|
|
|
|
|
by bsaul
2347 days ago
|
|
I really thank you for your comment, and the time you took to honestly assess the current situation. The one thing that i still have a hard time accepting is that a whole community of scientist accepts doing science on such a fragile basis (such as the fact that more than three quarter of the world wasn't recording temperature until 50 years ago). It seems impossible that so many people keep doing their work on fragile data (and code, apparently), while at the same time seeing their work used by politicians all over the world to advocate massive policy changes. I still haven't reached a personal conclusion, but i would be glad to read about the progress made by "hands deep in the dirt" people like you in their investigation. |
|
Mmmm. Surely it's the other way around, this is the expected and indeed only possible outcome.
I was in a different HN thread this week where I pointed out that there are some conclusions that might be right but which some sections of academia institutionally cannot reach, conclusions like:
1. We don't know enough to make predictions in this field.
2. Our datasets are inadequate for use.
3. Our research is unimportant and doesn't need to be done.
Note that commercial research can easily reach any of these conclusions; that's the function of senior management who are motivated by some fundamental ground truth goal rather than research for the sake of it.
Climatology is almost entirely driven by academia and other government institutions. They cannot reach a conclusion like, "old temperature datasets are of too low quality to derive models from" because then they'd invalidate the basis of their own careers. My impression is that climatologists have few transferable skills. Perhaps it's just small sample sizes, but it seems like their maths skills aren't really "hard" enough to outcompete physicists, mathmos or CS profs for jobs in finance or other exit routes.
Given that climatology has a single main theory (global warming), and that theory is based primarily on a single dataset (temperature), problems with that dataset have to be addressed by adjusting the data. Otherwise what's left for the field to do?