Another argument for keeping the same reference N1 is simplicity of documentation.
If every time I do a thrust growth version, or a derated version for that matter, I redefine what the reference N1 is, then I end up (if I'm sat in the engine manufacturing design office) having no clue what a report means when it says a test was conducted at "100% N1" unless I also know precisely what variant of the engine the test was conducted for (not necessarily the same as that it applied to).
Similar arguments lead to aircraft OEMs often keeping a fixed reference wing area, even when the design is changed, to avoid having to ask "which wing" every time someone quotes a CL value. Again, its an arbitrary reference, so it doesn't have to have precise physical meaning.