
Each time I come back to this it seems there is *another* bit of the real world that *somehow* was left out of the general circulation models people use. Failing that they are incorporated by some kind of fudge factor. My list to date is
1) Heat reflectivity due to poor (non existent apparently) cloud modelling. 1st I saw of this was Byte special edition on Forth (researcher using it to model clouds, but said its difficult) in c1984. 20 years later, still a bit tricky.
2) Cloud formation control by increased/decreased sunspot activity seeding upper atmsophere with solar particles acting as nucleation sites for water droplet formation.
3) Global dimming due to aerosols produced by aircraft engines at c40Ktf. Definitive comparision was only possible on 12th September 2001, when all US commercial jets were grounded. An impossible experiment to carry out under normal circumstances. So probably not included in any model before the next decade.
4) Disagreement on how to handle heat island FX caused by urbanisation of area surrounding weather stations. Are any stations immune? If so what happens when you run models with just those? Is it possible to do so and get any kind of sensible result?
5) and now dust levels also affecting heat retention and reflection.
And lets not forget most models seem to use a 3d rectangular grid. Only some use shapes which actually map onto a sphere in the first place and few if any (AFAIK) note the 21mile high equatorial bulge
*refining* general circulation models *implies* finding the sources of *systematic* error and modelling their underlying physics (and chemistry in some cases), not leaving them as *fudge * factors to be twiddled by *experts*. This refinement process does not appear to be happening. The Grids have got finer, the speed has increased (but is that due to the improved hardware?) but I get no sense of any eagerness to nail down the uncertainties. Why does Knuth's comment about premature optimisation being the root of most programming ills come to mind?
I keep hearing its a chaotic system, the butterfly effect, strange attractors blah blah. I can't shake the feeling they have created a "Doctrine of impotence" as Dr RV Jones put it. The modellers do not believe they can improve certain aspects of their models, so won't try.
I do believe we have an effect on the environment. Its not our scale, its the systems sensitivity. Like the PPM impurities in silicon that trigger orders of magnitude changes in resistance that make semiconductors possible.
But remember El Reg, London is a port city. If you end up going to work by canoe the actual causes might be a bit academic. I hope your not located in the basement.