Robert Wood and Dennis L. Hartmann

Submitted September 2004 to Journal of Climate

In part I of this study we focused upon models to describe the mesoscale spatial variability of cloud liquid water path LWP in low cloud. Here, we construct simple, exploratory, climatologies of the spatial variability characteristics of marine low clouds for one season and two regions of the East Pacific. The climatologies show interesting and important geographical variability. Large scale meteorological conditions are assessed for each MODIS scene using NCEP/NCAR reanalysis data, and regression analysis is used to assess relationships between large-scale forcings and mesoscale variability. Cloud fraction and LWP homogeneity are both strongly correlated with lower tropospheric stability LTS on timescales of days to weeks. While marine boundary layer (MBL) depth z_i is itself strongly correlated with LTS, cloud fraction is not found to correlate particularly well with z_i, suggesting that it is LTS itself, rather than z_i that is the primary determinant of low cloud fraction. However, MBL thermodynamic and cloud mesoscale variability does appear to be influenced by z_i. This result reflects the findings that mesoscale variability is an increasing function of the characteristic lengthscale of the mesoscale cellular convection (MCC), and that the characteristic lengthscale scales with z_i. The observed changes in cloud mesoscale variability from shallow stratus-topped MBLs through to trade cumulus can be accounted for using a physical framework that links MBL thermodynamic variability and cloud liquid water path variability using a pdf formulation. The pdf width is strongly tied to the MBL depth, and it is suggested that this framework, with scalings constrained by observations, is suitable for the parameterization of low cloud subgrid spatial variability in large-scale numerical models.