Buy

Books
Click images for more details

Support

 

Twitter
Recent comments
Recent posts
Currently discussing
Links

A few sites I've stumbled across recently....

Powered by Squarespace
« Tamsin 2 | Main | Awful astronomer astray »
Thursday
Feb022012

What trend would you like with your graph sir?

Thanks to reader William for pointing out Climategate 2 email 4578. The context seems to be a discussion of how to present temperature trends, and it is worth reading the full email thread. But Jones' contribution to the thread looks a bit problematic.

date: Mon Jul 18 14:25:52 2005
from: Phil Jones <p.jones@uea.ac.uk>
subject: Re: Text and CQ stuff
to: "Parker, David (Met Office)" <david.parker@metoffice.gov.uk>, Kevin Trenberth <trenbert@ucar.edu>

Kevin,

Even without smoothing it is possible to get a trend of nearer 0.75 if the trend starts around 1920 (especially if the cold year of 1917 is at the start). The periods chosen for Table 3.2.2 had some justification, so we need to be a little careful. As a schematic for CQ2 though, it will be a different way of showing the same data.


I'll talk it over with David.

Cheers

 

PrintView Printer Friendly Version

Reader Comments (26)

This is how post-modern climate "science" works:
Step 1: Decide what answer we want (0.75C warming).
Step 2: Decide what data to cherry-pick to get the answer we want.
Step 3: Decide what excuses you can use to explain the bits of cooling.
Step 3: Decide how to hide the uncertainty.
Step 4: Get some minion to plot the data because it is beneath our dignity and in any case we don't know how to run a spreadsheet.
Step 5: Publish the results and tell politicians it is settled science and something must be done.

Feb 2, 2012 at 8:13 AM | Unregistered CommenterPhillip Bratby

It's a powerful tool that Excel.

You can do just about anything with it.

Feb 2, 2012 at 8:25 AM | Unregistered CommenterGrantB

Matt Briggs, prescient as ever:

http://wmbriggs.com/blog/?p=5107

Feb 2, 2012 at 8:30 AM | Unregistered CommenterSean inglis

Wow, looking at that thread it just looks like they're chopping and dicing the record and serving it up like Delia in the most tasty presentably alarmist way they can there.

I love the bit where Trenberth goes off on a little interpretive jazz riff of what the data means:

In fact this could be linked to the modeling and attribution chapter to say that the warming in the first part of the 20th century was partly due to solar, the cooling from 1940 to 1970 to increased aerosol, and the warming after 1970 to the increasing GHGs.

They are so creative these climate scientists, the data speaks to them. ;)

Feb 2, 2012 at 8:31 AM | Unregistered CommenterThe Leopard In The Basement

To some extent, this is the way that normal science works. The data is never blindingly obvious. If you've never taken 3 data points, plotted them on a graph and seen a straight line, then added another 20 points and seen that one of the first 3 was an outlier - you won't understand.

It is natural to see patterns, and to interpret the data as you expect it to be. That is why it is so important to be critical of what looks like proof of a theory. Open data sets and rigorous statistical analysis, modelling etc. will all go some way to helping to get a better understanding of the real answer.

The quoted behaviour _is_ wrong, but it's not malicious.

Feb 2, 2012 at 8:42 AM | Unregistered CommenterSean Houlihane

Sean Houlihane

I agree with you there that is why the dread scepticism is useful to pull the dreamy eyed out of their reverie. I still think it gets a bit too creative when you get the attribution riff like Trenberth being considered useful as anything more than wishful thinking, it just sounds to me like "we'll allow a bit of solar there, sprinkle a bit of aerosol here, and only now comes the lovely big bad CAGW!"

Feb 2, 2012 at 8:50 AM | Unregistered CommenterThe Leopard In The Basement

Sean Houlihane - it is malicious when it's presented as certainty and critics are silenced.

Feb 2, 2012 at 8:50 AM | Unregistered CommenterCaroline

Well, we've all heard of "torturing the data" - but this is more like forcing it to sign a confession!

Feb 2, 2012 at 8:55 AM | Unregistered CommenterFoxgoose

> ... The curves
> which I fitted by eyeball using power point should be done more
> rigorously, perhaps using a cubic spline fit with strong tension., or
> a series of segments with divides at 1940 and 1970. Then a linear
> value with the given starting point could be determined for both the
> mean and both end of the error bars.

That for me encapsulates the whole approach. Well done, lads!

Feb 2, 2012 at 8:58 AM | Unregistered Commentermatthu

SH - It becomes malicious when you add in the attitude to FOI.

Feb 2, 2012 at 9:00 AM | Unregistered Commenternot banned yet

@GrantB

It's a powerful tool that Excel.

You can do just about anything with it.

Unless you're a professor of climate 'science' that is. Using the 'show trend line' function is beyond some of them.

@Sean Houllhane

Careful with that tone about outliers and 'you won't understand', please - there are a large number of contributors here who do understand and in excruciating detail. Meanwhile, in scientific usage, datum = singular; data = plural.

Feb 2, 2012 at 9:01 AM | Unregistered CommenterSayNoToFearmongers

Sean Houlihane

What is it when someone's egregious sleight of hand is used to justify policies which lead to unaffordable fuel bills and people die of hypothermia?

Feb 2, 2012 at 9:13 AM | Unregistered Commenterartwest

The more you torture the data, the more reality hurts you in return.

That is what is happening now.

Feb 2, 2012 at 9:24 AM | Unregistered CommenterNicholas Hallam

@artwest

Green zealots don’t actively set out to cause damage and misery; it’s rather that they don’t much care if it occurs (which it nearly always does because of their fantasyland schemes).

Their overwhelming concern and focus is on themselves; other people dying is just an irrelevance. Put another way: “A single death is a tragedy; a million deaths is a statistic.”

Feb 2, 2012 at 10:16 AM | Unregistered CommenterRick Bradford

Rick/artwest
The one pre-requisite qualification for being a proper Greenie is egotism. Everything else they do flows from that.
It is why it is your carbon footprint which needs reducing.
It is why they conspicuously don't run a car but inconspicuously make a 15-mile round trip to Waitrose rather than go to their local Tesco.
It is why they think nothing of jetting off to exotic places if it's to benefit "the planet" (there is no way that you can possibly be on the same flight for that reason!).
The list goes on.
They hanker for a simpler lifestyle without having a clue what that means. Try discussing it with them some time; you'll blow a fuse very quickly.
And yes, other people's suffering is of no concern to them. It's just a form of collateral damage in the struggle for environmental purity!

Feb 2, 2012 at 11:55 AM | Unregistered CommenterMike Jackson

All that matters is 'the cause ' anything done in its name is justfied .
And once again, as we seen approaches used that these individuals would not acppet from their own students , a rubbish acedmic standard

Feb 2, 2012 at 12:06 PM | Unregistered CommenterKnR

Deluded or malicious? Hey, one wasn't enough so they went for both!
=============

Feb 2, 2012 at 1:34 PM | Unregistered Commenterkim

Thanks to The Leopard In The Basement - that quote from Trenberth is an absolute cracker..
Seems to me that he could have said, with equal value, that the warming in the first part of the 20th Century was due to the arrival of horseless carriages; the cooling from 1940 to 1970 due to the birth of rock n' roll; and the warming after 1970 due to use of thicker duvets...

Feb 2, 2012 at 1:52 PM | Unregistered CommenterDavid

Sean Houlihane
There is an indistinct line between malicious and not. I have played with data the way you suggest just to see what it says, but in the end I analyzed it using my original research parameters. When I see

Even without smoothing it is possible to get a trend of nearer 0.75 if the trend starts around 1920 (especially if the cold year of 1917 is at the start).
I see malicious intent.

Feb 2, 2012 at 2:53 PM | Unregistered CommenterDon Pablo de la Sierra

Don Pablo

I have to disagree.

Trenberth says:

> However, the record is best characterized as level prior to about
> 1920, a warming to 1940 or so, leveling out or even slightly
> decreasing until 1970, and a fairly linear trend since then. Going by
> the low pass filtered data, the overall warming through 2005is 0.75ºC,
> with 0.5ºC increase occurring after 1970
.

And Jones responds:

Even without smoothing it is possible to get a trend of nearer 0.75 if the trend
starts around 1920 (especially if the cold year of 1917 is at the start). The
periods chosen for Table 3.2.2 had some justification, so we need to be a
little careful. As a schematic for CQ2 though, it will be a different way of
showing the same data.

In other words, the discussion is about presenting the actual warming accurately. Not about calculated misrepresentation.

Feb 2, 2012 at 11:17 PM | Unregistered CommenterBBD

BH writes:

Jones' contribution to the thread looks a bit problematic.

[O/T but related] Speaking of E-mail threads, Jones and his (problematic) contributions to annals of "climate science" ...

Readers might be interested in knowing that Phil Jones' "intuition" trumps examination of data and source code when he peer-reviews papers for journals:

Phil Jones keeps peer-review process humming … by using “intuition”

Feb 3, 2012 at 6:54 AM | Unregistered CommenterHilary Ostrov

In other words, the discussion is about presenting the actual warming accurately.
I would agree with you, BBD, except for two things.
1. A linear trend hides the periods when there was cooling (or at least no warming) and implies to those less au fait with these things than perhaps we are that that warming has been constant. It can also be used to match the apparent warming trend with CO2 increase which has not always been the case.
2. Jones' attempt to start the graph from "the cold year of 1917". Both sides cherry-pick so let's not pretend that they don't. The difference is that my cherry-picking is irrelevant, like yours; Jones' matters because he is feeding his results into government with a view to influencing policy.

Feb 3, 2012 at 1:41 PM | Unregistered CommenterMike Jackson

MJ

You haven't understood what Trenberth is saying (athough I did embolden it to make it clear). Look again.

There's no fraud here. Just misunderstanding fuelled by anti-science bias.

Feb 3, 2012 at 2:07 PM | Unregistered CommenterBBD

I understand perfectly what Trenberth is saying:
Flat to 1920; rising to 1940; flat or maybe a slight decline to 1970; rising from 1970. Produce a graph that shows that — with the appropriate trend lines — and I'm happy.
Produce one that starts at 1917 ('cos it was cold that year; thankyou,Phil!) and show a linear trend that ignores the levels and the downs and looks like a consistent 80+ year increase and I'll accuse you of cherry-picking and producing something that looks as if it is intended to mislead.
Simples

Feb 3, 2012 at 5:29 PM | Unregistered CommenterMike Jackson

MJ

Trenberth is clear: the low pass filtered data shows overall warming to 2005 as 0.75C with a 0.5C increase occurring after 1970. Think of low pass filter as just a multi-year average. It is a way of avoiding errors that might arise from sole reliance on linear trend fits. There is no dishonesty. Really. Just misunderstanding.

Feb 3, 2012 at 7:09 PM | Unregistered CommenterBBD

OK. I take your point. But ...
I know what Trenberth says. Can you explain in a nutshell what he means? Or what he is trying to achieve?
I'm not being funny. What do we all gain by the application of a low pass filter rather than simply let the data speak for themselves?

Feb 3, 2012 at 7:19 PM | Unregistered CommenterMike Jackson

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>