We have recently had some debate about the boundaries of a "Southern Ocean" and this led me to look at the sites assigned to the Pacific Ocean connection. Some were definitely incorrect but the question raises the issue about how to handle sites which are situated on "Seas" (or Gulfs, Bays, Straights, Estuaries etc). The Western Pacific has a large number of these - particularly around the East Indies.
It appears that "strictly" all seas which are "arms" of an Ocean are regarded as part of that ocean. This definition seems to exclude "Inland seas" even if connected to the "continuous ocean" (eg the Baltic, Mediterranean or Black).
We don't have an Atlantic connection yet but such a strict interpretation would place all sites on the North Sea, in the gulf of Mexico and the Caribbean as being "Atlantic".
This seems to me to lead to counter-intuitive results both culturally and geographically. it appears for instance that the "Arafura sea" south of PNG is regarded as an arm of the Pacific (and hence Lorentz Park should be assigned to it) whilst the adjacent "Timor sea" is regarded as part of the Indian Ocean (hence Kakadu should be assigned there). Now neither sea is "Oceanic" in the sense of that word meaning large and deep. Both lie in fact in the shallow waters which, within human history, have provided a land bridge between PNG and Australia. Why on earth one should be regarded as "Pacific" and another Indian" escapes me (both are east of the Wallace line which zoologically divides Australia from Asia.)
The reason I guess is that arbitrary boundaries lead to arbitrary decisions and the geographers felt they had gone far enough east by the time they reached the Arafura Sea!!
I would suggest that we don't connect any site which is on a "Sea" to an "Ocean". This would also mean that we could have eg a "Caribbean Connection" (amazing we haven't had it already!!) rather than lose them all under "Atlantic". At a quick glance we could also have a "South China sea" connection (If my memory serves me correctly - Vigan, Hue, Hoi An
plus a few more).
A geographical connection needs to balance geographic and cultural aspects and it seems to me we don't want connections which are too broad and numerous. What is it which truly connects every site around the "greater" Pacific? Certainly not culture, zoology, climate etc. A list which is too large just becomes amorphous and unfocussed.
Regarding "Bays" and Estuaries. We already have a "Port" connection which takes account of the impact of a site's connection to the sea/ocean on its culture/history. The existing "rule" for the ocean connection is "within 10kms" but what if it is up a bay or estuary?? I personally find this 10kms rather arbitrary. We should take into account whether the site has the "feel" of the sea/ocean?? I would suggest that sites located some way from the sea/ocean eg Melbourne are NOT assigned to the nearest Ocean or Sea and that the bay/estuary is counted as part of the 10kms if we maintain that!
What about Edinburgh if we had a North Sea connection - well the inscribed area isn't on the sea and the "sea" is an estuary ("Firth of Forth") - yet Edinburgh wouldn't be on the "Port" connection and its geographic position next to the sea has been important so - possibly!