Article: Climbate Change Irreversible - What to do now?

[Figment]

Originally Posted by Vreki

But there lies the problem. Those models are based on historical data, so of course the output of those models will match the historical data.

Not really true, is it? The historical data is the output ONLY if the model is correct. If the model is incorrect, the input will mess up the output and you'd get entirely different results.

So yes, the input is based on historical values, but that doesn't mean the model is historical accurate if the input values are poorly used. There's a big difference between just showing the historical data (which is what you claim it does) and computing and predicting the historical data.

In the latter case, you use historical data to verify if your model is correct, because that's what actually happened.

To be predictive, it will have to match future data, and there we again have the problem with the timeframe.

So you're saying weather models are completely inaccurate? I'm quite certain we're able to look some distance in the future.

Again, there's a degree of uncertainty, which increases as the future becomes more distant (as more variations in input could have happened). That doesn't mean you can't extrapolate your model.


What you typically get as outcome is a highly diverse range of options with a lower and upper level and it could be anywhere in between, in theory. However, based on data you have and likelier levels of input, certain trends become baselines.

Regardless of which trendline you pick, the trendlines aren't very positive under the existing models.



And the existing models have been right so far.