{"id":"de5e632a-e9c3-414a-83be-1bc98d8de3cb","arxiv_id":"1908.06509","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Baikal-GVD observations show year-to-year bursts of water luminescence in Lake Baikal, with a luminescent layer descending at speeds up to 45 m/day.","lead":"This paper reports measurements of background light in Lake Baikal using the Baikal-GVD neutrino telescope, showing that water luminescence rose sharply in 2016 and 2018 but stayed flat in 2017. It also describes a glowing layer that sank from the top toward the bottom of the lake and checks that two independent monitoring systems agree.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Baikal-GVD's year-to-year comparison (2016 vs 2017 vs 2018) is not controlled for PMT gain stability or hardware changes; both readout systems share the same PMTs, so their agreement cannot rule out a slow gain drift that would make 2017 appear quiet.","rationale":"The paper is a simple observational report; the reader's CONDITIONAL verdict is appropriate. My stress test targeted the inference from raw PMT rates to ambient luminescence over the full 2016-2018 period. The two-system agreement is strong evidence against trigger-threshold artifacts, and the stable charge distribution during 2016 is good evidence against a within-year gain change. But neither check constrains the long-term stability: the trigger and monitoring systems share the same PMT, so a slow gain drift or window fouling would appear in both. The fact that 2017 is reported as a quiet year sandwiched between two active years is exactly the pattern a non-monotonic hardware change (e.g., PMT replacement, HV adjustment, or a different string) could produce. This is not an accusation; it is a missing control that can be supplied from the experiment's calibration and run logs. A cross-year charge-spectrum ratio test would settle it. Therefore I retain the reader's CONDITIONAL verdict and do not elevate it to REJECT: the concern is specific and addressable, not a demonstrated error.","tokens_in":4340,"tokens_out":8675,"duration_ms":92947,"concrete_test":"For the same physical OM(s) (or at least same string/cluster), extract the single-photoelectron charge spectrum or the ratio of count rates at the 0.5 and 1.0 p.e. thresholds during stable background periods in each of 2016, 2017, and 2018. If the median charge or the rate ratio drifts by more than, say, 10%, re-normalize all rates to a fixed photo-electron threshold and redo the year-to-year comparison. Also report the deployed string/OM identifiers and HV settings for each data set. If the normalized rates still show the 2017 quiet year, the claim is robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that water luminescence was high in 2016 and 2018 but practically constant in 2017 rests on the assumption that PMT count rates are directly comparable across years. The paper provides within-2016 checks: the charge distribution remains unchanged during active and quiet periods (Sec. 2.1, Fig. 1), and the monitoring and trigger systems agree after rescaling (Sec. 4, Fig. 6). However, both systems count pulses from the same PMTs and share the PMT high voltage, gain, and optical-window state. A slow gain decrease, window fouling, or even a change of the chosen OM/string between years would lower the measured rate for the same physical light level, making 2017 look anomalously quiet and 2018 active if the PMT was replaced or re-biased. No cross-year calibration is reported: no stable LED/laser pulser, no single-photoelectron peak position vs time, and no high-voltage/current history. The paper also does not state explicitly that the same cluster/string/OM is used for all three years; new clusters were deployed in 2017 and 2018. The claim of agreement between the two readout systems does not address this shared-mode sensitivity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This conference proceedings paper reports optical background (luminescence) monitoring in Lake Baikal using the Baikal-GVD neutrino telescope. The authors present PMT count-rate data from two independent readout chains, the cluster trigger system and the online monitoring system, and claim that the ambient light field showed enhanced luminescence activity in 2016 and 2018, was practically constant in 2017, and that during active periods a highly luminescent water layer propagated from the top to the bottom of the lake at speeds up to about 45 m/day. They also report a lack of correlation between string tilts (torrent flows) and luminescence activity, and an agreement between the trigger-system and monitoring-system count rates after rescaling.","tokens_in":4535,"tokens_out":2084,"duration_ms":24443,"significance":"If the observational claims hold, the paper provides a useful, multi-year record of ambient light backgrounds for Baikal-GVD and a potential limnological data set on vertical transport in Lake Baikal. The strength of the paper is that its central assertions rest on direct count-rate plots, and the comparison between the trigger system and the online monitoring system is a genuinely independent cross-check of the same optical signal. The paper is self-contained and does not rely on fitting or model-dependent assumptions. However, the significance is limited by the absence of quantitative uncertainties, statistical tests, and cross-year calibration checks, so the temporal and spatial claims are presented more as visual interpretations than as measured results.","major_comments":[{"comment":"The central claim that 2016 and 2018 had increased luminescence while 2017 was practically constant rests on comparing PMT count rates across different years. This comparison requires that the optical modules, their gains, and their thresholds are stable or calibrated across years. The paper does not state whether the same strings/OMs were used in all three years, nor does it report any stability monitor such as a single-photoelectron peak position, an LED/laser pulser calibration, or a high-voltage history. Because the trigger and monitoring systems share the same PMTs and the same high-voltage and optical-window state, the agreement shown in Fig. 6 cannot exclude a slow common-mode gain drift, window fouling, or hardware change that would make 2017 appear artificially quiet. The authors should either provide evidence of hardware and gain continuity or weaken the cross-year claim accordingly.","section":"Sec. 2.2 and Conclusions"},{"comment":"The speed estimate of about 45 m/day is derived 'by comparing the maximum for different depths' (Sec. 2.2), but no definition of how the maximum is identified, no fitting procedure, and no uncertainty on the speed are given. This estimate is a load-bearing quantitative claim and should be accompanied by a precise method description and an error estimate, or be explicitly labelled as a rough visual estimate.","section":"Sec. 2.2"},{"comment":"The statement that the trigger-system and monitoring-system noise rates 'agree well' after rescaling is not quantified. The paper gives no error bars, no measure of scatter, no chi-square or correlation statistic, and no explicit description of how the rescaling factor is determined. Since this agreement is presented as a validation of the monitoring, the authors should provide a quantitative comparison, such as the residual distribution or the scale-factor uncertainty.","section":"Sec. 4, Fig. 6"}],"minor_comments":[{"comment":"The sentence 'The one photo-electron background is well correlated with the half photo-electron background' would benefit from a quantitative measure (e.g., correlation coefficient) instead of a qualitative description.","section":"Sec. 2.1"},{"comment":"The phrase 'we obtain a velocity profile of the flows' suggests a depth-dependent velocity field, but the paper then reports a single speed with two values; the authors should clarify whether a single speed or a full profile is meant.","section":"Sec. 2.2"},{"comment":"For the 2018 data, the notation 'High (Low) noise rates are presented in red (blue)' is not a quantitative scale; adding a color bar or numerical ranges would make the figure interpretable.","section":"Sec. 2.2 and Fig. 4"},{"comment":"The text states that two more clusters were deployed in 2019 and that the total number is five, but it is not clear whether any of the reported 2016-2018 data used the later clusters; specifying the exact data-taking periods and cluster configurations for each year would help.","section":"Introduction, Sec. 1"},{"comment":"The description of the monitoring-system counter as measuring 'nanosecond pulses' is vague; a sentence on the pulse-width acceptance and how it relates to the threshold settings would improve reproducibility.","section":"Sec. 4"}],"recommendation":"major_revision","confidential_remarks":"For a conference proceedings, the paper is within scope and reports a potentially interesting environmental data set. The main risk is the uncontrolled cross-year comparison, which is central to the paper's headline claim. The authors can likely address this by adding a clear statement of hardware continuity or a per-year calibration check, and by softening the claim if such data are unavailable. I would not recommend rejection, since the issue is fixable in the manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, this is a proceedings-style report from Baikal-GVD on ambient light in Lake Baikal, measured via PMT count rates. The new thing here is the GVD-era data: they see enhanced luminescence in 2016 and 2018, a quiet 2017, and a downward-moving bright layer with speeds up to 45 m/day. The two-detector cross-check (trigger vs. online monitoring) is a genuine plus; it shows the effect is not an artifact of a single readout chain.\n\nThe soft spots are what you'd expect in a 6-page ICRC paper. There are no error bars, no statistical tests, and the 45 m/day number is read off visual maxima. The bigger concern, which I think is legitimate, is that the year-to-year comparison lacks a cross-year calibration. Both readout systems share the same PMTs, high voltage, and optical windows, so their agreement cannot rule out a slow gain drift or window fouling that makes 2017 look quiet. Also, the paper doesn't state that the same string/OM was used across all three years; new clusters were deployed in 2017 and 2018. The within-2016 check (charge distribution unchanged) is good, but it doesn't cover inter-year stability.\n\nThat said, the central observation is not likely to be a phantom. The pattern resembles earlier NT200 results, and the depth-dependent propagation from top to bottom is spatially structured in a way that a simple gain drift wouldn't produce. So I read this as a credible, qualitative measurement report, not a rigorous quantitative study.\n\nWho is this for? The Baikal-GVD collaboration (for background modeling) and limnologists interested in Lake Baikal water dynamics. A serious referee could ask for calibration details, but the paper deserves to be in the conversation. I'd send it to review, not desk reject, and recommend the authors add at least a statement about PMT gain stability or a cross-year LED calibration.\n\nIn short: worth citing as evidence that GVD sees the same luminescence phenomena as NT200, but treat the 2017 quiet year as provisional until gain drift is ruled out.\n\nRecommendation: engage.","headline":"A useful but technically short conference report on Baikal-GVD optical noise; the year-to-year variability claim is real data but would be stronger with a cross-year calibration.","tokens_in":5450,"tokens_out":2195,"would_cite":false,"duration_ms":21621,"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 Baikal-GVD neutrino telescope tracks a luminescent water layer moving downward through Lake Baikal at up to 45 m/day in 2016 and 2018, with 2017 almost quiet.","keywords":["Baikal-GVD","neutrino telescope","optical noise","water luminescence","Lake Baikal","photomultiplier count rates","internal waves","environmental monitoring"],"falsifier":"Deploy an independent optical sensor (for example, a fluorometer or transmissometer) at several depths where the telescope operates and compare its readings with the PMT count rates during an active period; if the count-rate surges are not mirrored by the optical sensor, the claim that the rates represent water luminescence would be refuted.","tokens_in":4134,"feed_emoji":"🌊","tokens_out":3458,"duration_ms":37612,"temperature":0.7,"pith_summary":"This paper claims that the count rates of Baikal-GVD's optical modules, thresholded at half a photo-electron, faithfully track ambient light produced by luminescence in the deep water of Lake Baikal. It reports that this luminescence increased in 2016 and 2018, while 2017 showed practically constant background noise. During active periods, the maximum of the light field descended from the top of the detector to the lake bed, implying a luminescent layer moving downward at speeds up to about 45 m/day. The authors also find that two independent readout systems, the cluster trigger system and an online monitoring system, agree after rescaling for different pulse thresholds. If correct, the telescope becomes a continuous environmental sensor for the lake's hydrodynamics and biology, not just a neutrino detector.","feed_headline":"Lake Baikal light sensors catch glowing water layer sinking 45 m/day","feed_subtitle":"Neutrino detector's noise rates track water luminescence across 2016-2018, with 2017 calm.","key_machinery":"The central object is the per-optical-module photomultiplier noise rate: the number of registered pulses per second above half a photo-electron charge, collected both by the cluster trigger system and by an independent nanosecond-pulse counter in the OM electronics. Averaging these rates over all strings at the same depth produces a depth profile of the ambient light field, and tracking the profile's maximum over time reveals the vertical propagation of the luminescent layer. The half-photo-electron threshold suppresses photomultiplier dark noise so that the recorded pulses are dominated by single photo-electron signals from water luminescence.","core_discovery":"The paper establishes that the noise count rates of Baikal-GVD's optical modules, measured with a threshold of half a photo-electron and averaged over strings at fixed depths, track the ambient light field in Lake Baikal. During periods of high optical activity, the depth of maximum count rate moves from the top to the bottom of the detector, indicating a layer of luminescent water sinking at speeds up to about 45 m/day. Such active periods occurred in 2016 and 2018, whereas 2017 was nearly constant. The trigger system and the online monitoring system give rescaled noise rates that agree with each other.","pith_inferences":["If the luminescence is biological (for instance, bioluminescent organisms or dissolved organic matter), the telescope could be used as a coarse environmental sensor, and correlating the noise maps with chlorophyll or river inflow data could reveal the trigger of active periods.","A testable extension would be to compare the inferred 45 m/day layer velocity with direct current-meter measurements at the same depths; agreement would support advective transport, while disagreement would point to a biological or chemical propagation mechanism.","The quiet 2017 period provides a natural control year: if the same environmental drivers recur, one could look for a common seasonal pattern between 2016 and 2018 that is absent in 2017.","The 10-12 hour modulation could be checked against local internal-wave models or temperature-chain data, offering a way to validate the internal-wave interpretation independently of the optical measurements."],"forward_implications":["If the count rates indeed track water luminescence, the telescope provides continuous, depth-resolved monitoring of an environmental light source in Lake Baikal.","The agreement between the trigger system and the online monitoring system means the effect is reproducible and not an artifact of trigger conditions or thresholds.","The observed downward motion of the luminescent layer implies vertical transport in the lake at speeds up to about 45 m/day during active periods.","The regular 10-12 hour modulation of noise rates suggests internal waves at the depths of the optical modules, giving a new observable for lake hydrodynamics.","The year-to-year contrast (active in 2016 and 2018, quiet in 2017) indicates the luminescence bursts are episodic rather than strictly annual."],"supporting_citations":[{"why":"Supplies the detector layout, cluster geometry, and deployment timeline used throughout the analysis.","marker":"[1]"},{"why":"Describes the trigger system data acquisition that provides one of the two independent count-rate data sets.","marker":"[2]"},{"why":"Establishes that sunlight is negligible below about 700 m depth, so the measured noise is attributed to water luminescence.","marker":"[3]"},{"why":"Provides earlier NT200 observations of a similar moving-layer pattern, used for comparison and interpretation.","marker":"[4]"},{"why":"Gives the acoustic positioning data used to check whether string deviations correlate with luminescence activity.","marker":"[5]"},{"why":"Describes the online monitoring system that supplies the second, trigger-independent count-rate data set.","marker":"[6]"}],"fun_headline_variants":["Baikal neutrino telescope's noise tracks water's glow layer","Baikal sensors see glowing water sink 45 m/day in 2016, 2018","Glow layer sinks in Baikal; 2017 stays still, 2016/2018 move","Two independent Baikal systems align on water luminescence","Baikal detector sees luminescence layer descend; 2017 calm"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the photomultiplier count rates, after cutting pulses below half a photo-electron, are a faithful measure of ambient water luminescence rather than being dominated by detector electronics drift, PMT dark noise, or changes in trigger thresholds.","fun_headline_variants_meta":{"raw":{"variants":["Baikal neutrino telescope's noise tracks water's glow layer","Baikal sensors see glowing water sink 45 m/day in 2016, 2018","Glow layer sinks in Baikal; 2017 stays still, 2016/2018 move","Two independent Baikal systems align on water luminescence","Baikal detector sees luminescence layer descend; 2017 calm"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001276,"raw_usage":{"total_tokens":5122,"prompt_tokens":755,"completion_tokens":4367,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":371,"completion_tokens_details":{"reasoning_tokens":4267}},"tokens_in":371,"tokens_out":4367,"duration_ms":29939,"temperature":1.0,"reasoning_tokens":4267,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:41:59.579955+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Deploy an independent optical sensor (for example, a fluorometer or transmissometer) at several depths where the telescope operates and compare its readings with the PMT count rates during an active period; if the count-rate surges are not mirrored by the optical sensor, the claim that the rates represent water luminescence would be refuted.","supporting_citations":[{"cited_title":"Šimkovic et al., Baikal-GVD coll., these proceedings, Neutrino Telescope in Lake Baikal: Present and Future","cited_arxiv_id":null,"evidence_quote":"Supplies the detector layout, cluster geometry, and deployment timeline used throughout the analysis."},{"cited_title":"Dvornický et al., Baikal-GVD coll., these proceedings, Data Quality Monitoring system in the Baikal-GVD experiment","cited_arxiv_id":null,"evidence_quote":"Describes the trigger system data acquisition that provides one of the two independent count-rate data sets."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes that sunlight is negligible below about 700 m depth, so the measured noise is attributed to water luminescence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides earlier NT200 observations of a similar moving-layer pattern, used for comparison and interpretation."},{"cited_title":"Avrorin et al., Baikal-GVD coll., EPJ Web of Conf","cited_arxiv_id":null,"evidence_quote":"Gives the acoustic positioning data used to check whether string deviations correlate with luminescence activity."},{"cited_title":"Golubkov et al., Baikal-GVD coll., PoS ICRC2017 (2018) 1032 6","cited_arxiv_id":null,"evidence_quote":"Describes the online monitoring system that supplies the second, trigger-independent count-rate data set."}],"review_version":1}