Automatic identification system (AIS) on aprs.fi - Long Rx -> Tx distances

Question

I've been looking into some of the info available on aprs.fi, which also makes available AIS data, but most of the reports I see have such large Rx->Tx distances that I'm not clear on how things are actually working.

There is a document available here that details AIS data feeding into the network:

To receive AIS transmissions you need a 9600 bit/s data capable VHF receiver which can receive the marine VHF channel 87B (161.975 MHz) or 88B (162.025 MHz)

Once this transmission is received, how is it related to the wider network, in particular on APRS.fi?

Here is a concrete example that caused me some confusion. There are a little more than a hundred feeding stations listed here: https://aprs.fi/page/ais_sites but I found many boats from around the world are being listed as from https://aprs.fi/info/i/GM7HHB, outside of Edinburgh in Scotland in the IO85HV locator. A large number of boats on the Great Lakes have the GM7HHB station entry point. The city of Sault Ste Marie is a great example, and currently has a boat (Callsign: WYR4481, MMSI number: 366904940, IMO number: 729057) making this transmission:

Vessel	Last Heard	Longest	Rx -> Tx
PAUL R TREGURTHA	2021-11-13 09:53:58	EN36WR > IO85HV	3610.3 miles 300°

Well, all I'm comfortable saying at this point is $3600 \text{ miles} = 4.12*\sqrt{h}$ is a large antenna. Is there some i-gate procedure for AIS data, as well?

_{Note: The maritime mobile band is not an amateur band itself. I hope this question is not inappropriate here. I'm looking into this data as an amateur radio enthusiast and I do think this community is the most equipped to help.}

Answer 1

There's large AIS observer networks (think MaritimeTraffic.com and consorts), and chances are that Edinburgh station just feeds data from the internet into the system.
Furthermore, theres official and commercial satellite services for maritime observation, so that the AIS signal doesn't have to be captured by a coastal station, but gets just picked up in space, with a huge field of view.

Then: many AIS transponders are actually not transmitting their true positions (this is, in many but not all, cases a violation of maritime treaties, these things need to be functioning and well-maintained on all larger vessels, on some kinds of smaller ferries etc).

Lastly, AIS really is ancient, in technical terms, which means it's been specified with receivers in mind that can be beaten the hell out of by modern systems. GMSK at 9.6 kbd isn't really hard to detect and demodulate at all, and 12.5 W (class A devices, i.e. large ships) means it can be received very far, even through weak atmospheric bending paths¹.

(The only thing really annoying about receiving AIS is the lack of proper forward error correction. Had this system been designed state-of-the-art when it was conceived, that would've been part of the system, and would have greatly improved reliability/reach, or reduced required transmit power.)

---

¹ napkin calculation of free-space path loss: 160 MHz, that's a $\lambda=1.9\,\text{m}$, so over 100 km that's not even 120 dB FSPL.

The noise in a roughly 20 kHz equivalent noise bandwidth of that GMSK channel is -174 dBm/Hz + 43 dBHz = -131 dBm.

So if the transmitter has 12.5 W TX power, that's 100/2/2/2 W = (20-3-3-3) dBW = (20-9+30) dBm = 41 dBm. Therefore, assumin isotropic antennas at both ends, the 100 km distant receiver sees 41-120 dBm = -79 dBm.

The factor between noise power and RX power is then (again, no gain at the RX antenna!) 52 dB. For solid GMSK reception you need 6 to 9 dB, usually. So, things can get 42 dB worse and you'd still be able to receive successfully. Since every doubling of distance means a quartering of receive power (== -6 dB), the signal should therefore be decodable after 1400 km.

Now, as you can see, this is all a napkin-quality calculation, but it tells you why it's no big deal for a spaceborne receiver to cover huge swaths of ground/sea.