Santer's "17 years needed for a sign of climate change" compared against the IPCC models

I think most people here would agree that 17 years, or even 30 years, isn’t long enough to monitor any change in climate
– a 30 year period covers only half of what appears to be a 60 – 70 year cycle of highs & lows
– so it’s much better to look over, say, 66 years
The only problem with this is that we only have good data for 100 years at best, but it does show some sort of trend:
http://www.woodfortrees.org/plot/hadsst2gl/mean:792/plot/hadcrut3gl/mean:792
The nice thing about hitting the right cycle period is that it averages out intra-cycle variations, leaving just the underlying trend that we’re interested in.
– so, here, we’re left with an upward trend of about 0.4C per century…

To further illustrate my point, here’s a few more plots using different filter sizes
http://www.woodfortrees.org/plot/hadsst2gl/mean:360/plot/hadsst2gl/mean:480/plot/hadsst2gl/mean:600/plot/hadsst2gl/mean:720/plot/hadsst2gl/mean:840
Assuming the underlying data is correct, then it’s clear there’s an upward trend
– of about 0.4C per century
Climate models don’t really need to model the ‘multi-decadal’ oscillations, as long as they accept their limitations, and can correctly model & predict the long term trends
Also, it seems to me, that the long terms trends are fairly obvious, and not very difficult to model

Nick Stokes says:
November 20, 2011 at 5:54 am
I wrote a post here which compares them all. Except for the period 1930-1960, when there is some large oscillation, the comparison of the three with the models (graph here) shows that the model tracks the path of the observations better.
The global terrestrial temperature profile can be simulated simple with a good proxy with some (six to eleven) solar tide functions. The decade from 1960 to 1970 or the time from 1900 to 2020.
V.

NotTheAussiePhilM says:
November 20, 2011 at 10:19 am
“To further illustrate my point, here’s a few more plots using different filter sizes”
The 0.4 deg/century trend starts before massive industrial CO2 emissions in the 1950ies start… not a good omen for the CAGW theory and the future performance of their models. 😉

Camburn, Anthony
don’t make me laugh!
“that new fangled method of using stats” has to be “calculating in tongues” “calculating in forked forking tongues”

Systems theory and ecosystems theory all over again, welcome back to the 1970’s.

One reason for the “17 years” should be the COP “17” for this year.
Next year he will be shouting “18 years”, and in 2013 “19 years” ….. /sarc

“Bob Tisdale says:
November 20, 2011 at 1:37 pm
Werner Brozek: Here’s a further explanation that will complement that of Nick Stokes.”
Thank you for these replies. I have not digested them yet but plan to work on it. But in the meantime, unless I missed it, HADISST is not part of the options on the http://www.woodfortrees.org/plot/. Should it be there and if so, can you use your influence to get it added? Thank you!

For grins, I took the global SST 360 month figure above, superimposed solar cycle data from http://solarphysics.livingreviews.org/Articles/lrsp-2010-1/, screen-scraped and (linearly) scaled by eyeball so that the 240-month anomaly looks like it would be roughly proportional to the integrated area of the previous two cycles at the beginning, and added four vertical lines. The result is here http://www.phy.duke.edu/~rgb/combined.jpg (sorry, don’t know how to include it directly in a “reply in this interface).
The four vertical lines are:
a) 1945 — nuclear testing begins
b) 1957 — above ground nuclear testing exceeds 40 blasts a year, many in the megaton range.
c) 1963 — above ground testing peaks (1962 at 140 tests/year) followed by test ban treaty in 63.
c-d) 26 years over which there are at LEAST 40 tests/year, underground but (of course) vented to the atmosphere in such a way that there is still detectable fallout in the upper atmosphere. From d) on there have still been tests, but only a relative handful that with less total energy than was probably produced by volcanic activity over the same general timeframe.
I’ve often wondered why nuclear testing has never been considered (AFAIK) as a confounding parameter to simple climate models based on insolation alone as proxied by the solar cycle. As you can see, by aligning the solar cycle so that it merely eyeballs out normalized to the linear trend on a 20-previous year basis, one expects warming where it warms everywhere but the stretch from 1945-1967 where (I would postulate) we had a “nuclear winter” from the HUGE amount of aerosols and radioactive dust dumped into the stratosphere and beyond. If there is any truth to the idea that clouds are nucleated by GCRs as well as particulate matter, how much more strongly must they be nucleated by radioactive particulate matter, not to mention the substantial volume of exotic aerosols each blast no doubt produced.
If one inverts the graph of nuclear testing from wikipedia, it almost perfectly fits the “hole” in SSTs relative to 2-cycle average of solar output. This also explains why warming has been predominantly northern hemisphere — the anomalous COOLING was due to (mostly) northern hemisphere blasts, and the clearing of radioactive dust has had a stronger effect there as blasts were gradually cleaned further and further up by successive treaties (plus the fact that the tests themselves were smaller and smaller — numerical simulation was considered adequate to design most warheads from somewhere late 70’s on, and we had reliable designs for BIG bombs, all we would ever need on both sides of the cold war).
This also explains the part of the figure I haven’t drawn in — the last two solar cycles, if added, suggest that the anomaly SHOULD be more or less back on the pre-1945 pattern of depending mostly on the previous two solar cycles, which makes so very much sense, ESPECIALLY for SSTs where LOCAL warming and UHI effects are not easily cherrypicked and human-manipulated confounding effects.
Just a thought. I’d guess that the DoD or somebody has long-term measurements of the annual patterns of residual fallout, enough to estimate how they might compare and contribute to ordinary dust, aerosol/ghg and GCR modulation of insolation over the last sixty years.
rgb

I apologize — probably a result of laziness — I scraped the screen to build it, but I screwed up the resolution of the original figure in the process, I guess. Try again, I (should have) fixed it now. I also put up an eps version of the figure combined.eps at the same (otherwise) URL if the jpg doesn’t come through clean.
I’m not certain one should ignore the direct effect of aerosols and dust from nuclear blasts and go only with clouds. If clouds were strongly correlated, it might help support the GCR-cloud connection, but volcanoes can produce significant cooling without the radiation and with similar or even smaller total energies being released. It’s a multivariate process and hence it would be dangerous to assume that we know which of several possible mechanism dominate the effect (if any). However, everybody remembers that one of the predicted doomsday aspects of global nuclear war was “nuclear winter”.
It would be interesting to get a firmer figure on the total megatonnage blown per year — the wikipedia page has only the total number of tests, not the size of each test. The larger tests in the 50’s were up to 15 megatons IIRC, in a single blast right on the water (huge chunk of hot, radioactive water blasted up through all atmospheric layers); many of them (especially larger ones) were on islands and nearly all were above ground. The Apple 1 (no, not the computer) test blast part of the “teapot” series was set off on my birthday (3/29/55) at right about the time I was born (again, for grins); this series added up to around 150 kt set off in roughly 2 months, all of them ground blasts IIRC and hence they kicked up much dust.
From the figure it looks like even the very FEW blasts set off in 1945 very likely had an immediate effect followed by a cumulative (integrated) but gradually decaying longer term effect, just as the “right” way to deal with solar activity is almost certainly to use an e.g. exponential integration window that stretches back at least three or four cycles, not a straight up 30 month running average. This is probably true for Tisdale’s figures as well — he’s using a square window 240 or 350 months into the past, and for his purpose that is fine, but almost all of the “reasonable” ways to average past behavior to acquire a smoothed current behavior should be non-Markovian integrals with an exponential window in the past. Or maybe non-exponential, or (most likely) multi-exponential.
The timescale for turnover of the ocean is order of 1000 years, with the potential for non-Markovian behavior on longer scales still (see Global Physical Climatology by Hartman (1994). It’s not just what Mr. Sun is doing today, it is what Mr. Sun was doing ten, fifty, a hundred years ago. To quote: “The thermal, chemical, and biological properties of the deep ocean therefore constitute a potential source of long-term memory for the climate system for timescales up to a millenium.” Somehow I don’t think that the GCM people are including water cooled during the Maunder minimum (that is very likely still randomly upwelling at various points in the oceanic turnover), let alone water warmed during the MWP roughly 1000 years ago — but they should. Hartman’s book seems to have preceded Mann et al and the Hockey Team — he states quite clearly “The early Holocene from about 10000-5000 years ago was a time when Northern Hemisphere climates were warmer than today’s”. Gosh, I wonder how that happened, since it wasn’t anthropogenic CO_2. In fact, the facts as laid out TEXTBOOK style (as in everybody knew them) in 1994 offer scant support to the AGW due to CO_2 hypothesis.
rgb

November 20, 2011 at 11:06 am
Nick Stokes says:
I wrote a post here which compares them all. Except for the period 1930-1960, when there is some large oscillation, the comparison of the three with the models (graph here) shows that the model tracks the path of the observations better.
I think the relevant point of this subject is the physical nature of that oscillating global climate and the processes and geometries involved. As long as there is no satisfactory explanation of that, what is called ‘noise’ on a ‘holy’ increasing function to hell, this subject has no relevance.
Science ever has to show by arguments if there is an effect of relevance. Because the visible oscillating frequencies of temperature proxies are visible from 1/month to 1/millennium these analysed frequencies and its functions are the basis for a possible human bias. But in a first step the very nature of these oscillations have to be explained. But this is not jet done in the climate science community. Although there is some FFT work about the oscillations, it seems that the power frequencies do say nothing to the workers. By this a short look to possible sources of these oscillations easy can be found in the synodic tide profiles of the couples frequenting the Sun.
There is a cave in China in Wanxiang and deep in 1 km people have taken a stalagmite (WX42b) and analysed the delta O18 values from 195 AD – 2003 AD. It seems to be different in many ways whether one takes 14C data or 18O data, and moreover there are big differences in different locations on the globe. However, the very point of these data (for the oscillations) is the time coherence of changing values, and not in the first order the amplitude of the proxy temperature, because all these methods seems to be nonlinear in respect to seasons and/or temperature and atmosphere. And because there is no nonlinearity in the time, each movement around the Sun can be calculated in seconds of time, and can be a linear reference to the time calibrated scale from the decay data.
There is a possible interesting trend visible in the density of the planets in that way that first the density has an impact on the tide like functions visible in the samples and second that it seems that there could be a quantum effect in the densities. If there would be steps of integer density values and one subtract the integers, then the densities have a linar function with the range order.
http://volker-doormann.org/images/densityfraction.gif
However, it is evident that the synodic function oft the plutinos Quaoar and Pluto has a main impact on the amplitudes in the temperature proxies from all known samples in the last two millennia, followed by the synodic function of Neptune and Pluto.
http://volker-doormann.org/images/wx42b_1500_2000.gif
http://volker-doormann.org/images/wx42b_0_2000.gif
http://volker-doormann.org/images/wx42b_100o_1500.gif
http://volker-doormann.org/images/wx42b_500_1000.gif
http://volker-doormann.org/images/wx42b_1500_2000.gif
http://volker-doormann.org/images/wx42b_0_500.gif
It is not always clear from the data whether there is done a high frequency cut or not. Because of that it is possible that the amplitudes relating to the inner planets like Mercury and/or Venus are smoothed to death. But that these fast running objects make footprints in the hadcrut3 history can be shown.
From the given evidence the couple of Quaoar and Pluto makes its complicated period of 913.5 years, this means that the increasing heat profile since the LIA can be taken as a phase of a period.
For the subject of this thread it means that there is no basis for the hoax of human CO2 and no basis for a speculation.
V.