B7 - Identification of robust cloud and precipitation states via inverse methods
Other researcher: Nikolas Porz (PhD student)
Cloud states and cloud dynamics depend crucially on environmental conditions (atmospheric flows, thermo- dynamics, etc.), and in numerical simulations they also depend on algorithmic representations of the corresponding cloud processes. Although these dependencies can have a strong influence on the predictability of the atmosphere, it is not clear how sensitive cloud states respond to perturbations of initial or boundary conditions, respectively. In this project we study two typical environmental scenarios, namely (i) stratified flow over mountains, and (ii) warm conveyor belts, and we investigate the following questions:
- how predictable are cloud and precipitation states,
- how probable are different paths through the phase space leading to similar clouds and/or precipitation,
- how do competing ice formation pathways affect diabatic heat sources in the upper troposphere.
To tackle these questions we have used Phase 1 to design an up-to-date warm cloud model, which has similar complexity as operational parameterizations, but is consistently built on physical principles and thorough numerical algorithms. A novel key feature of this model is a reasonable handling of droplet formation in supersaturated regimes based on the well-known Köhler theory. The model is guaranteed to return nonnegative water densities and to conserve the total water mass up to losses due to precipitation. Like other cloud models it uses parameters that are not directly accessible to measurements, and those have been calibrated using precipitation time series as reference data.
In Phase 2 we will complement this model with an ice phase in the spirit of the so-called P3 scheme, but consistently simplified to our needs. In addition, we will include recent scientific findings concerning the interaction of liquid and ice particles as well as different ice formation processes. The final model will be a competitive alternative to existing operational cloud parameterizations.
Once the model is complete we study the impact of variations in environmental conditions, focusing on the two relevant scenarios mentioned above. Using statistical ensembles and inverse problems theory we analyze the relevance and robustness of different cloud pathways and precipitation states. Concerning the warm conveyor belts we derive correlations between the ice formation pathways and intensities of resulting diabatic heat sources. This will help to reduce uncertainties on cloud scale and to achieve better atmospheric predictability.