A finite source model fits public KMT-2025-BLG-3184 data better than a point source model in three site removal tests at Γ = 0.4
We test the rounded peak of the known microlensing event KMT-2025-BLG-3184 with six public KMTNet I-band pySIS streams. For 538 selected measurements and a fixed limb coefficient Γ = 0.4, a finite source model gives χ² = 890.96 for 522 degrees of freedom. A point source gives χ² = 1898.52 for 523 degrees of freedom. The source radius crossing time is 21.22 hours. Removing data from each site in turn gives 20.99–21.69 hours and reduces χ² by 239.38–648.74 relative to a point source. Shapes fitted on two sites also predict the third site better. The conditional Δχ² ≤ 1 profile is 21.04–21.40 hours; it is not a full uncertainty estimate. The source amplitudes exceed the baseline flux inferred from the supplied magnitudes by about 12–27 percent. Requiring nonnegative blend light gives 20.35 hours. Fixed limb laws with 0 ≤ Γ ≤ 1 give 20.41–22.47 hours. The full tested range is a model sensitivity range, not a confidence interval. Residual structure remains. This result supports a finite source shape under the tested models. It does not give a lens mass or establish a planet. The package preserves the original numerical data, provenance, search records, source code, fits, figures, and reproduction instructions.
microlensing · KMT-2025-BLG-3184 · MOA-2025-BLG-352 · finite source · source radius crossing time · KMTNet · blend light · reproducible astronomy
References
4- usesdoi:10.5303/JKAS.2016.49.1.37
We use public numerical observations from the KMTNet telescope network and include its required data credit.
- usesPublic alert data for KMT-2025-BLG-2093 give a source radius crossing time of 4.12 hours for a fixed limb law sci:sha256:8da8ce5aa295029fb36743a34ee6cee37a999fb54092f3ee3447dd1a4e92a21c
Its study of another event informed our site removal, robust baseline error scaling, fixed limb-law tests, and numerical controls. We extend the test design to KMT-2025-BLG-3184 with held-site predictions and blend-light constraints. We do not reuse its code or observations.
- doi:10.1088/0004-637X/695/1/200
Prior work on finite source integration provides the method context for our implementation and numerical checks.
- usesdoi:10.1051/0004-6361/202452648
We use VBMicrolensing 5.6.1 as a separate magnification check at all reference observation times.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesRemoving Chile data reduces the finite source fit gain from 368 to 3 in public KMT-2025-BLG-1134 data
We extend its catalog sample to all 28 events and reuse its MIT disk integration code. Its event 3184 result is prior work and is reproduced by our screen.
- usesSingle lens least squares fits at Γ = 0.4 give source radius crossing times of 21.12 and 23.07 hours in two reductions of 256 shared KMT-2025-BLG-3184 I-band exposures
We reuse its six public pySIS I-band files and its MIT finite source integration modules. We reproduce its six-stream reference fit, then add DIAPL and V data, a shared-exposure audit, and reduction sensitivity tests. This is a targeted extension of the known event study.
- usesCurves with one rise and one fall leave large residuals in three public KMT-2025-BLG-1760 streams with baseline error scales
We reuse its MIT finite-source integration code for exploratory lens fits and adapt its baseline and site control design. The final shape bound uses a separate exact regression calculation.