A description of the finite differencing used by the B-grid dynamical core

This paper describes the dynamical part of a global, hydrostatic grid point model developed from models described in Mesinger et al. (1988) and Wyman (1996). The model described here solves the same set of dynamical equations but has a modified vertical and horizontal grid. The vertical grid is a hybrid sigma-pressure coordinate system, sigma coordinate surfaces in the lower model levels may gradually transform to pressure coordinate surfaces in the upper model levels. The horizontal grid has been switched to the Arakawa B-grid (Arakawa and Lamb 1977). This has several advantages over the previously used E-grid configuration, especially in the global domain, where the convergence of diagonal fluxes near the pole of the E-grid has been eliminated in the B-grid. An additional benefit is the simplification of model output, a single field on the B-grid is essentially on its own rectangular A-grid. The basic time differencing schemes used has not changed from either of the preceding models. The step-mountain (eta) coordinate is no longer a supported option, although much of the code remains in place, very little effort has been made to ensure that it works correctly. This paper will describe the finite difference schemes used to solve the dynamical equations on the global B-grid. This will include descriptions of the time differencing, horizontal mixing, lateral boundary conditions, and polar filtering.