1999ApJ...514..869G

The Space Infrared Telescope Facility (SIRTF) will drift away from the Earth at ∼0.1 AU.yr^–1. Microlensing events will therefore have different characteristics as seen from the satellite and the Earth. From the difference it is possible in principle to measure v{d5}, the transverse velocity of the lens projected onto the observer plane. Since v{d5} has very different values for different populations (disk, halo, Large Magellanic Cloud), such measurements could help identify the location, and hence the nature, of the lenses. I show that the method previously developed by Gould for measuring such satellite parallaxes fails completely in the case of SIRTF: it is overwhelmed by degeneracies that arise from fact that the Earth and satellite observations are in different bandpasses. I develop a new method that allows for observations in different bandpasses and yet removes all degeneracies. The method combines a purely ground-based measurement of the ``parallax asymmetry'' with a measurement of the delay between the time the event peaks at the Earth and satellite. In effect, the parallax asymmetry determines the component of v{d5} in the Earth-Sun direction, while the delay time measures the component of v{d5} in the direction of the Earth's orbit.