It's a case of a misalignment of the pirtiin of the external mesh tied to the cockpit and used to fill in gaps in the external model that would be so annoyingly present without it.
It's possible more portions of the external mesh are included than are necessary, such as farther struts, which make a misalignment yet more apparent.
The first task is to identify this mesh in the cockpit's hier.him, then fiddle with the X, Y and Z coordinates (the last three numbers in the 12-elememt transformation matrix.) Anyone can do this with Notepad if the extracted hier.him is to hand. If the cockpit version of the external model is a direct copy, then an iterative adjustment will obtain a beautiful overlap.
There's no fixed standard across the board that sets the orientation of the axes for a cockpit mesh orientation. This can be determined by altering one of the aforementioned three numbers by, say, a meter. Note the direction the mesh moves after either a positive or negative change is applied. You now know which axis and the direction of same applies to that value. Do the same for the other two. Now you're set to iteratively bring the internal version of the mesh into congruence with the outside one.
I'm a WAW boy, otherwise I'd do this myself.