Most laws of physics have no absolute frame of reference; the oddly misnamed "relativity theory" was based on one quantity (i.e. the speed of light) unexpectedly being absolute. That is to say.. saying something like "constant motion" isn't merely inconvenient, it's meaningless.
The meaningful quantity is "constant relative motion", though in that case you have to ask, relative to
what. Apparently not hell.
The portals are actually under constant acceleration, though, which
is an absolute. If they weren't, they'd go flying off into space as the earth curves around the sun, never mind the rotation of the earth. I don't know how they're anchored; it couldn't be something as simple as sticking to a constant gravitational potential, since that'd allow them to move (approximately) laterally - they don't - and anyhow the earth's gravitational field is not actually a constant.
So, what's anchoring them..?
Well, let's see. First off, they are probably obeying gravity. That's a given with our laws of physics; you don't have a choice. Which is not to say that it
has to be the case, but if it isn't you have to throw out.. some bits of relativity that are pretty much known to be wrong. Heh.
They aren't on a free-fall curve, though. They stick to where they were opened. Obvious experiment: Try opening a portal in a moving vehicle.
At a wild guess, then.. some theories of gravity have space being more than four-dimensional, with the extra dimensions small (wrapping around), but nevertheless there. This makes gravity fall off faster than inverse squared for whatever the size of those extra dimensions are, and then revert to the usual inverse square law.
Now, those extra dimensions are also a good place to put additional planes, a plane being a four-dimensional universe such as ours.. or hell.
The portals would be higher-dimensional structures allowing translation along those extra axises.
This configuration could allow gravity to act much more strongly on the portals than normal matter, the portals having extent in extra dimensions. However, only
near-field gravity would be affected in this way, eg. from very close matter. The net effect could be to hook the portal to whatever matter is very close by (on the order of milli-to-micrometers). Unfortunately, that's mostly air, which moves quite a lot.
But wait! Such a one-to-one mapping between our world and hell would not allow for one other observed quantity of portals, namely that you can open a portal from any place in hell to any place in our world, or vice versa. To fix that, I hypothesize that the mapping is not one-to-one, but instead that both dimensions (or just one, doesn't really matter.. hm, hell would fit nicely) is crumpled up like an old wad of paper. I should probably mention fractals at this point; those are a good way to get a fitting structure with very small physical laws, and small laws = higher prior probability, thus more likely to be true.
So, in this case, the part of the portal entirely in our world would be affected only by nearby air; however, going very slightly along the higher-dimensional axis it's moving in would spread its gravitational shadow out to cover nearby matter such as, oh, the ground. Depending on the exact parameters here, that could work out to be just the top layer of soil, or kilometers of rock; it can't be very much more, because then the movement of the magma would start to matter.
This produces another interesting experiment. Are portals on water just as stationary?
Also, a portal opened in space would behave
very differently. It'd probably follow your spaceship around, but not necessarily in a convenient manner; the edges could get slice-y. You'd have to find a way to contain it for them to be very useful in-flight, but you could always just turn off the engines for a while to refuel; this does mean that you'd have to carry a sensitive around, unless you can completely mechanize it - I see no problem with this, however.
All in all, plenty of room for tests.
)
