Dynamic Initialization for Whole Atmospheric Global Modeling

Abstract

An iterative dynamic initialization method is presented to produce balanced initial conditions for whole atmospheric global modeling. In this method, a global hydrostatic numerical model is iteratively nudged toward ground-to-space wind and temperature profiles at specific date and time. Ground-to-space atmospheric profiles are obtained by fitting spline curves to reanalyses below the lower mesosphere and empirical model results in the upper atmosphere. An optimal nudging coefficient is determined by examining if reasonable structure of mesospheric gravity wave (GW) momentum forcing and residual mean meridional circulations can be obtained from balanced initial conditions. Estimated mesospheric GW momentum forcing is found to exhibit a distinctive structure with larger (smaller) values in the lower and upper mesosphere (in the midmesosphere), when compared with parameterized climatological forcing. The iterative dynamic initialization allows for dynamical balance among the model's prognostic variables and reduces excitation of spurious GWs and noises at initial time. However, theoretical imbalances, measured by the ellipticity of the nonlinear balance equation, are not completely eliminated in balanced flows, and they are found in narrow tropospheric frontal regions and over localized areas associated with the large-scale instability in the midlatitude middle atmosphere. These imbalances are discussed in the context of their potential relation to generation of planetary-scale and inertia GWs around the middle atmospheric and tropospheric jets.