It’s not entirely clear what data you are referring to here.
I don’t think that you can, with any assurance, ‘splice’ Argo data onto earlier bucket-and-intake data from a much smaller ocean area. If that is what you are pointing to then I would say it would be an intellectually wrong methodology with meaningless outcome. Indeed it would be somewhat reminiscent of what has come to be called ‘hiding the decline’.
If you are saying that what you have in Argo is sufficient, then I would be most confident of the data since 2003. There have been, as you will be aware a lot of debates about the earlier Argo data.