{"id":"b9c3393a-ff3a-41a6-9c11-12612648ae3a","arxiv_id":"2506.10743","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"External photoevaporation at moderate FUV fluxes (2 to 12 G0) truncates Upper Scorpius disks and explains the compact gas sizes seen by AGE-PRO, where pure viscous evolution fails.","lead":"This paper models the ten Upper Scorpius disks from the AGE-PRO survey with and without external photoevaporation, and finds that the pure viscous model cannot explain their small gas sizes, while adding photoevaporation reproduces 7 of 10 sizes within a factor of two. It estimates each disk's FUV flux from nearby massive stars (2 to 12 G0) and argues that even moderate UV environments reshape disk evolution.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 7/10 gas-size match rests on an FUV-interpolated CO-radius proxy (Eq. 9) whose systematic uncertainty is unquantified; a DALI-based recheck of R12CO should precede reliance on the count.","rationale":"The paper has two layers: a physical mechanism (external photoevaporation truncates disks and suppresses viscous spreading at 1-10 G0) and a quantitative demonstration (7/10 gas sizes reproduced within factor <2). The first layer is well supported by the surface-density evolution in Figures 3-5, by the FRIEDv2 mass-loss prescription, and by prior work; I do not object to it. The second layer, which is the headline result, depends on the R12CO proxy of Eq. (9) with an FUV interpolation that is not accompanied by a validated uncertainty. This is exactly the load-bearing uncertainty the reader identified. Secondary concerns, such as per-source selection over the alpha/M0/age grid and the lack of propagated uncertainties in the 7/10 statistic, reinforce the need for a statistical treatment but are less decisive than a possible systematic in the radius proxy itself. The proposed DALI recheck would settle whether the 7/10 count is robust; until then, CONDITIONAL is the right verdict, and no change to the reader's verdict is required.","tokens_in":31472,"tokens_out":6430,"duration_ms":80783,"concrete_test":"Re-run the size comparison after replacing the FUV-interpolated Eq. (9) with DALI post-processing on the stored DustPy snapshots: for each of the ten targets at its median FUV flux, and at the 16th/84th percentile fluxes for marginal sources, compute the 12CO 90% radius directly from the simulated gas surface density using the same CO abundance assumptions as Trapman et al. (2023). Compare these radii with the Eq. (9) + Figure 2 values. If the DALI-based radii differ by more than ~0.3 dex, or if the number of sources within a factor of 2 drops by more than two, the central quantitative claim should be revised to a qualitative statement and the CONDITIONAL verdict should require this check. A cheaper diagnostic is to recompute the success count with Eq.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 4.4, R12CO in the models is not computed from CO radiative transfer but from the semi-empirical critical column density fit of Eq. (9), calibrated with DALI at FUV = 1 G0, then adjusted to each source's 2-12 G0 field by interpolating Figure 2 of Trapman et al. (2023). The text itself notes this relation depends on initial carbon abundance and disk temperature. Since Eq. (10) sets R12CO by equating the model surface density to this critical value, any bias in log N_gas translates nearly one-to-one into a bias in the predicted CO size; for a typical exponential profile, a 0.3 dex error in Sigma_crit shifts R12CO by roughly 10-25 AU at R ~ 50-100 AU. The FUV correction is applied only in the external-photoevaporation runs, so it preferentially makes those runs look more compact. If the FUV interpolation overestimates N_gas(R12CO) at moderate FUV for low-mass disks, the headline '7/10 within factor <2' is inflated. No uncertainty is propagated from the carbon abundance, temperature, or interpolation, and the success count is quoted without error bars. The qualitative truncation effect is independently visible in the surface-density evolution (Figures 3-5); what is not yet secure is the quantitative match to the observed CO radii.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript studies whether external photoevaporation can explain the gas and dust properties of the ten AGE-PRO disks in Upper Scorpius. The authors compute the FUV flux at each disk from the surrounding OB stars, obtaining median fluxes between about 2 and 12 G0, and run 1D viscous evolution models with an external photoevaporation module based on the FRIEDv2 mass-loss grid. They compare the predicted gas masses, CO radii, millimeter fluxes, and dust radii at 2 and 6 Myr with the AGE-PRO observations. The central result is that pure viscous models fail to reproduce the observed small gas radii, while models including external photoevaporation reproduce the gas disk sizes for 7 of 10 sources within a factor of 2, assuming initial disk masses of 1-10% of the stellar mass. The dust masses and radii are not reproduced, which the authors attribute to the absence of dust substructures in the models.","tokens_in":31822,"tokens_out":6896,"duration_ms":74039,"significance":"If the quantitative claim is robust, the paper makes an important contribution: it shows that even moderate FUV irradiation of 1-10 G0 can suppress viscous spreading and truncate protoplanetary disks, so Upper Sco disks cannot be treated as a clean age-only comparison sample for disks in low-irradiation regions. The qualitative evolution shown in Figures 3-5, where external photoevaporation halts viscous spreading at moderate fluxes, is convincing and independently visible in the surface-density evolution. The paper is also transparent: it reports the dust-model failure clearly and identifies dust substructures as the likely missing ingredient. The main quantitative claim, however, rests on a semi-empirical CO-radius proxy whose FUV extension is not validated or assigned an uncertainty; this weakens the headline 7/10 statistic until the proxy is checked against thermochemical models at the actual FUV fluxes.","major_comments":[{"comment":"The headline success metric (7/10 gas sizes within a factor <2) is evaluated against a predicted R12CO obtained from the semi-empirical critical column density relation of Eq. (9), which is calibrated with DALI thermochemical models at FUV = 1 G0 and then adjusted to the source-specific 2-12 G0 fields by interpolating Figure 2 of Trapman et al. (2023). The text itself notes that Eq. (9) depends on the initial carbon abundance and disk temperature, and no uncertainty from those dependencies or from the FUV interpolation is propagated into the reported success count. Because Eq. (10) sets R12CO at the radius where Sigma = Sigma_crit, a systematic offset in log N_gas translates almost directly into an offset in R12CO, and because the FUV correction is applied only in the external-photoevaporation runs, any bias in the interpolation preferentially makes those runs appear more compact. Before the quantitative 7/10 claim can be relied upon, the authors should either recompute R12CO with DALI (or an equivalent thermochemical model) at the actual FUV fluxes and disk masses of the sample, or quantify and propagate the systematic uncertainty from carbon abundance, temperature, and the interpolation.","section":"Section 4.4, Eqs. (9)-(10)"},{"comment":"The paper states in Section 5.2 that UppSco 2 is predicted to be almost entirely dispersed by external photoevaporation in less than 6 Myr, and the caption of Figure 8 indicates that no model point is connected to this source; the observed source, however, has a measured 12CO radius of about 51 AU at an age of about 2.1 Myr. The conclusion counts 7 out of 10 sizes as reproduced and lists only UppSco 3, 4, and 5 as predicted too large, which implies UppSco 2 is counted as a success or is silently excluded. This inconsistency needs to be resolved explicitly: if UppSco 2 is a failure, the success rate is 6/10; if it is excluded, the rationale (e.g., large FUV flux uncertainty) should be stated and the headline count adjusted accordingly.","section":"Section 5.2 and Section 7, item 6"},{"comment":"The comparison is made only at the two fixed ages t = 2 and 6 Myr, while the individual isochronal ages in Table 1 range from about 1.8 to about 5.9 Myr with asymmetric uncertainties that in several cases span a factor of several in age. For sources whose predicted radius evolves strongly between 2 and 6 Myr, the statement that the radius is reproduced within a factor <2 depends on which endpoint (or intermediate age) is used. The authors should quantify how many of the ten sources have a model radius within a factor <2 of the observed value at an age consistent with the source's isochronal age and its uncertainties, rather than quoting a binary success count at two arbitrarily chosen ages.","section":"Section 5.2, Figures 6-8"}],"minor_comments":[{"comment":"The word 'protoplaneraty' should be 'protoplanetary'.","section":"Section 2"},{"comment":"The phrase 'a greater decrease in mass (hundreds of M_sun)' appears to use solar masses; in the context of protoplanetary disk masses it should presumably read 'hundreds of Earth masses' (hundreds of M_Earth).","section":"Section 5.1"},{"comment":"The text 'M_gas,0 = 0.1 M_sun' should presumably be '0.1 M_star', because the parameter grid in Table 3 defines M_disk,0 in units of stellar mass, and 0.1 M_sun would exceed the stellar mass for several of the low-mass targets.","section":"Section 5.3"},{"comment":"The diagonal labels 'Rdust', '5 Rdust', and '10 Rdust' are visually confusing; please relabel them with explicit subscripts or with text such as 'R_gas = 5 R_dust'.","section":"Figure 13"},{"comment":"A single representative luminosity L_star = 0.3 L_sun is used for all models even though the stellar masses in Table 1 range from about 0.13 to 0.56 M_sun; a short sensitivity test or an estimate of the resulting offset in the predicted R12CO would help the reader assess this approximation.","section":"Section 4.1, Eq. (3)"},{"comment":"In the sentence 'with the latest showing gas mass consistent with the model', 'latest' should be 'latter'.","section":"Section 6.1"}],"recommendation":"major_revision","confidential_remarks":"The qualitative conclusion that moderate FUV irradiation truncates disks and that pure viscous models fail for Upper Sco is well supported by the surface-density evolution and is likely robust. The quantitative 7/10 success count, however, depends on the unvalidated FUV extension of Eq. (9), and the counting of UppSco 2 needs clarification. I would recommend insisting on a thermochemical recheck or a full uncertainty propagation for the CO-radius proxy before the paper is accepted; the qualitative claims alone would justify publication after those points are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline is that external photoevaporation at moderate FUV fluxes (2–12 G0) does a genuinely good job of explaining why Upper Sco gas disks are compact, and the paper is honest about where it fails. The new content is the per-source FUV flux calculation for the ten AGE-PRO disks and the explicit model–observation comparison; Trapman et al. had identified the size discrepancy but had not included external photoevaporation. The simulation setup is standard, the FRIEDv2 prescription is the right tool, and the authors report the dust failure plainly instead of sweeping it under the rug.\n\nThe soft spots are real but not fatal. The headline '7/10 gas sizes within a factor <2' is measured against the semi-empirical CO radius proxy of Eq. (9), which is calibrated at FUV = 1 G0 and then interpolated to each source's FUV level. The interpolation only moves the external-photoevaporation runs, so a bias there would preferentially make those runs look more compact. The text notes the proxy depends on carbon abundance and temperature, but the success count carries no uncertainty from these. The paper itself flags this, but it stays a quantitative weakness. I also note the post hoc selection of display cases in Section 5.3 and the small sample for the correlations; neither is disqualifying.\n\nThat said, the qualitative conclusion does not rest on the 7/10 alone. The surface density evolution in Figures 3–5 shows truncation and suppressed spreading even at 1–10 G0, and the model fails for the most compact disks (UppSco 3, 4, 5) by predicting them too large rather than too small. The case for external photoevaporation being relevant in Upper Sco survives even if the proxy is off by a few tens of percent.\n\nWho is this for: disk evolution people, ALMA survey folks, and anyone comparing star-forming regions by age alone. It deserves a serious referee. The main request should be a DALI-based recheck of the radius proxy at the actual FUV fluxes, plus propagated uncertainties on the success count. I would accept peer review.","headline":"A solid, honest test showing moderate-FUV external photoevaporation can explain Upper Sco's compact gas disks; the qualitative result holds, but the 7/10 success count depends on an uncalibrated CO-radius proxy interpolation.","tokens_in":32471,"tokens_out":3142,"would_cite":true,"duration_ms":35229,"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":"External photoevaporation at moderate FUV fluxes reproduces the compact gas disks of Upper Scorpius, matching 7 of 10 observed CO radii within a factor of 2.","keywords":["protoplanetary disks","external photoevaporation","disk radii","disk masses","Upper Scorpius","FUV irradiation","viscous disk evolution","12CO emission"],"falsifier":"Run the same simulations but replace the analytic critical column density with full thermochemical post-processing at each source's actual FUV flux of 2 to 12 $G_0$ and recompute the modeled 12CO radii; if the 7-of-10 within-factor-of-2 agreement disappears, the central claim fails. A targeted version is to measure the 12CO column-density turnover radius directly in UppSco 3, 4, and 5, the sources the model overpredicts, and test whether the analytic relation overestimates the critical column there.","tokens_in":31256,"feed_emoji":"🪐","tokens_out":11807,"duration_ms":127232,"temperature":0.7,"pith_summary":"The paper sets out to explain a puzzle: the protoplanetary disks in Upper Scorpius, the oldest nearby star-forming region, are smaller than viscous evolution predicts at their ages. It argues that external photoevaporation by moderate far-ultraviolet radiation, 2 to 12 $G_0$ from nearby B-type stars, is the missing process. In one-dimensional viscous simulations with an outside-in photoevaporative wind, the modeled gas disk radii match the observed 12CO radii within a factor of 2 for 7 of the 10 AGE-PRO disks, provided the initial disk mass is between 1% and 10% of the stellar mass. Pure viscous models fail on the same comparison. The paper concludes that even moderate irradiation truncates disks and suppresses viscous spreading, so Upper Sco cannot be treated as an age-only endpoint of lower-irradiation regions.","feed_headline":"External UV explains 7 of 10 compact disks in Upper Sco","feed_subtitle":"Adding external photoevaporation to viscous models reproduces gas disk radii where pure viscous evolution fails.","key_machinery":"The central machinery is a one-dimensional viscous disk evolution model with an external photoevaporative wind added as an outside-in sink term. Mass loss is set by interpolating the FRIEDv2 grid of photoevaporation rates, a precomputed grid of mass-loss rates as a function of stellar mass, FUV flux, and disk surface density; the truncation radius is where that rate peaks, and gas exterior to it is removed weighted by the mass outside the truncation radius. The observable counterpart is the modeled CO radius, obtained not from full thermochemistry but from an analytic critical column density relation that gives the H2 column at which 12CO stops self-shielding, interpolated to each source's FUV flux. The FUV input fluxes are computed by summing blackbody FUV luminosities of all OB stars in a wide ~55-degree box, with parallax-based distances and Monte Carlo uncertainties.","core_discovery":"On the paper's own terms, the discovery is that the compact gas disks of Upper Scorpius are a signature of ongoing external photoevaporation, not a pure evolutionary stage. The authors estimate the FUV flux at each of the ten AGE-PRO targets from the summed contributions of all OB stars in a wide box around the region, obtaining median fluxes from about 2 to 12 $G_0$. They then evolve disks with viscous spreading plus a photoevaporative mass loss that removes gas outside a truncation radius. The resulting radius enclosing 90% of the 12CO flux agrees with the observed gas radius within a factor of 2 for 7 of 10 sources when the initial disk mass is 1 to 10% of the stellar mass, while the same models without external photoevaporation predict substantially larger disks. The paper does not claim viscosity is the accretion driver; it claims that environmental irradiation must be included before drawing conclusions about disk evolution from size measurements.","pith_inferences":["Beyond the paper: applying the same model to Lupus disks at their measured FUV fluxes would test the environmental term directly, since Lupus should show milder radius suppression and the Lupus-versus-Upper Sco size gap should close when both are run at the same flux.","Beyond the paper: the convergence of final disk size across initial radii suggests that in irradiated regions a measured CO radius may be inverted to estimate the present-day FUV field, offering a check of photoevaporation that does not rely on stellar catalogs.","Beyond the paper: the delayed-exposure test implies that disks carry no lasting memory of irradiation before roughly 1 Myr once they reach equilibrium, so cluster kinematics and migration history matter more than birth environment for observed sizes.","Beyond the paper: the marginal anti-correlation between FUV flux and disk size should strengthen in a larger sample; a survey of 50 or more Upper Sco disks could turn the trend into a statistically established constraint."],"forward_implications":["Gas disk size, not mass, is the earliest and clearest observable signature of moderate external photoevaporation: in the test simulations size suppression appears within 2 to 4 Myr, while mass differences become significant only after 6 to 8 Myr.","Upper Scorpius disks should be modeled as environmentally irradiated objects, not as the old-age descendants of regions like Lupus; comparisons of disk populations across star-forming regions must include irradiation level alongside age.","The initial disk radius cannot be recovered from final sizes in irradiated regions, because disks with initial characteristic radii of 10, 40, and 100 AU converge to nearly the same radius once the photoevaporative mass loss balances accretion.","The persistence of the gas-to-dust size discrepancy after photoevaporation is included implies that unresolved dust substructures or pressure bumps are still required to slow radial drift in these disks.","The model predicts that UppSco 2, exposed to roughly 12 $G_0$ around a 0.13 solar-mass star, should be nearly dispersed by 6 Myr; its observed detection points to a lower or time-varying FUV exposure."],"supporting_citations":[{"why":"Provides the FRIEDv2 grid of external photoevaporation mass-loss rates used to set the wind loss rate and truncation radius in every simulation.","marker":"Haworth et al. 2023"},{"why":"Supplies the outside-in depletion scheme that weights mass loss by the material exterior to the truncation radius.","marker":"Sellek et al. 2020"},{"why":"Gives the analytic critical column density relation used to convert modeled gas surface density into the observable 12CO radius.","marker":"Trapman et al. 2023"},{"why":"Supplies the empirical FUV luminosity-to-stellar-mass relation used to compute the incident flux at each target.","marker":"Winter & Haworth 2022"},{"why":"Documents the earlier finding that Upper Sco gas disks are smaller than viscous predictions, the baseline this paper addresses.","marker":"Trapman et al. 2020"},{"why":"Shows that MHD wind-driven models also struggle to reproduce Upper Sco sizes, motivating external photoevaporation as the alternative.","marker":"Trapman et al. 2022"},{"why":"Provides the membership catalog of Upper Sco OB stars from which the FUV field is computed.","marker":"Luhman & Esplin 2020"},{"why":"Implements the external photoevaporation module in the dust-and-gas code that this paper adapts to the FRIEDv2 grid.","marker":"Gárate et al. 2024"}],"fun_headline_variants":["External UV explains 7 of 10 compact disks in Upper Sco","Upper Sco disks' small sizes traced to external photoevaporation","Photoevaporation beats viscosity in shaping Upper Sco disks","UV from massive stars drives disk shrinkage: 7/10 match"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that 7 of 10 disks are reproduced hangs on the analytic relation that turns a simulated gas surface density into a CO radius, a relation calibrated at 1 $G_0$ and then interpolated to each source's higher FUV flux; if that proxy is biased for moderate-flux, low-mass disks, the match could be an artifact of the comparison rather than evidence for photoevaporation.","fun_headline_variants_meta":{"raw":{"variants":["External UV explains 7 of 10 compact disks in Upper Sco","Upper Sco disks' small sizes traced to external photoevaporation","Photoevaporation beats viscosity in shaping Upper Sco disks","UV from massive stars drives disk shrinkage: 7/10 match"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00027,"raw_usage":{"total_tokens":1664,"prompt_tokens":1023,"completion_tokens":641,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":639,"completion_tokens_details":{"reasoning_tokens":568}},"tokens_in":639,"tokens_out":641,"duration_ms":7566,"temperature":1.0,"reasoning_tokens":568,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:19:52.467706+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same simulations but replace the analytic critical column density with full thermochemical post-processing at each source's actual FUV flux of 2 to 12 $G_0$ and recompute the modeled 12CO radii; if the 7-of-10 within-factor-of-2 agreement disappears, the central claim fails. A targeted version is to measure the 12CO column-density turnover radius directly in UppSco 3, 4, and 5, the sources the model overpredicts, and test whether the analytic relation overestimates the critical column there.","supporting_citations":[{"cited_title":"2022, ApJ, 926, 61, doi: 10.3847/1538-4357/ac3ed5","cited_arxiv_id":null,"evidence_quote":"Shows that MHD wind-driven models also struggle to reproduce Upper Sco sizes, motivating external photoevaporation as the alternative."}],"review_version":1}