driven by data that is strongly rising. ” That comment specifically refers to the blue line which is still showing a strong monotonic climb at the end despite the “filter” managing to produce a downturn.

  16. Have the models Dr. Mann uses been calibrated to faithfully represent at least a part of the time frame in question?

  17. Outstanding work by Nic Lewis!

    I think many papers in many fields manage to slip thru with issues like this but after Mann, obviously and blatantly created the hockey stick graph, which replaced all previous global climate history prior to the Industrial Revolution, he drew attention to himself as a fraud. Now, all his work is carefully scrutinized by skeptics.

    If you catch somebody telling a big blatant lie, you scrutinize everything else they say after that point. This means your work had better be authentic because under the microscope many things show up that are not otherwise seen.

    With Mann, it’s even worse. After displaying his clear bias early on, it’s almost impossible to not see this in all his work since then by just looking for that bias.

    It’s likely, that even when he producers work that has authentic concepts, if it can be interpreted as subjective and intended to support CAGW, then that is what will be seen by skeptics.

    Once you destroy your credibility, it’s impossible to get it back.

  18. Yet another serious error is Mann’s use of multivariate regression “best fit” for his energy budget model.

    “The best fit to the observational NH series (82% variance explained) is achieved using
    an aerosol scaling factor of 1.2 (i.e. assuming that indirect aerosol forcing increases
    ERF by 20% relative to the direct forcing), linear scaling of volcanic optical depth
    with aerosol deposition, and a 0.25% Maunder Minimum-present solar forcing scaling
    assumption. These are the standard settings for the EBM experiments.”

    Linear regression can only be correctly applied to data with a well-defined x-variable, not one with error. Ignoring this will produce a lower value ‘slope’ than the real relationship. See:

    http://climategrog.wordpress.com/2014/03/08/on-inappropriate-use-of-ols/

    Doing this simultaneous over several variables and you get the same error in spades.

    The fact that you may be able to model your selected dataset (temperature) to within 80% is immaterial. This does not demonstrate (let alone prove) that you have got anywhere near the physically real ratios. If you throw in a similar number of red noise random series and regress them will likely to do nearly as well.

    Added to this, it is not even correct to fit any of the radiative” forcing” time series directly to the temperature data that you are assuming to be the end result. The surface (especially water) takes time to warm, this is an integration process, which also provokes at least the Plank negative feedback.

    The net result is to be found through a convolution with an exponential impulse response.

    Fitting temps directly to rad. forcings will get both the lag and the magnitude of the response wrong. The regression process is then left playing off one set of errors against another and fits some totally erroneous mismatch of conditions that have little physical reality.

    Mann, as ever the expert in ad hoc home spun methods, lacks any understanding of what he is doing or of basic science and engineering methodology.

    Sadly that is not unusual in climatology and his prominence probably just reflects the mediocrity of his peers.

    I fully expect he will be awarded another ;) Nobel Prize for this work.

  19. If you need this length of discussion to deal with it, it loses punch. A good elevator summary is much needed here. I’m afraid I come away with strawberry is better than chocolate. Is it safe to say that Nic’s point is that the AMO is a natural variation that contributes to NH or global temp (cause and effect) and Mann’s is that warming and aerosols creates the AMO (reversal of above cause and effect).

    It seems to me that the AMO and PDO were largely ignored by MS climate science (but not skeptics) up until skeptics spoiled the party first and the “hiatus” added an uppercut later that sent them spinning. With a couple of years hiatus in publishing by the bewildered team and a flurry of pent up skeptic papers turned loose, they had to come out with a fury. There will be a lot fanciful salvage going on as they come out of their punch drunkenness.

  20. Interesting article on NPR this am.

    Forgot what the segment was exactly but it seems some researchers have found out how modern research is a about reinforcing current scientific thought and not about a quest for knowledge.

    In short the researchers found that the process was flawed and the process would not allow articles into ‘peer reviewed’ publication if the research contradicted earlier research or if the research could not replicate the previous study.

  21. I have no problem with alarmists accepting this latest Mann-uver. Never get in the way when the opponent steps in the doo doo.

  22. from: http://arxiv.org/pdf/math/0305364v3.pdf (frequency analysis of quasi-periodic signals)

    “The Nyquist aliasing constraint means that to recover a given period, one needs
    to sample the data with at least two points per period. On the opposite, in order
    to determine precisely the long periods, one needs that the total interval length T
    is several time larger than these periods, in order to reach the asymptotic
    rates of theorems 1 and 2, or to be able to separate properly close frequencies.”

    There’s only 2 periods of AMO present. Not enough data to resolve low frequency signals with any sort of phase or amplitude accuracy.

    I also agree with the filtering problems noted above by Greg Goodman.

    I also note the use of boxcar averaging. , that’s 19th century analysis technique…

    Someone with a signal analysis background should be reviewing these kind of papers…

  23. Who is Nic Lewis? There should be a short paragraph which states, “This is a guest post by Nic Lewis, who is [professor of climate studies], [gifted amateur] who [regularly writes on AMO topics / does this stuff for fun].

  24. “I also note the use of boxcar averaging. , that’s 19th century analysis technique…”

    Nick seems to be the one to have brought runny means into it, maybe as an attempt to explain to those for who that is the only “smoother” they know.

    The paper’s filter is apparently based on a Butterworth function in Matlab. The problem is the padding I highlighted above and the flawed regression techniques.

  25. “The paper seeks to overturn the current understanding of the AMO, and provides what on the surface appears to be impressive evidence.” That is exactly what he tried to do with the temperature record of the past millennium, by the Hockey Stick.

  26. Gary Pearse says:
    May 19, 2014 at 10:07 am
    If you need this length of discussion to deal with it, it loses punch. A good elevator summary is much needed here.
    >>>>>>>>>>>>>>>>>>>

    My reading of it is that Mann:

    1. Fiddled with a variety of parameters until he got a curve fit that supported his hypothesis
    2. Hid contrary results produced by his own analysis by simply not presenting them

    The latter point could be easily fall into the category of “never attribute to malice that which can be explained by incompetence”. But given Mann’s track record ranging from the “Nature trick” to the Tiljander issue, to the use of known bad proxies such as strip bark pines to the original hockey stick graph being produced by code with a predilection for producing hockey sticks regardless of data, I personally would lean toward malice.

    As for the first point, this is an error that climate scientists seem to make over and over and over again. There are thousands, perhaps millions of parameters that could be tweaked this way and that way to produce a match to the past. He could have achieved his curve fitting by adjusting any number of different parameters in different ways, gotten an excellent fit to the past, and produced a completely different conclusion. Over and over again they tweak a few parameters, get a fit to the past, and announce they have a good model. Then the future doesn’t pan out the way the models predicted, and off they go tweaking a different set of parameters in a different way until they get yet another curve fit to the past that doesn’t predict the future. With the number of parameters there are to tweak, and the different ways there are to tweak them, they may eventually stumble upon the right combination. My own though on this is that given a race between this approach and a million monkeys with type writers attempting to reproduce by chance the collected works of William Shakespeare, my bet would on the monkeys.

  27. The more you wish to maintain your beliefs, the more you dig and thereby find ways to support for your beliefs: Its basis becomes increasingly complicated but the search is successful. Wherever negatives exists you can positives to counter them, even if they become exceedingly artful. The notion that a good or true belief is found with simplicity is abandoned: the universe pivots around the Earth through an intricate system of wheels and cogs.

    What I see here is an increasing technical sophistication that is, underneath, little more than wiggle-matching. This is not to say that the conclusion is incorrect – you can be right for the wrong reasons, after all. But what is happening is that the search for support is increasingly tenacious and its results, opaque. Simplicity which can be understood and supported or discredited directly has been tossed aside; only specialists can recognize the weakenesses or strengths of arguments used. Rebuttals have degenerated into one expert’s complaints about one thin thread or another, while the collective whole is no longer what counts. Apparently you cannot legitimately recognize a forest anymore if you are not personally familiar with every tree.

    We are in the times of scientists as lawyers. Legalese has infected climatology: the principles no longer matter but the particular words in reference to a particular case. If we were to see climatological issues as Constitutional rights, we could say that Freedom in its general sense has been lost while freedom to do this little thing in this particular place in this specific place at this noted time is trumpeted. We are the child with privileges, not one with rights.

    In climatology, CAGW is no longer the issue but whether two weather stations in Akron, Ohio, can be matched to eighty-three localized processes simulated in one researcher’s Cray computer. No longer does the whole determine the part, but the part now is claimed as an axiom to reflect the whole. We are being lead not by explainers of truth but by the pursuasive con man, the difference between a con man and a common liar being that a con man gives his lies the surface consistency of truth.

  28. Nic Lewis:

    I have read your article but have yet to read the subsequent thread. I write to congratulate you.

    You say

    I have shown that the evidence Mann claims disproves the detrended-AMO, and supports his differenced-AMO, is illusory. I have also shown that his code produces different results from those shown in his accepted paper. I have pointed out that graph lines produced by his code that would have made it much easier to spot the flaws in Mann’s evidence, although appearing in the figures in his Supplementary Information, were omitted from the figures in his main paper.

    A differenced-AMO approach has attractions in principle, but only makes sense if climate models are near-perfect, which is far from the case. The ease with which a simple EBM model can have its parameters adjusted to produce a nearly flat differenced-AMO shows the very low number of degrees of freedom involved, with only two full AMO cycles during the instrumental period. The very heavy, 50-year low-pass, smoothing applied by Mann arguably exacerbates this problem.

    Yes, you very clearly have shown what you claim. And, yes, your adjustment of parameters does demonstrate the effects of assessing only two cycles.

    Well done! Thankyou for sharing your assessment.

    And, as you say, your findings would have merited rejection of the paper if provided at review.

    Richard

  29. This is what happens when Mann shows ‘all his data’ and now you can understand why he tends to hide it instead. in the end he can only produce what he is capable off , and that is not much.

  30. Mods:
    I have been having problems posting to WUWT. Most recently I made a post to this thread which has vanished. Please let me know if it is not in the ‘bin’.
    Richard

  31. Can you imagine the negative press this would generate if this was a skeptical paper? The alarmists would be calling for the person to be stripped of their degrees or worse. The “f” word would be shouted from the highest levels.

  32. Is there a Climategate email where someone says “we really need to get rid of that AMO warm phase”?

  33. The grey lines are Mann’s new AMO.

    Notice how the 60 year cycle is gone. Notice how it started dropping around 1995.

    Notice how it bears absolutely no resemblance to north Atlantic sea surface temperatures.

    How can it be an AMO when it is completely different than what it is supposed to be measuring, the cyclic temperatures in the north Atlantic.

    Just-make-it-up-again-Mann.

  34. “””””….. It alternately obscures and exaggerates the global increase in temperatures due to human-induced global warming……”””””

    What is it about the AMO that enables it to “alternately obscure(s) and exaggerate(s)” ONLY the global increase in temperatures due to human induced global warming ??

    How can AMO differentiate human induced from natural ??

    Inquiring minds want to know.

  35. I really don’t understand why this site would even consider ANYTHING from this fraud, Even rebutting him is a waste of time and giving him attention he does not deserve please refer to the hockey stick fraud.

  36. “The true AMO signal, instead, appears likely to have been in a cooling phase in recent decades, offsetting some of the anthropogenic warming temporarily.”

    ‘appears likely to have been in a cooling phase in recent decades’

    …So even when referring to the past the claim is: ‘appears likely’

    …Its the past! It either was or it wasn’t!

  37. davidmhoffer says:
    May 19, 2014 at 10:25 am

    Right on the money. Mann is being Mann. He truly believes that if he can show some statistical connection that supports his conclusion then it is incontrovertible. He seems to think that he is King Henry VIII..

  38. I know I’m going to take a lot of flak for the following statement but please hear me out. Michael Mann is, in the end, nothing other than basically a numbers cruncher. And, for that reason, I’m quite glad he works in climate science. Otherwise, since he’s essentially a government employee anyway, he’d probably be doing research for the IRS.

  39. Mann makes an assumption that should be questioned. The historical data is nowhere near long enough establish the periodicity of the cyclical inputs. We talk about 11 year solar cycles but what’s the spread. Everyone concerned about amplitude but what about the error bars on periodicity?

  40. While we’re all being reviewers, I see that IOP has now published the second review of the Bengtsson paper. It’s devastating. There is no way an editor could have published, based on those reviews.

    But you won’t read about it in the Times.

  41. Nick Stokes says:
    May 19, 2014 at 12:03 pm

    So, the score so far is one review ignorant, unprofessional, and openly biased, and a second review that is serious and negative. I am waiting for the third review with great interest.

  42. Anthony this may be a bit of topic but could be said on any thread (it probably already has), thanks for your site it has and is an enlightening place to visit and learn from.( edit at will)

    The more I think about science in general, the more I see the larger problem. I am not sure how to express myself but what I see is a professor in front of a class teaching what he had been taught ( can I say indoctrinated with?) 30 years earlier by a professor who in turn was taught by his professor 30 years prior. So that to me seems to me that a whole lot of science today could be based on 60 year old thinking (indoctrination and paranoia) and methods. Although I am sure there are institutions very up to date and doing their work quietly.
    But then add in the relentless drive for grants, outside funding, patents, television time, etc. etc. and to me today’s students have been thoroughly set on a path of singular thinking as they are well aware that IF they stand up to their “teachers” it could deny them any kind of future in their chosen fields.
    That is a sad and terrible thing to see and witness. Hopefully some of them in this day and age of the internet will change that thinking process and can stay “anonymous” and break their shackles.
    That is my main reason for opposing any kind of regulation on the net. I am in my 60’s and I hope I will see a new age in my lifetime. Thanks, Asybot.

  43. And using pre-1979 HadCrut as real historic data ..

    Not a very sensible starting point.

    Failed at the first hurdle.

  44. Here we are a whole day after “Why don’t we all just agree on Global Warming?
    Posted on May 18, 2014 by Kip Hansen says in points 2 & 4 that some data is not very good and that scientists are positing much greater certainty than the data deserve.

    Golly gee, Mr. Wizard!

    I see no standard deviations plotted (by either of these folks), SST accuracies plotted to the closest tenth or hundredth of a degree C back to 1900 with probably as many as 1 sample per 1,000 – 10,000 – who knows how many square miles using liquid in glass devices with limits of observation equal to plus or minus 1 degree C more or less possibly…. with buckets

    As Van Halen sings “Woo! It’s got what it takes, So tell me why can’t this be love?”

  45. Greg Goodman provides some excellent commentary on why a 50 year filter on a graph ending in 2010 fails filtering 101. Nice work Greg.

  46. Nick Stokes;
    While we’re all being reviewers, I see that IOP has now published the second review of the Bengtsson paper. It’s devastating. There is no way an editor could have published, based on those reviews.
    >>>>>>>>>>>>>>

    I’ve read both reviews. The first amounts to a smear job I could discredit paragraph by paragraph if I had the time. The second review raises some concrete issues, but the issues it raises as “wrong” are actually quite debatable. Insisting that ECS is the right metric rather than TCS for example is not an issue of right or wrong as the reviewer suggests. Reality is that both metrics have merit. Labeling one as being completely wrong for this purpose would, BTW, invalidate an enormous number of papers supportive of the global warming meme.

    But all that is immaterial. If we accept your premise that the second review is “devastating”, then by the standards of that review, not a single thing that Michael Mann has ever published should have passed peer review. Not to mention swaths of other papers by other authors that anyone with 1st year physics and/or stats could see through.

    If you’re going to apply quality peer review standards Nick, by all means do so. But do it fairly, do it across the board. My prediction is you won’t, because your side of the debate would have precious little left to refer to.

  47. Mickey Mann-“The true AMO signal, instead, appears likely to have been in a cooling phase in recent decades, offsetting some of the anthropogenic warming temporarily.”

    =================================================================
    Soooo…………The missing heat that is hiding in the ocean is causing the ocean to cool?

  48. “… this is an error that climate scientists seem to make over and over and over again. There are thousands, perhaps millions of parameters that could be tweaked this way and that way to produce a match to the past. He could have achieved his curve fitting by adjusting any number of different parameters in different ways, gotten an excellent fit to the past, and produced a completely different conclusion. Over and over again they tweak a few parameters, get a fit to the past, and announce they have a good model.”

    This is a great point that needs to be made over and over again. They are just playing with numbers to create worthless charts and graphs to con the uninitiated. Since math education in this country is horrifically bad, they seem to get away with it. Could you imagine the difference it would make if honest Statisticians where reviewing these “papers” before publication?

  49. Mods

    Please explain why the above comment by me, or any other one, is not allowed to be posted on this site if I use my WordPress account. Have I been blocked for some reason? If so, just tell me and I’ll not bother you again.

    REPLY: Don’t know, but you aren’t “blocked”. Probably due to the fake email address you provide of name@twitter.example.com Try not to be so upset about the spam filter doing its job and have patience while moderators recover your comment. – Anthony

  50. IMHO Michael Mann missed his calling to riches by working in Climate Science than for Bernie Madoff in a hide the deception scheme.

  51. Mods.
    I think our host misunderstands. I have posted here of a long time using my WordPress account. Over the last couple of days the system just goes to a blank screen and tells me “sorry, this message can not be posted” or something like that when I try to do that. But after the first time I learned to save my comment and then log in with the Twitter button rather than the WordPress button. This works and the message was posted. It is the whole idea that I have been using the WordPress button for a long time and all of a sudden it thinks I can not post as long as I try using the WordPress button there in the lower right hand corner of the comment box, And yet the Twitter button works fine. This is weird.

  52. Good stuff Nic. very good
    I understood every word. every single word.
    But when you strung them together , I was totaly lost.

    So. Can somebody explain to a layman. based on what Mann has claimed

    will Manhatten be under water in 14 years from now ?

  53. Jason H says:
    May 19, 2014 at 10:46 am
    I’ll bet this paper was accepted for publication on the day it was received.
    ==============================================================
    Most likely it was already pre-approved

  54. Nic Lewis,

    It looks as if you may have something there. This should now be written up, if you have not already done so, and submitted to Geophysics Research Letters, as a “Comment on the Paper by Mann et. al.” which will force external referees to vet it and also will force Mann et. al. to write a reply to your comments. If he is truly wrong he will have to admit it.

  55. Meanwhile in the MSM today: “Read all about it! The Bigggggggggggg Mellllllllllllllllt Accccellllllllllllllllllerrrrrrates!”

  56. So what is the forecast over the coming 20-30 years. The AMO is flipping to cold now, and the result is the positive anomaly in the summer sea ice we see the CFSV2 seeing, a huge and unforecasted event ( it is the year of no ice cap you know as the 07 death spiral was supposed to be taking its toll). But Dr Mann should tell us, where this is going to go. It seems that if his forcing idea is correct why should this turn down. Factually Grays paper destroys this and his ideas are over 40 years old, by his method the amo is tracking right along his idea from the 1970s. But Dr Mann again refuses to make a forecast that we can test. Where does this go in 2020-2025-2030, instead preferring to explain what happened and simply say I have the explanation. So what is your forecast. I see your idea, take a stand and give us a product of that to test.

    There is no testing with hindcasting. Researchers may think so, but until the produce verifiable forecasts off their research that can be testing, there is no reason to accept as any more than speculation on the past. But when Bill Gray comes out and says he expects something from 1978 to happen in the coming 30-40 years, and it does, that explanation has merit!

  57. Nick Stokes says: While we’re all being reviewers, I see that IOP has now published the second review of the Bengtsson paper. It’s devastating.

    Indeed. The first reviewer asserts that the models were never intended to reflect reality. That’s pretty devastating! The second does a bait and switch with TSR / ESR…equally devastating!

  58. Dr Mann just said in his HuffPo article that “Part of science is recognizing your own errors, as is pointing out the errors of others”.

    So if he fails to engage with Nic Lewis about this topic, by his own definition Dr Mann becomes #antiscience !

  59. If your paper consists of respectable scholarship, why are you having this debate on a website instead of submitting it to Geophysical Research Letters for consideration and publication so that the original article – the one whose methods and conclusions you dispute – may be retracted?

    Please do submit your analysis for publication.

  60. Well Nicholas, I think that this pretty much guarantees that you will never be asked to review anything from any member of “the Team” for any journal that they submit papers to …

  61. ‘ Because the researchers know the true AMO signal for their synthetic data’

    Synthetic data? as opposed to real data? If it’s synthetic it ain’t real and it ain’t data.

  62. Awwww, aint that cute Mann is relevant again.

    I laugh to myself sometimes contemplating what relics of the Cagw media frenzy might stand the test of time as what not to do with a scientific issue. I often decide that youtube video “hide the decline” featuring mann will likely be towards the top of that list.

  63. hmm don’t know why that link won’t work here, sorry Mod., do with it as you will…..

  64. Initial conditions everywhere I look. Here the initial condition is that the model is right and reality is expressed as a delta. Wow.

    The detrending argument has a bit of merit, but the detrending does not change the structure or shape of the data, it just levels it. The real question is the genesis of the trend, which we do not know but the model purports to know.

    This guy fancies himself Mendelev, who could consult his model (the periodic table) and predict properties.

    Taking great liberties with a political quote: “Mike, I knew Mendelev, and you are no Mendelev.”

  65. For comments (e.g., W. Taylor) suggesting that the best or only place for Nic Lewis to seek to publish his analysis is in GRL, that journal apparently does not accept any comments or replies at all upon its published articles:

    http://publications.agu.org/author-resource-center/author-guide/grl/

    COMMENTS/REPLIES

    Owing to changing Editorial policies, GRL no longer considers comments and replies for publication.

    Oddly, “changing editorial policies” is not really a reason of any substance. It is like a statement “we no longer accept comments and replies because we no longer accept comments and replies.”

    uh huhhh that really explained a lot, GRL.

  66. Then, ofc, other journals would decline any comment or even full article from Nic on this topic because they would say it belongs in GRL.

    Also, ofc, there are generally extreme length limits extremely short) in all journals, which renders it nearly impossible to get this kind of detailed analysis off the ground.

    Nice way to hermetically seal the “scientific record” when you don’t like anything to upset your little applecart.

  67. Are we seeing a shift in the epistemology of climatescience where the burden of proof is lifted from the author to the reader? It is as if scientists put out a claim and say it is valid until you, the reader, proove beyond a reasonable doubt that the claim is impossible. I am aexaggerating a little to make my point understood, but I find it troubling. This adhoc “science” is not leading to progress, only politics and confusion.

  68. This paper by Mann fits in well with the Bengtsson paper rejection episode in highlighting a trend for the climate science establishment to reject real world observation in favour of models.

    Mann’s new version of the AMO does not even require the existence of the Atlantic ocean. All it requires is computer programmers and cultural expectations of climate which the models can be designed to satisfy.

    Likewise Bengtsson has become a marginalised heretic for daring to show evidence that climate models depart from reality. The Institute of Physics now no longer requires any real physics, it is just the Institute of Propaganda. Models represent self-contained and complete revealed truth, the requirement that they have any correspondence with real world observations has been quietly brushed aside.

    What this represents is very profound, it is an undoing of or a reversal of the renaissance. The “renaissance”, meaning re-birth, is the term used for scientific and cultural developments in Europe starting in the 1500s, which became forward looking and self-confident based on scientific progress, as opposed to backward looking nostalgia for the Greco-Roman classical civilization and a feeling of inferiority compared to the ancients. At its heart was the novel philosophy of the scientific method – led by individuals such as Galileo, Vesalius and Brahe – in which questions about how the world worked were addressed by carefully planned experiments and observations, rather than sitting in an armchair and pontificating about “reality” based on pre-existing societal predjudices. Galileo was a “classic” example – rather than accept the established opinion that a heavier object would “obviously” fall faster than a lighter one, he tested it experimentally at the leaning tower of Piza (a tower built on theoretical or model assumptions as to the strength of the underlying ground) – and found that they fell at the same speed.

    Galileo was imprisoned for his efforts. Now again, acts aimed at overturning predjudice with experimental logic-based reality are becoming equally hazardous for their proponents.

    So the real world is becoming a politically forbidden zone, a cloud of unknowing, the realm of skeptics and “deniars”, not of anyone who desires a career in climate “science”.

  69. Doug Proctor says:
    May 19, 2014 at 10:31 am

    “The more you wish to maintain your beliefs, the more you dig and thereby find ways to support for your beliefs..”

    A common rebuttal to skeptics is ‘it’s physics based’ and then they give a linear offering of the infrared absorption lines of CO2 in isolation from the cornucopia of other things happening including negative forcings created by the heating – essentially enthalpy of water in its various phases (a la Willis Eschenbach, Robert Lindzen and others). If this negative effect is as substantial as it appears, the ECS could 10 or 20 and it wouldn’t matter. The strongest data re this is that ocean surface temperatures have a maximum 30-31C, no matter what the strength of warming.

    In desperation, they probably are examining mini black holes and “strings” as supporting physics (although I believe these support fruitcake physics born of a very long doldrums in development of the science).

  70. Skiphil says:
    May 19, 2014 at 11:21 pm
    Did Mann coin the term “AMO”?
    =========
    In a world where Gore invents the Internet, all things are possible.

  71. Nic Lewis’s motivation for his “peer review” blog of Mr. Mann’s work comes from Mr. Mann’s defense of the free expression of those criticizing Lennart Bengtsson for joining the advocacy group, Global Warming Policy Foundation (GWPF). Mr. Bengtsson also recently had a paper of his rejected for publication by the science journal, Environmental Research Letters. As a result of Mr. Bengtsson’s criticism of that decision, ERL later released the reviews of Mr. Bengtsson’s paper so we could judge for ourselves if the critique was fair. This is a good opportunity to see actual peer review in action.
    I think it is a useful exercise to put Nic Lewis”s “peer review” of Mr. Mann’s paper next to the ERL reviews of Mr. Bengtsson’s paper. The tone of each give a good demonstration of what peer review is all about. The ERL reviews sound like three college professors providing useful comments to a student thesis. Constructive comments are given as a way to help make a better paper for possible resubmission.
    In contrast, Nic Lewis starts in to Mr. Mann first by failing to acknowledge the existence of the other two co-authors of the paper with Mr. Mann . This is to be ax-grinding about Mr. Mann first and the science of the paper second.
    The dispassionate nature of the ERL reviews stand in stark contrast to Nick Lewis’s passion. It would be interesting to see a dispassionate peer review of Mr. Lewis’s paper.
    This is the link to the ERL reviews of Mr. Bengtsson’s paper.

    http://ioppublishing.org/newsDetails/statement-from-iop-publishing-on-story-in-the-times

  72. Björn from Sweden says:
    May 19, 2014 at 10:34 pm
    This adhoc “science” is not leading to progress, only politics and confusion.
    ===========
    The underlying assumption of this paper is that the models are correct, and that aerosols explain the differences between models and observations.

    Then a sieve is applied to the data, until data that supports this assumption is located. This data is then published as evidence that the models are correct, aerosols are to blame.

    Using this approach one can prove black is white, up is down. There will always be some data to support any conclusion, no matter how wrong. All that is required is the time to search for it.

  73. If I study 1000 pieces of evidence, and 999 contradict my theory, is it OK to only publish only the 1 that supports my theory as evidence I’m right?

  74. matayaya says:
    May 20, 2014 at 7:22 am

    WOW, Im not sure how anyone could read both of these and come to the conclusions you did.

  75. EternalOptimist asks will Manhatten be under water in 14 years from now ?

    Well, the models probably think it is already

  76. From the premise that a warm AMO is directly associated with more negative NAO/AO conditions, there is no rational basis to any claim that increased forcing of any kind will give a warm AMO, as increased forcing produces more positive NAO/AO conditions.

  77. Nic,
    Climate science is being dragged kicking and screaming toward a reasoned view of climate sensitivity…. and you are pulling the hardest. Good work.

  78. matayaya says:
    May 20, 2014 at 7:22 am

    _____________________
    matayaya,
    Here are a couple of links which could really help you, a lot.

    http://carm.org/logical-fallacies-or-fallacies-argumentation

    http://www.nizkor.org/features/fallacies/index.html

    You see, your argument is weakened by your employment of a series of logical fallacies. In fact, there are so many defects in your reasoning that many here will unquestionably and completely discount anything which you might have to say.

    Why don’t you bone up on the information in those links and then come back and rewrite your post and next time, show us explicitly where the data, analysis and conclusions provided by Nic Lewis might be found wanting.

    After all, you don’t really want to be viewed as just another know- nothing, paid propagandist masquerading as a drive- by troll, do you?

  79. matayaya,

    I note in your link:

    The IPCC estimates of different quantities are not based on single data sources, nor on a fixed set of models, but by construction are expert based assessments based on a multitude of sources.

    You are no doubt not aware of the fact that the World Wildlife Fund, a self-serving NGO/QUANGO, has provided up to 40% of the IPCC’s ‘expert assessments’.

    Maybe you are not even aware that global warming stopped more than 17 years ago — a fact which, by itself, deconstructs the “carbon” scare.

    And maybe you don’t know that to date, the IPCC has been flat wrong in its all of predictions.

    Probably you don’t know that exactly none of the scary predictions by serial prevaricator Michael Mann have come to pass.

    I doubt that you know what the climate Null Hypothesis is, or that it has never been falsified — or even what that means.

    I don’t think you know that extreme weather events have been declining across the board, or that the sea level rise has not accelerated, as widely predicted.

    It is doubtful that you know much about the climategate emails, or that ‘Harry’ the programmer admitted that station data that the IPCC uses was largely fabricated.

    It is pretty clear that you don’t know that Arctic ice is rising fast, debunking another of the endless alarmist prophesies.

    And of course you don’t know enough to challenge any of Nic Lewis’ AMO arguments.

    So the question is: what do you know??

    What do I know? I know that Alan Robertson is probably correct in his final assessment above.

  80. The Navier Stokes equations describe fluid flow with changes in temperature and density. They are nonlinear, chaotic, with sensitive dependence on initial conditions. No finite set of data is sufficient to make a long term prediction. This has been known since the 1963 paper “Deterministic Nonperiodic Flow” by Edward Lorenz. If one attempts to predict distant futures from finite past data, one is either incompetent or a fraud.

  81. As far as I can see there is no satisfactory physical explanation for AMO. Futhermore, all graphical presetations of it that I have seen are a huge srtretch. It is time to stop imagining that it is an oscillation and realize that we are dealing with SST changes whose causes are unknown and periodicity non-existent. In contrast to that, ENSO does have a physical explanation and is traceable in all historic temperature records.

  82. stop imagining that it is an oscillation and realize that we are dealing with SST changes whose causes are unknown and periodicity non-existent.
    ========
    that would certainly be consistent with Mann having coined the phrase AMO.

  83. Referring to Nic’s “Fig 4“,

    Bill Illis (May 19, 2014 at 11:19 am) wrote:
    “The grey lines are Mann’s new AMO. [...] Notice how it started dropping around 1995.”

    See also the black line on Nic’s “Fig 5“.

    Bill (& others please take careful note),

    All Mann has done is extract the IPO (Interdecadal Pacific Oscillation). In spatiotemporal central limit, he’s admitting the models can’t do ENSO.

    I agree that IPO’s internal and that the NH (Northern Hemisphere) MD (MultiDecadal) wave is external, but Mann’s diagnosis of the external agents needs careful revision.

    Alert: Wyatt & Curry have conflated signals of a fundamentally differing nature in their “stadium wave”.

    …So Mann is both wrong & right.

    I should be able to provide some clarification days-to-weeks from now. (It won’t be accessible to a general audience, but I’ll share it anyway.)

    Regards

  84. “””””…..W Taylor says:

    May 19, 2014 at 9:00 pm

    If your paper consists of respectable scholarship, why are you having this debate on a website instead of submitting it to Geophysical Research Letters for consideration and publication so that the original article – the one whose methods and conclusions you dispute – may be retracted?

    Please do submit your analysis for publication……”””””

    Well it is very simple really Geophysical Research Letters; though it is indeed a respected and highly rated scientific journal, is printed in only small numbers, for a very restricted audience. I doubt that people without credentials in the field ; like me, could even qualify to receive a subscription to such a journal.

    Who knows; WUWT might actually have more readers than does GRL.

    Some very important publications may have extremely limited readership. I know of at least one major literature publication, that quite possibly NEVER had more than ONE person ever read it (all of it of course).

    That book, actually I recall, an eight volume set, in English, of the complete Prose Works of Richard Wagner, was in the music library of my alma mater; everything he ever wrote; that anybody knows about, and of course including the complete poems of all his music dramas; which he wrote for himself.

    While I was attending that place, I checked out each of those volumes, and read it from cover to cover. Almost 50 years later, a staff member checked in the library, and found that each volume still contained, the same signature card in the back, showing who all checked them out.

    Each card contained my signature on the very first line. Five of the cards contained no other signature, and none of the other three had more than five signatures; only one had more than two signatures.

    So the staffer, flogged those cards, and sent them to me. The books were finally entered into their computer data base. The well read volume (five readers) contains his very famous essay, on his concept of the Artwork of the Future…..(German: “Das Kunstwerk der Zukunft”) …..

    So people do tend to publish their stuff, where it is most likely to get read; such as here at WUWT.

  85. Like it or not.. Mann has an interesting argument. In very simple terms, the less pure the atmosphere from its optimal form the more the effect on the temperature of the system in question. That should be clear to any one a clear calm night, vs a night that is cloudy and calm. What is in the air does make a difference. In that aspect, he has “interesting points” But here is the problem. Its all fun and games until you have to make a forecast off it. I have argued with people that the test of their warming theories are here! The AMO is getting ready to flip.. my latest patriot post shows the linkage between the 1950s and now and how at the end of the warm amo, the water near the east coast is warmest http://patriotpost.us/opinion/26136 but the linkage between the late stages of the warm AMO, which was correctly forecasted by Bill Gray without Manns ideas in the late 1970s, ( at that time it was not the AMO) leaves Mann in the unenviable position of having to compete with someone who did not factor in any of his discoveries. Nevertheless if he wants to convince people, he needs to lead. What is his amo forecast.. using objective scales over the next 25 years.

    I suspect Mann has been boxed into a corner and is reacting accordingly on some other matters. The Hockey stick for instance, draws two major objections from me. If you are telling me the medieval warm period was local, you have to show me where it was cold. You simply cant say I am right, I dont need to, in science. Second the switching from tree rings because of pollution as I understand may simply be the correct answer given the circumstance. no different from a wrestler using one move the entire match, then switching to another that is better in the last 30 seconds. But SCIENCE IS NOT WRESTLING. You simply cant do that and then say its proof the way water boils at 212 or freezes at 32. That Liu’s study shows the current warm period, but also the global variation, simply says that the hockey stick is no better than Liu’s. It certainly says one of the two is wrong. But let me ask you this.. would you not defend your work? And then who among can safely say we are above walking away from the notoriety. That Mann will not debate anyone is a defensive reaction to the vitriol that is being hurled at people today. That it is no longer a matter where people can actually in good faith debate is because science is not in control anymore. Quite the contrary, it is a victim of what I think is a distasteful revolution in this nation that wishes to crush the individual. That Mann takes safety in numbers and power with people who find his work useful should be no surprise to anyone. That being said, if you can roll away the stone, you will find that he has some very interesting ideas. The key to become strongest is to make sure you also have your opponents strengths, and simply dismissing the possible out of hand does not make you stronger

Comments are closed.