You probably just need to look at the basic evidence steps from long ago.

The key to understanding that is to look at the isotopic ratios in atmospheric carbon. That provides a signature for carbon that has been recently exposed to ‘heavy’ radiation within the last few hundred years in the atmosphere and surface, and carbon that has been locked up for millions of years in rock protection. It is clear the the ‘extra’ CO2 in the atmosphere has been coming from fossil sources from that – and has been for more than 30 years – it was in my geochem texts in 1980.

Same for visible light energy getting dissipated to infrared, and the trapping routes for infra-red out of the atmosphere.

The insolation energy levels are reasonably well known from the 70s (you really can only do them from orbit). Same for the heat escaping the atmosphere.

But basically we’re seeing rapid changes in climate especially in the near polar regions that are the most sensitive to energy changes. They have been sustained for 50 years. There is no natural process that we know that could do that.

I’d suggest charting the Antarctica Peninsula temperatures. You’ll be able to look at run-away growth of about 0.5C per decade for the last 5 decades…..

One of the specific triggers for a increased energy climate effect may be to have some regions get significantly colder. There are several arguments about why the gulf stream periodically turns off and on. One of those involves increased temperatures on the lower US east coast, and a higher fresh water runoff. This dilutes the gulf stream and therefore stops the it (a major driver for the gulf stream is apparently the difference in salinity between the tropics and the polar areas).

Turn off the gulf stream, and the glaciers crank over the landscape in Northern Europe and to a lesser extent North Americia. So global warming can cause a regional climate change towards colder. The rest of the world continues on average to warm up because the heat in the gulf stream is dissipated closer to the tropics.

There are all sorts of similar effects in different regions. Like the effects of monsoons around the Indian subcontinent.

Basically anyone who uses “Global Warming” is generally regarded as a CCD. They are confabulating a cause with the effects and thereby demonstrating they don’t understand the scientific issue at all.

Just at present they will tell you that the current low temperatures in the upper northern hemisphere this year ‘prove’ that global warming isn’t happening. Personally if I was living there I’d be having a close look at why the polar air wasn’t as confined to the polar regions as it is normally. It could be a unexpected regional climate shift or simple statistical outlier. I’d also be measuring the gulf stream a lot. In both cases it could be a climate change due to global warming or it could be a simple variation.

You only see what you want to see.

An underlying assumption is that markets are ‘rational’ in that prices are reflecting the actual value of something”

Not really. I would only advocate using techniques that are valid for charts generally. There are a lot of indicators out there that clearly are only applicable to financial charts. I would not advocate broadening the use of those ones.

BTW, one of the reasons I am skeptical about the temperature going up indefinitely is that increased temperatures may activate processes that reverse the temperature. For instance, Jenny pointed to the gulf stream being interrupted as a possible mechanism that could achieve this that has been theorised about for awhile. Also, increased C02 might result in increased vegetation which could have the effect of sinking more carbon.

To use a financial analogy, there are an awful lot of people on the global warming side of the trade at the moment. However, it wouldn’t surprise me if nature turns around and surprises us all.

One thing you’re probably overlooking doing the charting using financial techniques.

An underlying assumption is that markets are ‘rational’ in that prices are reflecting the actual value of something. Ignoring that the people making markets frequently aren’t, that isn’t the case when you’re looking at natural processes. There are a lot of paths for extra energy to dissipate – some of which can take long times before the energy reappears.

One of those for example is the storage of heat in polar deep currents. Eventually these reemerge up around the tropics where a slightly higher tempature will result in a reduced ability to suck up tropical heat. I suspect we are short of information on how much heat is going into those energy transfers

Tamino is my hero. Mathematicians are the princes of the modern world, but so few of them can convey their magic to the rest of us clunkerbrains in the way that he can. Not all of his material is about climate, sometimes he just does math stuff like PCA analysis, Bayesian statistics, filters, error probabilities, central limit theorem and much more…

And if it turns out you are right about an upcoming period of cooling ts… you will likely see Tamino post a statistically rigorous proof of this on Open Mind before anyone else.:-)

I don’t know that we are actually that far apart. I certainly agree that the main warming trend is very strong and will take a major departure from this to indicate that the trend has changed.

I am genuinely unsure how much is due to human influence and how much is due to natural variables, but am open to being convinced either way. I tend to enjoy taking a contrarian stance in debates. However, that would overstate the degree to which I lean in the contrarian direction.

So far as technical analysis is concerned, I think the parabolic SAR is a useful tool for picking up changes in any trend. I don’t think it matters with this indicator if the data is from humans or from natural variables.

There are some indicators I don’t think would be appropriate for the climate change debate; for instance mean-reverting indicators such as Bollinger Bands, or indicators that make assumptions about overbought or oversold conditions..

Almost. I can see where you are coming from. Lets assume we are filtering a bunch of noisy data points in a time series. What determines the smoothness of the resulting curve is the parameter we choose for the filter. For instance if we are using a simple moving average filter, then obviously the period of the moving average is critical… ie the appearance of the result is determined by whether we choose a 5, 10 or 30 year period.

All filters have similar parameters that are chosen to reject the noise we want to discard, and reveal the underlying signal that we are interested in. This is a crucial point. You have to choose filter parameters that makes sense in terms of the signal you are trying to extract. There is no point for instance in trying to find a seasonal timescale variation… if you apply a 1000 year filter to the data. The signal you are looking for will be smoothed out of existence. Equally if you apply a 5 year filter to a process with an underlying 1000 year driver… the signal you are looking for will be hidden under noise.

By contrast including more data points does not necessarily change the shape of the curve, rather it increases your confidence in it’s accuracy.

However, such fitting calculations, by necessity, become progressively more reliant on historical data and less useful for predicting future trends as more data points are included.

True but still misses the point. Absolutely if you apply a 30 year filter to an historic time series, no-one would project the resulting curve into the future and pretend that it predicts for instance…tommorrow’s weather. (Or indeed anything else other than a smoothed trend over the next 30 years.)

The technical and political challenge with AGW is that the driver (human induced CO2 changes) have a timescale in the order of a century or so, which is why we usually apply 15 -30 year filters. But from a human political perspective this is very slow and we are apt to get bored or imagine that we can procrastinate doing something about it, so we get tempted to try apply filters with much less smoothing… and wonder why we get misleading results.

As for ‘technical analysis’… I accept that you know a lot about how it applies to making bets on various markets, and that at one level it makes sense as both markets and climate are complex, hard to predict systems … but I’m far from convinced that the mechanisms which drive planetary climate can be satisfactorily modelled as millions of ‘rational actors’ all intent on maximising their personal self-interest.

However I have to say that the nearly-a-hangover makes the reading a bit of a strain. I suspect that the drive to Auckland will be in hour hops.

It is all young labour and clare currans fault. The wine just appeared. And it was fun watching the crazy dancing. However I should have headed to bed earlier.

In contrast, too fewer data points in each calculation means a greater likelihood of picking up noise. That is why I made the point about not getting too excited about minor fluctions, and that a move downwards of around .2 of a degree would be needed before the direction of the major trend could be seriously questioned.

The wiki graph was probably not that useful because it did not have many data points, relying on annualised data. If it had used monthly data the calculations would have been more useful. One such indicator that is very useful for picking trend reversals and avoiding “fake-outs” is the parabolic SAR indicator:

http://stockcharts.com/help/doku.php?id=chart_school:technical_indicators:parabolic_sar

One thing that cannot be questioned is that the world has definitely been getting warmer.

Whether we are likely to get a reversal in world temperatures depends largely on the actual sensitivity of the climate to human emissions. This is the great unknown at the moment. If the sensitivity is high, then the trend will probably keep going up for the foreseeable future. If the sensitivity is low, then by definition, a substantial part of the previous warming will have been due to natural factors, and thus a change in those natural factors could have a substantial effect on reducing temperature.

What is needed is some good technical indicators. One very good indicator for picking up likely future

The current low temperatures in Europe are being matched by high ones in other parts of the Northern Hemisphere…related to yet another climate variable called the ‘Arctic Oscillation’. It’s nothing to do with CO2 and the last month or so this index has been almost off the bottom of the chart.

Frigid polar air is largely isolated from the temperate regions by strong circumpolar winds that keep it in place. When this AO index goes low, these winds move about dramatically, allowing the arctic and temperate regions to mix far more than usual.

Read Dangerous Curves again. What Tamino is saying is that just because a smoothing filter gives what looks like a good eyeball fit … does not necessarily mean that it actually explains anything.

Tamino’s point is that the type of function (linear, polynomial, sinusoidal and so on) that underlies the chosen filter… will determine how it behaves at the boundaries. And unfortunately for the purpose of trying to understand AGW temperature trends… it is the boundary at the present date which is the region we are most interested in… and exactly where most simple filters give misleading answers.

Nope… the technicalities is all. There is no avoiding them.

And because google is our friend, I found this link, which is non-climate page primarily about a Lowess Filter function for Excel… but scrolling down I found that the author has done some work for us and applied the filter to Global Temperature Data from NASA.

The resulting charts confirm Tamino’s results rather nicely, ie no ’rounding off’.

Looking at the graph I pointed too, I think the fit does look good. Tamino seems to assess goodness of fit on a similar basis. I think, on an eyeballing basis, the fitted line looks fairly good on the chart I refered to as well. Afterall, this is a Wiki article. I think it is safe to assume that there have been a number of iterations and corrections to the graph to get to this point, so I think it is fairly safe to trust the result.

If we get only two more points of relatively low temperature to complete the next data set, then the smoothed line on the fitted curve on Wiki is going to show a fairly dramatic step down. Given that there will be a lot of weather stations around the world taking fairly major hits for a month or two given the amount of snow and ice around at the moment, I think this year will show as quite low, so the probability of a down leg on a five year smoothed basis seems fairly strong.

I wouldn’t get too excited about it being a trend down though until amount of movement down was greater than any other movement down on the chart. Thus, I would want to see a move down of more than .2 of a degree before I could start to think about a change in trend on a serious basis.