Clouds provide one of the greatest uncertainties in climate projections. Covering fractions of mm to planetary scale waves, clouds are intrinsically challenging to forecast. Yet, the complex phase transition of water is so interesting, that many physicists, meteorologists, and engineers are keen to better understand the trubulent dynamics of clouds.

We approach this topic from two perspectives: one is the practical predictability of convection in a data-assimilation framework (Fig. 1), while the second perspective reveals insights from convective scale to planetary scale dynamics in two-dimensional turbulence on a shallow Earth.

Wind energy is one of the growing industries in Europe. Correct short term forecasts of the local wind field in combination with a detailed knowledge of the flow field around local wind turbines are essential to build efficient wind farms for optimum power production. We use a combination of high resolution large-eddy simulations (LES) and linear theory to study the stability of the wake flow behind an invidividual wind turbine. Linear theory can increase understanding in idealized configurations (Fig. 2).

High resolution simulations take into account meteorological effects such as wind shear, diurnal variation of wind strength, or direction. Incorporating turbulent inflow boundaries in a large-eddy simulation opens interesting research questions in state-of-the-art modelling of atmospheric boundary layers (ABLs).
Our novel approach with EULAG allows for fully developed turbulence upstream of a wind turbine.

Forests cover approximately 40 % of the earth's land surface and play a key role in our ecosystem as a sink of CO₂ and source of oxygen. In order to model exchange processes correctly, it is crucial to understand the flow phenomena that occur inside forest canopies (Shaw 1988). Traditionally, forests are represented as a horizontally homogenous porous medium in large eddy simulations (LES), as done in the pioneering LES by Shaw and Schumann (1992). We use a different approach and resolve individual porous trees inside the Pythagoras grove with the geophysical flow solver EULAG (Fig. 3). Thereby, we create a multiscale response on the flow structure.

Mean profiles of momentum and heat transport inside the resolved Pythagoras grove coincide with the earlier parametrizations. At the same time, flow phenomena can be investigated in detail at a scale of the individual trees.