We test a new scheme to study the magnetosheath. The scheme uses the solar wind measurements as the input into the gasdynamic convected field model, and the model output is compared with magnetosheath observations. In our four test cases there is a significant overall success in the model prediction. This scheme works better than other methods in magnetosheath studies and is potentially useful for space weather forecasts and nowcasts. The direction of the magnetic field is modeled most accurately. The prediction of the size of the magnetosphere is accurate within a few percent. The predicted thickness of the magnetosheath is accurate up to 90%. With a double-normalization procedure developed in this study, we are able to separate the processes intrinsic in the magnetosheath from those due to large-scale upstream temporal variations. The test cases confirm the existence of a compressional front one third of the distance from the magnetopause to the bow shock near the stagnation streamline. The magnetosheath density profile near the stagnation streamline is consistent with the models that add a compressional front between the two depletion processes described by the plasma depletion model. A major unexpected feature is that the magnetosheath flow pattern is very different from that described by the model and maybe by most other models, including MHD models. The magnetosheath flow near the stagnation streamline does not slow down gradually toward the stagnation point. It moves rapidly until reaching a very small region near the magnetopause. ¿ 1999 American Geophysical Union