{"id":"46be930c-574f-455c-a49a-65f5c0a5dc6f","arxiv_id":"1908.06741","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"An independent wavelet-based reanalysis of WASP43 b's Spitzer phase curves finds higher nightside flux, smaller hotspot offsets, and consistency with atmospheric circulation models and a prior reanalysis.","lead":"This paper reanalyzes Spitzer infrared observations of the hot Jupiter WASP43 b using a blind signal-separation technique, and reports warmer nightside temperatures and smaller hotspot offsets than the original analysis. The result matters because it resolves a mismatch between measured and modeled heat circulation on a key James Webb Space Telescope target.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The key 3.6 μm nightside claim hinges on discarding the BIC/AIC-preferred quadratic ramp for visit 2; if that ramp is correct, the nightside is near zero and the claimed inter-visit consistency disappears.","rationale":"The reader's weakest-assumption analysis identifies exactly the same hinge: the second 3.6 μm visit's ramp-model choice controls whether the nightside flux is high and consistent with visit 1, or near zero and in line with the original Stevenson et al. (2017) interpretation. My independent reading confirms this is the most load-bearing concern. Other potential issues, such as the unresolved limb-darkening degeneracy in transit parameters, affect the conversion to absolute temperatures only at the ~1% level and do not change the phase-curve flux ratios; they are secondary. The atmospheric-model comparison is also model-dependent, but the observational claim about higher nightside flux is the foundation on which the circulation-efficiency conclusion rests. The paper is transparent about the ambiguity and provides useful robustness checks, which is why the concern warrants a conditional verdict rather than rejection. However, because the claimed consistency between the two 3.6 μm visits is substantially weaker when the quadratic ramp is used (the two nightside-flux estimates are discrepant by roughly 3σ), and because all standard model-selection tools prefer the quadratic ramp, the central claim is not fully settled without a positivity-constrained or injection-recovery test. The reader's CONDITIONAL verdict already captures this, so I recommend no change to the verdict.","tokens_in":25363,"tokens_out":4416,"duration_ms":50897,"concrete_test":"Reanalyze the second 3.6 μm visit with the same wavelet pixel-ICA pipeline and the quadratic ramp, but impose a physical positivity prior on the nightside flux (e.g., parameterize the phase-curve minimum as exp(q) or add a hard prior F_night >= 0), then recompute the posterior and model comparison. If the constrained quadratic posterior for F_night centers near the linear-ramp value (≈3×10^-4) and agrees with visit 1, the paper's choice is likely acceptable. If the constrained posterior remains near zero and inconsistent with visit 1, the reported 3.6 μm nightside temperature and visit-consistency claim fail. An independent check would be an injection-recovery test: add a synthetic phase curve with known nightside flux to the actual visit-2 systematic noise and verify which ramp model recovers the injected value.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central assertion is that WASP43 b has higher nightside temperatures, smaller hotspot offsets, and more consistent 3.6 μm visits than Stevenson et al. (2017), implying circulation efficiency ε ~ 0.1–0.3. The load-bearing step is the model-selection choice for the second 3.6 μm visit: all information criteria (BIC, AIC, CAIC, DIC, Bayesian evidence) favor the quadratic ramp model, which yields F_night = (-1.6 ± 1.9)×10^-4, i.e., a near-zero or negative nightside flux. The paper instead adopts the linear ramp, giving F_night = (3.0 ± 1.5)×10^-4, and uses this to compute the weighted 3.6 μm nightside result and the visit-consistency claim. The justification is physical plausibility plus the strong correlation (PCC ~ 0.6) between the nightside flux and the quadratic ramp coefficients. However, that same correlation could indicate that the quadratic ramp is absorbing real long-timescale systematics, in which case the linear model's higher nightside flux is the biased quantity. The paper explicitly states it cannot rule out the quadratic solution as physically impossible (Section 5.2) and does not perform a Gaussian-process cross-check (Section 5.1). The half-phase-curve tests in Appendix C do favor constant/linear ramps for the second visit, which is mitigating but not decisive: selecting models on subsets after seeing the full-data result is not a blind test. Thus the central nightside-temperature and inter-visit-consistency claims, and the resulting circulation-efficiency estimate, rest on a model-selection decision that contradicts the paper's own statistical criteria. This is an internal-consistency issue, not merely a disagreement with prior work: the headline quantity changes sign and significance under a statistically preferred alternative.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an independent reanalysis of three Spitzer/IRAC phase curves of the hot Jupiter WASP43 b (two visits at 3.6 μm and one at 4.5 μm) using the wavelet pixel-ICA blind source-separation method. The authors report higher nightside temperatures, smaller hotspot offsets, greater consistency between the two 3.6 μm visits, and better agreement with atmospheric circulation models than the original Stevenson et al. (2017) analysis. They also examine the dependence of retrieved transit parameters on stellar limb-darkening models, perform half-phase-curve and transit/eclipse-only analyses, and derive an analytical scaling formula for transit-depth uncertainties as a function of observing duration. The central astrophysical conclusion is that the circulation efficiency of WASP43 b is roughly 0.1–0.3 rather than the near-zero value claimed by S17, which would resolve a long-standing mismatch between observations and circulation models.","tokens_in":25688,"tokens_out":4431,"duration_ms":47793,"significance":"If the main claims hold, the paper resolves a significant discrepancy in the hot-Jupiter literature: the extremely low circulation efficiency inferred by Stevenson et al. (2017) has been difficult to reconcile with 3D atmospheric circulation models, and the present analysis provides a credible alternative measurement that brings WASP43 b into line with empirical irradiation-temperature trends. The paper is methodologically transparent and careful in several important ways: it tests multiple ramp models and two ICA implementations, analyzes half phase curves as a consistency check, inflates weighted-average error bars when individual visits disagree, and explicitly discloses limitations, including the possibility that the true uncertainties on the 3.6 μm peak offsets exceed the nominal error bars. These strengths make the analysis a valuable contribution regardless of the ultimate resolution of the model-selection question.","major_comments":[{"comment":"The central nightside-temperature and inter-visit-consistency claims depend on discarding the BIC/AIC/CAIC/DIC/Bayesian-evidence-preferred quadratic ramp model for the second 3.6 μm visit in favor of a linear ramp model. The justification given is physical plausibility (the quadratic solution yields F_night = (-1.6 ± 1.9) × 10^-4) and the strong correlation between the nightside flux and the quadratic ramp coefficients (PCC ~ 0.6). However, the paper explicitly states in Section 5.2 that the quadratic solution 'cannot [be] rule[d] out as physically impossible,' and it does not perform a Gaussian-process cross-check, which it acknowledges as a possible alternative. Because the positive nightside flux (3.0 ± 1.5) × 10^-4 and the resulting weighted 3.6 μm result in Table 4 are direct consequences of this model-selection decision, the claim that the two 3.6 μm visits are consistent at the ~1σ level is not yet established with the level of confidence claimed. I recommend adding a quantitative test—for example, a Gaussian-process detrending of the second visit, or synthetic-injection experiments showing that a known phase-curve signal is recovered without bias under the linear-ramp assumption—or alternatively presenting the nightside flux as model-dependent and substantially weakening the inter-visit consistency claim.","section":"Section 5.1, Table 6"},{"comment":"Even if one accepts the linear-ramp model for the second 3.6 μm visit, the reported uncertainties do not account for model-selection uncertainty. The difference between the linear-ramp and quadratic-ramp nightside fluxes is 3.0×10^-4 minus (-1.6×10^-4) = 4.6×10^-4, which is roughly three times the quoted uncertainty on the individual linear-ramp measurement and larger than the 1.26× inflation factor applied to the weighted F_night^MIN in Table 4. The paper should add a systematic error term representing the difference between plausible ramp models (or report the nightside flux as a range spanning both models), because the current Table 4 understates the total uncertainty in the headline 3.6 μm nightside result.","section":"Section 4.1, Table 4"},{"comment":"The claim that the observations show 'greater similarity with the predictions of the atmospheric circulation models' is overstated relative to the evidence presented. The 3.6 μm dayside flux is higher than all models in the grid by roughly 2–4σ (the paper reports discrepancies of ~150 ppm at 2σ and ~300 ppm at 4σ for solar and 3× solar metallicity), and no single model reproduces both channels within the quoted errors: the best 3.6 μm model has 10× solar metallicity with P_cloud = 2×10^-2 bar, while the best 4.5 μm model has 1× or 3× solar metallicity with P_cloud = 10^-3 bar. The statement that 'a range of models ... can reproduce all of the measured phase curve parameters within less than 2σ' does not support the stronger conclusion of good agreement. The paper should either qualify the atmospheric-model comparison as a loose consistency check or provide a formal statistical comparison (e.g., a joint likelihood over both channels) before claiming that the model grid supports the inferred circulation efficiency.","section":"Section 5.3, Figures 9-10"}],"minor_comments":[{"comment":"There is a typo: 'unphyisical' should be 'unphysical' in the sentence 'We discarded the (unphyisical) results obtained for the second 3.6 μm visit.'","section":"Section 4.1"},{"comment":"The text cites 'Mendonça et al. (2018b)' but does not distinguish it from the first Mendonça et al. (2018) reference at the point of citation; please introduce both papers explicitly (e.g., M18 vs. Mendonça et al. 2018b) in the text so the reader can connect the citations to the reference list.","section":"Section 5.4, References"},{"comment":"The statement that the approximation Fin ≈ Fout affects Delta p2 by less than 0.1% for transit depths up to 3% would benefit from a brief derivation or a reference; as written, the reader cannot verify this bound without re-deriving it.","section":"Appendix A, Equation A5"},{"comment":"The legends and captions use similar color shades (blue and dodger blue) for the quadratic and linear ramp models of the second 3.6 μm visit; adding distinct marker styles (e.g., filled vs. open symbols) would improve accessibility, since the shades are hard to distinguish in printed grayscale.","section":"Figure 4 and Figures 18-20"}],"recommendation":"major_revision","confidential_remarks":"This is a careful and genuinely useful reanalysis; the wavelet pixel-ICA pipeline is a valuable independent check on the Spitzer/IRAC phase curves, and the authors are admirably transparent about the weaknesses of their own model-selection step. The main concern is that the paper's headline claims (higher nightside temperatures, ~1σ consistency between the 3.6 μm visits, and consequent circulation efficiency) are all downstream of a model-selection decision that the authors themselves flag as not definitive. I would encourage the editor to request additional tests rather than reject: a Gaussian-process cross-check or synthetic-injection experiment for the second 3.6 μm visit would go a long way toward making the central claim robust. The paper's scope is a good fit for the journal, and the authors' candid discussion of limitations is a strong positive signal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this. It's a genuinely careful reanalysis of the three Spitzer phase curves of WASP43 b, and it probably gets closer to the truth than Stevenson et al. (2017). The main parameters agree with Mendonça et al. within 1 sigma, which is a good sign. What's actually new here: the limb-darkening degeneracy study (ATLAS vs PHOENIX coefficient sets shift transit depth by 400-700 ppm), the transit-only bias quantification (~100 ppm from ignoring phase-curve modulation), the half-phase-curve consistency checks, and the error-scaling formula in Appendix A. The authors also deserve credit for testing two ICA implementations, multiple ramp models, and for inflating the 3.6 um error bars when the two visits disagreed. They're transparent about the weak spots.\n\nThe soft spot is exactly where the reader's stress test points. For the second 3.6 um visit, every information criterion (BIC, AIC, CAIC, DIC, Bayesian evidence) prefers the quadratic ramp, but that solution gives a negative nightside flux. The authors adopt the linear ramp instead, citing physical plausibility and the strong correlation (PCC ~ 0.6) between nightside flux and quadratic ramp coefficients. That's a reasonable judgment call, but it is a judgment call. If the quadratic ramp is right, the two 3.6 um visits are inconsistent at ~3 sigma and the weighted nightside flux drops, though it stays positive. The paper's broader conclusion — warmer nightside than S17, smaller hotspot offsets — does not collapse, because the first 3.6 um visit and the 4.5 um visit already point that way, and M18 independently found similar results. Still, the Table 4 values at 3.6 um should be read as conditional on that ramp choice.\n\nThe limb-darkening degeneracy is demonstrated but left unresolved; the paper doesn't choose a preferred coefficient set, which is honest but limits the transit parameter claims. The atmospheric model comparison has several tuned parameters (cloud top pressure, metallicity, absorption) and is more illustrative than decisive.\n\nThis deserves a serious referee. I'd send it out, and I'd bring it to the reading group. The main referee request: a Gaussian-process cross-check on the second 3.6 um visit, or at least a sensitivity table showing the weighted 3.6 um results under both ramps. I'd cite the transit-only bias and the error-scaling formula in my own work.","headline":"Careful reanalysis that probably fixes the S17 WASP43 b result, but the 3.6 um nightside/visit-consistency headline rests on discarding the statistically favored ramp model — treat as plausible, not settled.","tokens_in":26323,"tokens_out":4452,"would_cite":true,"duration_ms":44523,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A blind-source reanalysis of WASP43 b's infrared phase curves raises the nightside temperature by hundreds of kelvins and sets the circulation efficiency near 0.1–0.3, not near zero.","keywords":["planets and satellites: individual (WASP43 b)","planets and satellites: atmospheres","planets and satellites: fundamental parameters","stars: atmospheres","techniques: photometric","techniques: spectroscopic","exoplanet phase curves","wavelet pixel-ICA"],"falsifier":"Take a new 4.5 μm phase curve of WASP43 b and reduce it without committing to a ramp model: if the nightside flux again lands near zero, or the hotspot offset returns to a large eastward shift, the warmer nightside reported here is a detrending artifact. A cheaper check is to apply a Gaussian-process detrending, which marginalizes over ramp shapes, to the archived second 3.6 μm visit and see whether the nightside flux remains positive when the quadratic ramp is not set aside by hand.","tokens_in":25104,"feed_emoji":"🪐","tokens_out":17800,"duration_ms":162839,"temperature":0.7,"pith_summary":"WASP43 b is a hot Jupiter on a 19.5-hour orbit, and its infrared phase curves — the planet's brightness tracked continuously around its star — had seemed to show almost no day-to-night heat transport. This paper reanalyzes the same space-telescope data with a blind signal-separation method called wavelet pixel-ICA and argues that the original conclusion was an artifact of how detector systematics were removed. The reanalysis finds a nightside hundreds of kelvins warmer, smaller eastward hotspot offsets, and two previously irreconcilable 3.6 μm visits that now agree at about the 1σ level. If these results stand, WASP43 b's circulation efficiency is roughly 0.1–0.3 rather than near zero, which puts the planet back on the empirical trend connecting stellar irradiation to circulation efficiency and ends a long-standing clash between these phase curves and atmospheric circulation models.","feed_headline":"WASP43 b's nightside is hundreds of kelvins warmer","feed_subtitle":"A blind-source reanalysis of the Spitzer phase curves puts the planet's heat circulation at 0.1–0.3, not near zero.","key_machinery":"The machinery that carries the argument is wavelet pixel-ICA, a blind source-separation method: it applies a one-level Daubechies-4 discrete wavelet transform to each of the 25 pixel time series in the stellar aperture and then decomposes the set into statistically maximally independent components, so that one component carries the transit/eclipse/phase-curve signal and the others absorb the instrumental systematics without a prescribed functional form. The second, equally decisive piece of machinery is the ramp-model choice for the second 3.6 μm visit: BIC, AIC, CAIC, DIC, and Bayesian evidence all prefer a quadratic ramp, but the authors override that selection because the quadratic solution forces the nightside flux to negative values and its two ramp coefficients correlate with the inferred nightside flux at Pearson coefficient ≈ 0.6, whereas the linear-ramp solution is physically plausible and only weakly correlated (PCC ≈ 0.1). A grid of 2D-ATMO circulation models — with 1×, 3×, and 10× solar metallicity, varying cloud-top pressure, and a fixed 4 km/s substellar zonal wind — then provides the comparison surface on which the re-derived parameters succeed where the original parameters failed.","core_discovery":"The central claim is that the earlier portrait of WASP43 b as a nearly stagnant atmosphere — almost no nightside emission and a large eastward hotspot — was an artifact of the detector-ramp modeling. When the raw pixel time series are separated into astrophysical signal and instrumental components by wavelet pixel-ICA, the nightside brightness temperature comes out near 1016 K at 3.6 μm in the first visit and near 837 K in the second visit when a linear ramp is adopted, against the 2σ upper limits reported originally, and near 700 K at 4.5 μm; the hotspot offset at 4.5 μm shrinks to 11.3° ± 2.1° east of the substellar point, and the two 3.6 μm visits become consistent within about 1σ. The statistically preferred quadratic ramp for the second 3.6 μm visit is set aside because it drives the nightside flux unphysically negative and its coefficients are strongly correlated with the inferred nightside flux (Pearson coefficient ≈ 0.6), whereas the linear-ramp solution is physically plausible and only weakly correlated. Interpreting the fluxes as blackbody emission gives a circulation efficiency of about 0.1–0.3 instead of the near-zero value claimed by the original study, and a grid of 2D-ATMO circulation models with a high, thin cloud deck (cloud-top pressure near $10^{-3}$–$10^{-2}$ bar) reproduces the measured phase-curve parameters, while cloud-free models are rejected at 4–8σ in nightside flux.","pith_inferences":["A generalizable procedural lesson follows: when a detector-ramp parameter correlates strongly with an astrophysical parameter (the quadratic-ramp coefficients versus nightside flux, PCC ≈ 0.6), information-criterion rankings can select unphysical models; reporting these correlations alongside BIC/AIC would make future reanalyses of phase-curve data much easier to adjudicate.","If the warmer nightside is real, the apparent differences between visits and between wavelengths become evidence of evolving nightside clouds — a prediction that can be tested directly by a new 4.5 μm observation, which should catch the nightside flux varying between epochs.","The same blind-separation pipeline could be run on archived phase curves of other hot Jupiters previously classified as low-efficiency; some of those classifications may also be detrending artifacts rather than real atmospheric states.","The analytic error-scaling relation of the paper, $\\Delta p^2 \\approx (\\sigma/F)\\sqrt{N_{\\rm tot}/(N_{\\rm in}N_{\\rm out})}$, can be inverted for scheduling: beyond roughly three transit durations of out-of-transit baseline, extra observing time buys little transit-depth precision, so the marginal time is better spent on eclipse and phase coverage."],"forward_implications":["If the reanalysis is right, WASP43 b moves from the canonical example of a non-circulating hot Jupiter to a planet with moderate heat redistribution (ε ≈ 0.1–0.3), in line with the empirical irradiation–efficiency trend established for other planets.","The 4.5 μm data become consistent with circulation models of near-solar metallicity carrying a high, thin cloud deck (cloud-top pressure ≈ $10^{-3}$ bar), and cloud-free models are excluded at 4–8σ in nightside flux.","The two 3.6 μm visits being consistent at about 1σ removes the basis for discarding one of them, so future analyses of this planet can use the full data set.","The retrieved transit parameters are degenerate with the stellar limb-darkening model at the 400–700 ppm level, a systematic floor that will matter for the next generation of space observatories.","Transit-only analyses of these data are biased by roughly 100 ppm in transit depth because of the flat-baseline assumption, and the provided scaling relation $\\Delta p^2 \\approx (\\sigma/F)\\sqrt{N_{\\rm tot}/(N_{\\rm in}N_{\\rm out})}$ lets observers choose baselines that keep this bias below the noise."],"supporting_citations":[{"why":"The original phase-curve analysis whose near-zero nightside fluxes, large hotspot offsets, and discarded 3.6 μm visit this paper revises and compares against.","marker":"S17"},{"why":"The prior reanalysis of the same data whose phase-curve parameters the paper reproduces within 1σ and whose disequilibrium-chemistry explanation the paper replaces with a lower cloud-top pressure.","marker":"M18"},{"why":"Introduces the wavelet pixel-ICA pipeline, the blind source-separation method that carries the entire detrending analysis.","marker":"Morello et al. 2016"},{"why":"Supplies the SPARC/MITgcm circulation models that failed to reproduce the original low-efficiency results, defining the mismatch the paper resolves.","marker":"Kataria et al. 2015"},{"why":"Presents the 2D-ATMO code used to compute the grid of model phase curves that the re-derived parameters match.","marker":"Tremblin et al. 2017"},{"why":"Provides the code and stellar-atmosphere grids used to compute the four limb-darkening sets whose degeneracy with transit parameters is quantified.","marker":"Espinoza & Jordán 2015"},{"why":"Asserts the irradiation–circulation-efficiency trend predicting ε ≈ 0.5 for WASP43 b, the expectation the new estimates move toward.","marker":"Keating & Cowan 2017"},{"why":"Supplies the PHOENIX stellar models used to compute limb-darkening coefficients and to convert measured flux ratios into brightness temperatures.","marker":"Husser et al. 2013"}],"fun_headline_variants":["WASP43 b's nightside gets a thermal boost in reanalysis","New blind-source analysis heats up WASP43 b's nightside","Hot Jupiter WASP43 b's nightside is warmer than reported","Reanalysis of WASP43 b shows a more active atmosphere"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The revision rests on one modeling choice: for the second 3.6 μm observation the paper sets aside the quadratic detector-ramp model that every statistical test prefers, because it drives the nightside flux to unphysical negative values, and adopts the linear ramp instead; if the quadratic ramp is the true detector behavior, the second-visit nightside is consistent with zero and the 3.6 μm nightside temperature claim loses its main pillar.","fun_headline_variants_meta":{"raw":{"variants":["WASP43 b's nightside gets a thermal boost in reanalysis","New blind-source analysis heats up WASP43 b's nightside","Hot Jupiter WASP43 b's nightside is warmer than reported","Reanalysis of WASP43 b shows a more active atmosphere"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000545,"raw_usage":{"total_tokens":2727,"prompt_tokens":1185,"completion_tokens":1542,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":801,"completion_tokens_details":{"reasoning_tokens":1467}},"tokens_in":801,"tokens_out":1542,"duration_ms":12228,"temperature":1.0,"reasoning_tokens":1467,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:35:57.002093+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a new 4.5 μm phase curve of WASP43 b and reduce it without committing to a ramp model: if the nightside flux again lands near zero, or the hotspot offset returns to a large eastward shift, the warmer nightside reported here is a detrending artifact. A cheaper check is to apply a Gaussian-process detrending, which marginalizes over ramp shapes, to the archived second 3.6 μm visit and see whether the nightside flux remains positive when the quadratic ramp is not set aside by hand.","supporting_citations":[{"cited_title":"P., & Tinetti, G","cited_arxiv_id":null,"evidence_quote":"Introduces the wavelet pixel-ICA pipeline, the blind source-separation method that carries the entire detrending analysis."},{"cited_title":"P., Fortney, J","cited_arxiv_id":null,"evidence_quote":"Supplies the SPARC/MITgcm circulation models that failed to reproduce the original low-efficiency results, defining the mismatch the paper resolves."},{"cited_title":"J., et al","cited_arxiv_id":null,"evidence_quote":"Presents the 2D-ATMO code used to compute the grid of model phase curves that the re-derived parameters match."},{"cited_title":"2015, , 450, 1879","cited_arxiv_id":null,"evidence_quote":"Provides the code and stellar-atmosphere grids used to compute the four limb-darkening sets whose degeneracy with transit parameters is quantified."},{"cited_title":"& Cowan, N","cited_arxiv_id":null,"evidence_quote":"Asserts the irradiation–circulation-efficiency trend predicting ε ≈ 0.5 for WASP43 b, the expectation the new estimates move toward."}],"review_version":1}