> I swear by the end of Covid19 we're all going to look back at the models and determine what climate skeptics have been saying for decades: Models are easily manipulated by bad inputs - and generally trash
Humans often do over-react to threats, and they usually have inaccurate data, and often, predictions of dire catastrophes are proven incorrect. If you look back at the past 80 years, most predictions of large-scale catastrophe have been wrong. But that also doesn't show it couldn't have happened, or why it didn't.
Most people who were supporters of those predictions later dismiss their inaccurate predictions or apologize for them. And everyone collectively ignores when their predictions of woe don't come to pass, because they're in the past now, and they don't have any tangible effect on our present or future, and also, it might be a little embarrassing.
But none of that is the point of the predictions, really. If we perceive an impending and potentially serious problem, we have to at least try to raise an alarm and do something to try to avert it, even if it turns out later there wasn't a great threat, or it wasn't caused by what we thought. We use the best guesses we have, often over-estimating on purpose, because under-estimating would have a worse effect.
Saying the models (or inputs) are "trash" is like saying covering your face with cloth to prevent a virus is "trash". Is putting a thick cotton cloth over your face going to prevent you getting the virus? No, of course not; it's not filtering nanoparticles, things can still get in or out of the corners, it doesn't stop you from touching your face or an open sore, etc. But it is one of several imperfect tools that we can use to mitigate the threat.
And ultimately, logic doesn't win in cases like these. I saw someone reasonably intelligent who was sharing a single paper a few weeks back with projected US death tolls to COVID-19, based on models with data that the researchers directly stated were incomplete. They were sharing the paper because the numbers were very scary, and on the off chance that the numbers were right, they wanted to scare people into action. The pertinent question to me in that case was, would scaring people into action cause more help or harm?
So I don't agree that the existing inputs are troubling. They may be inaccurate, and if so, they should be improved. But what I do find troubling is the possibility that people are suggesting their inaccuracy is a reason to remove them, rather than improve them. Basically, I think that to ignore the warning because the data is flawed is missing the bears for the woods.