| Summary: | Mixed-phase stratus clouds are the prevalent cloud type in the Arctic during the winter and transition seasons. Despite their important role in various climate feedback mechanisms they are still not well understood and are difficult to represent accurately in large-scale models. In this study the role of ice nucleation mechanisms, the influence of the coastally-generated circulations and parameterized ice crystal habit on the longevity and structure of Arctic mixed-phase clouds were examined using detailed mesoscale, cloud- and eddy-resolving model simulations. The structure and the lifetime of simulated Arctic mixed-phase clouds were found to be highly sensitive to the concentration of ice-nuclei (IN) acting in deposition/condensation-freezing nucleation mode. Contact nucleation could not produce significant ice amounts unless the contact IN concentrations were increased to unrealistically high values. Local, coastally induced circulations were found to be responsible for maintaining the continuous ice precipitation along the coastline through transport of deposition/condensation-freezing IN from above the cloud layer. It was demonstrated that incorrect partitioning of the liquid and ice phase can produce errors in the surface radiative budget of up to 90 Wm-2. Simulated IN sensitivity and liquid/ice phase partitioning were found to depend critically on the assumed ice crystal habit. It was demonstrated that a large range of liquid or ice water path can be produced by reasonable changes in ice crystal habit mass-dimensional and terminal fall-speed relations based on data reported in the literature. The changes in ice crystal habit were shown to be related to liquid layer formation, splitting of liquid layers, and cloud dissipation mechanisms in multi-layered Arctic mixed-phase clouds. These results suggest that predicting changes in crystal habit is of significant importance for correct model representation of mixed-phase clouds. Three additional ice nucleation mechanisms (evaporation IN, evaporation freezing, and immersion freezing) were examined regarding the ability of our model to more accurately simulate liquid and ice water content, ice concentrations, and observed cloud structure. All of these mechanisms were found to be capable of producing ice crystal concentrations similar to the observed values, while maintaining the liquid content of the cloud. The observed cloud structure was also correctly reproduced for extended periods of time.
|
|---|