On 29 May 2006 08:13:13 -0700, "basskisser"
wrote:


Jack Goff wrote:
On Sun, 28 May 2006 23:41:45 -0700, -rick- wrote:

Jack Goff wrote:
Therefore, unlike your circuit
simulator, there is no way to check the output of their climate
simulator against real-world results to verify its accuracy.

So we've apparently misplaced all records of the past?

Of course not. But those records are woefully incomplete to enable an
accurate model to be constructed. How many weather satellites did we
have 100 years ago?

Maybe they didn't have weather satellites then, but they had weather.
They also had people quite competent in keeping data.

They sure did. Unfortunately, the instruments they used were crude
compared to today's, and many measurements needed weren't made because
the science to allow their measurement didn't exist.


You seem to be thinking that climate is like an NPN transistor. It's
not. Think of a black box with 200 inputs and 10 outputs. We know
what the ouputs are, and can measure them. We know what most of the
inputs are, and are pretty sure about the rest. It's reasonable to
assume that there's a few that we don't know about, and may never
know. Of the inputs we understand, we've just recently identified and
have been able to measure many of them (in the climate timeline scheme
of things). We've seen that there is a huge time lag inside of this
box, sometimes years, sometimes decades. Finally, we have virtually
no control of any of the inputs, so we can't change just one and
observe the outputs. Most of the inputs are totally out of our
control, and are constantly changing. So once again, unlike your
simple circuit on the bench, the climate computer model can not be
verified against the real world.

So answer this, Rick. As previously discussed, weather models can't
tell us with any decent accuracy what it will be like in 5 days. Are
you really telling me that you believe a climate model's prediction
for 94 years into the future?

Flawed analogy. Very flawed. the model for recent events (5 days in
your case is much more detailed and refined than the 94 year model. The
more detailed and the more refined a model is, the more instances of
error. Ergo, while a 5 day model might not be accurate in your eyes, if
it were the same detail as the 94 year model, it would be spot on.

No more flawed than the logic you just used. Weather is *simple*
compared to climate. Point is that we can't predict something
relatively simple for a short 5 day window, but you expect to be able
to predict something extremely complex for a 100 YEAR window?

I can predict our climate 5 days from now... barring a huge meteor
impact, it'll be virtually the exact same as today's. The meteor
points out one of the problem with climate prediction... we don't know
what we don't know. No matter how you dress it up, any prediction
from the experts is no better than a guess. Maybe in another hundred
years, we'll get better at it. Or maybe we'll be extinct.

No one knows. :-)