"Our greatest responsibility is to be good ancestors."

-Jonas Salk

Tuesday, October 30, 2007

The code is not the model

I have a new essay up on Correlations, wherein I try to get the reader comfortable with the idea of a "model".

In writing this, I came to understand that the way the word "model" is used in the climate sciences is confusing. An executable software package (a "program") is often called a "model" but this overvalues the code and undervalues the model. The code is an attempted embodiment of the model. The model is the science. The realization of the model ("running the code") is the prediction. The code itself is just an instrument.

It's hopeless to demand that we stop calling it a "model". It's just too ingrained. We should be aware, though, that this is sloppy thinking. The code is just code.

5 comments:

Anonymous said...

In the good old days, if someone doubted the numerical results of your line printer's output, you just added "ALL RESULTS CHECKED BY THIS COMPUTER" at the bottom of each page and that took care of it :-)

Anonymous said...

This is slightly off-topic, but I think it timely.

If large, high-yield wind farms were established over a significant fraction of the planet, could they be expected to have an effect on general circulation, due to the change to frictional qualities at the bottom of the ekman spiral?

If so, could the changed effects be included in current climate models?

Michael Tobis said...

AK, yes the friction effect of large windfarms could be measurable on small scales. It is unlikely to be systematic, I'd guess, and very unlikely to be large enough to worry about.

It could easily be put into models as a modified surface roughness term, but I don't know that anyone has found the topic interesting enough to actually pursue it.

The energetics of greenhouse gases are such that each unit of greenhouse gas adds much more energy to the lower atmosphere than any direct energy effects.

Anonymous said...

Michael:

I always thought you were an XP guy - you know the only complete specification for a model is the executable code.

Best,
Jim Manzi

Michael Tobis said...

Jim, yes, in a way that is the ideal I'd work toward if given the chance.

It would be awfully hard to explain that in the circles I move in in the daytime, though. It certainly doesn't describe the codes we have now.