Identification of cirrus formation regimes using cluster analysis of back trajectories and satellite data

In recent years our understanding of cirrus cloud processes has been significantly advanced. However, a large uncertainty regarding the influence of cirrus formation mechanisms on the microphysical properties, and hence radiative properties of cirrus clouds still remains. This leads to uncertainty in global climate models and climate change projections. In this work we aim to identify different cirrus formation regimes and analyze their influence on cirrus microphysical properties. We combine DARDAR-Nice satellite observations with Lagrangian back trajectories of meteorological and aerosol reanalysis data on the Northern Hemisphere. Our goal is to classify observed cirrus clouds by means of their trajectories and investigate the trajectories' influence on observed cirrus microphysical properties. With our data-driven nested clustering approach we identify different meteorological regimes that lead to cirrus formation. We are also able to isolate the effect of dust ice nucleating particle (INP) exposure along the trajectory from meteorological variability.

We identify four different meteorological clusters that lead to characteristic cirrus cloud microphysical properties and can be associated with liquid origin and in-situ formed cirrus clouds. Furthermore, we find that dust concentrations in cirrus cloud back trajectories are significantly higher compared to cloud free trajectories with comparable meteorological conditions. This indicates the importance of dust acting as INP during heterogeneous nucleation. The magnitude of the dust concentration, however, has only a negligible effect on cirrus microphysical properties.