Hacker News Reader

Comment by gadders | Hacker News Reader

Doesn't that apply to climate science/ climate change though?

What are you implying? Scientists can't make stars either, but they can make models describing stars and make new observations about these stars that test the validity of these models. Plenty of climate models have been disproven by this exact way. What has been remarkably consistent is that man-made greenhouse gases are causing climate change. For example, models predict that man-made global warming would cause the stratosphere to cool, which is exactly what they measured. By contrast, increased solar activity would have also heated up the stratosphere.

Why? Climate science works with quantative models and makes concrete predictions. So if the predictions don‘t come true the model is false. E.g. this random blog I found googling compares IPCC predictions with actual outcomes: https://johncarlosbaez.wordpress.com/2012/03/27/the-1990-ipc... [the link is just supposed to show how climate science is falsifiable. I have no idea whether the numbers in this specific source are trustworthy]

More comparisons here[0]. As is important, it's noticed that models even from the 80's made fairly accurate predictions. That's the real kicker: models accurately predict observable outcomes 20-40 years later (and have only improved since then).

I'm not sure why there's so much contention about climate science given that we've been observing this trend for nearly 50 years now. Scientist makes prediction about something 50 years later and 50 years later the prediction is shown to be accurate. It feels weird to not trust a source with such a good and observable track record...

[0] https://www.carbonbrief.org/analysis-how-well-have-climate-m...

The thing is - there are so many predictions made and a lot that don't come through, it's hard to find which are the "real" ones.

I also read somewhere that there are several different scenarios available for the IPCC climate models and a lot of the more "scary" predictions are based on the least likely model.

Finally - we saw the limitations of modelling in COVID in the UK, with adverse consequences.

Reply on news.ycombinator.com