More importantly, given a two dimensional border, "punching through" isn't even necessary since the putative fleet could just as easily go around them (at a larger scale, that is).
But what I do is outline the heretofore undiscussed reasons why the map does not need to suggest two-dimensionality - it just has Outposts from 1 to 7 (and perhaps 8) lined up close to Romulus, because of special bias. The, say, fifty thousand other installations (many of them no doubt automated) would cover the other dimensions; since this stretch of the border is so very special, it has the big and important crewed Outposts clustered right there, with their low numbers.
Indeed, the scattershot attack of the Romulans "out of numerical sequence" might suggest three-dimensionality right there and then, with OP8 actually closer to OP4 than to OP7, say.
The Enterprise's assumption -- and it is a good one -- is that the Romulans are testing the border to see what the Federation will actually do, and if the Enterprise doesn't engage the bird of prey NOW, it will embolden the Romulans to escalate further attacks.
...Which is about as stupid as the Xindi launching a warning attack before their main one.
If this cloakship (and perhaps a couple of sister ships) were engaged in breaching the Federation siege where it's the most convenient, at the Earth-patrolled sector close to Romulus, then it would seem prudent for them to immediately signal the Praetor about the success, even at the risk of exposing themselves. But "immediate" is relative - it would take at least days for another starship to arrive, apparently.
What may instead have held back the "immediate" Romulan attack may be that the attack against the siege was in fact a dismal failure. Neutralizing just four outposts might not suffice: if the ship we saw got two of them, and two other ships each got one, but the other nine ships involved in the attack all failed in their missions, then there could be no war.
That's classic "special forces" or "minisub" warfare: the attackers themselves will not know whether their mission was a success or a failure until their (unlikely) return to home base, regardless of whether they themselves managed to fire their torps or set their charges.
The time of year is just incidental. If the Romulans had tried this stunt two months later, they would have hit outposts 12, 13, 14 and 19.
But statistically that would mean there are only a few dozen of them. By making this region special, there can be millions. And the region is special in the "large" model only.
Maybe I missed something, but how is this obvious?
Because it's the stretch where Romulus lies. If we assume the curving RNZ is a segment of a ring, that is, a local representation of a spherical shell imprisoning the Romulans, then obviously Romulus is very distant from the center and basically brushing up against this specific wall segment. If we assume the RNZ is the shape of a salamander or a swastika for some reason, the map still shows Romulus at least equidistant to the visible RNZ and the putative invisible bit "just to the right of the map" - and it would be difficult to justify why the map would omit this putatively proximal "right side RNZ" and instead show so much uninteresting empty space to the left of the chain of outposts.
The only reason that this portion of the RNZ is shown is that it's the section that's under attack at that particular moment.
But take any other portion and it would lack Romulus! Surely that makes this one very special indeed?
Timo Saloniemi