Investigating the Scale Dependence of SCM Simulated Clouds by Using Gridded Forcing Data at SGP

Large-scale forcing data, such as vertical velocity and advective tendencies, are required to drive single-column models (SCM), cloud-resolving models and large-eddy simulations. Previous studies suggest that some model discrepancies could be attributed to the lack of horizontal inhomogeneity in the large-scale forcing data. This study investigates the spatial variability of the gridded large-scale forcing data derived from a three-dimensional constrained variational analysis (3DCVA) method and its impact on SCM simulated clouds and radiation. The analysis of March 2000 intensive operational periods at the ARM SGP site shows that the large-scale forcing data have large spatial variability especially when there are frontal systems passing through. SCM simulations are performed in each sub-column and the combination of these sub-columns captures some characteristics of the front structure. However, physical parameterizations are still the most important error source in SCM, since most model discrepancies remain in the sub-column runs. Overall, with the spatial variability of forcing, the average of sub-column SCM simulations of clouds and radiation are more consistent with the domain-averaged observations compared to the coarse column simulation