Spurious Cloud-Edge Supersaturation in Eulerian and Lagrangian Microphysical Schemes

This presentation focuses on a problem with numerical modeling of cloud processes near sharp cloud–environment interfaces. The problem is known as the spurious cloud-edge supersaturation and it comes from the advection of cloud water into an unsaturated gridbox. It arises in models due to the lack of representation of partially cloudy gridboxes in most microphysical schemes, and the nonlinear coupling between the temperature, water vapor and supersaturaion. Since accurate prediction of the supersaturation field is critical for detailed microphysical representation, the spurious supersaturation problem can lead to unphysical microphysical properties of the cloud water, cloud droplet concentration and particle size spectrum in simulated clouds (e.g. Stevens et al., 1996; Grabowski and Morrison, 2008).

In the presentation, I will compare behavior of various microphysical schemes in a 1-dimensional test. The test includes Eulerian schemes, but also particle-base Lagrangian representation of warm-rain microphysics from the libcloudph++ library (http://libcloudphxx.igf.fuw.edu.pl/; Arabas et al., 2015). The Lagrangian and Eulerian (e.g. double-moment) schemes treat the processes of cloud droplet activation/deactivation, condensational growth and evaporation in a different ways. Therefore, the spurious cloud-edge superaturation affects the cloud field properties differently. In addition, some implementations of the Lagrangian scheme apply local values of the variables instead of the gridbox means. This can improve the simulation and possibly eliminate the spurious cloud-edge supersaturation.

I will also discuss how the approach based on advection of the supersaturation instead of the water vapor mixing ratio (Grabowski and Morrison, 2008) can significantly improve the situation even in the schemes that use only the gridbox means.