About / method

A search, with the evidence attached.

We look for distant Solar System objects: trans-Neptunian objects and possible dwarf planets. This notebook records recoveries, sensitivity tests, rejected links and unfinished work alongside any future discoveries.

Rubin alerts, through Lasair

The main search uses Rubin Observatory / LSST alerts supplied by the Lasair-LSST broker. A moving body can appear as a different diaObject each night. We search single-night objects, link them across nights, and fit their apparent motion. Credit for the observations belongs to Rubin Observatory / LSST; credit for broker access belongs to Lasair. Image grids contain Rubin cutouts.

1. Orbit-space hill climbing. Seed cross-night pairs using parallax geometry, refine trial orbits, and apply fit-quality, photometric-consistency and bound-orbit gates. A tight short-arc fit alone does not establish a real body.
2. Injection tests. Plant synthetic objects in each field's real cadence and clutter. A recovery of x/4 describes those four plants at their stated brightnesses, not a general completeness measurement.
3. Non-detection test. Extrapolate onto other covered nights and check whether a source expected above the detection limit is absent. Missing or unfinished checks are not passes.
4. Known-object matching. Cross-check the MPC distant-object catalogue and JPL small-body identification. Poorly determined catalogue orbits can be far from their predictions, as 2025 OO737 illustrates.
5. Image inspection and MPC acceptance. Inspect the images, verify the linkage and identification checks, and obtain MPC acceptance before adding a genuinely new object to Discoveries.

Archival DECam

The human-reviewed write-up in docs/SWEEPS.md records 14 blind archival searches in NOIRLab NSC DR2 / DECam data: Ophiuchus, nine campaign fields and four Scorpius sub-fields, covering about 29 square degrees of unique sky in 2013–2019. No credible new object was found. The Ophiuchus weak links and the cell4710 four-night link were rejected as chance alignments; the latter failed checks on motion, colours and covered-night non-detections.

The write-up reports 33/40 planted objects recovered in Ophiuchus plus campaign 1 and 16/16 in Scorpius. The Scorpius sub-fields reused the same planted orbits, so these are not independent tests. Later reruns differ: each campaign page preserves its own measurements. Historical reports may retain links rejected in the write-up.

Limits of the evidence

Injection and fitting share approximate two-body dynamics and Earth ephemerides. These tests probe linkage in real clutter and cadence, but do not test every physical or astrometric systematic. Real known-object recoveries provide a separate check. Short arcs can leave semi-major axis and eccentricity very uncertain; diameter estimates depend on assumed albedo. Reported rms is radial, not per-coordinate.

Catalogue pre-clustering can hide moving objects near stars or stationary points. Sensitivity depends on brightness, distance, cadence and footprint. A null result applies only to the searched data and tested sensitivity. Planned, downloaded, skipped and incomplete fields do not establish a null result.

This is a static snapshot of files that may still be growing. Logs and report JSONs are separate records and can describe different reruns. Curated findings take precedence for reviewed identifications; nothing is automatically promoted to a discovery.

Source code: EuanHoll/distant-object-survey (private repository). Method and archival summary adapted from docs/SWEEPS.md.