Paul,
Yes, I realise it would be reduced in amplitude, but it still seems that it's not clear that a Marcott like analysis necessarily misses variability on scales smaller then their resolution.

Martin,
Tamino has done something like that. His analysis seems to suggest that century-scale variability like what we've experienced in the last 100 years or so would not be missed. However, it is a blog post and I certainly haven't done something similar to check it myself.

aTTP, isn't this a question of signal processing? Some variability may be picked out but not the correct amplitude. The sampling frequency is low compared to the possible signal frequency. I really do feel that to make progress here real effort needs to be expended in several key areas, mostly technical. I'd highlight several:

(i) development of robust palaeothermometers. Current proxies are limited because of the requirement to determine distribution coefficients between two phases only one of which is usually present in a proxy archive.

(ii) improvement in measurement accuracy, precision and sensitivity to allow greater temporal resolution.

(iii) Significant steps in determining independent accurate and robust chronologies based on radiometric clocks.

I'm not sure that until we do this we will move forward with these multiproxy studies, at least on time scales of the Holocene and beyond.

Paul,
I don't know the answer. My question was genuine. My feeling is that this isn't quite as simple as a sampling issue, given that the proxies are not simply temperature measurements separated by a time of 300 to 400 years, they seem to be more like averages over that kind of timescale. Hence, a sufficiently large perturbation, that lasts for sufficiently long, might be expected to influence the proxy result. Tamino's analysis suggests that this is the case, but I haven't really seen this presented elsewhere.

aTTP,
I understand the point you are making and think it may well be necessary to refer to the original studies to understand in detail the sampling strategy. If it is as you describe then my professional judgement is that anomalies of the amplitude of the current warming will not be detectable beyond the noise. For example the sensitivity of the oxygen isotope thermometer is between 4 and 5 degrees C per per mille change in d18O. Typical external precisions for a routine measurement of d18O are about +/-0.05 to 0.1 per mille at 1 sigma. That's equivalent to a 0.25 to 0.5 degree temperature shift. Of course one can do many repeat measurements to reduce the se but in my experience this is rarely done, or at least not as often as it should be.

I don't think any of the other proxies have any better sensitivity or resolution and most a lot worse.

A residual signal may still occur as suggested by Tamino but it may well to be beyond the ability of the proxy studies to detect.

Martin A,

thank you for that description. I think aTTP's point is subtle and likely still comes under signal processing. Instead of point sampling a signal at a particular sampling frequency one is averaging the signal into boxes with a width of the sampling wavelength. Intuitively this might contain some more information than single point sampling but my sense is the signal will be strongly attenuated and we will have no knowledge of it's correct frequency and amplitude. I also think it will be buried in the measurement noise.

Paul,

A residual signal may still occur as suggested by Tamino but it may well to be beyond the ability of the proxy studies to detect.

Quite possibly. As I say, I'm not an expert on this. Something to bear in mind, though, is that we've had ~100 years of warming (with some variability) and it's not likely to suddenly drop back to pre-industrial levels. Hence if our current warming is simply a perturbation that will return to a pre-perturbation equilibrium, we can't assume that an equivalent perturbation in the past would only be ~100 years long, since that ignores the later cooling phase (which we have not yet experienced, and almost certainly will not), which can't be instantaneous. So, if something equivalent has occurred in the past, but wasn't picked up by the Marcott analysis, it implies that even fairly large perturbations on >100 year timescales are missed.

Breath of Fresh Air,

that very nicely illustrates the point about attenuation and loss of information.

I've just gone back to the paper, which does say

The results suggest that at longer periods, more variability is preserved, with essentially no variability preserved at periods shorter than 300 years, ~50% preserved at 1000-year periods, and nearly all of the variability preserved for periods longer than 2000 years.

However, they were testing this using a white-noise time series in place of their proxy data. Hence, it's still not clear that this shows that it would completely miss a large perturbation with a timescale slightly smaller than their resolution. Again, I don't know the answer. It's clear that there analysis would attenuate variability on timescales shorter than their resolution, but whether this means that it would completely miss something comparable to what we've experienced in the last 100 years is not clear. As you say, though, the precision of their proxies may mean that it would be hard to distinguish this from the noise.

aTTP,
we are possibly getting to a key point. The proxy record is not sufficient to tell us about the frequency and amplitude of natural climate variability, certainly when the amplitude of variation is on the order of a kelvin or so. Thus it is difficult to put the current warming into context. Are temperatures and rates of change abnormal?

Certainly there are many studies that suggest on a local scale temperatures at those sites have been warmer than present and that significant climate fluctuations with rapid transitions have been recorded. One only has to look at the central Greenland climate record. Similarly, the 8.2ka event has a significant presence at many sites globally but barely registers on the Marcott et al reconstruction.

aTTP,

thanks for checking back to the Marcott et al paper. Mysteriously my data base has somehow lost the pdf! It explains why the 8.2ka event barely registers in their reconstruction.

Paul,

Thus it is difficult to put the current warming into context. Are temperatures and rates of change abnormal?

To put some physics into this, I think it would be very difficult to construct a physically plausible scenario in which something comparable to what we've experienced in the last 100 years or so is driven primarily by some kind of internal process; especially in the absence of large ice sheets. Dansgaard-Oeschger events might be comparable in terms of magnitude and rate, but seem to be associated with an ice sheet instability. This is especially true if you also consider that we've had ~1K of warming and still have a planetary energy imbalance (i.e., we need to warm further to return to quasi-equilibrium). So, it may be that we haven't ruled out that something comparable might have happened previously in the Holocene, but if it has, explaining it will be extremely difficult in the absence of some kind of external influence.

aTTP,

I am not as certain as you seem to be that the present warming, or at least a component of it cannot be the result of some form of natural process. Though I grant you we don't have any handle on what that might be. Certainly the Eemian interglacial appears to have some characteristics that suggest a climate that might be somewhat more variable than the present interglacial. This is without an anthropogenic influence. The extent that these might be linked to ice sheet dynamics or Milankovich cycles is not known. The climate characteristics in terms of diurnal patterns etc. also seem to be very different from todays even though mean temperatures are similar, or maybe fractionally warmer.

I deliberately used the term natural and not internal so as to include other external effects (non-anthropogenic) that may contribute to the warming that we presently don't fully understand including solar effects and solar modulation of cosmic rays etc. I'm not so keen to readily dismiss these as having no, or an insignificant role to play in addition to rising GHG levels.

Paul,

I am not as certain as you seem to be that the present warming, or at least a component of it cannot be the result of some form of natural process.

That isn't quite what I said and - technically - I'm not really certain about anything. On the other hand, it is actually difficult to construct a physically plausible scenario in which a majority of our recent warming is natural (either internal, or forced, or a combination of both). One real stumbling block for a significant natural contribution is how you can explain that we've warmed by as much as we have while still retaining a planetary energy imbalance. This tell us that something is producing a radiative perturbation (it can't simply be energy released from the oceans, for example). However, the physical processes that can do so are essentially the same processes that act as feedbacks to anthropogenically-driven warming. Hence it becomes somewhat inconsistent to argue for a large response to internally driven warming, while - at the same time - arguing for a weak response to externally-forced warming.

I deliberately used the term natural and not internal so as to include other external effects.

Yes, I realise, but I was also using internal specifically. If there were some external influence that produced a change in forcing comparable to what's happened through anthropogenic influences in the last 100 years or so, a similar warming would be expected. In a sense, this would simply be further evidence for anthropogenically-driven warming today. The complication is, of course, what this might be. Solar variations are expected to be an order of magnitude or so smaller, and although volcanoes can produce quite large changes in forcing, they tend to produce cooling on much shorter timescales

aTTP,

I think we both chose our words very carefully. I have no problem with that. We also agree that the warming is likely produced by an external forcing. What I don't know and have not been convinced by my reading yet is what proportion is anthropogenic and what proportion is natural. Yes solar irradiation variability is small but this doesn't rule out a role for solar modulation of the cosmic ray flux and it's impact on clouds and albedo. Here some of the proxy data points to strong covariation of proxy response and, for example, cosmogenic nuclide production rates that is hard to ignore.
I'm not sure how we can test hypotheses with the extant data or indeed if that is possible. Witness this discussion at BH. I respect your position but it isn't clear to me that the past centuries warming is outside the range of natural variability in terms of both amplitude and wavelength. Scientifically this is exciting and for my own research interests there are opportunities for novel methods to contribute to the debate.

Paul,
We should obviously not reject anything (well, unless it is obviously wrong). However, an issue with some kind of significant alternative (cosmic rays, magnetic fields,...) is that not only do you have to find some kind of plausible mechanism and show that it can actually operate, you also have to show why - in this case - anthropogenically-driven warming isn't operating as we expect. Even ignoring feedbacks, we'd expect anthropogenic emissions to be responsible for about half of the observed warming (given the change in forcing). We expect feedbacks to be positive, so that makes it more than half overall. If cosmic rays (or some other alternative) are responsible for a significant fraction, then what has happened to the expected anthropogenically-driven warming?

So, there may be some viable alternative, but it is hard to see what it can be if we still have to find an actual mechanism, show that it actually operates, and then show why our current understanding of anthropogenic warming is wrong (or that we misunderstand something quite fundamental). It's a tall order. I'm not saying it shouldn't be done, just that it's going to take some quite amazing breakthroughs for it to become viable.

aTTP,

I wouldn't quibble with anthropogenic warming driving about 50% of the 20th century rise in temperature. I think this is in line with other studies, notably Shaviv's work. I don't see it as an either/or case. Of course it does imply a much lower climate sensitivity which is again in line with the reducing estimates of climate sensitivity. I think the Svensmark's work and the CLOUD experiment are turning up some interesting results. Of course there a long way from experimental data to show cosmic rays can produce precursor nucleii that will grow into CCN's. None the less I'm sure you agree these are interesting experiments none the less. Positive or negative results we will learn something.

On a different tack. If you ever find yourself in the vicinity of East Anglia please feel free to call by my lab and see some of the work we are doing with new isotopic proxies for temperature.

aTTP says:-

"the later cooling phase (which we have not yet experienced, and almost certainly will not) and:-

"it is actually difficult to construct a physically plausible scenario in which a majority of our recent warming is natural and in which something comparable to what we've experienced in the last 100 years or so is driven primarily by some kind of internal process."

These opinions suggest aTTP is certain such warming as we have experienced since the nadir of LIA is unique and its cause is unequivocally anthropogenic and that temperatures will not fall in the future in response to non anthropogenic forcings. This might be a tenable view if the paleoclimate record did not record any comparable warmings, followed by coolings in the pre- industrial era. This is not the case however as a few examples from the record of VOSTOK, EPCA and NORTHGRIP ice cores demonstrates.There are numerous instances of coolings following warmings comparable or greater than that of the last 150 years even - considering only back as far as the peak of the Eemian interglacial.
As examples:-

130- 126 ka a fall of 5.5ºC
15 - 13 ka a fall of 6.0 C ( peak of Allerod interstadial to trough of Younger Dryas0
116- 115 ka a rise of 2ºC (during a period of 11 ka of falling continually temperatures)
9 - 8 ka a fall of 2.5 ºC ( 8200 y event)

Although the causes of these events is not fully understood it is clear that none of the cited temperature shifts can be attributed in any way to anthropogenic forcings - but all are in the range of the recent temperature shifts which aTTP apparently believes can not be other than anthropogenic.

The opinion that " there will almost certainly not be another cooling phase" can not be justified in the present state of knowledge of the causes of earlier cooling phases ( and indeed earlier warming phases) of similar magnitude. It can only rely on complete faith in the predictive ability of the present generation of GCMs - which faith has yet to be empirically justified.

Martin A, that is also my understanding of the earth radiation budget experiment. It seems to be a failed experiment, producing data with such implausibly large and uncertain values that everybody on all sides agree that it is effectively useless in the short term. Shit happens in real science, whether it comes from a satellite, or not. But it explains why we see so little coverage, comment or publicity about it. And what they do glean suggests that changes in short-wave radiation may have been responsible for recent warming. Ouch.

Obviously that's a bit of a bummer for the CO2 IR-catastrophists, but it has never stopped Entropic Man from making sweeping claims about the net earth radiation budget.

Martin A

The atmosphere and ocean surface are warming at 0.16C/decade.

Ocean heat content is increasing by 3*10^22Joules/year.

Ice melt from Greenland and Antarctica is 400 cubic kilometres per year.

Sea level is rising by 3.3mm/year of which 40% is thermal expansion.

All of these processes require energy to be absorbed by the system. If it was in balance the energy content of the system would be constant, and, apart from internal energy exchanges, so would these.

There you go.