Best Techniques for Protecting Faint Nebulosity in Signicant Light Pollution

Well, that means the data by itself is already very tricky to begin with. So to correct that with software is indeed trickier then as well. What I normally do is to take my integration, stretch it to the max preview setting (30%) and oversaturate the image as well. When that blows out the entire image, I just set 5 boxes around the edges and see what happens after that, if that reveals an image that is a bit more clear to correct, you can clear the boxes and place them on the right places around the image in the preview and then try again. It will become an iterative workflow.

Vincent's suggestion is a good one. I do something similar, kind of, but use a tickling approach especially if I cannot find areas of obvious nebulosity, or when the background and nebulosity are so close that large stretching doesn't work. You can choose a good region, run lpc tool and then keep adding boxes. That sometimes works for me because unlike DBE in puxinsight or gradient terminator in PS (and siril and others of course,) you don't need to undo completely the process and search for a new background.

Alternatively, you can add all boxes you thi k are correct, run lpc tool, then if some regions look wrong you can remove boxes and run again, reverse iterations I call them. With a low tolerance slider, it is likely you will always have some red boxes to remove when signal is detected.

I always check online to see where background sky vs signal is located and try to find those regions in my images as a good starting point, then refine

Checking online for where to expect nebulosity is a very good one as well indeed.