GrumpyRabbit on Nostr: _Why Numerical Climate Models Fail at Long-term Climate Prediction_ Short answers: 1) ...
_Why Numerical Climate Models Fail at Long-term Climate Prediction_
Short answers:
1) Extreme sensitivity to initial conditions.
Planetary climate is a complex, dynamical system with multiple feedback loops—so it's subject to mathematical chaos. Per Wolfram, the only way to validly model it is to just observe what it does—which can only be done in real time, not in advance.
2) Incomplete models that don't account for all factors and/or feedback loops. There is no proof that the global climate models actually model ALL factors that matter. And there's a history of new factors that matter being discovered. Repeatedly.
3) As is required for all "investment advisors" to inform their customers: "Past performance is no guarantee of future results." The global climate models make future predictions by simply extrapolating from past history. That fails for the same reason that trading strategies based on past performance do. There's a term for that phenomenon used by trading strategy professionals: "Curve fitting." It's a pejorative term.
https://wattsupwiththat.com/2024/11/08/why-numerical-climate-models-fail-at-long-term-climate-prediction/
Short answers:
1) Extreme sensitivity to initial conditions.
Planetary climate is a complex, dynamical system with multiple feedback loops—so it's subject to mathematical chaos. Per Wolfram, the only way to validly model it is to just observe what it does—which can only be done in real time, not in advance.
2) Incomplete models that don't account for all factors and/or feedback loops. There is no proof that the global climate models actually model ALL factors that matter. And there's a history of new factors that matter being discovered. Repeatedly.
3) As is required for all "investment advisors" to inform their customers: "Past performance is no guarantee of future results." The global climate models make future predictions by simply extrapolating from past history. That fails for the same reason that trading strategies based on past performance do. There's a term for that phenomenon used by trading strategy professionals: "Curve fitting." It's a pejorative term.
https://wattsupwiththat.com/2024/11/08/why-numerical-climate-models-fail-at-long-term-climate-prediction/