Premium
Why are mixed‐phase altocumulus clouds poorly predicted by large‐scale models? Part 2. Vertical resolution sensitivity and parameterization
Author(s) -
Barrett Andrew I.,
Hogan Robin J.,
Forbes Richard M.
Publication year - 2017
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
eISSN - 2169-8996
pISSN - 2169-897X
DOI - 10.1002/2016jd026322
Subject(s) - sensitivity (control systems) , environmental science , liquid water path , liquid water content , radiative transfer , atmospheric sciences , supercooling , resolution (logic) , boundary layer , phase (matter) , geology , meteorology , mechanics , physics , cloud computing , optics , electronic engineering , artificial intelligence , quantum mechanics , computer science , engineering , operating system
Single‐column model simulations of mixed‐phase altocumulus clouds were shown to have a strong vertical resolution sensitivity in Part 1 of this paper. Coarse‐resolution models were unable to simulate the long‐lived supercooled liquid layer at cloud top, typically only 200 m thick. In this paper, the sensitivity to vertical resolution is investigated using idealized simulations. Vertical gradients of ice water mixing ratio and temperature near cloud top are found to be inadequately represented at coarse resolution. The vertical discretization using grid box mean values, rather than the full vertical profile, leads to biased calculations of mixed‐phase microphysical process rates and affects the diagnosis of thin liquid water layers. As a result, the liquid water layer becomes quickly glaciated and altocumulus cloud lifetime is underestimated. Similar impacts are expected for mixed‐phase boundary layer clouds commonly observed at high latitudes. A novel parameterization is introduced that accounts for the vertical gradients of ice water mixing ratio and temperature in the microphysics calculations and the diagnosis of liquid near cloud top. It substantially improves the representation of altocumulus layers in coarse vertical resolution single‐column model simulations and reduces the bias identified in Part 1. The new parameterization removes the large underestimate in supercooled water content caused by the resolution sensitivity for temperatures warmer than −30°C. Given the radiative importance of mixed‐phase altocumulus clouds, their underestimation by numerical weather prediction models, and their potential to act as a negative climate feedback, there is a need to reevaluate the global climate sensitivity by implementing the findings in these two papers in a climate model.