Comment on “Opinion: Can uncertainty in climate sensitivity be narrowed further?” by Sherwood and Forest (2024)

Abstract

This comment addresses assertions made by Sherwood and Forest (2024) (SF24) regarding the narrowing of the range of equilibrium climate sensitivity (ECS). SF24 challenged a previous study by Lewis (2022) (L22) that found a narrower and substantially lower ECS level. This comment clarifies that, contrary to SF24’s claims, L22 did not rule out a high ECS level based on historical evidence and did identify and correct errors in Sherwood et al. (2020), in particular in relation to its likelihood estimation; their method, ironically, substantially underestimated likelihood for their historical evidence at high ECS levels. It also appraises L22’s revisions to S20’s methods and input assumptions and considers how these have contributed to the lowering and narrowing of the ECS range. This comment also discusses the role of priors in Bayesian ECS estimation and explains why the subjective Bayesian approach favoured by SF24 can often produce unreliable inference for uncertain parameters such as ECS. Finally, the importance of considering structural uncertainties in climate models, particularly regarding tropical warming patterns, is extended beyond the points raised by SF24. Such uncertainties could affect ECS estimation, not only from historical period evidence, but also from climate process understanding and emergent constraints. They seem more likely to suggest that existing ECS estimates are too high rather than too low.