Band contrasts differ by more than three formal errors in 147 of 355 Gaia DR3 and SMASS II asteroid pairs when covariance is set to zero
We test a fixed, signed spectral contrast near 0.7 micrometres in public Gaia DR3 and SMASS II data. The sample has 355 numbered asteroids with Bus classes B, C, Cb, Cg, Ch, or Cgh. The Gaia score ranks Ch/Cgh against the other classes with area under the receiver operating characteristic curve 0.830. The 95% object bootstrap interval is 0.786 to 0.871. Gaia minus ground contrast has median +1.513 percentage points. Its 95% interval is +1.299 to +1.751. Differences exceed three combined formal errors in 147 of 355 pairs, or 41.4%. This calculation adds the reported marginal variances with covariance set to zero. After subtraction of the sample median difference, 104 pairs still exceed three errors. A separate fixed Gaussian score gives similar class ranking. A threshold set from development controls recovers 24 of 67 test Ch/Cgh objects and selects 7 of 116 test controls. The formal errors alone do not describe the survey differences. The result does not establish their cause, new mineral detections, or surface change. The selected sample does not give a main-belt hydration fraction. We preserve code, aggregate results, figures, the analysis plan, exact source requests, response checksums, and commands to rebuild the test from public downloads.
asteroids · Gaia DR3 · SMASS II · reflectance spectra · spectral classification · hydration · measurement errors · reproducible benchmark
References
10- usesdoi:10.5270/esa-qa4lep3
We use Gaia DR3 mean asteroid reflectance spectra, marginal errors, and quality flags.
- usesdoi:10.26033/fj1d-vb37
We use the PDS SMASS II V1.0 ground spectra and their reported errors.
- usesdoi:10.26033/e1p3-xm59
We use the Bus class field in the PDS Asteroid Taxonomy V6.0 table. Provenance records specify the PDS4 bundle bytes.
- doi:10.1051/0004-6361/202243791
This work describes Gaia DR3 asteroid spectra, quality flags, and prior comparisons with ground spectra.
- doi:10.1051/0004-6361/202348752
This prior work measures the 0.7 micrometre band in Gaia spectra. Our fixed signed scores test a different procedure.
- doi:10.1051/0004-6361/202245134
This prior work corrects blue Gaia samples. Our scores use wavelengths at 550 nm and above.
- doi:10.1051/0004-6361/202556127
This prior work classifies asteroids with Gaia spectra and other data. Its experiment differs from our fixed-score benchmark.
- doi:10.1051/0004-6361/202245311
This prior Gaia and SDSS comparison reports survey color differences, including at high signal to noise.
- doi:10.1006/icar.2002.6857
This paper describes the SMASS II observations and measurement limits.
- doi:10.1006/icar.2002.6856
This paper defines the Bus classes used for the target and control groups.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- uses95% bounds require unequal contrasts in 14 of 355 Gaia DR3 and SMASS II asteroid pairs under Gaussian errors with any dependence
We use its fixed contrast, source requests, and sample rules. We reproduce its 147-of-355 independent-error count. We add exact simultaneous bounds for any error dependence and identify 14 separate pairs.
- uses95% bounds exclude a common four-band correction for 355 Gaia and SMASS asteroid pairs under the stated Gaussian model
We reuse its fixed contrast, sample rules, and exact source request record with attribution. The new analysis tests all common four-band corrections.