A few years ago some people at Motorola hatched the idea of a cell telephone network in the sky involving some 77 satellites in low earth orbit. They called it Iridium after the element which has atomic number 77. Well, now they have revised things to use 66 satellites (plus 6 on-orbit spares) but they still call it Iridium. Lots of very bright flares have been seen from these satellites, and they are fun to observe.
The sun has visual magnitude about -27, and the full moon about -13. Venus right now is magnitude -4.4, Jupiter -2.3, and Saturn +0.4, according to Sky and Telescope. The brightest star is Sirius at about -1.5. Iridiums have been observed (and predicted) to flare to visual magnitude as bright as -9. You can even see them flare in the daytime if you know exactly where to look!
If a satellite measuring a meter or more in some dimension had a mirrored surface and pointed right at you, it would be like a kid with a mirror in the sun pointing the spot in your eye--dazzling. We don't usually see this in satellites because the only big flat thing on many satellites is the solar panel, and this tends to point toward the sun, which doesn't often put it in the right position to flash at us. (Geostationary satellites are the exception to this, and can be seen flashing at some times of year.) Dead satellites can happen to flash like this, but this is unpredictable.
I recently saw a video of the Iridium assembly line; the satellites have incredibly shiny mirrored flat antenna panels with small antenna elements spaced out on them. These point down toward earth, and are responsible for the flashes. (You can find pictures of these satellites on the net.) For more information on the Iridium flashes, and software to predict them under DOS/Windows, check out the VSOHP Iridium page.
So what does it mean for my predictions? Well, I don't yet have an accurate prediction program running under Linux; the iridflar program available from the VSOHP runs under DOS or Windows. I have had some success in getting it to run under the Linux DOS emulator, but it is presently insufficiently automated for me to use, and it seemed to bomb out at the end. For now, simple logic suggested to me that high passes (but not too high) a little to the east might give flashes from the right side antenna panel. Indeed, the kids and I have seen bright flashes which I estimated to be at magnitude -3 to -4 using this method. There has been one post to the seesat mailing list to this effect also. Accordingly, I have added an Iridium pass filter to my prediction filter previously based only on magnitude. An Iridium which seems to be a candidate for flashing due to the geometry is passed through even though its predicted magnitude might be large (dim).
There are some Iridiums which are dead, as discussed in the October 24, 1997 posting to the seesat mailing list by Paul D. Maley, who is at the NASA Johnson Space Center. He says that Iridiums 11, 21, 27, and 36 are not in the operational orbits and so probably will not flare predictably. Those are being deleted from my predictions. Note the apogee and perigee of the predictions; if this is around 770 then the satellite has reached operational orbit. If they are around 500-600, then it has not. It could be a new launch, or it could have a problem. (As of this writing, 15 Nov 1997, we have close to 40 of these things in orbit, and we should be up to 72 within 6-8 months, if all goes according to plan). There are also a couple of simulated Iridium payloads which were launched to test the launcher. These have no deployed antennas and are also deleted.