B4 - Radiative interactions at the NWP scale and their impact on midlatitude cyclone predictability
Diabatic heating by radiation is a main driver of the atmospheric circulation. This project addresses the rep- resentation of radiative transfer in numerical weather prediction (NWP) models, with the aim to identify ways for improvements in terms of computational speed and accuracy, and to study the dynamical impact of three- dimensional (3D) radiative effects that are currently neglected in NWP models. The work is motivated by earlier work on the impact of 3D radiative effects on cloud organization and the importance of interactions between radiation and cloud microphysics. The work is further motivated by the current trend towards high-resolution models, for which 3D radiative effects become increasingly important compared to models with coarser resolution.
The work is structured in three work packages (WP), with each WP employing one PhD student.
WP1 aims at substantially speeding up radiative transfer in NWP models and will address the impact of how often radiative transfer is calculated in NWP models. Currently, radiation is called less often than dynamics: the radiation time step in ICON is around 15 minutes, even for ICON-LEM operated at 300 m. This is problematic because clouds vary much faster, and our previous studies showed that correct spatial distribution of radiative heating rates matters for cloud organization. At the same time, substantial computational resources are currently spent on spectral integration. We will optimize the trade-off between calling frequency versus spectral accuracy, and will explore new ways to more effectively implement 3D radiative transfer computations on modern computer architectures, e.g., by treating radiation "dynamically" in each model time step.
WP2 will address implications of radiation for midlatitude dynamics and predictability. WP2 is motivated by the recent finding of Aiko Voigt that cloud-radiative effects weaken idealized midlatitude cyclones. Baroclinic life cycles will be simulated with the ICON-NWP model at convection-permitting resolution of 2.5 km, with subdomains simulated by ICON-LEM with 300 m resolution. This will allow us to consider 3D radiative effects of grid-scale clouds using existing parameterization for ICON-LEM developed by Bernhard Mayer. The synoptic impact of cloud-radiative effects will be analyzed by means of a potential vorticity framework developed in Phase 1, and the life cycle results will be put into context through a NAWDEX case study.
WP3 will study 3D radiative effects in the "grey zone" of the ICON-NWP model (resolutions of 1-10 km), for which both grid-scale and subgrid-scale effects are important, as shown in Phase 1. The grey-zone requires accurate treatments of cloud overlap for subgrid-scale clouds, grid-scale 3D effects, and horizontal gradients of water vapor and temperature at front lines. This work will study at which resolution and under which circumstances radiative effects from sharp spatial contrasts of clouds, water vapor, and temperature are relevant, and how they can be parameterized.
The project is a twin-project between LMU and KIT. It combines the expertise in radiative transfer of Bernhard Mayer (LMU) with the expertise in modeling and cloud-radiation-circulation coupling of Aiko Voigt (KIT). The project will further draw on the dynamical and microphysical expertise of W2W colleagues.