Blending, Source-Flux Fraction and the Systematic Floor in Microlensing Parameter Estimation

This paper computes how unrelated light blended with the source, common in crowded bulge fields, biases the impact parameter, timescale and peak magnification recovered from a light curve. For a representative lens of 0.4 solar masses at 5.0 kiloparsecs, in front of a source at 8.0 kiloparsecs, the angular Einstein radius is 0.494 milliarcseconds, the event timescale is 36.1 days at a relative proper motion of 5.0 milliarcseconds per year, the microlensing parallax is 0.152, and the physical Einstein radius projected at the lens is 2.47 astronomical units. The central finding is that unmodelled blending at the thirty per cent level biases the recovered timescale by a comparable fraction, a systematic larger than the statistical error of a well-sampled event. All quantities are computed from stated assumptions and published survey values used as comparison points; the study is simulated and computed, not measured. Where the computation shows that an inference commonly drawn from such events is not supported at the stated precision, that negative result is reported in full.