Options are your book is larger or your cover is smaller than DIN A5. Also look into the ‘resolution’ of your image, or pixels per inch. When Scrivener and PDF interpret the resolution differently, you may get different results.
In an e-book you could ‘stretch’ the image using 100% width and Auto Height, but this is for PDF, which uses exact measurements.
There are no features in Scrivener that would create a cover page, like you’re thinking of. You can stick a photo into a text file, but that’s going be inside the text margins like normal figures would be.
That said, it doesn’t explain why the image is smaller than the paper size, and there are issues with how the text engine works with images. It has a very interesting relationship with reality, let’s say. Best I can tell, it confuses the concept of display resolution with print dimensions. In print of course, 420 points is precisely how wide a sheet of A5 is, no more no less, and as fixed a measurement as saying it is 148mm wide. But if you insert a 420pt wide image into the editor it comes to the mistaken conclusion that the image should be 517pts wide (~182mm)! But that’s just the start of the confusion, as even at 517pts wide, the image is somehow far smaller than even the margin area—never mind the overall paper size—which is what you’re seeing in your screenshot. You’d have to probably double the image width to 840pts (296mm in reality, 148mm in the universe of the text editor) to get it the right size.
But at that point, if you have Page View turned on in the editor, you’ll see the problem I’m referring to initially. Since if the image is the right size (or fooled around with to trick the engine into displaying it the right size, incorrectly) it would be offset by the top and left margin width so that it is cropped off the page to the right and bottom, anyway.
In short, don’t use Scrivener to make books (for one, because it can’t), that’s not what it is for! I wouldn’t myself think to put a final quality image into it, nor any text editor for that matter. I would at most use its image placement feature as a way of inserting a proxy that will later be replaced in production—i.e. a thumbnail that I can write around for my own reference, to know what I’m talking about.