{"id":"8185a394-f7c4-4423-b183-4fe56b24ac4f","arxiv_id":"2411.14071","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"May 2024's superstorm was followed by an unusual thermospheric density undershoot in the north, attributed to record nitric oxide radiative cooling.","lead":"Using Swarm satellite data, the authors report a rare post-storm density dip of up to 23% in the northern thermosphere on May 12, 2024, after the strongest geomagnetic storm in 20 years. They link the dip to record-high infrared cooling by nitric oxide, which matters for satellite drag forecasting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The NO-overcooling attribution for May 2024 is not quantitatively supported because Joule heating for this event was never computed; on May 12 the density undershoot could reflect recovery dynamics rather than excess radiative cooling.","rationale":"The load-bearing assumption is exactly the one the reader identified: attributing the May 12 density undershoot to NO overcooling requires that Joule heating and particle precipitation had decayed to prestorm levels by May 12, but Section 3.4 explicitly declines to compute Joule heating for this event. I agree with that assessment. The Swarm density depletion and SABER NO enhancements are direct observations and are genuine supporting evidence; the paper should get credit for showing the depletion across three satellites and for reporting the SABER cooling enhancement. What is missing is the quantitative energy-balance step that would separate 'a density decrease during recovery' from 'an overcooling caused by NO.' The reader's CONDITIONAL verdict is appropriate: the observational core can be accepted, but the causal headline requires an additional computation of the May 2024 heating budget (or a physics-based simulation) before it is demonstrated. I do not see an internally inconsistent step that would force rejection; the gap is a missing analysis, not a contradiction. I also note the abstract-only SABER density claim at 105-110 km has no supporting figure in the body, but that is a secondary reporting issue, not the load-bearing concern.","tokens_in":14224,"tokens_out":6855,"duration_ms":65490,"concrete_test":"Estimate the high-latitude Joule heating (and, if possible, particle heating) for 10-12 May 2024 using the Knipp et al. (2004) parameterization with provisional Dst and PCN/AE indices, and compare the daily integrated heating with the SABER-derived NO radiative power for May 12. If the May 12 heating is at or above prestorm levels while density undershoots, the NO-overcooling attribution fails; if heating is clearly below prestorm while NO cooling remains 2-4 times prestorm, the claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the -23% Swarm-C density depletion on May 12 is caused by enhanced NO infrared cooling during recovery—requires the premise that all storm-time heating sources had returned to prestorm levels by May 12. The paper itself states in Section 3.4: 'The Joule heating rates were not calculated for the May 2024 geomagnetic superstorm because of the unavailability of definitive Dst-index and PCN-index.' The premise is therefore transferred from Halloween 2003 rather than checked for this storm. This matters because 'overcooling' is an energy-balance statement: the observed density decrease is real, but it only demonstrates overcooling if net radiative cooling exceeds net heating during the recovery phase. Without a Joule/particle heating estimate for 10-12 May, the undershoot could equally reflect the normal recovery of thermospheric circulation/composition or a temporary reduction in high-latitude energy input, with NO cooling present but not necessarily the controlling term. No quantitative link is made between the SABER NO cooling power on May 12 and the magnitude of the density deficit; the quoted daily radiative power (12.35 x 10^11 W) is for May 11, while the depletion is reported for May 12, and the text only states qualitatively that NO IRF stayed 2-4 times prestorm on May 12. The peak-flux comparison with Halloween 2003 is interesting but does not establish cause.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an observational analysis of thermospheric neutral density (Swarm-A, -B, -C) and thermospheric NO infrared radiative flux (TIMED/SABER) during the May 2024 geomagnetic superstorm. The authors identify a post-storm density depletion in the northern polar region on May 12, reaching about -23% relative to May 9 in Swarm-C, despite rising solar EUV flux. They attribute this depletion to enhanced NO radiative cooling during the recovery phase, calling it a rare thermospheric overcooling event. The paper also reports an all-time high NO IRF of 11.84 ergs/cm2/s and compares the event with the Halloween 2003 storms.","tokens_in":14486,"tokens_out":4940,"duration_ms":45367,"significance":"If the attribution to NO radiative cooling is correct, this would be the first documented thermospheric overcooling event in solar cycle 25 and would strengthen the case that NO cooling plays a controlling role in the recovery of the thermosphere after extreme storms. The analysis uses publicly available, independent datasets, and the density depletion is internally consistent across Swarm-A and -C, which is a notable strength. The comparison with Halloween 2003 is useful for context. However, the central causal claim is not quantitatively supported because the paper does not compute Joule heating for May 2024, does not connect the NO cooling power on the depletion day to the observed density deficit, and provides no uncertainty estimates.","major_comments":[{"comment":"The attribution of the May 12 density depletion to NO overcooling rests on the premise that Joule heating and particle precipitation had returned to prestorm levels by May 12. The paper itself states, 'The Joule heating rates were not calculated for the may 2024 geomagnetic superstorm because of the unavailability of definitive Dst-index and PCN-index.' Without a Joule heating estimate (or at least an upper bound) for 10-12 May, the density undershoot could equally reflect the normal recovery of thermospheric circulation and composition, or a temporary reduction in high-latitude energy input, rather than excess radiative cooling. The analogy to Halloween 2003, where Joule heating was computed and shown to decay, does not establish the premise for this event because storm morphology differs. Since 'overcooling' is fundamentally an energy-balance statement, this missing term is load-bearing for the central claim.","section":"3.4"},{"comment":"No quantitative link is made between the NO radiative cooling power and the magnitude of the observed density depletion. The daily NO radiated power of 12.35 x 10^11 W is quoted for May 11, while the density depletion is reported on May 12; for May 12 the text only states qualitatively that NO IRF remained 2-4 times prestorm. The authors should provide either a time-resolved energy budget (cooling power versus heating power over 10-12 May) or a scaling argument that relates the radiative energy loss to the density change. As written, the causal statement is unsupported by the data shown.","section":"3.2-3.3"},{"comment":"The abstract claims that 'TIMED/SABER observed thermospheric density between 105 and 110 km altitude shows an enhancement during this thermospheric overcooling,' but this observation is not presented or analyzed anywhere in the main text, figures, or tables. If this SABER density measurement is intended as evidence for the overcooling mechanism, it must be added and discussed; otherwise, the claim should be removed from the abstract.","section":"Abstract"},{"comment":"The reported depletion percentages (-23%, -22%, etc.) are given without uncertainties or significance testing. The density data are derived from accelerometer and precise orbit determination, which carry known errors, and the chosen reference day (May 7, 8, or 9) changes the depletion magnitude from about -9% to -23%. The authors should quantify the statistical significance of the depletion relative to natural variability (e.g., quiet-day density variation) to support the claim that this is a rare and notable event.","section":"3.3"}],"minor_comments":[{"comment":"There is a typo in the data description: 'datasets ate utilized' should be 'datasets are utilized.' Also, 'Solar Heliospheric Observatory (SOHO) Solar EUV Monitor (SEM)' should likely be 'Solar and Heliospheric Observatory.'","section":"2"},{"comment":"The text states 'The high pressure solar wind of about 50 hPa' - the unit hPa is inconsistent with typical solar wind dynamic pressure values (nPa). Please correct the unit or the value, and ensure it matches Figure 1.","section":"3.1"},{"comment":"The sentence 'The Joule heating rates were not calculated for the may 2024 geomagnetic superstorm' contains a capitalization error ('may' should be 'May').","section":"3.4"},{"comment":"The claim of an 'all time high' NO radiative cooling flux of 11.84 ergs/cm2/s is compared only with the Halloween 2003 storms. Since SABER has been operating since 2002, the authors should verify this against the full SABER record or cite a study that does so, rather than relying on a single historical comparison.","section":"Abstract / 3.2"},{"comment":"In the Figure 1 caption, 'magneta' should be 'magenta.'","section":"Figure 1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is an interesting data report on a high-impact event, and the density depletion is worth documenting. However, the overcooling attribution is currently an interpretation rather than a demonstrated energy balance. The authors should either add a Joule heating estimate (even provisional) or reframe the conclusion as 'consistent with NO cooling' rather than 'caused by NO overcooling.' The abstract's unsupported SABER density claim should also be fixed. This is repairable within the scope of the paper, hence major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things. First, the May 2024 thermospheric density depletion is real: Swarm-A and -C both show a northern-hemisphere post-storm undershoot of roughly -23% relative to May 9, and the Swarm-B data roughly corroborates it. That is a solid observational result, and the direct comparison with Halloween 2003 is genuinely useful. Second, the paper's attribution of that undershoot to NO overcooling is not quantitatively established: the authors state in Section 3.4 that Joule heating was not calculated for May 2024 due to missing definitive Dst and PCN indices. Overcooling is an energy-balance statement, and without a heating estimate for 10-12 May you can't show that radiative cooling exceeded net heating. The undershoot could partly reflect recovery dynamics or a temporary lull in high-latitude energy input. That is the load-bearing soft spot, and it is not a minor one.\n\nWhat the paper does well: the density analysis is careful and internally consistent. The authors normalize with NRLMSISE-00, consider EUV flux, and show the depletion appears despite rising EUV. The SABER NO infrared flux data show enhanced cooling persisting 2-4x above prestorm levels through May 12, which is exactly what you would want to see for an overcooling candidate. The comparison with the 2003 Halloween storms, for which Joule heating was actually computed, provides a useful template even if it can't substitute for a May 2024 heating estimate.\n\nThe softer spots: the abstract includes a claim about SABER density at 105-110 km that I could not find in the body—that needs to be fixed. The daily NO radiated power quoted for May 11 is not the same day as the reported May 12 depletion, and the text only qualitatively connects the two. There are also no error bars on the density depletion percentages, which matters when the peak depletion is -23% and the reference-day choice shifts it to -8% or -16%.\n\nWho is this for: space-weather and thermosphere-density people tracking storm response, and anyone working on satellite drag models. It is a useful case study, not a definitive mechanism paper. A serious referee should engage with it—the data are public, the question is important, and the missing heating estimate is addressable. My recommendation: send to peer review, with major revision requiring a Joule heating estimate for May 2024 (even from provisional indices or a proxy) and a reconciliation of the abstract's SABER density claim.","headline":"A real and well-documented density undershoot for the May 2024 superstorm, but the paper doesn't quantitatively close the case that NO radiative cooling caused it—Joule heating for this event is never computed.","tokens_in":15052,"tokens_out":1905,"would_cite":true,"duration_ms":19991,"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":"The May 2024 geomagnetic superstorm produced a rare post-storm thermospheric overcooling, with Swarm-C measuring a 23% density depletion on May 12 despite rising solar EUV, likely driven by record nitric-oxide infrared cooling.","keywords":["thermospheric overcooling","nitric oxide radiative cooling","geomagnetic superstorm","thermospheric density depletion","Swarm satellites","TIMED/SABER","Halloween storms","solar EUV"],"falsifier":"A direct calculation or retrieval of Joule heating for May 10-12 using alternative magnetic-activity proxies (for example provisional AE, SuperMAG, or assimilated indices) that shows heating still above pre-storm levels on May 12 would undermine the overcooling attribution, as would a finding that NO cooling at the sampled local times and latitudes on May 12 was not actually elevated.","tokens_in":13990,"feed_emoji":"🛰️","tokens_out":6653,"duration_ms":55240,"temperature":0.7,"pith_summary":"The paper argues that the May 2024 geomagnetic superstorm, the strongest of solar cycle 25, caused a rare post-storm overcooling of the upper atmosphere, an event previously documented only during the 2003 Halloween storms. Using Swarm satellite density data, it finds that northern-hemisphere thermospheric density on May 12 fell about 23% below its May 9 pre-storm level even while solar extreme-ultraviolet radiation was rising. It attributes this undershoot to elevated infrared radiative cooling by nitric oxide (NO) during the storm's recovery phase, and reports the highest NO cooling flux ever observed by TIMED/SABER, about $11.84\\ \\mathrm{ergs\\,cm^{-2}\\,s^{-1}}$, slightly exceeding the 2003 Halloween-storm peak. If correct, this confirms that overcooling is not unique to the Halloween storms and that NO radiative cooling can dominate the thermospheric energy budget in the recovery phase of an extreme storm.","feed_headline":"May 2024 superstorm cooled the thermosphere by 23%","feed_subtitle":"May 2024 superstorm left a 23% density deficit despite rising solar EUV, a sign of NO-driven overcooling.","key_machinery":"The central mechanism is NO infrared radiative cooling: NO molecules, produced in abundance by auroral particle precipitation, are vibrationally excited by collisions with atomic oxygen and then emit $5.3\\,\\mu\\mathrm{m}$ photons that escape to space, converting thermospheric thermal energy into radiation. The paper combines TIMED/SABER limb radiances in the $5.3\\,\\mu\\mathrm{m}$ channel, integrated over 115--250 km to give NO infrared flux and then globally integrated to give daily radiative power, with Swarm accelerometer-derived mass densities normalized to a reference altitude via the NRLMSISE-00 model and referenced to pre-storm days (May 7, 8, or 9) to quantify the density undershoot.","core_discovery":"The central claim is that the May 2024 superstorm produced a post-storm thermospheric density depletion of roughly $-23\\%$ (Swarm-C, northern polar region, May 12) relative to the pre-storm day May 9, observable across Swarm-A, -B, and -C in the northern hemisphere but absent in the southern winter hemisphere. Because solar EUV flux kept increasing through the event, the expected density response was a rise, not a fall; the paper attributes the observed undershoot to NO $5.3\\,\\mu\\mathrm{m}$ radiative cooling that remained 2--4 times above pre-storm levels through May 12, long after the storm's Joule heating had subsided. It also reports the largest NO infrared cooling flux on record, $11.84\\ \\mathrm{ergs\\,cm^{-2}\\,s^{-1}}$, slightly above the $11.74\\ \\mathrm{ergs\\,cm^{-2}\\,s^{-1}}$ peak of the 2003 Halloween storm, and a higher average NO radiative power (about $9.27\\times10^{11}$ W) than in either phase of the Halloween storms.","pith_inferences":["If the attribution holds, thermospheric density models used for satellite drag predictions should be checked for a systematic post-storm low bias after storms of Dst near $-412\\ \\mathrm{nT}$; overcooling could translate into roughly $23\\%$ less drag in the northern hemisphere a day after storm peak.","The paper did not compute Joule heating for May 2024 because definitive Dst and PCN indices were unavailable, so the claim that heating had returned to pre-storm levels by May 12 is inferred by analogy with Halloween 2003; a direct reconstruction using provisional indices or other magnetic-activity proxies would test whether the residual density deficit is truly all NO cooling.","The SABER density enhancement at 105--110 km during the overcooling hints that the cooling may extend below the usual 115--250 km integration range, so a lower-altitude energy-budget analysis could refine the magnitude of of the overcooling.","Comparing May 2024 with the Halloween storms suggests that peak NO cooling flux may scale more with storm duration or energy deposition geometry than with Dst magnitude alone, a relation that could be tested against other severe storms."],"forward_implications":["If correct, post-storm density forecasts for extreme geomagnetic storms must account for NO radiative cooling that persists 2--4 times above quiet levels for days, not just during the main phase.","The May 2024 storm provides a second, well-observed case of thermospheric overcooling, strengthening the case that overcooling is a real phenomenon for storms of comparable strength and duration.","The reported all-time-high NO cooling flux of $11.84\\ \\mathrm{ergs\\,cm^{-2}\\,s^{-1}}$ and daily radiative power exceeding 1 TW imply that NO cooling can dissipate more than a terawatt of storm energy in a single day.","The hemispheric asymmetry—overcooling in the northern summer hemisphere but not the southern winter hemisphere—suggests that seasonal composition and dynamics modulate the net density response, a constraint for thermosphere-ionosphere circulation models.","The recovery phase showed density returning to pre-storm levels by 2--4 UT on May 12 in northern polar regions while NO cooling stayed elevated, supporting a direct causal chain from persistent cooling to the density undershoot.","These results imply that satellite drag models tuned to storm-time heating may systematically overestimate post-storm drag for severe storms, a practical concern for LEO operations."],"supporting_citations":[{"why":"First reported thermospheric overcooling during the 2003 Halloween storms and provides the -23% to -26% density-depletion benchmark this event is compared against.","marker":"Lei et al. (2012)"},{"why":"Numerical study establishing NO radiative cooling as the cause of Halloween-storm overcooling, the interpretive model used for the May 2024 event.","marker":"Chen & Lei (2018)"},{"why":"Defines the TIMED/SABER NO 5.3-µm infrared flux retrieval and integration method used to compute radiative cooling.","marker":"Mlynczak et al. (2010)"},{"why":"Provides the Swarm-derived thermospheric neutral mass density datasets used for the density depletion analysis.","marker":"Siemes et al. (2023)"},{"why":"Documents the Swarm GPS-based density retrieval procedure underlying the density observations.","marker":"van den IJssel et al. (2020)"},{"why":"Identifies the 26-34 nm EUV band as a key solar driver for thermospheric density, supporting the claim that rising EUV alone would lift density.","marker":"Dudok de Wit & Bruinsma (2011)"},{"why":"Supplies the Joule-heating estimation formula used for the Halloween-storm comparison, though not applied to May 2024 due to missing indices.","marker":"Knipp et al. (2004)"},{"why":"Reports the >1 TW daily NO radiative power for May 11, corroborating the all-time-high cooling claim.","marker":"Mlynczak et al. (2024)"}],"fun_headline_variants":["May 2024 superstorm overcools thermosphere by 23%","Record NO cooling follows May storm's 23% density dip","Thermosphere defies solar EUV with 23% post-storm drop","Superstorm leaves 23% density deficit, NO cooling at record","May 2024 storm: NO radiative cooling peaks, density dips 23%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"By May 12, the storm's heating sources—chiefly Joule heating and auroral particle precipitation—had returned to pre-storm levels, so the observed density undershoot can be attributed almost entirely to elevated NO infrared cooling; the paper did not directly calculate Joule heating for May 2024.","fun_headline_variants_meta":{"raw":{"variants":["May 2024 superstorm overcools thermosphere by 23%","Record NO cooling follows May storm's 23% density dip","Thermosphere defies solar EUV with 23% post-storm drop","Superstorm leaves 23% density deficit, NO cooling at record","May 2024 storm: NO radiative cooling peaks, density dips 23%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000165,"raw_usage":{"total_tokens":1290,"prompt_tokens":1022,"completion_tokens":268,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":638,"completion_tokens_details":{"reasoning_tokens":172}},"tokens_in":638,"tokens_out":268,"duration_ms":3504,"temperature":1.0,"reasoning_tokens":172,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:33:47.583316+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct calculation or retrieval of Joule heating for May 10-12 using alternative magnetic-activity proxies (for example provisional AE, SuperMAG, or assimilated indices) that shows heating still above pre-storm levels on May 12 would undermine the overcooling attribution, as would a finding that NO cooling at the sampled local times and latitudes on May 12 was not actually elevated.","supporting_citations":[{"cited_title":", Burns, A G","cited_arxiv_id":null,"evidence_quote":"First reported thermospheric overcooling during the 2003 Halloween storms and provides the -23% to -26% density-depletion benchmark this event is compared against."},{"cited_title":"\\ Lei, J","cited_arxiv_id":null,"evidence_quote":"Numerical study establishing NO radiative cooling as the cause of Halloween-storm overcooling, the interpretive model used for the May 2024 event."},{"cited_title":", Borries, C","cited_arxiv_id":null,"evidence_quote":"Provides the Swarm-derived thermospheric neutral mass density datasets used for the density depletion analysis."},{"cited_title":", Doornbos, E","cited_arxiv_id":null,"evidence_quote":"Documents the Swarm GPS-based density retrieval procedure underlying the density observations."},{"cited_title":"\\ Bruinsma, S","cited_arxiv_id":null,"evidence_quote":"Identifies the 26-34 nm EUV band as a key solar driver for thermospheric density, supporting the claim that rising EUV alone would lift density."}],"review_version":1}