{"id":"0f151868-4415-4752-8d59-1f8fabfb55b7","arxiv_id":"2608.13464","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A star and asteroid calibration combination produces 48 point fixed pattern corrections that remove MIRI MRS artifacts and raise achievable S/N above 1000 for fully corrected point source spectra.","lead":"This paper constructs a new set of fixed pattern corrections for JWST's MIRI medium-resolution spectrometer, using standard stars and asteroids to remove instrument artifacts from point source spectra. When applied, the corrections raise the practical signal-to-noise ceiling from a few hundred to over 1000 in some wavelengths and remove features that could be mistaken for real astrophysical signals.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Single-PFPC-per-dither assumption is the least secure link: sub-pixel pointing variations can leave or imprint fringes, and the paper's own Fig. 6 shows epochs with negligible improvement.","rationale":"The reader's weakest_assumption identifies exactly this issue, and the paper's own text supports it. In good faith, the paper is well documented, the MRS-PFPC code is released, and validation on HD 163466 and asteroids is independent of the PFPC construction. However, the strongest claim is stated as applying uniformly to point sources with the default dither pattern, while Section 4.3 shows conditions under which the PFPC improvement is negligible. The only quantitative support for the single-correction assumption is a qualitative citation to Rigby et al. (2023); no measurement of delivered dither-position repeatability or its effect on PFPC residuals is provided. The Gasman et al. intra-pixel mini-grid data already exist and would settle whether the assumption holds. The CALSPEC-model concern is real but partially mitigated: PFPCs derived from stars and asteroids, which make independent model assumptions, agree, and the CALSPEC models use no JWST data. The offset issue directly affects the headline S/N numbers because those numbers are an upper envelope from visually clean regions, not a worst-case guarantee. Therefore the CONDITIONAL verdict remains appropriate; this stress-test does not move it.","tokens_in":15537,"tokens_out":5319,"duration_ms":62244,"concrete_test":"Use the Gasman et al. (2024) 10 Lac intra-pixel dither observations: for each nominal dither position, shift the observed spectra by the known sub-pixel offsets, apply the corresponding nominal PFPC, and measure the residual fringe amplitude (e.g., RMS of Obs/PFPC after continuum removal). Compare this with the distribution of delivered pointing offsets from JWST engineering telemetry for the calibration programs. If the residual exceeds roughly 0.5% for offsets within the observed TA scatter, the single-PFPC-per-dither assumption fails and the product should include offset-dependent corrections or a per-wavelength uncertainty map.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing vulnerability is the assumption in Sections 1 and 5 that one static PFPC per nominal dither position is sufficient for all TA-acquired point sources. The paper itself cites sub-pixel non-repeatability of the MRS grating wheels and the Gasman et al. (2024) intra-pixel work showing position-dependent residuals, and Section 4.3 reports epoch-to-epoch S/N scatter with negligible PFPC improvement in some channel-1 epochs. If delivered pointings differ from the calibration-set positions by even a fraction of a pixel, the PFPC will not remove residual fringes and can imprint new ones at the 0.5-1% level. The claimed factors of 2-8 and S/N values above 1000 come from visually selected clean regions and are averaged over epochs, so they do not establish performance for worst-case sub-pixel offsets. This is acknowledged as future work, but until the offset sensitivity is quantified, the broad applicability claim for all observations with the default dither pattern is not secure. The CALSPEC model accuracy assumption is less concerning here because the stars and asteroids, which use independent model assumptions, show good agreement in the PFPC measurements; the offset issue has no such independent bound.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript derives and validates a set of 48 wavelength-dependent Point Fixed Pattern Corrections (PFPCs) for JWST/MIRI MRS point-source spectra observed with the default four-point dither pattern. The corrections are constructed from sigma-clipped averages of observed-to-model ratios for O/A/G dwarf CALSPEC standards and two asteroids, with masking around stellar lines and low-S/N regions. The authors validate on the independent A dwarf HD 163466 and five asteroids not used in constructing the PFPCs, demonstrate removal of known artifacts such as the 5.8 micron feature, and report S/N improvements up to factors of about 8 (without residual fringe correction) and about 3 (with it), with coadded S/N values exceeding 1000 at short wavelengths. A public Python package (MRS-PFPC) and the associated data are provided, and the corrections are explicitly tied to jwst pipeline version 2.0.0.","tokens_in":15783,"tokens_out":8522,"duration_ms":81819,"significance":"If confirmed, this is a high-value calibration product for the MIRI MRS community: it addresses the dominant fixed-pattern noise for point sources, extends coverage to all four channels, and provides a practical tool with public code and machine-readable outputs. The empirical construction is sound in outline, and the authors are careful to validate on sources excluded from the fit and to combine stars and asteroids to separate stellar-line masking from continuum correction. The explicit coupling of the corrections to a specific pipeline version and the acknowledgment that future pipeline updates will require new PFPCs are appropriate. However, the headline quantitative claims (factors of 2-8, S/N above 1000) rest on an incompletely specified S/N estimator, on visually selected clean regions, and on PFPC uncertainties that are not propagated; these need to be tightened before the numbers can be taken at face value.","major_comments":[{"comment":"The S/N estimator is not defined precisely enough to support the quantitative claims. The text says S/N was measured by fitting a line to the continuum and measuring the standard deviation in visually clean regions, but it does not state whether the reported values are per native spectral sample, per spectral resolution element, per binned wavelength interval, or per fitted continuum segment. The reported improvement factors (up to about 8 without residual fringe correction, up to about 3 with it) and the S/N above 1000 claim depend directly on this choice. Please specify the estimator, the binning (including the width of the clean regions), the number of independent samples per segment, and the criteria used to select regions, and make the region list available in the package. Because the regions are visually selected, the factors are best-case values; please also provide the full per-segment and per-epoch distribution of improvement factors.","section":"Sec. 4.3, Fig. 6"},{"comment":"The paper's own data show that the static single-PFPC-per-dither assumption is not sufficient for a subset of observations: the bottom-left panel of Fig. 6 shows channel-1 epochs with negligible improvement when the residual fringe correction is applied, and the text attributes this to sub-pixel differences in the delivered dither locations and grating-wheel non-repeatability. This qualification is directly relevant to the abstract's claim that the PFPCs improve spectra 'regardless of their S/N' and to the summary's general applicability statement for the default dither pattern. Please quantify the fraction of epochs and segments for which the PFPC improvement is less than 1.1, report the worst-case performance, and restate the applicability claim as typical rather than best-case behavior. This is load-bearing because a user needs to know when the correction will fail.","section":"Sec. 4.3, Fig. 6 (bottom-left); Sec. 5"},{"comment":"The PFPCs are described as setting the limiting S/N (100-300 without residual fringe correction, 180-430 with it), and Sec. 4.3 concludes that the coadded S/N is limited by the PFPC S/N, but no per-wavelength uncertainty is provided for any of the 48 PFPCs. The sigma-clipped average in Fig. 1 has a natural standard error of the mean from the individual target measurements; please compute and plot this uncertainty, report the number of independent measurements surviving clipping in each segment and dither, and propagate the PFPC uncertainty into the corrected spectra. Without this, the statement that the coadded S/N 'gives the PFPC S/N' and the headline S/N above 1000 value cannot be verified.","section":"Sec. 3, Fig. 2; Sec. 4.3"}],"minor_comments":[{"comment":"The phrase 'targeted stars stars' contains a duplicated word and should be corrected.","section":"Sec. 2, paragraph 2"},{"comment":"The caption contains the typo 'quadratic fitf'; it should read 'quadratic fit'.","section":"Sec. 3, Fig. 1 caption"},{"comment":"There are two typos in this section: 'neglibable' should be 'negligible' and 'expect for segment 4A' should be 'except for segment 4A'.","section":"Sec. 4.3"},{"comment":"The sentence 'This emission is seen in the pipeline PFPC spectra before residual fringe correction' is ambiguous; please clarify whether the HI 7-6 emission appears in both the pipeline and PFPC reductions before the residual fringe correction is applied.","section":"Sec. 4.2"},{"comment":"The step that multiplicatively corrects the overall level of each dither position to the average of the four dithers should be justified, since it removes any dither-dependent throughput differences; please state whether this is intended and discuss its effect on absolute flux calibration.","section":"Sec. 4.1, step 2"},{"comment":"The text states that observations from different cycles are shown with different symbols, but the figure caption and legend do not specify the symbol-to-cycle mapping; please add this information.","section":"Fig. 6"},{"comment":"The statement that 'S/N values exceeding 1000' are achieved should explicitly note that this refers to the coadded spectrum, not a single epoch, and that the coadded S/N is stated to be limited by the PFPC S/N rather than by photon noise.","section":"Sec. 5"}],"recommendation":"major_revision","confidential_remarks":"The paper is well suited to AJ's instrumentation and techniques scope, and the public code and data links are a strength. The requested additions--per-wavelength PFPC uncertainties, an exact definition of the S/N estimator, and an epoch-resolved performance distribution--are feasible with the data already in hand and would materially strengthen the calibration product. I have no concerns about citation or disclosure issues."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Give the paper its due: it produces the first full-wavelength PFPC set for MIRI MRS covering all 12 segments and the four default dithers as one product, and it ships an open-source package with the calibration files. The validation on HD 163466 and five asteroids not used in the construction is the right way to avoid circularity, and the agreement between stellar and asteroid corrections strengthens the case that the PFPCs are instrumental rather than model artifacts. The improvement in channel 2 and the removal of the 5.8 um artifact are real gains.\n\nThe soft spots are the ones you'd expect. The single-PFPC-per-dither assumption is load-bearing, and the paper does not quantify how sensitive the corrections are to sub-pixel pointing offsets. The authors cite grating wheel non-repeatability and Gasman et al.'s intra-pixel work, and their own Fig. 6 shows epochs with negligible improvement in channel 1. That doesn't sink the product, but it does mean the 'S/N above 1000' claim applies only to the best-behaved observations, not to every point source. The abstract and summary could be more careful with that qualification. Also, the S/N improvements are measured in visually selected clean regions, so they likely overstate the global improvement, and there are no per-wavelength uncertainty maps for the PFPCs. These are routine reviewer requests.\n\nThe paper's logic holds up. The circularity concern is correctly handled. The CALSPEC/asteroid model accuracy is not a serious worry because the independently derived corrections agree.\n\nVerdict: solid calibration paper, worth referee time. I'd send it out. Anyone working with MRS point source spectra should know about this product.","headline":"A solid, practically important calibration paper for MRS point sources, with correct validation and a useful package; the main caveat is the unquantified sensitivity to sub-pixel pointing offsets.","tokens_in":16323,"tokens_out":3637,"would_cite":true,"duration_ms":31238,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper constructs 48 point fixed pattern corrections for JWST MIRI MRS point-source spectra and shows that they remove fixed-pattern artifacts and raise signal-to-noise limits above 1000 at shorter wavelengths.","keywords":["JWST","MIRI","Medium Resolution Spectrometer","fixed pattern noise","spectral calibration","signal-to-noise","infrared spectroscopy","point sources"],"falsifier":"Take an independent MRS point-source observation not used to build the PFPCs, reduce it with and without the corrections, and compare residuals across many repeated target-acquisition visits at the same dither positions; if the PFPC-corrected spectra still show the same 1-2% dither-dependent fringes or the 5.8 µm feature, the single-correction-per-dither assumption fails.","tokens_in":15348,"feed_emoji":"🔭","tokens_out":9223,"duration_ms":78895,"temperature":0.7,"pith_summary":"The paper seeks to establish that the fixed-pattern noise capping the signal-to-noise of JWST MIRI Medium Resolution Spectrometer point-source spectra can be largely removed by a set of 48 wavelength-dependent corrections, built from observations of flux-calibration O, A, and G dwarfs plus two asteroids. Combining stars and asteroids lets the corrections stay high signal-to-noise across the full 5-28 µm range and fills in regions masked near stellar lines. The authors claim that applying these PFPCs removes both narrow and broad artifacts and raises S/N by up to almost 8x when the pipeline residual fringe correction is not used, and up to 3x when it is used, reaching S/N values above 1000 at shorter wavelengths. The result matters because the fixed pattern currently limits the weakest spectral features that can be detected, and because broad artifacts such as a 5.8 µm feature resembling dust absorption are shown to be instrumental.","feed_headline":"New corrections push JWST spectra past S/N 1000","feed_subtitle":"Applying 48 fixed-pattern corrections removes artifacts and lifts MRS signal-to-noise to about 8x.","key_machinery":"The central object is the Point Fixed Pattern Correction (PFPC), a wavelength-dependent multiplicative correction measured separately for each of the 12 MRS channel/grating segments and each of the 4 dither positions, yielding 48 corrections in all. It is built from ratios of reduced calibration spectra to models: CALSPEC synthetic spectra for the O, A, and G dwarf standards and quadratic continuum fits for asteroids, with stellar lines and low-S/N regions masked and a sigma-clipped average taken across the ensemble of targets. The PFPC acts as a transfer function that divides out the fixed-pattern noise tied to the specific pipeline version and reference files, and the accompanying MRS-PFPC package applies it to individual dither spectra before averaging and optional residual fringe correction.","core_discovery":"The central discovery is that the MRS point-source fixed-pattern noise, which sits at roughly 1-2% amplitude and includes residual fringes, sampling artifacts, flat-field uncertainties, and interpolation artifacts, is stable enough across the default four-point dither pattern to be captured by a single correction per dither position. Dividing each observed spectrum by the appropriate CALSPEC model for a star or a quadratic continuum for an asteroid, masking lines and low-S/N regions, and sigma-clipping across targets yields 48 high-S/N PFPCs. Applied to independent observations, these corrections remove both narrow and broad artifacts: the pipeline's continuum 'wiggles' disappear, the 5.8 µm carbonyl-like feature vanishes, and measured S/N rises by up to ~8x without residual fringe correction and up to ~3x with it, reaching above 1000 at short wavelengths. The paper also finds that the S/N of coadded spectra is limited by the S/N of the PFPCs themselves, not by photon noise, implying that more calibration observations would raise the ceiling.","pith_inferences":["If the PFPC S/N is indeed the limiting factor for coadded spectra, then adding more diverse calibration targets to future versions should keep raising the maximum achievable S/N, a trend the paper already sees when extra HD 163466 observations are included in the PFPC construction.","The single-correction-per-dither assumption should be tested on observations taken without target acquisition or after grating wheel resets; channels 1 and 2, where dither-to-dither PFPC variation is largest, are the most likely places for residual fringes to reappear.","A natural extension is to fold the PFPCs into the pipeline reference files themselves, which would let the entire archival MRS data set be reprocessed to this calibration level rather than requiring users to run the custom package.","The removal of the 5.8 µm feature suggests that other broad MRS spectral features currently attributed to astrophysical dust or molecular carriers should be re-examined with PFPC-corrected spectra before being published as detections."],"forward_implications":["For point sources whose spectra have molecular or emission-line structure that makes the pipeline fringe correction unsafe, PFPC-corrected spectra gain up to almost 8x in S/N, so weak features in such sources become detectable that would previously be buried in fixed pattern noise.","For sources where the fringe correction can be applied, PFPCs push achievable S/N past 1000 at short wavelengths and up to 3x higher than the pipeline, bringing bright-star observations into the sub-1% calibration regime.","The broad 5.8 µm feature and other continuum wiggles that survive pipeline reduction are shown to be instrumental and are removed, meaning MRS surveys should not interpret those features as interstellar dust absorption.","Because the PFPCs are tied to the specific pipeline version and reference files, any future pipeline update requires regenerating them; the MRS-PFPC package is designed to select the matching PFPC version automatically."],"supporting_citations":[{"why":"Supplies the prior dither-dependent fixed-pattern correction based on one A star that this work generalizes to more targets and all channels.","marker":"D. Gasman et al. 2023"},{"why":"Provides the asteroid-based correction approach for longer MRS wavelengths and the use of per-dither corrections.","marker":"K. M. Pontoppidan et al. 2024"},{"why":"Defines the MRS flux calibration and flat-field construction that set the fixed-pattern residuals the PFPCs remove.","marker":"D. R. Law et al. 2025"},{"why":"Quantifies residual fringes as the dominant fixed-pattern noise source in MRS spectra.","marker":"I. Argyriou et al. 2020"},{"why":"Supplies the CALSPEC models used as the comparison spectra for the stellar flux standards.","marker":"R. C. Bohlin et al. 2014"},{"why":"Documents sub-pixel grating-wheel positioning variations that motivate testing single-correction-per-dither assumptions.","marker":"P. Patapis et al. 2024"},{"why":"Supports the pointing-repeatability argument for using one correction per dither position.","marker":"J. Rigby et al. 2023"},{"why":"Defines the jwst pipeline version 2.0.0 to which the PFPCs are tied.","marker":"H. Bushouse et al. 2026"}],"fun_headline_variants":["JWST MRS fixed pattern corrections boost S/N to 1000+","48 fixed-pattern corrections lift JWST MRS S/N by 8x","New MRS corrections erase artifacts, S/N reaches 1000","JWST MRS S/N jumps to 1000 with new PFPCs","Fixed pattern corrections give JWST MRS 8x S/N boost"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the JWST pointing at each dither position repeats well enough that a single correction per dither position captures the fixed pattern, and that the CALSPEC stellar models and quadratic asteroid continua used to measure the corrections are accurate enough that the residuals being fit are purely instrumental.","fun_headline_variants_meta":{"raw":{"variants":["JWST MRS fixed pattern corrections boost S/N to 1000+","48 fixed-pattern corrections lift JWST MRS S/N by 8x","New MRS corrections erase artifacts, S/N reaches 1000","JWST MRS S/N jumps to 1000 with new PFPCs","Fixed pattern corrections give JWST MRS 8x S/N boost"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000526,"raw_usage":{"total_tokens":2550,"prompt_tokens":969,"completion_tokens":1581,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":585,"completion_tokens_details":{"reasoning_tokens":1481}},"tokens_in":585,"tokens_out":1581,"duration_ms":9949,"temperature":1.0,"reasoning_tokens":1481,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:26:34.510366+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take an independent MRS point-source observation not used to build the PFPCs, reduce it with and without the corrections, and compare residuals across many repeated target-acquisition visits at the same dither positions; if the PFPC-corrected spectra still show the same 1-2% dither-dependent fringes or the 5.8 µm feature, the single-correction-per-dither assumption fails.","supporting_citations":[],"review_version":1}