Where your antenna hears
Two days ago the first AIS station outside our own roof came online. It sits at the Port of New Orleans and it has been delivering about half a million messages a day, 277 different hulls in its first two days. That water was dark for us until then. Our upstream provider has no receiver on the Lower Mississippi, and we had measured that: the daily city sweep probed New Orleans twice and found nothing both times, while Houston in the same request came back with 175 vessels.
One volunteer with an antenna fixed it in a day. But the only thing our pages could say about it was a count of hulls and messages, because the station's own numbers had no geography in them at all. The positions arrived, went into the live vessel on the map, and were thrown away.
They are kept now, and the station page has a Coverage tab that draws them.
What the map shows
The grid is squares of a tenth of a degree, roughly eleven kilometres on a side, shaded by how many messages arrived from inside each one. Darker is a fairway heard solidly. Faint is the edge of sensitivity, where a hull gets through once in a while and then doesn't.
It is measured, not predicted. There is no range circle and no calculation from your aerial height, because VHF over water does not work in circles. It runs up the river and stops at the levee; it reaches forty kilometres in one direction and eight in another. A square nothing has been heard in is left dark rather than shaded in from its neighbours, which is the same rule the reception rose has always kept.
It is not where your antenna is. The squares are where the ships were. That distinction matters beyond politeness: it is why the map works for the several receivers whose owners never entered coordinates, and for anyone feeding raw NMEA over UDP without registering at all. The reception rose needs to know where the antenna stands. The grid does not, so nobody is left off it for having skipped a form field.
There is a network coverage map too, with every ships receiver merged into one grid. It sits behind a link rather than on the AIS landing page, the same way the aircraft side keeps tar1090 one click away from the feeder landing. That view deliberately never says which antenna heard which square. Naming a receiver per cell would narrow its location to a few kilometres, and the feeder page promises that never leaves your own station's page.
What else moved
The reception rose now works on the ships band. It was written for aircraft and the plumbing was band-neutral already, so once positions started being kept it filled in on its own; the captions just needed to say "vessel" where they said "aircraft". If your AIS station has no coordinates on file, the tab now tells you that in one line, along with where to add them, instead of showing an empty plot.
Both instruments read the last 30 days and keep the best each has ever measured. One quiet night should not erase a horizon that took weeks to build.
If you run an AIS receiver, the connect line is on the AIS network page, and it is still one line of config.