Editing Openai/6938382f-f9e4-8003-a158-6d812c1bce84 (section)

===== 4.1 Comparison with mainstream practice =====

Mainstream climate science:
* Recognizes clouds and cloud feedbacks as among the largest sources of model uncertainty, but does not conclude that no anthropogenic signal is detectable. IPCC<ref>{{cite web|title=IPCC|url=https://www.ipcc.ch/site/assets/uploads/2018/02/WG1AR5_Chapter09_FINAL.pdf|publisher=ipcc.ch|access-date=2025-12-10}}</ref>
* Treats uncertainty via: - multi-model and perturbed-physics ensembles, - internal variability sampling (multiple initial-condition runs), - scenario uncertainty, and - observational and forcing uncertainties. IPCC<ref>{{cite web|title=IPCC|url=https://www.ipcc.ch/site/assets/uploads/2018/02/WG1AR5_Chapter09_FINAL.pdf|publisher=ipcc.ch|access-date=2025-12-10}}</ref>
* Uses optimal fingerprinting and other statistical detection methods to show that anthropogenic forcing is very likely the dominant cause of observed warming since mid-20th century. IPCC<ref>{{cite web|title=IPCC|url=https://www.ipcc.ch/site/assets/uploads/2018/02/WG1AR5_Chapter09_FINAL.pdf|publisher=ipcc.ch|access-date=2025-12-10}}</ref>

Frank:
* Asserts that IPCC evaluations do not propagate physical errors and use only precision measures. Propagation_of_Error_and_the_Re… - This is overstated. AR5 explicitly discusses multiple uncertainty sources (internal variability, model structural differences, forcing uncertainty) and how ensemble spread relates to them. IPCC<ref>{{cite web|title=IPCC|url=https://www.ipcc.ch/site/assets/uploads/2018/02/WG1AR5_Chapter09_FINAL.pdf|publisher=ipcc.ch|access-date=2025-12-10}}</ref> - It’s true that IPCC doesn’t run a literal Taylor-Kuyatt “propagate lab-instrument errors through a model” exercise, but GCMs are not simple algebraic devices; uncertainty quantification in complex nonlinear PDE systems is necessarily different.
* Uses cloud error diagnostics from Jiang et al. and Lauer & Hamilton, but reinterprets them in a way those authors never suggest, and that is inconsistent with follow-up work which views them as biases to be reduced and understood, not as per-year random forcing noise. AGU Publications<ref>{{cite web|title=AGU Publications|url=https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2011JD017237|publisher=AGU Publications|access-date=2025-12-10}}</ref>