Ross, B. B., and I. Orlanski, 1982: The evolution of an observed cold front. Part I: Numerical simulation. Journal of the Atmospheric Sciences, 39 (3), 296-327.

Abstract: The 48 h evolution of an observed cold front is simulated by a three-dimensional mesoscale-numerical model for a typical springtime synoptic situation over the southeastern United States. The model used in this study employs anelastic equations of motion on a limited-area domain with locally determined inflow/outflow side boundaries.
Both the observed and simulated characteristics of the weather system indicate a mature front which intensifies and then weakens over the 48 h period. Moist convection occurs in the form of intermittent squall lines in the observed case; in the numerical simulation, convection develops above and somewhat ahead of the surface front after 24 h as an ensemble of convective cells.
An investigation is made of the mesoscale and subsynoptic-scale features of this solution to determine their sensitivity to the inclusion of moisture and to the magnitude of the eddy viscosity used in the numerical simulation. The primary effect of increased eddy viscosity is to reduce somewhat the propagation speed of the front. The major changes due to moisture inclusion occur when convection develops along the cold front; these convective effects, which are apparent in the subsynoptic as well as the mesoscale features of the solution, include increased low-level convergence, reduced surface pressure due to diabatic heating, and the deflection of winds due to upper-level divergence. In addition, small temperature changes occur in the middle troposphere between the jet stream and the surface front when either viscosity or moisture is varied; these disturbances are a clear manifestation of the effect which changes in the cross-stream circulation intensity have upon the frontal system.
A fundamental feature of the mesoscale structure of the front in all cases is the tendency of the line of maximum horizontal convergence at the surface to move ahead of the line of maximum vertical vorticity. This phase shift appears to be related to the propagation characteristics of the frontal system. Also, the mesoscale moist convection develops a cellular structure throughout the convective zone in the low-viscosity solution; the use of higher viscosity tends to suppress these cells, particularly near the surface.