Climate Change Feedbacks in Aquaplanet Experiments With Explicit and Parametrized Convection for Horizontal Resolutions of 2,525 Up to 5 km

• Main Authors: M. H. Retsch, T. Mauritsen, C. Hohenegger
• Format: Article
• Language: English
• Published: American Geophysical Union (AGU) 2019-07-01
• Series: Journal of Advances in Modeling Earth Systems
• Subjects: ICON; climate change feedbacks; aquaplanet; high resolution; explicit convection; ECS
• Online Access: https://doi.org/10.1029/2019MS001677

Abstract Earth's equilibrium climate sensitivity (ECS) is the long‐term response to doubled atmospheric CO2 and likely between 1.5 and 4.5 K. Conventional general circulation models do not convincingly narrow down this range, and newly developed nonhydrostatic models with relatively fine horizontal resolutions of a few kilometers have thus far delivered diverse results. Here we use the nonhydrostatic ICON model with the physics package normally used for climate simulations at resolutions as fine as 5 km to study the response to a uniform surface warming in an aquaplanet configuration. We apply the model in two setups: one with convection parametrization employed and one with explicit convection. ICON exhibits a negative total feedback independent of convective representation, thus providing a stable climate with an ECS comparable to other general circulation models, though three interesting new results are found. First, ECS varies little across resolution for both setups but runs with explicit convection have systematically lower ECS than the parametrized case, mainly due to more negative tropical clear‐sky longwave feedbacks. These are a consequence of a drier mean state of about 6% relative humidity for explicit convection and less midtropospheric moistening with global warming. Second, shortwave feedbacks switch from positive to negative with increasing resolution, originating foremost in the tropics and high latitudes. Third, the model shows no discernible high cloud area feedback (iris effect) in any configuration. It is possible that ICON's climate model parametrizations applied here are less appropriate for cloud resolving scales, and therefore, ongoing developments aim at implementing a more advanced prognostic cloud microphysics scheme.