We can add lossless compression to the encoding dict so that the netCDF files are smaller. I think this makes sense to enable across the board...?