{"id":"8afb2178-3b71-4f99-8c5f-f7afaaed2d38","arxiv_id":"2608.00193","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A new open-source exposure time calculator simulates the full photon path for the proposed WST telescope's three spectrograph modes and gives provisional signal-to-noise and limiting-magnitude baselines.","lead":"This paper presents pyetc_wst, an open-source exposure time calculator that simulates light's full path from star to detector for the planned 12-meter Wide-field Spectroscopic Telescope, returning signal-to-noise spectra and limiting magnitudes for its three spectrographs. It matters because its performance baselines will inform survey planning and design trade-offs for a proposed next-generation facility.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Headline limiting magnitudes inherit unverified throughput tables; robustness to realistic throughput uncertainty is untested.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: all performance predictions scale directly with unmeasured, preliminary throughput tables that are internal engineering estimates. The paper itself discloses this limitation, which is why the verdict is CONDITIONAL rather than ACCEPT. My analysis agrees: the tool's framework is standard and the code appears functional, but the headline limiting magnitudes and the 'validated framework' claim in §6 outstrip the evidence shown. The concrete test of sensitivity to throughput uncertainties would determine whether the concern actually moves the numbers enough to matter. Since the reader already conditioned the verdict on this issue, my read does not change the verdict; it reinforces it. I do not see a more fundamental flaw in the ETC equations, the noise model, or the software architecture as described. The missing validation evidence and lack of error bars are real but secondary; the throughput dependence is the single point of failure.","tokens_in":10339,"tokens_out":3236,"duration_ms":42612,"concrete_test":"Propagate a wavelength-dependent ±10% and ±20% band around the T_ins(λ) curves in Fig. 8 through Eq. (1) and recompute the limiting magnitudes in Figs. 10–11 (§5.3). If the dark-sky μ_r or r values shift by more than ~0.2 mag, the headline numbers are not robust to the stated engineering uncertainty and need error bars; if shifts are <0.1 mag, the concern is minor. As a secondary check, independently re-derive one SNR curve (e.g., IFS Blue V=19 in Fig. 9) from the published equations and tables to confirm the code implements Eq. (2) without hidden normalization factors.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central numerical results — IFS surface-brightness limit μ_r ≈ 25.4 and MOS-LR point-source limits r ≈ 22–23.4 (Figs. 10–11, §5.3) — are directly proportional in Eq. (1) to the instrument throughput T_ins(λ). These tables (Fig. 8) are explicitly preliminary, 'derived from the throughput model delivered by the WST system engineering team' (§3.1), and the paper itself warns they 'will be updated as the design matures, and all performance numbers reported in this paper should be interpreted accordingly' (§5.1). The coatings are only 'expected to reach' given values (§2.1); detector RON/dark are assumed (Table 1). No error bars or sensitivity analysis accompany the limiting magnitudes. Consequently, the headline survey-depth claims rest on single-point engineering estimates that have not been independently validated. The Conclusions' statement that the throughput model has been 'verified against first-principles expectations' is not supported by any shown verification. This does not undermine the tool's existence or its standard photon-budget formulation, but it does mean the quantitative performance baselines should be treated as provisional, with uncertainties propagated from the throughput assumptions.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents pyetc_wst, an exposure time calculator for the proposed Wide-field Spectroscopic Telescope, covering IFS, MOS-LR, and MOS-HR. It implements a source-to-detector photon budget (Eq. 1), static and SkyCalc sky backgrounds, PSF/fibre injection, a noise model (Eq. 2), four computation modes, and web/REST/CLI interfaces. Example outputs include SNR spectra, noise decomposition, and limiting magnitudes: IFS surface brightness μ_r ≈ 25.4 AB arcsec^-2 and MOS-LR point-source limits r ≈ 22–23.4 AB for 1 h dark-sky observations.","tokens_in":10549,"tokens_out":5572,"duration_ms":64082,"significance":"The main value of this work is a public, modular ETC implementation built on standard photon-budget and variance equations, with useful features such as SkyCalc integration, a rich SED/template library, flexible computation modes, and noise decomposition. These are concrete strengths that make the tool immediately usable for WST survey planning. However, the quantitative performance predictions are not independent measurements: they scale linearly with preliminary system-engineering throughput tables that are explicitly subject to revision. The paper's claim that the throughput model and performance numbers are 'verified against first-principles expectations' goes beyond what is actually shown and should be tempered or supported by a sensitivity analysis.","major_comments":[{"comment":"The claim that GLAO improves the IFS surface-brightness limit by reducing the sky background per spaxel is physically incorrect for a fixed 0.25″ spaxel. For a uniform extended source, both source and sky counts per spaxel are independent of PSF. If instead the extraction aperture is matched to the PSF, a smaller PSF reduces the aperture area and thus the SNR per resolution element. Fig. 10 and the headline μ_r ≈ 25.4 AB arcsec^-2 rest on this step. Please specify the exact spatial binning/resolution element and provide a derivation; as written, the GLAO-specific SB limiting magnitudes are unsupported.","section":"§5.3, §2.1"},{"comment":"All limiting magnitudes and SNR curves scale linearly with the preliminary throughput tables T_ins(λ) in Eq. (1), which are unmeasured engineering estimates ('expected to reach' 79.8%/91.3% telescope throughput, assumed detector RON/dark) with no quoted uncertainty. The paper itself warns in §5.1 that 'all performance numbers reported in this paper should be interpreted accordingly,' yet §6 states the throughput model has been 'verified against first-principles expectations' and presents the limits as confirming WST performance. No verification or sensitivity analysis is shown. Please add an uncertainty propagation or sensitivity study for the headline numbers, or explicitly reframe the conclusions as provisional engineering estimates.","section":"§5.1, Fig. 8, §6"}],"minor_comments":[{"comment":"The package name is typeset inconsistently as 'pyetc wst' in the text and 'pyetc_wst' in code/URLs; unify the notation.","section":"Throughout"},{"comment":"The caption says 'rebinned by 5 Å for display' while the text describes SNR per spectral pixel; clarify whether the values are per native pixel or per rebinned bin.","section":"Fig. 9 caption"},{"comment":"The definition 'per resolution element of 1.4 Å co-added over 3 spectral pixels' specifies only the spectral bin; the spatial element used for the IFS surface-brightness limit should be stated explicitly (e.g., 1 spaxel, N×N bin, or PSF area).","section":"§5.3"},{"comment":"References [5], [6], and [7] cite 'Proc. SPIE This conference' with paper numbers but no page/article details; if available, add full bibliographic information or a DOI.","section":"References"},{"comment":"The MOS-HR channels are named Blue, Green, Yellow, Red, which may be confused with the MOS-LR channels of the same names in Fig. 8. Consider adding the central wavelengths or a prefix (e.g., HR-Blue) to the channel labels.","section":"Table 1"}],"recommendation":"major_revision","confidential_remarks":"This is a software/tool paper with provisional performance numbers. The physical issue with the GLAO surface-brightness computation is a correctness problem in the modeling, not just a caveat, and should be fixed. The throughput sensitivity issue is acknowledged in the body but contradicted by the conclusions; the authors should either add the requested analysis or tone down the claims. Neither issue invalidates the tool itself, but both affect the headline numerical results."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe useful thing here is the tool and the honest caveats, not the headline numbers. pyetc_wst is a real ETC codebase with web, API, and CLI access, four computation modes, Sersic morphology support, SkyCalc integration, and full noise decomposition. First WST performance baselines are now public, which matters for survey planning. Equations 1–2 are the standard photon-budget and variance sums and they are internally consistent; Table 1 and the figure captions are explicit enough that the SNR curves are reproducible from the paper itself. The §5.1 warning that every number is preliminary because the throughput model is unmeasured is also unusually straightforward.\n\nThe soft spots are where you'd expect. The load-bearing inputs are the wavelength-dependent throughput tables T_ins(λ) from the WST system engineering team, one of whom is a co-author and the March 2026 source. Those tables are engineering estimates — coatings 'expected to reach' given values, CMOS detectors assumed at RON 1.0–1.4 e−. Limiting magnitudes in Figs. 10–11 inherit those assumptions with no error bars and no sensitivity analysis. The stress-test note is right about that. It is not a fatal flaw, because the paper says the numbers will change as the design matures, but it means the quantitative survey-depth claims in the abstract and conclusions should carry a 'provisional' label, not be presented as results.\n\nThe one thing I'd push on is the conclusions sentence calling the framework 'complete, validated' and stating the throughput model has been 'verified against first-principles expectations.' No verification is shown anywhere. Either show the check or drop the word. That is an overreach and it is easy to fix.\n\nMinor: no commit hash or version pin for the code, and the example outputs don't include a reproducibility package. For a tool paper that's worth asking for.\n\nBottom line: this is a legitimate instrument-software paper, not a science-breaking one. The community it serves — the WST collaboration and anyone building ETCs for multi-mode spectrographs — will get real value. Send it to peer review with the request that the validation claim be substantiated or softened and that sensitivities be added. I'd cite it if I were working on WST survey planning.","headline":"A solid, honestly caveated ETC tool paper for WST whose survey-depth numbers are provisional; the 'validated' claim in the conclusions overreaches, but the code and method deserve refereeing.","tokens_in":11287,"tokens_out":2201,"would_cite":true,"duration_ms":25207,"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 presents pyetc_wst, an end-to-end exposure time calculator for the Wide-field Spectroscopic Telescope that simulates the full photon path from source to detector and predicts concrete survey depths for all three spectrograph mode","keywords":["Exposure Time Calculator","Wide-field Spectroscopic Telescope","Integral Field Spectroscopy","Multi-Object Spectroscopy","Signal-to-Noise Ratio","End-to-end simulation","Throughput model","Limiting magnitude"],"falsifier":"Take one IFS blue channel and one MOS-LR channel, measure their end-to-end throughput on the ground using a calibrated continuum source, and compare the ETC's predicted SNR for a standard star of known flux against the SNR actually recorded on a detector with known read-out noise. A discrepancy larger than the stated detector noise contributions would falsify the model's predictive accuracy; a laboratory measurement of total instrument efficiency at 500 nm would already bound the main limiting-magnitude claims.","tokens_in":10081,"feed_emoji":"🔭","tokens_out":5830,"duration_ms":63276,"temperature":0.7,"pith_summary":"The paper introduces a new exposure time calculator for the proposed Wide-field Spectroscopic Telescope, a 12-meter facility with three simultaneous spectrograph modes. Rather than using simple scaled efficiencies, the calculator simulates the complete photon path: source spectrum, atmospheric transmission, telescope and instrument throughput per channel, fibre coupling, sky emission, and detector noise. From these it produces wavelength-dependent signal-to-noise spectra, noise decompositions, simulated raw spectra, and limiting magnitudes. The headline performance numbers are that the integral-field mode reaches a surface-brightness limit of about 25.4 AB magnitudes per square arcsecond in the blue and 25.5–26.0 in the red, while the low-resolution multi-object mode reaches point-source limits of roughly 22 to 23.4 AB magnitudes, all for one hour of exposure in dark sky. A sympathetic reader would care because such a tool makes the project's survey capabilities concrete and testable before construction.","feed_headline":"Simulation puts WST's one-hour depth at r≈23.4","feed_subtitle":"Full photon-budget model from source to detector lets survey planners set exposure times before the telescope is built.","key_machinery":"The central object is the photon-budget equation S(λ) = F_λ(λ) (λ/hc) A_tel τ_atm(λ) T_ins(λ) f_fib(λ) t_exp, together with the per-pixel noise variance σ² = N_DIT [S_src + S_sky + N_pix (d t_DIT + σ_RON²)]. The instrument throughput tables T_ins(λ), one per channel and currently the only unmeasured input, carry most of the physical content. The fibre injection fraction is computed from a Moffat PSF integrated over a circular aperture, with an object–fibre displacement parameter for pointing errors. Sky emission and atmospheric transmission come from a static table set or a live external sky service. The four computation modes invert these equations to find SNR, exposure time, or the optimal","core_discovery":"The paper's central claim is that pyetc_wst implements a true end-to-end model of the WST instruments. For each wavelength element, the detected signal is the product of the source flux, telescope area, atmospheric transmission, a wavelength-dependent total instrument throughput table, fibre injection fraction, and exposure time; the noise variance is a sum of source photon noise, sky photon noise, dark current, and read-out noise, with coadding and DIT/NDIT combinations handled explicitly. The tool supports point sources, uniform surface brightness, and Sérsic profiles, and generates Monte Carlo realisations of observed 1D spectra. The author's conclusion, on the basis of this model and the","pith_inferences":["Editorial inference: because the throughput tables are the only unmeasured input, the same tool can be rerun with revised tables at every design iteration, turning the ETC into a living survey forecast.","Editorial inference: the fibre-injection treatment could be sharpened by adding atmospheric dispersion and 3-D slit losses, which would most affect the blue end of each channel.","Editorial inference: making the calculator public invites independent reproduction of the headline depths, which would test the design baseline without needing the telescope."],"forward_implications":["If the model holds, WST's IFS will be able to obtain SNR=3 per resolution element on surface brightness of about 25.4 AB mag/arcsec² in one hour in dark sky, which defines the accessible regime for studies of faint diffuse emission.","MOS-LR point-source limits of r≈22–23.4 AB in one hour imply that a single 2-degree pointing can deliver spectroscopy for tens of thousands of targets down to those magnitudes.","The noise decomposition at V=19 shows source photon noise dominated below 850 nm and read-out/dark noise below 5%, so those detector assumptions are not currently limiting the predicted performance.","The exposure-time inversion modes allow survey planners to convert a target SNR directly into a DIT/NDIT schedule, which is exactly what is needed for designing an observational campaign."],"fun_headline_variants":["WST exposure calculator: end-to-end photon budget","Simulating WST from source to detector for survey planning","WST ETC models every photon path to set exposure times","Full noise model predicts WST depth for any target","WST's ETC: complete simulation for survey architects"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Everything hinges on the unmeasured, wavelength-dependent instrument throughput tables supplied by the system engineering team; if the as-built optics and detectors underperform those curves, every SNR and limiting magnitude in the paper falls in proportion.","fun_headline_variants_meta":{"raw":{"variants":["WST exposure calculator: end-to-end photon budget","Simulating WST from source to detector for survey planning","WST ETC models every photon path to set exposure times","Full noise model predicts WST depth for any target","WST's ETC: complete simulation for survey architects"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000183,"raw_usage":{"total_tokens":1194,"prompt_tokens":833,"completion_tokens":361,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":281}},"tokens_in":577,"tokens_out":361,"duration_ms":5167,"temperature":1.0,"reasoning_tokens":281,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T01:01:52.448372+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one IFS blue channel and one MOS-LR channel, measure their end-to-end throughput on the ground using a calibrated continuum source, and compare the ETC's predicted SNR for a standard star of known flux against the SNR actually recorded on a detector with known read-out noise. A discrepancy larger than the stated detector noise contributions would falsify the model's predictive accuracy; a laboratory measurement of total instrument efficiency at 500 nm would already bound the main limiting-magnitude claims.","supporting_citations":[],"review_version":1}