Wednesday, August 17, 2011

In a galaxy long ago

Nullius in Verba. the motto of the prestigious Royal Society in London, is usually translated as ‘on the word of no one’. When it was coined back in 1663, it was intended to distance science from the methods of the ancient universities, which relied heavily on the personal authority of the scholars. 

So, Dave Cohen over at Decline of the Empire got his shorts in a bit of a knot over the myopia of scientists.

Now, in my prior life, I was an alpha geek, in charge of a gaggle of scientists.  I finally said screw it and left for a lower paying but rational job.  Might have been one of my good decisions that was.

Anyway, after years of spending time within this subculture, I am in no way surprised by the lack of diligence in data set selection that leads them down their scholastic path as the idea of actually going out and generating a data set is an irritating byline for most scientists.  But I was mostly surprised by this statement.
  Stupidity (or obtuseness) on this scale is almost enough to convince me to dismiss all climate scientists as total idiots whom nobody should listen to about anything. I do however trust the research they do within their fields of expertise, and of course the Earth is warming due to CO2 emissions. 
I guess that I was a bit stunned here.  When I was in back in college during the dark ages (1970's), the "climate scientists" were preaching a new ice age, complete with apocalyptic warnings.  Then they got hold of computers and the interesting work of Edward Lorenz and went to town with their spiffy computer models and came out with a completely fund-able new set of theories complete with models to certify that they weren't completely full of shit.

I had buddies who were part of the crew at NCAR who came up with the models.  Great folks who were great drinkers of Newcastle Brown Ale and lovers of the high desert.  But mostly their love was transferred to the Cray YMP in Boulder and the models swirled inside that they saw as defining the weather.

I spent time in Southern Utah, camping with these folks and listening to their descriptions of their work.  What I was always surprised at is how casually they assumed their initial data set.  It was never to be questioned, it was just to be fed into the Cray.  A digital offering to a silicon Moloch.

Where this comes around to my point of contention with Mr. Cohen.  I agree with him that global warming is real.  I will even allow that human activity is a contributor to the phenomenon.  What I don't agree is that one should trust the research that they do in their own fields, unless you are willing to check their datasets, statistical inferences, principle assumptions, and understand the bounds of validity set by these constraints.

In weather and atmospheric science, in its current guise, one must also understand the math behind the model.   Difficult stuff this, I am not at all for certain that I completely understand it, and my math chops are better than most.

But the final issue that I wish to bring up is the nature of funding.  Funding is done primary through government grants, and the only way that you get this is to pay homage to the current model.  In other words, the results generated in the current funding regimen are as suspect and potentially myopic as that generated by a candidate drug whose research is funded by the company that owns the patent of the compound being tested.

There is something there, our world appears to be getting warmer.  Is this increase an artifact of enhanced temperature monitoring systems?  Can it be unequivocally tied to CO2 increases?  Is it reversible?  How can we describe the world system that allows temperature increase?

What I fear more than anything is that we are not holding the climate scientists to a rigorous enough standard.  I think that we are putting them in the same box that we put nuclear scientists in the 50's and 60's.   We see the climate scientists as handsome young idealists, standing in front of their array of gleaming computers, showing us pretty pictures on a CRT.  Saying simply that you trust people in their fields of expertise is to allow them to control the conversation.  To allow them to pull off white-lab-coat omniscience can lead to things like the Hanford reservation, Chernobyl, and Fukishima Dai-Ichi.


Craig Cavanaugh said...

Funding is the key word. I work amongst "scientists", and even worse, masters and PhD students. Their "science" always seems to support their profs' work, and please their financiers. It's a racket...

Oh, computer models. Too many ASSumptions on far too many variables. Just look at hurricane models for an example of how imperfect computer models are. said...

The [at least] half-century long roller-coaster ride of mutually exclusive certainties by people practicing at science ought to serve as a warning flag whenever trusting certainties even within their own areas of expertise finds itself in the same context with declarations and pronouncements influencing decision making.

It would be refreshing to see the honest answer, "Based on current models, data, and our level of understanding of [fill in blank] our best guess is [fill in blank].

Nice post. Thanks for sharing it

russell1200 said...

Even if they don't have a vested interest in the a particular end result, there is pressure to find "something". Papers based on tests that show no interaction, or corelation are not usually going to get published.