High-resolution simulations to improve and tune the boundary-layer cloud parametrizations – HIGH-TUNE

Because of their large coverage and cooling effect of the Earth system, boundary-layer clouds are key elements of the climate system. They modulate the water and energy cycles of the atmosphere and strongly impact surface temperatures at various scales. These clouds are often much smaller than a grid cell of global weather forecast and climate models and must thus be “parameterized” through a set of approximate equations that aims at representing the collective behaviour of an ensemble of clouds and their impact on the large-scale model variables. The approximate nature of parameterizations and the diversity of approaches for the choices of associated free parameters through tuning are responsible for important biases in global models. Moreover, the spread of the boundary-layer cloud radiative effects dominates the spread of climate change projections for global warming, in response to a given perturbation of greenhouse gases.
The main objective of HIGH-TUNE is to improve the representation of boundary-layer clouds focusing on the boundary-layer dynamics and cloud-radiation interaction. Important progresses have been made in the last decades in boundary-layer cloud parameterizations, based on the comparison of single-column versions of the global models (SCM) with explicit 3D high-resolution Large-Eddy Simulations (LES) of the same scene of boundary layer clouds. To go one step further, we will build on this approach based on SCM/LES comparison, thanks to two important methodological breakthroughs that benefited from recent advances in other scientific disciplines: i/ estimation of the radiative effect of clouds from LES results using efficient Monte-Carlo algorithms will be used as a reference for parameterization evaluation and tuning; and ii/ state-of-the-art statistical tools for automatic tuning will be adapted to the SCM/LES comparison. Combining automatic tuning tools and full radiative computations will allow us: 1/ to address the energetic tuning of climate models on process-based studies and propose parameter ranges for the final global tuning and 2/ to progress in the representation of clouds themselves and in the understanding of how they depend on boundary-layer dynamics and radiative approximations. To reach these objectives, the consortium gathers applied mathematicians, radiative transfer experts, climate and atmosphere modelling experts, which guarantees a real and significant outcome of the project in state-of-the-art global weather forecast and climate models.
Beyond the range of acceptable parameters values, the automatic tuning with radiative metrics will provide a more comprehensive documentation and understanding of the parameterization behaviour in several cloud regimes. It will be used to revisit several aspects of cloud parameterizations and consider new developments as: (i) adding the representation of dry air intrusion (key process in the transport of water vapour) (ii) improving the representation of the subgrid horizontal and vertical heterogeneities of clouds, the overlap assumption for cloudy grids as well as the solar zenith angle dependency at high latitudes and (iii) refining the assumptions made on cloud optical properties and microphysics. The improved and tuned parameterizations will be systematically tested in full 3D configuration and compared with satellite and in-situ observations.
The main outcomes of the project will be: 1/ the first demonstration of a tuning strategy at the process scale, with in particular, the use of cloud radiative effects as a central metrics; 2/ the availability of an efficient code for computing radiation on 3D cloud scenes from LES; 3/ improved representation of boundary layer clouds for global weather forecast and climate models, through improved parameterizations and better tuning of free parameters.