{"id":"ccd57d02-def3-4bf1-adf8-07bf9d293113","arxiv_id":"2608.09990","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"The iSTORM instrument, a 32-crystal CeBr3 gamma-ray spectrometer, met its design energy range and detected glows, TGFs, and flickering gamma-ray flashes during the ALOFT campaign.","lead":"iSTORM is a new gamma-ray spectrometer flown on NASA's ER-2 aircraft to measure thunderstorm radiation. The paper describes the instrument and reports glows, terrestrial gamma-ray flashes, and flickering gamma-ray flashes observed during the ALOFT campaign.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The uniform redistribution of BBB-dropped events can bias FGF/TGF temporal structure, and the paper never quantifies losses or validates the reconstruction against the independent BGO instrument.","rationale":"The paper's central claim is that iSTORM achieves a stated energy range and geometric area and successfully recorded glows, TGFs, and FGFs during ALOFT. The instrument description itself is credible, and the existence of the events is independently supported by the co-flying BGO instrument. The weakest point in the argument is the reconstruction of lost events in high-count-rate conditions. The reader identified this same assumption, and I agree it is load-bearing: the FGF duration and pulse-spacing analysis in Fig. 16 relies on millisecond-level timing, and the uniform insertion of lost events is an untested approximation that could systematically reshape burst profiles. This does not invalidate the paper's instrument-level claims, but it does justify the conditional verdict: the science highlights, especially Fig. 16, should not be taken as finalized until the loss-correction is quantified and cross-validated. Because the reader already reached CONDITIONAL, no verdict change is needed; the specific validation path is to use the A5202 event counter and BGO coincidences to test whether uniform redistribution preserves the observed temporal structure.","tokens_in":14203,"tokens_out":2884,"duration_ms":38618,"concrete_test":"For all 24 FGFs and the 7 TGFs in the July 24, 2023 case study, compute the lost-event fraction per event from the A5202 event counter and compare iSTORM's time-tagged count profile with the independent BGO instrument's time-tagged profile for the same bursts. Then recompute Fig. 16 using only FGFs with a lost fraction below 2% and, separately, using only raw recorded events with no uniform insertion. If the slope or intercept of the duration-spacing fit changes by more than the fit uncertainty, or if the correlation loses significance, the reported relation is not robust to the loss-correction assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2 states that when the BBB drops packets during high-count-rate events, the A5202 event counter recovers the number of lost events, but not their energies or accurate times, so lost events are distributed uniformly between two recorded events. This uniform-insertion correction is load-bearing for the paper's showcase science: the FGF pulse timings in Section 3.3, the per-pulse durations, the 15 ms average spacing, and the Fig. 16 duration-versus-spacing correlation all depend on recovered event times being approximately correct. If packet loss is not uniform but is concentrated in the high-rate phases of a burst, as would be expected if the BBB or Ethernet link is overwhelmed precisely when the count rate peaks, then uniform insertion will systematically stretch or smooth the recovered pulse profile. A TGF or FGF pulse could be broadened and its apparent duration inflated, and the spacing between pulses could be distorted because lost events are smeared into inter-pulse gaps. The paper gives no estimate of the lost fraction for the FGF events, no demonstration that the loss pattern is time-uniform, and no comparison with the co-flying BGO instrument, which observed the same FGFs with a different readout. Without such a check, the Fig. 16 correlation—the only quantitative relation presented in the paper—could partially reflect the packet-loss correction rather than the underlying atmospheric process.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript describes the design, environmental testing, and first in-flight results of iSTORM, a 32-element CeBr3 scintillator array with SiPM readout flown on the NASA ER-2 during the ALOFT campaign. The authors report that the instrument achieves a useful energy range of about 250 keV to 5 MeV under flight conditions with a total geometrical area of 157 cm^2, and they show examples of gamma-ray glows, TGFs, and the newly reported flickering gamma-ray flashes (FGFs). The paper also presents a linear relation between FGF duration and average pulse spacing in its Fig. 16.","tokens_in":14497,"tokens_out":2798,"duration_ms":31029,"significance":"If the performance claims hold, iSTORM provides a useful airborne capability for thunderstorm gamma-ray observations, and the reported detections of glows, TGFs, and FGFs are scientifically important. The paper is honest about key limitations: spectra are stated not to be deconvolved, and the calibration is based on standard Cs-137 ground measurements. The direct comparison with the co-flying BGO instrument is promised but not shown. The central concern is that the temporal structure of FGFs, which underpins the only quantitative relation in the paper, depends on a packet-loss correction that is not validated. This is a load-bearing issue, not a cosmetic one.","major_comments":[{"comment":"The description of the event-counter correction is not sufficient to support the FGF timing results. The paper states that when the BeagleBone Black drops packets during high-count-rate events, the lost events are 'distribute[d] uniformly between two recorded events.' The FGF pulse durations, spacings, and the correlation in Fig. 16 are derived from these corrected event times. If packet loss is concentrated at the high-rate phases of a burst, uniform insertion will systematically stretch pulse durations and fill inter-pulse gaps, directly biasing the reported 15 ms spacing and the duration-versus-spacing relation. The paper gives no estimate of the lost fraction for the FGF events, no test of the uniformity assumption, and no comparison with the independent BGO readout that observed the same phenomena. Please quantify the loss rate for the events shown and validate the reconstruction against the BGO data, or explicitly state how the main timing results are robust to non-uniform loss.","section":"Section 2, paragraph on the A5202 event counter; Section 3.3"},{"comment":"The linear fit in Fig. 16 is the only quantitative relation in the paper, but it is reported without uncertainties: the fit parameters Duration[ms] = 5.4 × Average FGF Pulse Time Difference [ms] + 14.2383 have no error bars, no coefficient of determination, and no statement of how many points were used or whether the correlation is statistically significant. The error bars shown are the standard deviation of pulse times within each FGF, which does not include the uncertainty introduced by the packet-loss correction discussed above. Please provide fit uncertainties, a goodness-of-fit measure, and a clear description of the sample size and selection criteria for the 24 FGFs.","section":"Section 3.3, Fig. 16"},{"comment":"The manuscript states that 'The BGO’s observation is consistent with that of iSTORM' but shows no comparison. Because the BGO instrument uses a different readout and would be affected differently by dead time or packet loss, a direct overlay or time-tagged comparison of the FGF pulse structure is a natural and powerful validation of the iSTORM timing. Adding this comparison would also address the concern that the observed FGF morphology is an artifact of the uniform event-redistribution procedure.","section":"Section 3.3"},{"comment":"The energy-range claim needs more support in the flight environment. Figure 7 shows that the Cs-137 photopeak is cropped into the overflow bin at -5 °C, and the paper states that -5 °C is the lower limit for science operations. The pod is described as having blowers to maintain non-freezing temperatures, but no in-flight gain calibration or temperature record is reported to show that the 250 keV to 5 MeV range is actually maintained during typical cruise altitudes. In addition, the in-flight count-rate time series (Figs. 11 and 14) and the glow/background spectra in Fig. 13 are presented without statistical uncertainties, which is problematic for a paper that asserts the detection of specific transients. Please add error bars to the count-rate and spectral quantities and describe how the in-flight energy scale was monitored.","section":"Section 2.1 and Section 3.1.1"}],"minor_comments":[{"comment":"Reference [3] is a duplicate of Reference [1]; the same paper is listed twice with identical bibliographic information.","section":"References"},{"comment":"The caption contains a typo: 'measurmed' should be 'measured'.","section":"Fig. 13 caption"},{"comment":"The caption writes 'Time Different' where 'Time Difference' is meant.","section":"Fig. 16 caption"},{"comment":"The acknowledgment contains a typo: 'Earh Science Division' should be 'Earth Science Division'.","section":"Acknowledgments"}],"recommendation":"major_revision","confidential_remarks":"The paper is a straightforward instrument description with early science highlights. The core concern is not with the hardware itself but with the unvalidated packet-loss correction that directly affects the FGF timing results, which are a primary claim of the manuscript. The requested additions (loss fraction, BGO comparison, fit uncertainties) are feasible within the scope of the paper and would substantially strengthen it. The manuscript also omits statistical uncertainties on all in-flight measurements, which should be addressed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid instrument paper, not a discovery paper. What is genuinely new is iSTORM itself—construction, environmental testing, and flight performance—and that material is handled carefully. The FGF science is already reported in the Nature companion papers; here it mostly provides context plus one descriptive fit.\n\nThe environmental testing is the strongest part: vacuum exposure to 90 kft equivalent, rapid evacuation/pressurization, thermal cycling down to −15 °C, and Cs-137 spectra at different temperatures, with an explicit −5 °C lower limit for science operations. The authors are also honest about what they did not do—no spectral deconvolution, no real-time telemetry, heaters left unused—and they say so plainly.\n\nThe real soft spot is the packet-loss recovery. Section 2 states that when the BBB drops packets, the lost events are distributed uniformly between two recorded events. That assumption is load-bearing for the FGF pulse timings, per-pulse durations, the 15 ms average spacing, and the Fig. 16 duration-versus-spacing correlation. If losses cluster during the high-rate phases of a burst, as one might expect when the data link is saturated exactly when counts peak, uniform insertion will smear events into inter-pulse gaps and systematically inflate apparent pulse durations. The paper gives no estimate of the lost fraction for the FGF events, no evidence that loss is time-uniform, and no cross-check against the co-flying BGO instrument, which observed the same phenomena with a different readout. Without that, the Fig. 16 relation—the only quantitative result in the paper—could partially reflect the correction rather than the storm physics. This is a legitimate concern, but it does not undermine the central instrument claims, which rest on calibration-source spectra and direct flight detections.\n\nMinor issues: no data release, no error bars on count rates, and no uncertainty on the Fig. 16 fit parameters. The citation pattern is fine; the authors cite the companion ALOFT/Nature papers for the discoveries, and self-citations point to actual prior results.\n\nWho should read this: anyone building or planning airborne scintillator instruments, and anyone using the ALOFT dataset who wants the readout details of one of the two gamma-ray detectors. It deserves a serious referee. My recommendation: send it to review, and ask the authors to quantify the packet-loss fraction for the FGF events, justify the uniform-distribution assumption, or compare the recovered timing with the BGO data. I would not desk-reject, and I would not accept the Fig. 16 correlation as quantitative until that is addressed.","headline":"A credible instrument paper whose main soft spot is the unquantified packet-loss recovery used in the FGF timing analysis; the instrument claims themselves hold up.","tokens_in":14998,"tokens_out":1818,"would_cite":true,"duration_ms":21641,"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 iSTORM, a 32-crystal CeBr3 gamma-ray spectrometer for ER-2 flights, and reports its 250 keV-to-5 MeV range, 157 cm2 area, and detection of glows, TGFs, and new flickering gamma-ray flashes during ALOFT.","keywords":["terrestrial gamma-ray flash","gamma-ray glows","flickering gamma-ray flashes","CeBr3 scintillators","silicon photomultipliers","airborne gamma-ray instrumentation","ALOFT campaign"],"falsifier":"Take a bright, packet-loss-affected FGF seen by both iSTORM and the BGO instrument and compare the recovered iSTORM pulse spacings with the independently timed BGO pulses in the same window; systematic disagreement concentrated in loss intervals would show that assigning lost events uniformly between recorded events distorts the FGF timing results. Alternatively, replay known high-rate pulse trains with controlled bursty losses through the A5202 data stream and check whether the recovery algorithm reproduces the injected spacings.","tokens_in":1866,"feed_emoji":"⚡","tokens_out":3229,"duration_ms":125654,"temperature":0.7,"pith_summary":"iSTORM is a gamma-ray spectrometer built to fly on a high-altitude ER-2 aircraft and look down into thunderstorms. This paper claims that its 32 one-inch CeBr3 scintillator crystals, read out by silicon photomultipliers, cover roughly 250 keV to 5 MeV under flight conditions with 157 cm2 of collecting area. During the ALOFT campaign the instrument recorded gamma-ray glows, terrestrial gamma-ray flashes (TGFs), and the newly named flickering gamma-ray flashes (FGFs). On the July 24, 2023 flight over the Bay of Campeche it saw almost 100 TGFs and multiple glows and FGFs in a single active storm system. If these measurements hold, airborne arrays at 20 km can resolve a much richer and more common population of thunderstorm gamma-ray transients than space telescopes see.","feed_headline":"Airborne detector catches nearly 100 TGFs and new flickering flashes","feed_subtitle":"iSTORM's 32 CeBr3 crystals cover 250 keV to 5 MeV from the ER-2 and spot a new storm phenomenon.","key_machinery":"The load-bearing design is an array of 32 one-inch-diameter CeBr3 scintillator crystals, each read out by a custom silicon photomultiplier (SiPM) array biased at 26 V, with signals digitized by a CAEN A5202 front end using a CITIROC 1A ASIC in dual-gain, 12-bit spectroscopy mode. Segmentation reduces pulse pileup, the dual gains and zero suppression enable the 250 keV to 5 MeV range, and an event counter in the data stream recovers the number of lost events during high-count-rate intervals, with those lost events redistributed uniformly between recorded events. A hermetically sealed one-atmosphere box and a GPS-disciplined clock allow the instrument to operate at 65-68 kft cruise altitude.","core_discovery":"The paper's central claim is that a compact, segmented CeBr3/SiPM spectrometer can do science-quality gamma-ray transient measurements from an ER-2 aircraft. The instrument achieves an energy range of approximately 250 keV to 5 MeV with 157 cm2 total geometrical area, and during ALOFT it detected glows, TGFs, and FGFs. In the 10-minute case study from July 24, 2023, iSTORM saw 7 TGFs embedded in glow episodes plus 3 FGFs, and the paper reports a linear relation between FGF total duration and average pulse spacing, $\\mathrm{Duration\\,[ms]} \\approx 5.4 \\times \\mathrm{average\\,spacing\\,[ms]} + 14.2$. The companion BGO instrument saw the same events, which the paper takes as cross-validation that FGFs are a distinct, real phenomenon rather than an instrumental artifact.","pith_inferences":["A next test beyond the paper is to replay recorded ALOFT high-rate intervals through a bench setup with known injected event times and check whether the uniform-loss recovery reproduces the injected pulse spacings; this would directly test the FGF timing results.","If the FGF duration-spacing relation holds in more events, it would suggest a common regulator such as a persistent avalanche region modulating pulse rate, rather than a sequence of independent TGF-like bursts.","Because FGFs lack optical and radio counterparts, archival airborne and balloon gamma-ray datasets may contain unrecognized multi-pulse structures in the 20-250 ms range; a blind search could test that possibility.","An imaging instrument such as the planned coded-aperture telescope could map FGF pulse origins in the cloud, connecting the measured timing correlation to the spatial structure of the electric field."],"forward_implications":["A 157 cm2 airborne detector can measure thunderstorm gamma rays from about 250 keV to 5 MeV while cruising at 20 km, so near-storm aircraft surveys are a viable complement to space-based TGF observations.","The July 24, 2023 flight's nearly 100 TGFs indicate that very active storms produce far more flashes than satellite-based rarity estimates implied.","The measured FGF relation, $\\mathrm{Duration\\,[ms]} \\approx 5.4 \\times \\mathrm{average\\,pulse\\,spacing\\,[ms]} + 14.2$, gives models a quantitative target for the electric-field conditions that generate flickering flashes.","The absence of optical or radio counterparts to FGFs implies that gamma-ray observations may be the only remote way to detect those events.","With a faster readout and near-real-time count-rate telemetry, the upgraded iSTORM can extend the same measurements to pyrocumulonimbus storms during the INSPYRE campaign."],"supporting_citations":[{"why":"Defines the ALOFT campaign and the ER-2 flight context within which iSTORM's detections were made.","marker":"[22]"},{"why":"Reports gamma-ray glows at 20 km altitude from the companion BGO instrument that cross-validates iSTORM's observations.","marker":"[28]"},{"why":"Provides the broader ALOFT glow analysis that the paper treats as reference for glow phenomenology.","marker":"[39]"},{"why":"Reports the discovery of flickering gamma-ray flashes, the new phenomenon iSTORM also records.","marker":"[46]"},{"why":"Is the original discovery paper for the terrestrial gamma-ray flashes that iSTORM observes.","marker":"[41]"},{"why":"Establishes the weak TGF population observable from aircraft altitude that motivates the instrument's sensitivity.","marker":"[11]"},{"why":"Is cited as the ongoing spectral deconvolution work needed to interpret iSTORM's measured glow spectra.","marker":"[40]"},{"why":"Documents the CITIROC 1A ASIC that performs the SiPM readout and sets the energy response.","marker":"[32]"},{"why":"Describes the A5202 front end whose event counter is used to recover lost events in high-count-rate data.","marker":"[31]"}],"fun_headline_variants":["Airborne detector on ER-2 spots new flickering gamma-ray flashes","iSTORM gamma-ray instrument discovers novel storm flash phenomenon","New gamma-ray phenomenon 'flickering flashes' seen from NASA ER-2","Compact airborne spectrometer finds new thunderstorm gamma-ray bursts"],"cache_read_input_tokens":17152,"weakest_assumption_plain":"The load-bearing premise is that, when data packets are lost in the brightest flashes, the missing events were spread evenly in time between the two recorded events around each loss; if losses arrived in bursts instead, the reported pulse timings and count rates would be biased.","fun_headline_variants_meta":{"raw":{"variants":["Airborne detector on ER-2 spots new flickering gamma-ray flashes","iSTORM gamma-ray instrument discovers novel storm flash phenomenon","New gamma-ray phenomenon 'flickering flashes' seen from NASA ER-2","Compact airborne spectrometer finds new thunderstorm gamma-ray bursts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00022,"raw_usage":{"total_tokens":1428,"prompt_tokens":910,"completion_tokens":518,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":526,"completion_tokens_details":{"reasoning_tokens":444}},"tokens_in":526,"tokens_out":518,"duration_ms":6045,"temperature":1.0,"reasoning_tokens":444,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:49:28.977399+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a bright, packet-loss-affected FGF seen by both iSTORM and the BGO instrument and compare the recovered iSTORM pulse spacings with the independently timed BGO pulses in the same window; systematic disagreement concentrated in loss intervals would show that assigning lost events uniformly between recorded events distorts the FGF timing results. Alternatively, replay known high-rate pulse trains with controlled bursty losses through the A5202 data stream and check whether the recovery algorithm reproduces the injected spacings.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the ALOFT campaign and the ER-2 flight context within which iSTORM's detections were made."},{"cited_title":"Ostgaard, H","cited_arxiv_id":null,"evidence_quote":"Reports gamma-ray glows at 20 km altitude from the companion BGO instrument that cross-validates iSTORM's observations."},{"cited_title":"Østgaard, A","cited_arxiv_id":null,"evidence_quote":"Reports the discovery of flickering gamma-ray flashes, the new phenomenon iSTORM also records."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Is the original discovery paper for the terrestrial gamma-ray flashes that iSTORM observes."},{"cited_title":"Bjrge-Engeland, N","cited_arxiv_id":null,"evidence_quote":"Establishes the weak TGF population observable from aircraft altitude that motivates the instrument's sensitivity."},{"cited_title":"Sarria, N","cited_arxiv_id":null,"evidence_quote":"Is cited as the ongoing spectral deconvolution work needed to interpret iSTORM's measured glow spectra."},{"cited_title":"FERS-5200: a distributed Front-End Readout System for multidetector arrays","cited_arxiv_id":"2010.15688","evidence_quote":"Describes the A5202 front end whose event counter is used to recover lost events in high-count-rate data."}],"review_version":1}