Quote:
Originally Posted by gunnut
But the whole point is that the earth is not a constant, closed, static system. We can come up with a model to predict the weather IF we had all the variables. But we don't. Our computers are powerful but our models suck. How else do you explain that not a single climate model was able to predict the future. They can't even predict the past. All they can do is predict the present.
|
You are correct that the models do not "predict" (if you mean accuracy down to any arbitrary number). And we have a system in which many of the inputs are not static (insolation, net CO2 output, volcanoes, El Nino events, etc.)
The value in the models is not to say "there will be no clouds, and a temperature of 68.5 degrees at (some arbitrary point on the planet) at (an arbitrary point in time.)" The value is to identify long-term trends.
For example, in semiconductor deposition machinery, you *cannot* say that at (a particular point) on the substrate at (a particular nanosecond) there will be 4589 particles of germanium arsenide at a particular energy level in contact with the substrate. You can say 'when I increase (or decrease) the net energy in this system, my gross energy flow vectors will look roughly like this over time, and I can expect (to within a degree of probability) the energy in (this volume) to be x; +-y'
Interestingly enough, while these models cannot give 'specifics', the main value in them is over "macro-time", in which the models give quite decent results in a macro-sense.
So, some models have given a quite decent outcome to outputs. But, of course, no model could ever "predict" sporadic and not well understood events (like a volcanic eruption, or El Nino, or for that event, "projected" anthro drivers).
My best guess is that the despised models are continuously being updated to reflect newer understandings, such as the nth-order effects of land use on the system, or that of the secondary effect of particulates on the system.
And you are correct in your assumption that the "models do not predict" the current apparent downturn in temperature. The items I have read tend to implicate this as being a result of a "downtick" in the PDO; I do not know how anyone could have temporally "predicted" this event (if, in fact, the specualtion is correct as to the PDO being the root.) But, given that it *might* be happening, nothing prevents the modelers (and models) for attempting to input this "signal" into the systems that they are describing.