REVIEW 3 major objections 4 minor
WFC3/IR keeps losing sensitivity at about 0.06–0.13% per year, but more slowly than earlier studies reported.
Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →
T0 review · grok-4.5
2026-07-15 02:24 UTC pith:XJ5B3RE2
load-bearing objection Useful incremental WFC3/IR sensitivity update with clear rates; the persistence reinterpretation is plausible but not yet demonstrated from the abstract alone. the 3 major comments →
Updates to the WFC3/IR Photometric Stability Stellar Cluster Study
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Using F160W photometry of 47 Tucanae the WFC3/IR channel continues to lose sensitivity at −0.07% ± 0.01% per year (full data) and −0.06% ± 0.02% per year (first exposure only); using M4 F110W the corresponding rates are −0.13% ± 0.02% and −0.07% ± 0.03% per year. The loss rate has decreased since the previous study, attributed to newer visits designed to mitigate persistence; earlier observations may have overestimated the decline.
What carries the argument
Repeated photometry of uncrowded regions of the star clusters 47 Tucanae and Messier 4 in F110W and F160W, with both full-visit and first-exposure-only analyses, used to isolate the secular instrumental sensitivity trend after persistence-mitigation choices.
Load-bearing premise
That photometry of uncrowded cluster regions, after the chosen persistence-mitigation steps, cleanly isolates true instrumental sensitivity change from residual astrophysical variability, flat-field evolution, or uncorrected persistence.
What would settle it
An independent reduction of the same or additional WFC3/IR cluster visits that applies a different persistence correction or different selection of uncrowded stars and recovers a statistically higher (or lower) linear sensitivity-loss slope inconsistent with the reported −0.06% to −0.13% per year rates.
If this is right
- Users of WFC3/IR photometry spanning many years can adopt the updated loss rates of roughly −0.06% to −0.13% per year depending on filter and data-selection choice.
- The pipeline time-dependent sensitivity reference file (IMPHTTAB) remains the same as before this report.
- Earlier sensitivity-loss estimates based on programs not designed to control persistence are likely overestimated.
- New monitoring visits that deliberately mitigate persistence provide a cleaner baseline for future rate measurements.
Where Pith is reading between the lines
- If the reduced rate is real and continues, long-baseline infrared photometry programs may need smaller cumulative corrections than previously budgeted.
- A natural next test is to re-reduce the pre-2022 visits with the same persistence-mitigation strategy used for the new visits, to quantify how much of the earlier higher rate was residual persistence.
- The filter dependence (F110W appearing steeper than F160W in the full-data case) could be checked with simultaneous multi-filter observations of the same fields.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript updates the long-term photometric sensitivity-loss rate of HST WFC3/IR using photometry of uncrowded regions of 47 Tucanae (F160W) and Messier 4 (F110W), extending Bajaj et al. (2022) by 4–7 visits per target/filter into 2026. Reported rates are −0.07% ± 0.01% yr⁻¹ (F160W, full dataset) and −0.06% ± 0.02% yr⁻¹ (first exposure only); for F110W on M4, −0.13% ± 0.02% yr⁻¹ (full) and −0.07% ± 0.03% yr⁻¹ (first exposure). The authors conclude that the loss rate has decreased relative to the prior study, attribute this decrease to newer visits designed to mitigate persistence (implying earlier work overestimated the loss), and leave the IMPHTTAB time-dependent sensitivity correction unchanged.
Significance. If the updated empirical rates and the persistence-based interpretation hold, this is a practically useful calibration update for WFC3/IR users and for the IMPHTTAB product. Distinguishing full-visit versus first-exposure samples is methodologically careful and strengthens the empirical side of the work. The result is incremental relative to Bajaj et al. (2022) but of clear operational value for HST photometry. The machine-readable rates with uncertainties and the explicit decision on IMPHTTAB are strengths if the isolation of instrumental sensitivity change is adequately demonstrated.
major comments (3)
- [Abstract (persistence / rate-decrease claim)] The central interpretive claim—that the decrease in measured loss rate relative to Bajaj et al. (2022) “likely stems from” newer visits designed to mitigate persistence, and that earlier observations therefore overestimated the IR sensitivity loss—is load-bearing for both the reliability of the updated rates and the decision to leave IMPHTTAB unchanged. The abstract supplies no residual-persistence metrics, no formal pre- vs post-mitigation slope comparison with significance, and no quantitative tests that residual flat-field evolution or low-level astrophysical variability in the uncrowded 47 Tuc / M4 fields are negligible. These controls must be shown explicitly for the causal interpretation to stand.
- [Abstract (full vs first-exposure rates)] The milder first-exposure-only rates (−0.06% ± 0.02% yr⁻¹ F160W; −0.07% ± 0.03% yr⁻¹ F110W) are consistent with persistence in later frames but do not by themselves demonstrate that the secular instrumental rate was previously inflated rather than that true degradation slowed. A quantitative decomposition or controlled comparison (e.g., slopes restricted to pre-mitigation vs post-mitigation visits, with uncertainties) is needed to support the overestimation claim.
- [Abstract (method / isolation premise)] The isolation premise—that photometry of uncrowded cluster regions after the authors’ persistence-mitigation choices cleanly isolates instrumental sensitivity change—underpins both the new rates and the claim that earlier rates were overestimated. Star-selection criteria, light-curve construction, residual-persistence tests, and any checks against flat-field evolution or astrophysical variability are not assessable from the abstract alone and must be documented with sufficient detail for independent scrutiny.
minor comments (4)
- [Abstract] Wording “may have overestimated the IR channel’s sensitivity due to the effects of persistence” is slightly ambiguous: the intended meaning appears to be overestimation of the sensitivity *loss rate* (i.e., measured fluxes too high early on), not overestimation of absolute sensitivity. Clarify.
- [Abstract] Cluster name is written inconsistently as “47 Tucanae”, “47Tuc”, and “47 Tuc”; standardize.
- A compact comparison table of the new rates versus Bajaj et al. (2022), with time baselines, sample definitions (full vs first-exposure), and uncertainties, would make the “rate has decreased” claim immediately checkable.
- State explicitly how many visits and what date range enter each fit (full vs first-exposure; F160W vs F110W) so the leverage of the new 4–7 visits is transparent.
Circularity Check
No significant circularity: empirical sensitivity-loss rates are fitted outputs from new photometry, not forced by construction or load-bearing self-definition.
full rationale
This is an abstract-only review of an empirical calibration update. The claimed rates (−0.07% ± 0.01%/yr F160W full; −0.06% ± 0.02%/yr first-exposure; −0.13% ± 0.02%/yr F110W full; −0.07% ± 0.03%/yr first-exposure) are ordinary least-squares (or equivalent) slopes fitted to new photometry of 47 Tuc and M4 plus 4–7 additional visits since Bajaj et al. (2022). The prior work is the natural previous epoch of the same monitoring program, not a uniqueness theorem, ansatz, or definition that forces the new slopes. The statement that the loss rate has decreased and that earlier observations may have overestimated sensitivity is an interpretive claim about residual persistence, not a quantity that is definitionally equal to an input parameter. IMPHTTAB is reported unchanged, which is a null decision rather than a circular prediction. No self-definitional loop, fitted-input-as-prediction, or uniqueness import is present in the available text. Residual scientific risk (whether persistence mitigation fully isolates instrumental sensitivity) is a correctness concern, not circularity. Score 1 reflects only the ordinary self-citation of the prior epoch of the same program, which is not load-bearing for the numerical rates themselves.
Axiom & Free-Parameter Ledger
free parameters (2)
- sensitivity loss rate (F160W 47 Tuc, full) =
-0.07% ± 0.01% per year
- sensitivity loss rate (F110W M4, full) =
-0.13% ± 0.02% per year
axioms (3)
- domain assumption Uncrowded stars in 47 Tuc and M4 are photometrically stable at the level needed to track sub-percent-per-year instrumental changes.
- domain assumption Recent visits designed to mitigate persistence successfully reduce residual charge effects relative to earlier epochs.
- domain assumption Linear (or piecewise-linear) time dependence adequately describes the sensitivity loss over the full baseline.
read the original abstract
The WFC3/IR channel has been shown to lose sensitivity over time. The sensitivity loss rate has been quantified in past works and we extend those measurements into 2026 using photometry of uncrowded regions of star clusters 47 Tucanae and Messier 4 in the F110W and F160W filters. This study adds 4-7 additional visits since Bajaj et al. (2022) for each target/filter combination. Using F160W data of 47Tuc, we find that the IR channel continues to lose sensitivity at a rate of -0.07% $\pm$ 0.01% per year using the full dataset, and -0.06% $\pm$ 0.02% per year using only the first exposure of every visit. Using M4 data observed in F110W, we find a sensitivity loss rate of -0.13% $\pm$ 0.02% per year using the full dataset, and -0.07% $\pm$ 0.03% per year using just the first exposure of every visit. We find that the rate of sensitivity loss has decreased since the previous study. This likely stems from the inclusion of more recent visits that were designed to mitigate persistence effects. These results suggest that earlier observations, which were not intended for sensitivity measurements, may have overestimated the IR channel's sensitivity due to the effects of persistence from bright sources. The IMPHTTAB reference file that contains the correction to the time-dependent sensitivity remains unchanged from the sensitivity measurements of this report.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.