Estimation of the Vertical Velocity Leading to the Formation of Cirrus Using Ultra-High Resolution Global Simulations

Cirrus clouds significantly impact the radiative and transport processes of the upper troposphere and the lower stratosphere. State-of-the-art global models parameterize the formation of cirrus explicitly linking ice nucleation events to the aerosol properties and the cloud-scale dynamics. However most GCMs cannot resolve the scale at which cloud formation occurs. Thus subgrid scale dynamics is typically parameterized by relating the vertical velocity variance, σw, to grid-scale fields. These parameterizations are typically validated against field campaign data for specific locations. However an assessment of the global spatial distribution of σw is lacking, limiting the ability of GCMs to describe cirrus formation. Here the non-hydrostatic version of the NASA Goddard Earth Observing System model (GEOS-5) is used to estimate the variance of vertical velocity in GCMs. GEOS-5 was run at cloud-resolving resolutions (~7 km), allowing the explicit calculation of σw. Our results indicate that σw is determined by orographic drag and local convection, and higher over the continents than over the ocean. A recently developed parameterization of σw is also evaluated. Compared to the model results the parameterization is able to reproduce the global distribution of σw for warm cirrus clouds but tends to overestimate σw near the tropopause. Our work provides for the first time an assessment of the global variability in the subgrid scale dynamics leading to the formation of cirrus.