Get PJ Media on your Apple

PJM Lifestyle

Climate Change and the Problem of ‘Press Release Science’

Press releases are meant to sell, not report.

by
Charlie Martin

Bio

August 1, 2013 - 9:00 am
Page 1 of 4  Next ->   View as Single Page

When I agreed to write a weekly science column, I have to admit that it was intimidating at first — the thought of having to find a topic every week.

I guess I wasn’t thinking clearly. The supply of people saying dumb stuff about science, and the opportunities to explain why the things they’re saying are dumb, is just endless.

It’s kind of sad when it’s NASA though.

Yesterday’s Astronomy Picture of the Day is a video from NASA’s Goddard institute of Space Science (GISS) in New York City. GISS is directly upstairs from Tom’s Restaurant on Broadway, famous in song and TV.

None of which is particularly important in this story.

So, if you watch the video (embedded below), it’s an animation of the computed distribution of surface temperatures changing from 1891 to 2011. The video itself is an excellent example of press-release science. Let’s just list some of the issues:

  • The scale on the temperature variation, which is colorized to run from deep blue to fiery red, is a mere ±2°C. Sure looks impressive to go from ooooh-cold blue to OMG red though.
  • The use of a Mercator projection means that the visible area of the map is completely dominated by the Northern Hemisphere above about 40°N (which happens to be right where I’m sitting now) and in particular by northern Canada, Greenland, and Alaska. This will be important, because
  • As the video runs, those vast areas of the northern and southern extremes turn that fiery OMG red giving the impression of a world wide conflagration.

This is why it’s “press release science”: they’re reporting real results — although questionable for various reasons — but doing so in a TV-movie dramatization.

YouTube Preview Image

The APOD blurb is interesting too.

Explanation: How has the surface temperature of Earth been changing? To help find out, Earth scientists collected temperature records from over 1000 weather stations around the globe since 1880, and combined them with modern satellite data. The above movie dramatizes the result showing 130 years of planet-wide temperature changes relative to the local average temperatures in the mid-1900s. In the above global maps, red means warmer and blue means colder. On average, the display demonstrates that thetemperature on Earth has increased by nearly one degree Celsius over the past 130 years, and many of the warmest years on record have occurred only recently. Global climate change is of more than passing interest — it is linked to global weather severity and coastal sea water levels.

Let’s look at it piece by piece.

Comments are closed.

All Comments   (17)
All Comments   (17)
Sort: Newest Oldest Top Rated
Whether it is Mercator or Equirectangular (and I agree it's the latter) it's still a pretty huge distortion towards the poles. In this representation Greenland appears to be roughly the same size as or maybe even a bit bigger than Australia. In reality, at 836.3K square miles, Greenland is a bit less than 1/3rd of Australia's 2.97 million square miles.
51 weeks ago
51 weeks ago Link To Comment
That's not a Mercator. If it were, Greenland would appear to be as tall as Africa.

Rather, it looks like an Equirectangular projection, a map commonly used in science because of its mathematical tractability. It's not an equal-area map, but its distortion is nothing like as bad as a Mercator.
51 weeks ago
51 weeks ago Link To Comment
You might take a look at how many stations are north of 80*N today. And when many of them were cut out of the data set. Today the high latitude temps are not determined by actual data but by statistical smearing of the temps from stations up to 1200 KM distant. GIGO

Re: CO2 forcing, may I assume that you know that the global temp increase for the last 16 years has been statistically insignificant? But the CO2 input has continued to increase. Therefore, the assumed CO2/temp link is/has been wrong for a long time. Not to say that it's non-existent, only that it ain't what most people think it is. There have been a number of recent media articles bemoaning the lack of knowledge wrt this phenomenon.

But then, there are also a number of other problems with the science that are normally swept under the rug by the media release science crowd.
51 weeks ago
51 weeks ago Link To Comment
The *assumed* CO2 link is wrong. That's correct. The physics that says there *exists* a CO2 forcing is well demonstrated. The problem is that CO2 isn't the only forcing that influences temperature, and we're seeing more and more reason to think it's not dominant. I talked about this back in my 25 June column: http://pjmedia.com/lifestyle/2013/06/25/can-we-actually-even-tell-if-humans-are-affecting-the-climate/
51 weeks ago
51 weeks ago Link To Comment
Correction: The world map is an equirectangular projection. It does exaggerate area near the poles, but not as much as a Mercator projection would.
51 weeks ago
51 weeks ago Link To Comment
Yeah, sorry, I was on a roll and didn't check. Thanks for the correction.
51 weeks ago
51 weeks ago Link To Comment
Let's not forget the reason their data begins around 1880 -- that's roughly the time that The Little Ice Age ended. They CHOSE 1880 as the starting point knowing that the Earth has naturally warmed up since then, i.e., true climate change, which has nothing to do with any effects due to man-made CO2.
If they'd started the graph in 1750, for example, the first half of the video would have shown the Earth getting dramatically colder (even as more CO2 was being pumped into the atmosphere as the Industrial Revolution got underway).
51 weeks ago
51 weeks ago Link To Comment
Modelling past history is just curve fitting, no matter what the model looks like.

This fact is usually not known to the modeller, who thinks of it as science.
51 weeks ago
51 weeks ago Link To Comment
You know, my PhD work involved a lot of modeling, and I really don't believe that. Most people doing modeling do actually know that, and climate models are *hard*. The problem is that when you start getting distinguished chairs for your dissertation, even when the stats are eventually proven wrong (see, eg, Mann) then it becomes harder and harder to say "Damn, I guess that hypothesis is blown."
51 weeks ago
51 weeks ago Link To Comment
I've had some fairly minor experience with hydraulic modeling of a water system. IMHO if the number of unknowns and variables are limited, and the system is not terribly complex, the models can be used to make educated guesses at how a system will behave if certain changes happen.

The global climate system, however, is infinitely complex. I am highly skeptical that the models are accurate enough to justify huge changes to the way energy is produced and distributed. I would have that same skepticism even if hucksters like Mann had not been playing games with the data.

We're talking about changes that if implemented world wide would literally starve millions to death. I think we better have a pretty damn high accuracy probability before going down that road.
51 weeks ago
51 weeks ago Link To Comment
Well think of it this way. Your model has a bunch of parameters, say "effective viscosity" and a thousand other things you've pulled from your ass because you can't solve the Navier Stokes equations.

You've got past data.

You "tune the model."

Now think of Kalman Filtering. It tells you, for each added data point, what the best way to adjust unobserved parameters is. This is meant for things like target tracking in radar systems and the like, updating things you don't observe, like drag, position, velocity, lift, and so forth.

That's exactly what the modeller does, perhaps inefficiently but conceptually the same. Match the model to the past data to determine "effective viscosity."

Kalman Filtering gives you a least squares fit. So it's curve fitting.
51 weeks ago
51 weeks ago Link To Comment
Oh, I'm not disagreeing about the curve fitting -- I mean, that's what modeling is -- one way or the other, you end up constructing an empirical that necessarily is effectively a polynomial with a small number of coefficients compared to the real system. I just think that modelers are, in general, aware in their heart of hearts that that's what they're up to.
51 weeks ago
51 weeks ago Link To Comment
I think Pielke is wrong twice. First, his flat trend lines are unbelievable, there should be some noise. Perhaps he did this on purpose as a first approximation, but the RMS variation is data, too, that he loses this way. Second, I don't believe in CO2-forced anything, and believe he needs to review his science. At best it's unproved, at worst it's absurd.

One has to remember that APOS is a NASA site, and NASA has been run by a "climate change" lunatic under the Obamanation, it's a wonder they show only this. What it does not establish is what to make of a one degree Celsius change. I haven't heard of any climate change lunatic model that predicts that small a change, though that doesn't stop them from getting hysterical over a change even a hundred times smaller.

.. and for a model to not correctly forecast a PAST history, is pretty dang weak.
51 weeks ago
51 weeks ago Link To Comment
don't be silly. The flat trend lines are linear best fit, of *course* there's no noise.

If you don't believe in CO2-forcings in greenhouse effect, you need to review yours.
51 weeks ago
51 weeks ago Link To Comment
Charlie, with all that RED, we must be DOOMED....but at this point, what difference does it make, Man has a 100% mortality rate!
51 weeks ago
51 weeks ago Link To Comment
I have had a career in science, albeit checkered.

Two things I know about intersect AWG, and on both counts AGW is claiming to know what it cannot know.

1. We can't solve the Navier Stokes equations. They govern the atmosphere. In three dimensions, flows go to shorter and shorter scales, which means that no resolution at all is adequate to represent the flow. You need small scale flows because they act back on large scale flows as sort of an ersatz viscosity. So AGW modellers pull an equation out of their ass and solve that instead. That isn't science of any kind. So there is no theory for AGW.

2. You can't distinguish a trend from a cycle with data that's short compared to the cycle to be excluded. The eigenvalues of the distinguishing matrix explode, and not just slightly explode. So no measurement at all is good enough to tell you anything, and not just by a little. Mathematical truth. So there is no AGW data.

AGW people claim to know both. They can't. Furthermore they have to know that.

If just those two things I happen to know about are wrong, how many thousand other things are wrong.

I do know about grants and sociology, which seem to offer a better explanation.
51 weeks ago
51 weeks ago Link To Comment
http://www.nasa.gov/topics/earth/features/2011-temps.html

This is the link to NASA's page providing the first video. It actually offers two videos... the misleading color movie, and the misleading color movie WITHOUT a colorbar.

You know, just in case you want to spread the video and make sure the viewers don't dare try to quantify the inferno.

I seriously cannot see any reason to take a legend off of a 'scientific' graph for any reason other to assist misinformation.
51 weeks ago
51 weeks ago Link To Comment
View All