NavList:

Greg, you wrote:

I have a hunch Frank has an easier way to do this though.

I liked your idea of comparing angular sizes. That would certainly give you an approximate angular height which might be nice to have in advance. But the general technique for a satellite fix depends on the fact that we can calculate the exact position of the satellite on the celestial sphere for any given latitude and longitude on the Earth. This is very similar to traditional celestial navigation: we calculate or simulate the altitudes of celestial objects from an assumed position and then compare the simulation with our observations. For a known instant of UT (which we have to assume is correct here), we can calculate the position of the Sun and the ISS "among the stars". These positions would be RA (or SHA) and Dec. Then you just fiddle with the observer position until the ISS ends up superimposed on the Sun exactly as shown in the photo. With the right software or web site app, this can be worked out very quickly --mere minutes of work. There's a catch though: many of the popular tools available will filter out daytime passes since you can't see satellites in daylight! One exception is "calsky" which even has a specific tool for calculating ISS solar transits. So enter an observing location in the Netherlands for the date and UT and narrow down until you get a match for the photo. Easy! For navigational use, we need a tool that does this sort of work without a continuous Internet connection (maybe downloading orbital elements daily or weekly) and that's also purpose-built for offsetting from an "assumed" position. I'm working on it... 

As I noted previously, for this composite photo we should really be getting an error ellipse of some sort since we don't know if the UT as given applies to the central image or one of the others. I would run the process for the two ISS images closest to the edge of the Sun's disk (at opposite ends of the pass) and then draw the long axis of the error ellipse so that it extends somewhat beyond the two resulting points on the ground. The width of the error ellipse may be quite narrow, possibly as little as 100 feet if we're careful. When we have our error ellipse, I'll email Marco Langbroek, and we'll find out how close we are.

-FER