In theory, you could use the same regridding approach you plan to use for gridded emissions to regrid these other files.
However, just as it would be better to regenerate gridded emissions from the underlying inventories and gridding surrogates for your nested domain by running SMOKE so as to better capture the spatial texture of features like roadways, population distribution, etc., it would also be better to do so for the files you listed. For BEIS_NORM_EMIS, this would mean running SMOKE’s normbeis program for your intended domain using higher-resolution BELD input data which should be available, though you’d probably have to ask the SMOKE community for help. To generate a higher resolution OCEAN file, follow this tutorial. To generate higher-resolution input files for the NH3 bi-directional option from scratch, you’d have to run FEST-C.
In the end, you’ll have to think through what you want to get out of the higher resolution nested simulations. It seems you initially considered interpolating meteorology, and here you mention interpolating emissions. What are the features that you did not consider represented at a fine enough resolution in your current domain that make you consider running a higher resolution nested domain? These are the features you will want to prepare “from scratch” for your finer resolution domain, rather than just regridding from your coarser domain.