Attributing the behavior of low‐level clouds in large‐scale models to subgrid‐scale parameterizations

Abstract

AbstractThis study explores ways of establishing the characteristic behavior of boundary layer schemes in representing subtropical marine low‐level clouds in climate models. To this purpose, parameterization schemes are studied in both isolated and interactive mode with the larger‐scale circulation. Results of the EUCLIPSE/GASS intercomparison study for Single‐Column Models (SCM) on low‐level cloud transitions are compared to General Circulation Model (GCM) results from the CFMIP‐2 project at selected grid points in the subtropical eastern Pacific. Low cloud characteristics are plotted as a function of key state variables for which Large‐Eddy Simulation results suggest a distinct and reasonably tight relation. These include the Cloud Top Entrainment Instability (CTEI) parameter and the total cloud cover. SCM and GCM results are thus compared and their resemblance is quantified using simple metrics. Good agreement is reported, to such a degree that SCM results are found to be uniquely representative of their GCM, and vice versa. This suggests that the system of parameterized fast boundary layer physics dominates the model state at any given time, even when interactive with the larger‐scale flow. This behavior can loosely be interpreted as a unique “fingerprint” of a boundary layer scheme, recognizable in both SCM and GCM simulations. The result justifies and advocates the use of SCM simulation for improving weather and climate models, including the attribution of typical responses of low clouds to climate change in a GCM to specific parameterizations.