Kwinten Van Weverberg — Met Office - UK
Cyril Julien Morcrette — Met Office - UK
Category
General topics – Clouds
Description
Joint PDFs of water content and cloud fraction for (top row) observed liquid clouds (Liq-Obs) and simulations using a diagnostic and prognostic cloud scheme (Liq-Diag and Liq-Prog) during MC3E.
(2nd row) Joint PDFs for liquid clouds for stepwise improvements to the diagnostic cloud scheme,
including linking the sub-grid variance to the boundary-layer scheme (Liq-Diag-BL), introducing a cubic power law for sub-grid variability (Liq-Diag-BL-N3) and introducing a skewness profile (Liq-Diag-BL-N3-SK).
Histograms to the bottom and the right of the panels show distributions of water content and cloud fraction respectively.
(3rd row) As for the top row, but for ice clouds (Ice-Obs, Ice-Diag and Ice-Prog).
(4th row) Joint PDFs for improvements to the diagnostic clouds scheme,
including linking the sub-grid variance to the BL scheme (Ice-Diag-BL), changes to generating ice cloud fraction (Ice-Diag-BL-NW), and introducing a cubic power law for sub-grid variability (Liq-Diag-BL-NW-N3).
Representing partial cloud cover within a grid box has a large impact on radiative transfer and microphysical processes in NWP and climate models.
However, with such models starting to approach the km scale and to resolve more cloud features, the model resolution to which such a parameterization is beneficial is currently uncertain.
This study explores the added value of using a cloud fraction parameterization for a range of resolutions (up to 1 km grid spacing) in an operational NWP model.
Experiments with diagnostic and prognostic cloud fraction schemes are performed, as well as more advanced versions of these schemes linking the cloud scheme to moisture and temperature variances from the boundary-layer parameterization, and an experiment with no cloud fraction parameterization at all.
Using joint probability density functions of cloud fraction and water content, an evaluation of each of these schemes is performed against ARM Microbase data for the period of the MC3E field campaign (April-June, 2011) over the U.S. Southern Great Plains.
Observations suggest that even at 1 km grid spacing, a cloud fraction scheme would still be beneficial.
However, it is also found that current cloud fraction schemes tend to have cloud fraction distributions that are not binary enough compared to observations, and struggle to converge to an all-or-nothing scheme at the highest resolution.
A second part of this study uses the ARM Microbase data to explore a range of assumptions in the diagnostic cloud scheme applied in the first part of this study.
It is found that applying a higher-order power-law moisture and temperature distribution, as well as introducing a realistic skewness profile, largely improves the cloud properties even at the highest resolution.
A different approach for generating ice cloud cover is also proposed that leads to much improved and more binary cloud cover at high resolution, more in agreement with observed ice cloud fractions.
Joint PDFs of liquid and ice cloud water content and fraction for various experiments are shown in Figure 1.
Findings from this study will be used to develop a new scale-aware prognostic cloud fraction scheme that can be applied at a range of resolutions in convection-permitting simulations.
