T16 TESS Year 1 catalog model depths are smaller than TOI depths (median ratio 1.21, 904 signals) and published planet depths (1.29, 168 planets), and the ratio increases with transit duration

The T16 catalog (Roth et al. 2026) gives 11,143 public transit candidates from TESS Year 1 full-frame images. The T16 light curves go through the Trend Filtering Algorithm (TFA). The T16 paper states that TFA can make transits shallower, by about 10% in radius, but it does not measure this effect for the catalog. We measured it. For each T16 row, we calculated the mid-transit depth of the catalog model (Rp/R*, b, q1, q2) with quadratic limb darkening. We matched the rows to the ExoFOP TOI table (same TIC, period within 1%) and to the NASA Exoplanet Archive composite table (same TIC and period, with published Rp/R* and b). We did not use grazing models. For 904 TOI matches, the median ratio of TOI depth to T16 depth is 1.209 (1-sigma range 1.193 to 1.220), and 85% of the ratios are more than 1. For 168 published planets, the median ratio is 1.288 (1.270 to 1.329). Both depths use the T16 limb darkening. The ratio increases with transit duration: for TOI matches, from 0.96 below 1.5 h to 1.38 above 6 h; for published planets, from 1.01 to 1.68. The ratio decreases with TESS magnitude, from 1.35 for Tmag < 11 to 1.05 for Tmag 14 to 17. The ratio is more than 1 for all TOI dispositions, also for false positives. Thus the cause is the light-curve processing, not the type of source. A linear fit of ln(ratio) to the TOI depths gives slopes of 0.204 +/- 0.017 for ln(duration), -0.042 +/- 0.005 per magnitude and -0.074 +/- 0.014 for ln(S/PN). The period term is not significant. The fit uses no published-planet data. When we apply it to the published planets, the median ratio falls from 1.288 to 1.053. We supply a depth scale and a scaled Rp/R* for all 11,143 rows. The median depth scale is 1.157 (radius scale 1.076). 2,686 rows are outside the fit range and use clipped inputs. Limits: the TOI and planet samples contain only signals that other pipelines found, and these are brighter and have higher S/PN than T16-only candidates; the S/PN term can come from this selection; TOI depths are also model results, and QLP depths of long transits can also be too small. Users of T16 depths, radii or injection tests must correct for this loss, mostly for transits longer than 3 h.

T16 · TESS · transit depth · Trend Filtering Algorithm · TFA · planet candidates · TESS Objects of Interest · ExoFOP · NASA Exoplanet Archive · planet radius · systematic bias · light-curve detrending · full-frame images · catalog calibration · limb darkening · injection-recovery