The Dependence of SCM Precipitation and Clouds on the Spatial Scale of Large-scale Forcing at SGP

Tang, S., Lawrence Livermore National Laboratory

Atmospheric Thermodynamics and Vertical Structures

Convective Processes

Tang S, M Zhang, and S Xie. 2017. "Investigating the dependence of SCM simulated precipitation and clouds on the spatial scale of large-scale forcing at SGP." Journal of Geophysical Research: Atmospheres, 122(16), 10.1002/2017JD026565.


Time series of surface precipitation and high cloud.  Black line represents the domain-mean observation (Obs.).  Red line represents the SCAM5 simulation with domain-mean forcing (SCM_DM).  Blue line and blue shade represent the mean of the SCAM5 simulations with sub-column forcing (SCM_SC) and the spatial standard deviation, respectively.



Time series of surface precipitation and high cloud.  Black line represents the domain-mean observation (Obs.).  Red line represents the SCAM5 simulation with domain-mean forcing (SCM_DM).  Blue line and blue shade represent the mean of the SCAM5 simulations with sub-column forcing (SCM_SC) and the spatial standard deviation, respectively.

Science

Previous studies have suggested that some errors in the single-column model (SCM) simulations could be attributed to the lack of spatial variability in the specified domain-mean large-scale forcing.  A recently developed, gridded, large-scale forcing data set describes the spatial variability of the forcing fields, which provide an opportunity for single-column models to better simulate spatially inhomogeneous frontal systems.

Impact

The gridded large-scale forcing data describe the spatial variability of the large-scale environment.  During the passage of a frontal system, the large-scale forcing fields show different features in different part of the domain; therefore the domain-mean forcing fields may not be able to represent the actual large-scale conditions. Gridded large-scale forcing data describe the spatial variability of the forcing fields, which helps SCMs capture these spatially inhomogeneous features. By configuring the gridded large-scale forcing data into different resolutions, we can investigate the scale mismatch between model simulations with point measurements, and investigate the scale dependence of parameterizations.

Summary

We run the single-column model of CAM5 in each sub-column of the analysis domain for a deep frontal case in 2-3 March 2000 at ARM's Southern Great Plains (SGP) observatory, then average the sub-column simulations over the analysis domain. The domain-averaged sub-column simulations are then compared with SCM simulation with domain-mean forcing. The results show that the simulation with sub-column forcing reduces the delay of the precipitation onset and captures the secondary peak, which is missing in the simulation with the domain-mean forcing.  It also captures the transition feature of the domain-mean high clouds. A simple example is also given to demonstrate that the gridded large-scale forcing data can be used towards testing scale-aware parameterizations.