Steve Allen wrote:
A better check would be to find an official timetable for a recent solar eclipse in Mozambique. The times in that table would say whether it was produced using the actually different contact times in each city or whether the contact times for one city were incorrectly shifted according to the longitude of other cities.
I doubt whether there'd be anything official in Mozambique. I did look for an unofficial time, and found this: http://www.folhademaputo.co.mz/pt/noticias/nacional/mocambique-podera-presen... An English translation is here: http://clubofmozambique.com/news/solar-eclipse-tomorrow-mozambique/ It quotes Mozambican astronomer Hélder Geraldes as saying "the sun will also be visible in the form of a ring in the city Pemba at about 11h58m tomorrow". According to NASA <https://eclipse.gsfc.nasa.gov/SEgoogle/SEgoogle2001/SE2016Sep01Agoogle.html>, at Pemba that annular solar eclipse reached its maximum at 09:19:20 UT. It's hard to square that with Geraldes's "11h58m" if tzdb is correct and Mozambique was two hours ahead of UTC. If we go instead by Pemba's local mean time, which I calculate as UTC +02:42:12, the greatest extent would have been at 12:01:32 LMT which is pretty close to Geraldes's stated time. So this is some evidence that an astrononmer in Mozambique uses local mean time. That being said, NASA says that eclipse was only partial in Pemba and you'd have to go north into Quirimbas National Park to get to the eclipse's central line, so Geraldes's prediction doesn't entirely square up with what NASA says, even if we assume Pemba observes LMT. With your contacts in the astronomical world, maybe you can track down Hélder Geraldes and see what he has to say. (Don't all astronomers know each other? :-)