Diagnostics for and Evaluations of New Physical Parameterization Schemes for Global NWP Models

Book excerpt: The procedures and results of a study undertaken to evaluate and assess the impacts of three new parameterization schemes for the GL global spectral model as a 3-4 day range forecast model are described. The tree parameterization schemes are one each for the boundary-layer physics, moist convection and heating due to solar and terrestrial radiations. These schemes are incorporated either singly or jointly into a rhomboidal-30, 12-layer global spectral model for four-day simulations using FGGE III- a data as input. Evaluation and assessment are made on the basis of two kinds of global statistics: mean and root-mean-square errors, and on their magnitudes and distributions. The statistics are generated for both the primary, that is, prognostic, variables, and supplementary variables such as zonal-mean and zonal-eddies energy densities. The new moist convection scheme has been found to increase convective activity significantly and maintain it throughout the four-day period. It also warms and dries the middle troposphere, but produces rainfall far in excess of the climatology. The radiation parameterization has been found to cool the atmosphere and reduce its specific humidity. It counterbalances enhanced heating and moistening brought about by the new boundary-layer and moist convection schemes and eliminates the systematic warming of the old model.