Implementation and evaluation of a prognostic TKE turbulence scheme in the ECMWF IFS

The ECMWF Integrated Forecasting System (IFS) currently relies on a first-order turbulence closure that is robust at coarse resolution but increasingly limits the representation of atmospheric boundary-layer processes as model resolution increases. To better capture turbulence–surface–cloud interactions, a prognostic turbulence kinetic energy (TKE) scheme has been implemented in the IFS, providing a higher-order closure while maintaining numerical stability and affordable computational cost. By prognosing TKE, the scheme introduces memory of turbulence intensity in space and time, enabling a more physically consistent evolution of mixing and boundary-layer structure.

The formulation builds on the TKE scheme used in the global ARPEGE model and has been adapted for application in the IFS across a wide range of flow regimes. Developments include an extended turbulence length-scale formulation, revised stability functions, the inclusion of prognostic cloud fraction in stability diagnostics, partial equilibrium assumptions for selected source terms, explicit advection of TKE, and improved identification and treatment of stratocumulus regimes.

The impact of the scheme is assessed using high-resolution global simulations at 4.4 km horizontal resolution. Results demonstrate clear improvements in near-surface temperature over complex terrain, with particularly strong performance over mountainous regions, and a general improvement in 10 m wind speed. Upper-air forecast scores remain largely neutral. Further impacts on boundary-layer structure, low-level clouds, and nocturnal jets will be described.