Soren wrote on Nov 14
th, 2010 at 3:55pm:
A point of reference for what? It's completely irrelevant to illustrating the difference between big poicture (river, climate) and detail (rapids, weather). How long the paper boat takes from here to there is utterly, utterly irrelevant to this illustration. You need absolutely no statistical point of reference to demonstrate the difference between seeing only a tree or the whole wood.
It's an analogy. The position if the boat represents where we are at any given time. We can get a broad picture of where the boat is located at any given time quite easily by using some fairly simple calculations.
Once you narrow in on the position of the boat a higher resolution, the mathematics becomes more difficult - ok?
Let's drop it, because you clearly understand the difference.
Most climatic data is averaged over at least a 20 year period. If we take the temperature maps for Australia, you'll find very little difference over the last 50-100 years. For example:

and

As you can see - not much difference. Maybe a bit of an increase in most areas. So blind Freddy can see that it's relatively easy to predict reasonably accurately what that chart will look like in another 20 years given a relatively steady state scenario.
No complex models needed.
However, if you try to predict what the precipitation in in Bankstown next Sunday, it's a far more difficult thing to do.
Given that we're talking about average global temperatures and precipitations, the argument that it's incredibly complex becomes silly.
Another analogy is predicting the shape of an baby elephant when it grows to adulthood. I think we agree that we could make a reasonably good prediction of that, however predicting the shape of the wrinkles in its skin would be extremely difficult by comparison.
High resolution= weather; low resolution = climate.