An analysis of cloud cover in multiscale modeling framework global climate model simulations using 4 and 1 km horizontal grids

Over the past few years, a new type of global climate model (GCM) has emerged in which a two-dimensional or small three-dimensional cloud resolving model is embedded into each grid cell of a GCM. This approach is frequently called the multiscale modeling framework (MMF) but is also known as a cloud-resolving convection parameterization or a superparameterization. In this article, we compare joint histograms of cloud top height and optical depth from the MMF with those being produced by the International Satellite Cloud Climatology Project (ISCCP) and from the Multiangle Imaging Spectroradiometer (MISR). While the form of the ISCCP and MISR data sets is conceptually similar, the satellite sensors and the algorithms differ, with the result that the joint histograms can differ quite significantly even when viewing exactly the same clouds. The analysis takes advantages of the strengths of each data set, as well as the differences in these data (which, for example, allow one to characterize the amount of some multilayer clouds). MMF simulation runs with three different resolutions are analyzed. One simulation is run using a 4 km horizontal grid with 26 vertical levels (on a stretched grid), a second simulation is run with a 1 km horizontal grid and the same 26 vertical levels, and a third simulation is run with a 1 km horizontal grid and 52 vertical levels. The analysis shows that the MMF reproduces the broad pattern of tropical convergence zones, subtropical belts, and midlatitude storm tracks as observed by ISCCP and MISR. However, the model has several significant shortcomings. Perhaps most seriously, it significantly underpredicts the amount of low-level cloud in most regions. The simulation with a 1 km horizontal grid and 52 vertical layers is found to improve modestly several aspect of the MMF low-level cloud cover. The model output is obtained using ISCCP and MISR instrument simulators and the role of horizontal resolution in the instrument simulators is examined.