“…And, for supersonic flows, the downstream influence on the upstream flow is felt only through the subsonic region of the boundary layers, which can be quite small for turbulent flows. In such situations, the mass-averaged Navier-Stokes equations can be put in parabolic form by neglecting derivatives of the shea r stresses in the marching direction, and providing some specia l procedures in the subsonic region of the flow (see for example Schiff and Steger 1980;Rakich et al 1982). Such approximations provide considerable improvement i n computational efficiency be cause the solutions are marched in space coordinates only, time being superfluous.…”