6.1      Partitioning between deep and shallow convection in MPAS.

Fowler, Laura D., and Mary C. Barth, Mesoscale & Microscale Meteorology Laboratory, National Center for Atmospheric Research

The impact on moist processes of a variable-resolution mesh over the Tropical Pacific Ocean is used to analyze the partitioning between shallow and deep convection in two scale-aware convection schemes implemented in the Model for Prediction Across Scales (MPAS). The first convection scheme is the Grell and Freitas (GF, 2014) scheme which comprises separate formulations of deep and shallow convection. The second convection scheme is the Multi-Scale Kain Fritsch (MSKF; Alapaty et al. 2014) scheme which more simply distinguishes between shallow and deep convection as functions of the cloud depth.

Each convection scheme is used in two 31-day experiments for December 2015. The first simulation employs a 30-km uniform resolution mesh, while the second simulation uses a 30-6 km variable resolution mesh where the refined grid is located over the warmest sea-surface temperatures in the Western Tropical Pacific Ocean.

Results show that because GF and MSKF assume that shallow convective clouds do not precipitate, the formation of precipitation in deep convective updrafts contributes a major part to increased simulated precipitation compared to TRMM Precipitation Analysis data in the uniform resolution experiments. In the variable resolution experiments, decreased deep convection over the refined area of the mesh produces increased relative humidity and excess grid-scale precipitation.

Results indicate that the GF convection scheme strongly overestimates the cloud liquid water path (LWP) compared to CERES data and to the MPAS experiments using the MSKF convection scheme, due to increased mixing from shallow convection. The GF and MSKF convection schemes yield similar cloud ice path (IWP) along the Inter-Tropical Convergence Zone and Western Pacific warm pool.