Was Crichton right and Hansen wrong, as claimed by amancalledchuda?

Question

I'm trying to figure out what chuda is going on about in this question:

http://answers.yahoo.com/question/index;_ylt=Ahcv_Bibgu9s.zetjaawxfPty6IX;_ylv=3?qid=20070917050525AAjH0xK&show=7#profile-info-a85bd4d28281b52b7bdecfc33cd65da3aa

This is why I wanted to discuss this via email - his additional details are so long that I can't even figure out what he's talking about anymore. He seems to have ignored my response and continues to claim that I'm wrong when I say Hansen's 1988 predictions were quite accurate, and he's even gone so far as to say that Crichton was not wrong when he claimed that Hansen was off by 300%.

Here is my evidence to support my claim. As you can see, up through 2005 Hansen's scenarios B and C were both quite accurate:

http://www.logicalscience.com/skeptic_arguments/models-dont-work.html

Therefore Crichton's claim that Hansen was off by 300% (lie by omission of Scenarios B and C) is wrong.

Can anyone clarify what the heck chuda is arguing?

Answer 1

The trends in Hansen's 1988 prediction are what is being used to assess it's success. If you examine the primary literature it is clear that Hansen's prediction was never meant to be a precise year to year prediction of future climate. Climatologists are more aware than anyone that the precise year-to-year future climate can't be predicted perfectly due to natural variability due to the emergence of unforeseen external events such as volcanoes and El Nino's - just two of the events affecting Hansen's predictions.

To avoid this problem of model variability never having a hope of precisely reproducing observed variability scientists examine the decadal trends. This is all about the whole difference between weather and climate. A seasonal cycle is still affected by natural variability yet a 30 year period is deemed to be climate by the world meteorological organisation. By examining a trend line fitted through the data in Hansen's model and the observations a legitimate comparison can be achieved, and this has been carried out here:

http://www.realclimate.org/index.php/archives/2007/05/hansens-1988-projections/

The trends agree remarkably well, hence Hansen's model prediction was successful.

Why are people playing disingenuous games by cherry picking start and end points to make a temperature comparison? This effectively changes the goal posts to something Hansen's prediction was never supposed to achieve.

Dana, I would recommend the real climate link for future discussions of this point because it contains more data, it is more up to date and it has some great discussion too.

EDIT:

Some people are now claiming that Hansen's inputs were right for model A yet his output was much higher than reality. This is wrong. If people look at the real climate post then they will see a comparison between model input (the radiative forcing) and the observed radiative forcing:

http://www.realclimate.org/images/Hansen88_forc.jpg

The real forcings are somewhere between B an C for the most part, and the trends are again somewhere between B and C for the most part. Who can honestly say that the real forcings look like model A?

There are several ways to misrepresent Hansen's results and they have all been attempted now....shameless.

EDIT:

Amancalledchuda,

You can't use year to year comparisons to formulate a trend. A trend line fitted to all of the data points give the only legitimate and unbiased means of comparing these data sets. Your method is entirely sensitive to when the start and end points are. You criticised Hansen for doing this by accusing him of deliberately ending his plot in 1998 until you realised that he wrote it in January 1999. Why is choosing 1999 any fairer and less biased? Bearing in mind hat Hansen, or any other legitimate scientist, never made such a comparison why are you fixated on this bizarre method?

Example:

A biased person could choose 1998 as the end point and calculate a trend from 1988 to 1998 based solely on the differences in the temperature anomalies to get warmer trend. A second biased person could select 1999 as the end point to create a perceived trend that was cooler than the first persons. Which is correct? Neither, they are both equally idiotic. Scientific results which are this sensitive to when the trend runs aren't worth diddly squat. To calculate a meaningful trend you need to use a statistical technique such as least squares to fit a trend line to the data. The trend lines are far less sensitive to the start and end criteria.

This is truly bizarre. You, yourself, took a pot shot at Hansen for supposedly being guilty of something which you then proceeded to do. You're being entirely hypocritical.

Please read the real climate post because it shows how you can compare these data sets in a legitimate fashion. There is also a much longer comparison period to look at.

Thank you for using my Deltoid link and for acknowledging it, but you need to read it more carefully, especially the last paragraph which mentions "disingenuous baseline games". You may not realise it, but your comparison method is effectively a "disingenuous baseline game". Use a best fit trend line, please.

Answer 2

I pretty much said this to chuda's question:

Just like a stopped clock is still correct twice a day, let's not be too eager to congratulate Mr. Hansen. A rolling prediction is useless. Let's compare Hansen's predictions to the last finalized year - 2006. We'll even give them home court advantage by using the beloved and oft cited wiki instrumental graph instead of satellite numbers:

Wiki:
http://en.wikipedia.org/wiki/image:instr...

Hansen:
http://www.giss.nasa.gov/edu/gwdebate/00...

As you see, in 2006, the instrumental record shows an anomaly of slightly over 0.4C. When you compare Hansen's predictions for 2006, you get this:

A (most dire) ~1.0C, a 0.8C jump from 1988, 4 times (400%) the actual jump of ~0.2C.

B (Hansen's "preferred") ~0.8C, a 0.6C jump, 3x (300%) the actual jump.

C (CO2 emissions frozen at 2000 levels) ~0.7C, a 0.5C jump, 2.5x (250%) the actual jump.

But, maybe I'm not being generous enough. The Wiki graph show 2006 as slightly above 0.4C and 1988 as slightly below 0.2C...so let's say that the actual jump was 0.25C (super generous!) We get:

A 300% off
B 240% off
C 200% off

I don't know what people consider "ballpark figures", but if your ballpark is the Grand Canyon, then yeah, Hanson MAY be close.

Just how generous am I being? Well, if I instead used Hansen's actual numbers from 1988 it would give us a starting anomaly of ~0.35C giving us this:

A 1600% off
B 1200% off
C 1000% off

Or even more so, since Hansen himself submitted this 1998 evidence of the "accuracy" of his predictions, then he was, in effect, standing by it's post-1998 accuracy. As those predictions were entirely opposite the observed temperatures through 2006, the predictions would be invalidated.

Even if we bent the rules and used the arbitrary 0.0C baseline of the anomaly, we would still get extremely large errors:

A 150%
B 100%
C 75%

Anyone who wants to frame Hansen's prediction in gold is totally welcome. You may be able to trade it in on some Enron stock...

Answer 3

Crichton should stay in his lane on this issue, nothing wrong with writing a science fiction book on global warming, but end it there. Crichton is an M.D. not a physicist like Hansen. If Crichton wants to comment on scientific issues he should stay in the genetics (jurassic park) and infectious disease( andromeda strain) realm where he belongs. When the skeptic movement uses a doctor as its main voice; it is just showing its weakness*.

http://www.realclimate.org/index.php?p=74

hansen did well, especially considering the year was 1988 when computers sucked and far less was understood about climate.

* no one is relying on algore for global warming science

Answer 4

Aren't you in the slightest bit annoyed that he starts his climate at the peak of the cold period, the 1960s. Of course it looks like exaggerated warming since then. If you extend the time scale to a more useful scale of a thousand years, you will see a general warming trend with some minor ups and downs including the down we just came out of. I read Crichtons book and I thought it was kind of boring and not one of his better ones, but he was just analyzing data. Just because his honest assessment wasn't in agreement with the alarmists, he tends to get shunned and insulted (not that you were). I don't know the answer to the question though. I am skeptical of those who pretend to be Nostradamus.

Answer 5

You’re a right dodgy bloke on the quiet, aren’t you dana?

“I'm trying to figure out what chuda is going on about”

You know exactly what I’m going on about! You disagree, that’s very clear, but you understand what I’m saying perfectly well.

Your suggestion that I’m *such* an idiot that you can’t hope to lower yourself to my level in an effort to understand, is just slightly derogatory – and I take great offence.

I’ll explain what I’ve said to others and we’ll see if they can “figure out” what I’m “going on about”, shall we?

**********
My argument is very simple: that James Hansen’s 1988 predictions of temperature rise by 2000 were wrong. Here’s why…

In the hot summer of 1988 James Hansen made three guesses of what temperature would do by the year 2000. He gave three Scenarios: A= 0.45°C, B=0.3°C & C=0.25°C.

I quoted the MSU satellite data (which anyone can check for themselves here… http://www.co2science.org/scripts/CO2ScienceB2C/data/temperatures/msu.jsp ) which shows that temperatures actually rose by only 0.14°C between 1988 and 2000.

I therefore conclude that Hansen was wrong.
**********

There, anyone having any trouble figuring out what I’m going on about?

Dana supports his view by referring to Figure 3 in his link, but wait a minute… That figure is by Hansen himself! So dana believes Hansen was right, because Hansen himself says so! And Hansen can only make the claim that he was correct when he uses the GISS temperature data set, that he has a hand in maintaining – and a temperature data set that is showing temperatures that are much higher than any other data set.

When you use any other data set Hansen was wrong. Take this link, for example… http://scienceblogs.com/deltoid/2006/08/climate_fraudit.php (Thank you to Paul H from my question for this link) It compares Hansen’s estimates with the HadCRUT3 data set and, as you can see, the observed temperature is just about following Hansen’s Scenario C (but, I would argue, observed temperature is falling behind even that). Hansen’s Scenario C assumed that the rise in greenhouse gasses would stop in 2000. Anyone here going to claim that that has happened? No? But temperatures are behaving as Hansen predicted they would if the greenhouse effect stopped rising.

So, it seems to me that this leaves us with two choices. Either…

Greenhouse gasses are not having the effect on temperatures that we expected.

Or…

Hansen was wrong.

Which is it?

:::EDIT:::

Paul H: OK, I’ll confess that you may be losing me a little bit here, so let me qualify what I’m about to say by accepting that I may be misunderstanding the points you are making.

You say…

“You can't use year to year comparisons to formulate a trend”

Does this mean that you believe that I have compared the absolute values for 1988 and 2000? If so, I didn’t. The absolute values are: 1988 = 0.118317, 2000 = 0.020142. Subtract the former from the latter and you get a difference of -0.098175. Rounding to two decimal places gives us -0.01°C. That’s different to the 0.14°C figure I quoted, so clearly that is not what I did.

You then say…

“A trend line fitted to all of the data points give the only legitimate and unbiased means of comparing these data sets.”

I thought that’s what I did. I quoted the result of the “best fit slope” across the trend.

Are you perhaps saying that selecting just the years 1988 – 2000 is wrong and I need to select more years to get a better averaged trend? Okay, from what I’ve seen, most temperature graphs have smoothed lines which are “five-year means”. So, would it be better if I started and ended my sample two years earlier/later? So taking 12 years of temperature rise based on the “best fit slope” of the years 1986 – 2002 we get: 0.18°C.

Is that wrong too?

How about ’85 – ’03: 0.22°C
’84 – ’04: 0.23°C
’83 – ’05: 0.21°C
’82 – ’06: 0.21°C
Or the whole data set, perhaps- ’79 – ’06: 0.16°C

I say again, these are 12 years warming based on the “best fit slope” of the years selected. So which is the “best” sample length?

Whichever you think is best, they are all below Hansen’s lowest estimate.

You then say…

“You criticised Hansen for doing this by accusing him of deliberately ending his plot in 1998”

I have *never* criticised Hansen for doing that. When I first saw the graph I was unaware that it was Hansen’s; I assumed it belonged to the alarmist site that dana linked to and thought, correctly I feel, that they should have a more up to date graph to support their claims. I stand by my statement that the reason they haven’t, is because the one they are using looks better. I accepted Hansen was not guilty of the same criticism as soon as I discovered the date that he created it. I think I was very clear about this point.

I would ask, however, whether Hansen smoothed his observed temperature line on that graph, or used the absolute values? And that’s a rhetorical question, because the answer is plain to see, so perhaps you should be directing you criticism towards him? Or do different rules apply to Hansen?

You then waffle on a bit, talking about things I never did, and then say…

“To calculate a meaningful trend you need to use a statistical technique such as least squares to fit a trend line to the data.”

I must confess that I haven’t looked into the actual mathematics of how they create their trend slope, I’m not *that* interested in all this (and I have no idea what “least squares” is), I simply trusted that they were doing it correctly. If you wish to try and prove that their technique is invalid, then you go ahead, I’m sure they’ll be very pleased to hear from you. I say again, though, I *was* quoting the trend.

Next you say…

“This is truly bizarre. You, yourself, took a pot shot at Hansen for supposedly being guilty of something which you then proceeded to do. You're being entirely hypocritical.”

You are correct, it *is* bizarre, because, as I pointed out above, I never took a pot shot at Hansen, and, as far as I can understand, I didn’t do what you’re accusing me of. Thus, I dispute the claim that I’m being hypocritical.

And in answer to your final comment, just in case anyone has missed it, I *did* use the “best fit slope” for my figure of 0.14°C for the rise from 1988 to 2000.

Answer 6

Crichton has shown that random numbers entered into Hansens formulas create a hockey stick.

"But still this increase is steep and unusual, isn’t it? Well, no, because actually you can’t trust it. It turns out that Mann and his associates used a non-standard formula to analyze his data, and this particular formula will turn anything into a hockey stick---including trendless data generated by computer."

http://www.michaelcrichton.net/speech-ourenvironmentalfuture.html
http://www.michaelcrichton.net/NPC-NewVersion_files/image010.jpg

You should take the time to read the top link. It provides information you won't be able to refute. Just be warned, after reading this link, you may never look at global warming the same way!

Answer 7

Michael Crichton's views on climate carry the same weight with me as those of Barry Bonds or Jay Leno.

Answer 8

Hansen's OUTPUTS were accurate.

His INPUTS were CRAP!

His model uses the WRONG inputs to give the RIGHT outputs!

Read the text - don't just look at the pretty pictures!

Model A most closely matched the real world vis-a-vis the CO2 level rise, but predicted much more warming than we got. Models B and C most closely matched the real world temperature, but the CO2 levels were in fantasy-land.

Therefore, it remains dubious to base public policy on coincidence.

Answer 9

I can't, but then I can't explain why there is this incredible fascination with Nostradamus either.

Suggest to chuda he read this:

http://en.wikipedia.org/wiki/Dunning-Kruger_effect