Global Distribution of Key Features of Streamer Corona Discharges in Thunderclouds

1. Introduction Radio, optical, and some indirect chemical recordings since the early 1980s suggest that cold, non-thermal streamer corona discharges are common in thunderstorms worldwide (Bandara etal.,2019; Bozem etal.,2014; Le Vine,1980; Li etal.,2021; Liu etal.,2018; Neubert etal.,2021; Soler etal.,2020; Wiens etal.,2008). Fast breakdown (Rison etal.,2016; Tilles etal.,2019) seems to underlie streamer coronas that cause the so-called Narrow Bipolar Events (NBEs) originally detected by Le Vine(1980) in the form of strong Very High Frequency sources from in-cloud discharges. Typical light spectra of streamer corona discharges in air are strongly dominated by near-ultraviolet blue emissions (300–450nm) corresponding to the Second Positive System (SPS) of molecular nitrogen (N2) (Ebert etal.,2010; Gallimberti et al., 1974; Grum & Costa, 1976) with the strongest transition at 337 nm (Gordillo-Vázquez etal.,2012; Hoder etal.,2016; Malagón-Romero & Luque,2019) and generally undetectable oxygen atom 777.4nm emissions typical of lightning. Optical signals from lightning stroke flashes are characterized by including both 337nm emissions and, especially, strong 777.4nm optical emissions (Blakeslee etal.,2020; Christian Abstract We present nighttime worldwide distributions of key features of Blue LUminous Events (BLUEs) detected by the Modular Multispectral Imaging Array of the Atmosphere-Space Interaction Monitor. Around 10% of all detected BLUEs exhibit an impulsive single pulse shape. The rest of BLUEs are unclear (impulsive or not) single, multiple or with ambiguous pulse shapes. BLUEs exhibit two distinct populations with peak power density <25µWm −2 (common) and ≥25µWm −2 (rare) with different rise times and durations. The altitude (and depth below cloud tops) zonal distribution of impulsive single pulse BLUEs indicate that they are commonly present between cloud tops and a depth of ≤4km in the tropics and ≤1km in mid and higher latitudes. Impulsive single pulse BLUEs in the tropics are the longest (up to ∼4km height) and have the largest number of streamers (up to ∼3×10 9). Additionally, the analysis of BLUEs has turned out to be particularly complex due to the abundance of radiation belt particles (at high latitudes and in the South Atlantic Anomaly [SAA]) and cosmic rays all over the planet. True BLUEs can not be fully distinguished from radiation belt particles and cosmic rays unless other ground-based measurements associated with the optically detected BLUEs are available. Thus, the search algorithm of BLUEs presented in Soler etal. (2021), https://doi. org/10.1029/2021gl094657 is now completed with a new additional step that, if used, can considerably smooth the SAA shadow but can also underestimate the number of BLUEs worldwide. Plain Language Summary The presence of corona electrical discharges in thunderclouds has been suspected for a long time. These thunderstorm coronas can be observed as Blue LUminous Events (BLUEs) formed by a large number of streamers characterized by their distinct 337nm light flashes with absent (or negligible) 777.4nm component (typical of lightning leaders). The Modular Multispectral Imaging Array of the Atmosphere-Space Interaction Monitor has successfully allowed us to map and characterize BLUEs. The results presented here include a global analysis of key properties of BLUEs such as their characteristic rise times and duration, their depth with respect to cloud tops, vertical length and number of streamers. This study also includes two different global annual average climatologies of BLUEs depending on considerations about the rise time and total duration of BLUEs worldwide. SOLER ETAL. © 2022 The Authors. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. Global Distribution of Key Features of Streamer Corona Discharges in Thunderclouds S. Soler1 , F. J. Gordillo-Vázquez1 , F. J. Pérez-Invernón1 , A. Luque1 , D. Li2 , T. Neubert2 , O. Chanrion2 , V. Reglero3, J. Navarro-González3 , and N. Østgaard4 1Instituto de Astrofísica de Andalucía (IAA-CSIC), Glorieta de la Astronomía s/n, Granada, Spain, 2National Space Institute, Technical University of Denmark (DTU Space), Kongens, Denmark, 3Image Processing Laboratory, University of Valencia, Valencia, Spain, 4Department of Physics and Technology, Birkeland Centre for Space Science, University of Bergen, Bergen, Norway Key Points: • BLUEs are found between ∼1 and ∼4km below cloud tops in the tropics and ≤1km in mid and higher latitudes • Two distinct populations of BLUEs with peak power density <25μWm −2 (common) and ≥25 μWm −2 (rare) are observed • Fast rise time (<0.05ms) BLUEs occur very superficially (<1km) near cloud tops with high power density ≥100μWm −2 Supporting Information: Supporting Information may be found in the online version of this article. Correspondence to: F. J. Gordillo-Vázquez, vazq[email protected] Citation: Soler, S., Gordillo-Vázquez, F. J., Pérez-Invernón, F. J., Luque, A., Li, D., Neubert, T., etal. (2022). Global distribution of key features of streamer corona discharges in thunderclouds. Journal of Geophysical Research: Atmospheres, 127, e2022JD037535. https://doi.org/10.1029/2022JD037535 Received 21 JUL 2022 Accepted 5 DEC 2022 Author Contributions: Conceptualization: S. Soler, F. J. Gordillo-Vázquez Data curation: S. Soler Formal analysis: S. Soler, F. J. Gordillo-Vázquez Funding acquisition: F. J. Gordillo-Vázquez Investigation: S. Soler, F. J. GordilloVázquez, F. J. Pérez-Invernón, A. Luque, D. Li, T. Neubert, O. Chanrion, V. Reglero, J. Navarro-González, N. Østgaard 10.1029/2022JD037535 RESEARCH ARTICLE 1 of 16 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 2 of 16 etal.,1989,2003; Montanyà etal.,2021). However, BLUEs exhibit strong optical emissions in the 337nm and negligible (at the noise level) 777.4nm emissions (Soler etal.,2020). Corona streamers can appear alone as leaderless corona discharges and/or in combination with hot leaders of lightning, blue jets (Wescott etal.,1995) and blue starters (Edens,2011; Wescott etal.,1996). The term Blue LUminous Events (BLUEs) has been recently applied to in-cloud and to partially emerged transient electrical discharges that emit pulses of light mostly blue, that is, they could include a small fraction of red (from the first positive system of N2) and infrared optical emissions (777.4nm from atomic oxygen). This definition includes blue jets and starters, but also positive and negative NBEs, which are the VLF/LF radio manifestation of in-cloud leaderless streamer coronas (Cooray etal.,2020; C. Kuo etal.,2015; Li etal.,2021; Liu etal.,2018,2019; Neubert etal.,2021; Rison etal.,2016; Soler etal.,2020; Tilles etal.,2019). Apart from distinct optical emissions directly associated with corona streamers, the dissimilarity between BLUEs and lightning discharges is also manifested by the fact that BLUEs can occur individually with no associated lightning discharges or they can be the initial event of lightning discharges (Li etal.,2022; López etal.,2022; Soler etal.,2020,2021). BLUEs optical detections from space have been reported from the limb-pointing Imager of Sprites/Upper Atmospheric Lightning onboard FORMOSAT-2 (J. Chou etal.,2011; J.-K. Chou etal.,2018; C.-L. Kuo etal.,2005; C. Kuo etal.,2015;; Liu etal.,2018). A variety of BLUEs including kilometer-scale blue discharges at the cloud top layer at ∼18km altitude, blue starters and a pulsating blue jet propagating into the stratosphere were color photographed from the International Space Station (ISS) (Chanrion etal.,2017). BLUEs have also been recently observed by the nadir-pointing MMIA onboard ASIM in the ISS since April 2018 (Husbjerg etal.,2022; Li etal.,2021; Neubert etal.,2021; Soler etal.,2020). However, it is a known issue that powerful NBEs are highly likely to be misclassified or simply missed by lightning locating systems (Zhu etal.,2022). For 1022 NBEs reported by Leal etal.(2019), the misclassification rates were 78% and 56% for the National Lightning Detection Network (NLDN−GLD360) and the Earth Network Total Lighting Network, respectively. They also found that the percentage of misclassified NBEs becomes even higher for higher intensity events. Connected to this, Chanrion etal.(2017) saw BLUEs from the ISS without any GLD360 detections. Thus, ground based detection of BLUEs by lightning networks remains a challenge. A number of recent thunderstorm case-based works have detailed studied properties of BLUEs associated with positive NBEs (Soler etal.,2020), negative NBEs (Li etal.,2021; Liu, Zhu, etal.,2021; Liu, Lu, etal.,2021; Neubert etal.,2021) and multiple pulse BLUEs (Li etal.,2022) using MMIA data of thunderclouds in several locations of the world. Additionally, a recent study presented the first worldwide nighttime climatology of BLUEs in thunderclouds derived from 2years of BLUEs data recorded by MMIA in ASIM (Soler etal.,2021). Here we focus on analyzing key properties of worldwide nighttime BLUEs detected by MMIA between 1 April 2019 and 31 March 2021 (note that ASIM can only observe during the night). Our study also discusses BLUEs depending on considerations about the rise time and total duration of BLUEs in the planet. We investigate intrinsic properties of BLUE events including temporal features like pulse shape (single, multiple or irregular), rise times and total duration times. We analyze the 337nm Peak Power Density (PPD) and total brightness, and the depth below thundercloud tops where BLUEs occur as a function of the latitude and longitude. Finally, we also present worldwide zonal and meridional distributions for the vertical lengths and approximate number of streamers of single pulse BLUEs. 2. Observations and Data Observations of BLUEs were carried out with the MMIA high sampling rate (100 k samples/s) photometers in the near UV (337nm/4nm), tuned to the strongest line of the N2 SPS, and in the near infrared band (777.4nm/5nm) for recording the atomic oxygen triplet line of lightning. MMIA also incorporates a high-speed photometer in the UV (180–230nm), capable of recording part of the N2 Lyman–Birge–Hopfield band, and a pair of 337nm/4 nm and 777.4nm/5nm filtered cameras (at 12 fps) with ∼400m/pixel spatial resolution (Chanrion etal.,2019). BLUEs exhibit strong features in the 337nm/4nm photometer with negligible (or very minor) signal in the 777.4nm/5nm photometer, which is continuously monitored (Soler etal.,2020). Once a true positive BLUE detection is confirmed, the cameras are checked for possible associated images. Methodology: S. Soler, F. J. GordilloVázquez, F. J. Pérez-Invernón, O. Chanrion, N. Østgaard Project Administration: F. J. Gordillo-Vázquez Resources: F. J. Gordillo-Vázquez Software: S. Soler Supervision: F. J. Gordillo-Vázquez Validation: S. Soler, F. J. Gordillo-Vázquez Visualization: S. Soler, F. J. Gordillo-Vázquez Writing – original draft: F. J. Gordillo-Vázquez Writing – review & editing: F. J. Gordillo-Vázquez 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 3 of 16 The two worldwide annual average distributions of nighttime BLUEs presented and discussed here (see Figures1 and2) were obtained with global ASIM-MMIA level 1 (calibrated) data in a period of two years (1 April 2019–31 March 2021) shifted 7 months ahead (see Section4 for details) with respect to the earlier 2-year period explored in Soler etal.(2021). The BLUEs shown in Figures1 and2 can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses or any other irregular pulse shape. The first global average distribution (GD-1) of BLUEs presented was derived using the algorithm described in Soler etal.(2021). In order to explore the possible influence of Radiation Belt Particles (RBP) and Cosmic Rays (CR) on our dataset, the second global average distribution (GD-2) of nighttime BLUEs discussed here includes the condition that 337nm events are removed in the entire planet when their rise times (τrise) are ≤40μs and their total duration (τtotal) times are ≤150μs (see Figure2). This is a new (optional) step in the Soler etal.(2021) algorithm that by default allows events with any duration above 50μs. In spite of this, there are still some hundreds of BLUEs with duration below 50μs (see GD-2 in the right column of Figures10 and11). This is fully compatible with point 4 of our BLUE search algorithm that reads “create groups defined by five or more consecutive blue counts (10μs each) above the 337nm photometer threshold” (Soler etal.,2021), and with the criterion on how the total duration of an event is calculated (from the fitted signal) getting the position in a time of the blue peak and then go backward and forward until the signal drops below 10% of the maximum. When our algorithm searches for a BLUE event it needs to find at least five counts above the threshold, so apparently, we can assume that all events should last at least 50μs. However, when we apply these criteria to very sharp BLUEs with a high Peak Power Density we can see that, in some cases, the total time can be less than 50μs (as mentioned above). It is worth mentioning here that ultrahigh energy (>10 19eV) CRs are being monitored from the ISS by the nadir-facing Multiwavelength Imaging New Instrument for the Extreme Universe Space Observatory (Mini-EUSO) telescope in operation since October 2019 (Bacholle etal., 2021; Miyamoto etal.,2021). Mini-EUSO observes the nighttime Earth in the near ultraviolet (UV) - blue range (290−430nm), with a spatial resolution of about 6.3km (FOV of 300×300km) and a temporal resolution of 2.5μs. According to Mini-EUSO observations, ultrahigh energy CRs cross one or a few pixels of the photocathode detector releasing a high-intensity light curve that can last a maximum of ≤150μs with a sharp increase (≤40μs) (see figure 5 in Miyamoto etal.[2021]). The global distributions of nighttime BLUEs in Figures1 and 2 include ∼46,000 events and 26,500 events, respectively. One can note that it is then quite possible that not all the ∼20,000 events removed in GD-2 with respect to GD-1 are RBPs and CRs. The presence of the South Atlantic Anomaly (SAA) can be distinguished within a rectangle with borders in latitudes 5°S and 45°S, and longitudes 0° and 100°W in Figure1 for GD-1. Note that the SAA's shadow is mostly removed in Figure2 for GD-2 but with the drawback of probable underestimation of the number of BLUEs worldwide in GD-2. The seasonal nighttime average distributions of BLUEs according to GD-1 and GD-2 are further shown in Figures3 and4, respectively. In both seasonal distributions, BLUEs are more common in the boreal summer closely followed by the boreal autumn. The main visual difference between Figures1 Figure 1. Two-year average (1 April 2019 through 31 March 2021) nighttime climatology of global Blue LUminous Events (BLUE) electrical activity in thunderclouds (GD-1) showing ∼46,000 BLUEs. The map is generated using 2°×2° grid cells. The BLUEs shown can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses or any other irregular pulse shape. The annual global rate of GD-1 BLUEs peak at 9.5 events s −1 in the local midnight (00.00 local solar time), and show a decreasing global rate as local daytime approaches (and there is less Modular Multispectral Imaging Array observation time). On average, the global annual average rate of BLUEs in GD-1 is 6.0 events s −1. Note that Atmosphere-Space Interaction Monitor can only observe during the night. Figure 2. Two-year average (1 April 2019 through 31 March 2021) nighttime climatology of global Blue Luminous Events (BLUE) electrical activity in thunderclouds (GD-2) removing events in all planets (including the South Atlantic Anomaly) with rise time (τrise)≤40μs and total duration (τtotal)≤150μs. The map is generated using 2°×2° grid cells. Note that it is quite possible that not all the removed ∼20,000 events with respect to GD-1 are radiation belt particles and cosmic rays. Consequently, the number of BLUEs (∼26,500) shown in this GD-2 distribution is most probably underestimated. The BLUEs shown can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses, or any other irregular pulse shape. The annual global rate of GD-2 BLUEs peak at 5.5 events s −1 in the local midnight (00.00 local solar time), and show a decreasing global rate as local daytime approaches (and there is less Modualr Multispectral Imaging Array observation time). On average, the global annual average rate of BLUEs in GD-2 is 3.5 events s −1. Note that Atmosphere-Space Interaction Monitor can only observe during the night. 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 4 of 16 and3, and the distributions of BLUEs in Soler etal.(2021) is the disappearance of events occurring in high latitudes during the annual average and during the SON, DJF, and MAM. 3. Methodology The key features of the BLUEs were obtained by applying different types of fits (depending on whether the light source is considered point-like [see Figure S1 of the Supporting InformationS1] or extended [see Figure S2 of the Supporting InformationS1]) to the impulsive 337nm light curves among the ∼46,000 and ∼26,500 BLUEs registered worldwide according to the global annual average distributions shown in Figures1 and2. However, as commented below, the two types of fits considered here are valid provided that the 337nm light curve of the BLUEs exhibits a relatively clear (single pulse) impulsive shape, which occurs in only ∼12% and ∼10% of the ∼46,000 and ∼26,500 BLUEs detected by MMIA in the investigated period (1 April 2019–31 March 2021) and distributed according to GD-1 and GD-2, respectively. Clear impulsive single pulse BLUEs are considered when the fittings have R 2>0.75, being R 2 the so-called coefficient of determination (used as a metric of fit goodness), which can change between negative values and 1 (perfect fit). A number of correlations have been established between the above mentioned characteristics of BLUEs. By assuming a point-like source for the light source deep in the cloud, the first hitting time (FHT) fit (Soler etal.,2020) provides a first approximation to the single pulse BLUE depth with respect to cloud tops within thunderstorms, 337nm peak power density (and total brightness), rise and total times (Luque etal.,2020; Soler etal.,2020). The tail of the light curve from an extended source (see Figure S2 of the Supporting InformationS1) that spans altitudes close to the cloud top to a maximum distance L0 inside the cloud can also be fitted to obtain the bestfit cutoff (characteristic photon diffusion) time τD= 𝐴𝐴𝐴𝐴 2 0 /4D, and the total number of source photons (N) (Li etal.,2021), with D being a diffusion coefficient (Soler etal.,2020). For short times after the optical emission and assuming that the mean absorption time (τA) of the photons inside the cloud is much larger than τD, the fit discussed in Li etal.(2021) predicts a ∼(𝜏𝐷𝑡)−1∕2 dependence for the photon flux exiting the cloud top.From N and τD one can obtain the number of streamers in the BLUE as well as its maximum length L0 (Li etal.,2021) (for more details, see the Supporting InformationS1). We have assumed that extended sources are located in the perfect nadir (no angle with vertical). However, this might not always be the case and such an assumption could underestimate the total optical energy of all the 337nm photons emitted by a BLUE. Finally, we found that ∼2,700 events (with R 2>0.75) in GD-1 and ∼262 events (with R 2>0.75) in GD-2 can be fit by both the point-like and the extended-source models (see Table1). Likely these are cases where the optical source is both small and close to the cloud top. Table1 shows the total number of BLUE events in GD-1 and GD-2 discriminating according to their type (point-like or extended). Note that for the point-like sources the total time is derived from the FHT model fitting, while for the extended sources the total time is obtained from the raw data (since the fitting used for the extended sources does not include the rise). Figure 3. Nighttime seasonal distribution of Blue LUminous Events (BLUE) electrical activity in thunderclouds associated to GD-1 in Figure1. The global nighttime seasonal BLUE rates are: 6.7 (SON), 4.7 (DJF), 5.9 (MAM), and 6.8 (JJA) BLUEs s −1. These maps are generated using 2°×2° grid cells. The BLUEs shown can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses or any other irregular pulse shape. 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 5 of 16 4. Results and Discussion In order to provide some perspective to the reader, Figures5 and6 show four maps that display, in four intervals of peak power density (PPD), the geographical distribution (per 2°×2° grid cell) of the ∼46,000 and ∼26,500 BLUEs detected by ASIM-MMIA corresponding to the GD-1 and GD-2 annual average distributions. Panels (a), (b), (c), and (d) of Figures5 and6 show BLUEs with PPD <25μWm −2, between ≥25 and 50μWm −2, between ≥50 and 100 µWm −2, and ≥100μWm −2, respectively. In regard to the ∼53,000 BLUEs earlier reported by Soler etal.(2021), there are 3183 with PPD ≥100μWm −2. Of these 3,183 and 2,992 events, that is, ∼94% (mostly in latitudes above 35°N and below 35°S) concentrate between 1 September 2018 and 31 March 2019 during the first 7 months of the 2-year period evaluated by Soler etal.(2021), while in the following 17months there are only ∼6% of the 3,183 BLUEs worldwide with PPD >100μWm −2. Interestingly, we only found 282 BLUEs with PPD >100 μWm −2 when searching (with exactly the same algorithm as in Soler etal.[2021]) in the two years 1 April 2019–31 March 2021. It is important to note that on March 2019 there was an update of the ASIM-MMIA cosmic ray rejection algorithm software (ON only over the SAA before March 2019, ON everywhere after March 2019) that could have influenced the above findings. The trend described above for the PPD ≥100μWm −2 range is also identified (using exactly the same algorithm as in Soler etal.[2021]) in the other three lower PPD ranges between 1 September 2018 and 31 March 2019. However, the difference in the number of BLUEs (relative importance) of each PDD range with respect to successive periods of 7 months (1 September 2019–31 March 2020, and 1 September 2020–31 March 2021) is ∼6%. Consequently, in our study, we decided to move forward 7months the period of time chosen to carry out the analysis of characteristics of BLUEs presented here. 4.1. Distribution of BLUEs According to Altitude and Peak Power Density The worldwide distribution of BLUE altitudes in terms of latitude is displayed in Figures7a and7c for GD-1 and (b, d) for GD-2. All panels in Figures7 and8 except panels (c, d) of Figure7 display 5,374 (GD-1) BLUE events (∼12% of ∼46,000) and 2,242 (GD-2) BLUE events (∼10% of ∼26,500) with good quality fitting (R 2>0.75), that is, those that are closest to impulsive single pulse point-like sources. However, panels (c, d) of Figure7 shows most of the ∼46,000 BLUEs of GD-1 (panel (c)) and ∼26,500 BLUEs of GD-2 (panel (d)) detected by ASIM-MMIA that is fittable independently of the fitting quality, that is, for any value of R 2. Thus, Figure7 (c, d) shows the altitude distribution of BLUE events which 337nm light curves might not be fully explained by single pulse point-like sources. The approximate altitude distributions of BLUEs (see Figure 7 a–d) are derived by assuming that the global cloud top height distribution can be approximated by the annual mean variation of the tropopause heights with the latitude (Heumesser et al., 2021; Offroy et al., 2015). BLUEs can, however, occur associated with deep convection scenarios and are often detected in overshooting cloud tops extending across the local tropopause (Liu etal.,2018), which would break down the above-mentioned assumption and can introduce an altitude uncertainty of ∼1–3km (Liu etal.,2018). Figure 4. Nighttime seasonal distribution of Blue LUminous Events (BLUE) electrical activity in thunderclouds associated to GD-2 in Figure2. The global nighttime seasonal BLUE rates are: 3.7 (SON), 2.6 (DJF), 3.7 (MAM), and 4.0 (JJA) BLUEs s −1. These maps are generated using 2°×2° grid cells. The BLUEs shown can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses or any other irregular pulse shape. GD-1 GD-2 Eliminate events worldwide with (τrise≤40μs and τtotal≤150μs) TOTAL 46,283 26,355 Point-like source (R 2>0.75) 5,374 2,242 Extended source (R 2>0.75) 4,258 553 Common (R 2>0.75) 2,717 262 Note. That for the Point-Like Sources the Total Time Is Derived From the First Hitting Time (FHT) Model Fitting, While for the Extended Sources the Total Time Is Obtained From the Raw Data. Table 1 Total Number of BLUE Events in GD-1 and GD-2 Discriminating According to Their Type (Point-Like or Extended) 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 6 of 16 In order to compare and/or validate the geographical (Figure2) and altitude (Figures7b, 7d) distributions of BLUEs obtained in this paper for GD-2 we compared them (see Figure S17 and Figure S18 of the Supporting InformationS1) with several case-based studies of BLUEs already published (provided the events occurred within the dates of our climatology). Figure S17 of Supporting InformationS1 shows local maps with numbers 1, 2, 3, and 4 on them indicating the centroid of the set of BLUEs investigated in the case-based studies reported in (a) Li etal.(2021) and Liu, Lu, Neubert, etal.(2021) over Southern China, (b) López etal.(2022) over Colombia, (c) Soler etal.(2020) over Indonesia, and (d) the study by Li etal.(2022) about multiple pulse BLUEs over nearby Malaysia. The numbers 1 through 4 in Figure S18 in Supporting InformationS1 are placed in the mean altitude and in the centroid (in terms of latitude and longitude) of the reported BLUE events in the mentioned local studies. A number with a ′ indicates the mean height obtained by radio (VLF/ LF) for the same set of BLUE events. The red dashed lines (above and below the solid red line) indicate the ±3km uncertainty associated with our approximation to globally compute the tropopause in each latitude. It can be seen that strong thunderstorms with overshooting tops penetrating into the lower stratosphere (cases 1 and 4) can create some discrepancies between the heights obtained by optical (applied globally in our analysis) and radio but, in any case, within the uncertainty of the method chosen to represent the tropopause at a global scale. The red line in panels (a)–(d) in Figure7 marks the annual mean variation of the tropopause heights with latitude. Two zonal layers of BLUEs (see greenish zones) can be distinguished: a top layer of shallow BLUEs (≤0.5km depth), and a deeper (2.5–5km depth) layer of BLUEs mostly located among tropical latitudes (10°S to 20°N). Figures7e, 7f and Figure8 show detailed GD-1 (left column) and GD-2 (right column) latitudinal and longitudinal distributions of impulsive single pulse BLUEs with respect to their different depths below thundercloud tops around the globe and peak power densities. A thin and shallow layer of BLUEs at a depth ≤1km below cloud tops is visible across all latitudes (see Figures7e and7f). This usually corresponds to the location of negative NBEs (Li etal.,2021; Wu etal.,2014). A second thicker and deeper layer between ∼1.25km and ∼3.25km below thundercloud tops is also distinguishable within the tropics and part of the subtropical regions (Figures7e and7f). The longitudinal distributions of depths clearly exhibit three BLUE chimneys (see Figures8c and8d) with the American chimney in GD-1 (see Figure8c) being more populated due to the South America’s contribution. A fourth dim chimney is also visible in the Pacific Ocean. These changes in GD-2 where the most populated chimney is Asia/Australia (see Figure8d) and with the Pacific chimney remaining. Figures7e, 7f and8 show detailed GD-1 (left column) and GD-2 (right column) latitudinal and longitudinal distributions of impulsive single pulse BLUEs with their peak power density (μW/m 2). Large (≥50μW/m 2) peak power BLUE events appear scattered across all latitudes and longitudes. Most single pulse BLUEs exhibit peak power densities <25μW/m 2 (as shown in Figures5 and6 for BLUES with all sorts of pulse shapes). BLUEs with peak power densities below 25μW/m 2 are common within the tropics, within the Indian subcontinent, and near the ocean of the Asia/Australia BLUE chimney. The three BLUE chimneys appear in the meridional distribution of BLUEs shown in panels (c) through (f) of Figure8. Figure 5. Geographical distribution (in 2°×2° grid cells) of the ∼46,000 nighttime GD-1 Blue LUminous Events (BLUEs) detected by Atmosphere-Space Interaction Monitor-Modular Multispectral Imaging Array in the two-year period from 1 April 2019 to 31 March 2021. Panels (a), (b), (c) and (d) show BLUEs of any shape with peak power density (PPD) between ≥3 and <25μWm −2, between ≥25 and <50μWm −2, between ≥50 and <100μWm −2, and ≥100μWm −2, respectively. There are 39,980 BLUEs in panel (a) (19990/year), 4,367 in panel (b) (2,183.5/year), 833 in panel (c) (416.5/year), and 282 in panel (d) (141/year). The BLUEs shown in these maps can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses or any other irregular pulse shape. 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 7 of 16 4.2. Global Distribution of BLUEs According to Rise and Total Times Figure9 presents zonal (a)-(d) and meridional (e)-(h) distributions of the rise and total duration times of impulsive single pulse BLUEs of GD-1 [left column] and GD-2 [right column]), respectively. Note that the rise and total times are calculated as the elapsed times since the raw (or fitted) signal is above 10% of the maximum until it reaches the maximum (rise time), and until it passes the maximum and decreases again to 10% of the maximum (total time). For point-like sources, the rise and total time are derived from the FHT model fitting, while for extended sources the total time is obtained from the raw data. The presence of the SAA can be clearly seen between 5°S and 20°S in panels (a, c) of Figure9 for GD-1, and between 0° and 100°W in panels (e, g) of Figure9 for GD-1. Single pulse BLUEs with fast (≤30μs) rise times and short (≤0.5ms) total times occur across all latitudes and longitudes. Zonal distributions show that the single pulse BLUEs concentrated within the tropics exhibit rise and total times ranging from 50μs to 0.5ms, and from 0.8ms to ∼4ms, respectively. Complementarily, meridional distributions displayed in Figures9e and9h show that single pulse BLUEs in the three main chimneys exhibit roughly the same rise and total duration times with the Europe/ Africa chimney being the one with slightly faster and shorter rise times and durations. The left/right columns of Figure10 for GD-1/GD-2 show the connection between the peak power density (μWm −2) and total brightness (μW mm −2) of impulsive single pulse BLUEs and their rise and total duration times. Both magnitudes (peak power density and maximum brightness) exhibit two clear populations. The most numerous population includes single pulse BLUEs with fast (≤50μs) rise times and short (≤0.5ms) durations reaching peak powers and total brightnesses of up to 200 and 30–40μWmm −2, respectively. The second group of single pulse BLUEs reaches longer rise times (up to ∼0.8ms) and total times (up to ∼4ms) associated with lower peak powers (≤50μWm −2) and total brightnesses (∼25μWmm −2). The two groups of BLUEs with fast (≤50μs) and slow (>50μs and up to ∼0.8–1.0ms) rise times are consistent with those mentioned in Husbjerg etal.(2022). 4.3. Correlations Between BLUE's Depth and Their Rise and Total Times The left/right columns of Figure11 for GD-1/GD-2 represent how light scattering affects key features (rise time and total time) of the 337nm light curves of BLUEs. In order to appropriately visualize the relationships between the rise and total times we have plotted the total and rise times as a function of the depth below the cloud top (panels (a, c) for GD-1 and (b, d) for GD-2), and the total time versus rise time for GD-1 (panel (e)) and for GD-2 (panel (f)). In general, the closer the BLUE source is to the cloud top, the faster rise times and shorter duration times due to the weak scattering (as reported by Li etal.[2021]). On the contrary, when BLUE sources are deeply buried in storm clouds (like the ones in Soler etal.[2020]), scattering by cloud droplets and ice crystals blurs their image as observed from above and produced 337nm light curves characterized by relatively long rise times (0.2−0.5ms) and total duration times (>1.5ms). Therefore, shallow (≤1km depth) single pulse BLUEs exhibit fast rise times below ∼50μs and total duration times below ∼0.5ms. Deeper single pulse BLUES occurring between ∼1 and ∼4km below cloud tops are characterized by rise and total times that increase up to about 0.8ms (for 4km) and 5ms (for ∼3km), respectively. The bottom branch of Figure11a and11c for GD-1 also shows some few BLUEs that, even occurring deeper (1–8km below cloud tops) in thunderclouds, still exhibit fast rise times of ≤60μs and Figure 6. Geographical distribution (in 2°×2° grid cells) of the ∼26,500 nighttime GD-2 Blue LUminous Events (BLUEs) detected by Atmosphere-Space Interaction Monitor-Modular Multispectral Imaging Array in the two-year period from 1 April 2019 to 31 March 2021 removing events in all the planet (including the SAA) with rise time (τrise)≤40μs and total duration (τtotal)≤150μs. Panels (a), (b), (c), and (d) show BLUEs of any shape with peak power density (PPD) between ≥3 and <25μWm −2, between ≥25 and <50μWm −2, between ≥50 and <100μWm −2, and ≥100μWm −2, respectively. There are 25,720 BLUEs in panel (a) (12,860/year), 380 in panel (b) (190/year), 52 in panel (c) (26/year), and 204 in panel (d) (102/year). The BLUEs shown in these maps can have any temporal shape, that is, no distinction has been made here among BLUEs with a single (impulsive or not) pulse, multiple pulses or any other irregular pulse shape. 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 8 of 16 short durations (≤200μs). This result is not physically possible and should be disregarded. These few unphysical events in Figure11a, 11c for GD-1 disappear almost completely in Figures11b, 11d for GD-2. Figure11 (e) for GD-1 and Figure11f for GD-2 show the connection between the total time duration of single pulse BLUEs and their rise time. For fast (≤40μs) rise times, the total duration ranges from ∼100 to ∼300μs. However, as the rise time increases beyond ∼40μs, panels (e, f) of Figure11 show the main branch in the top associated with single pulse BLUEs with relatively long (0.4–5ms) total durations. It is also interesting to note that, as shown in panels (a) for GD-1 and (b) for GD-2 of Figure10, fast (≤40μs) rise time BLUEs come with a 337 peak power density that can reach values of up to ∼200μWm −2. When rise times are beyond 50μs the BLUEs' peak power density stays below ∼50μWm −2 and exhibits a decreasing trend as rise time increases from ∼0.1ms to ∼1ms. Figure 7. Approximate altitude distributions of GD-1 (left column) and GD-2 (right column) impulsive single pulse Blue LUminous Events (BLUEs) (with R 2>0.75) (a), (b), all sort of first hitting time (FHT) fittable BLUEs in GD-1 and GD-2 with any value of R 2 (c), (d). Zonal distributions of GD-1 and GD-2 impulsive single pulse BLUE depths (e), (f). Note that, always, GD-1 is in the left column and GD-2 is in the right column. The colorbar indicates number of events. 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 9 of 16 4.4. Global Zonal/Meridional Distributions of BLUEs' Streamers and Lengths Figure12 presents the zonal and meridional distributions of the lengths (L0) and the number of streamers (k) for GD-1 (left column) and GD-2 (right column) of single pulse BLUEs, respectively. The lengths of most single pulse BLUEs roughly vary between ∼100 and ∼1,500m in GD-1 (see panels (a) and (c) in Figure12). However, the group of tropical BLUEs with lengths between 2km and up to ∼5km dominate in GD-2 (see panels (b) and (d) in Figure12). According to our analysis, the number of streamers in single pulse BLUEs ranges between 10 8 and 10 9 in agreement with previous results (Cooray etal.,2020; Li etal.,2021; Liu etal.,2019). BLUEs with most streamers concentrate between 20°S and 20°N (see panels (e) for GD-1 and (f) for GD-2 of Figure12) with the tropical band including some BLUEs with up to ∼2–3×10 9 streamers. The meridional distribution of BLUEs' streamers (see panels (g) for GD-1 and (h) for GD-2 of Figure12) exhibits a three chimney structure with the Asia/Australia chimney including some BLUEs with up to ∼2–3×10 9 streamers. Figure 8. Zonal distributions of GD-1 and GD-2 impulsive single pulse Blue LUminous Events (BLUE) peak power density (a), (b). Meridional distributions of impulsive single pulse BLUE depths (c), (d) and peak power density (e), (f). Note that, always, GD-1 is in the left column and GD-2 is in the right column. The colorbar indicates number of events. 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Journal of Geophysical Research: Atmospheres SOLER ETAL. 10.1029/2022JD037535 16 of 16 Minschwaner, K., Kalnajs, L., Dubey, M., Avallone, L., Sawaengphokai, P., Edens, H., & Winn, W. (2008). Observation of enhanced ozone in an electrically active storm over socorro, nm: Implications for ozone production from corona discharges. Journal of Geophysical Research, 113(D17), D17208. https://doi.org/10.1029/2007jd009500 Miyamoto, H., Fenu, F., Barghini, D., Battisti, M., Belov, A., Bertaina, M., etal. others (2021). Simulations studies for the mini-euso detector. arXiv preprint arXiv:2112.12150. Montanyà, J., López, J. A., Morales Rodriguez, C. A.,van der Velde, O. A., Fabró, F., Pineda, N., etal., others (2021). A simultaneous observation of lightning by asim, Colombia-lightning mapping array, glm, and iss-lis. Journal of Geophysical Research: Atmospheres, 126(6), e2020JD033735, https://doi.org/10.1029/2020jd033735 Neubert, T., Chanrion, O., Heumesser, M., Dimitriadou, K., Husbjerg, L., Rasmussen, I. L., etal. (2021). Observation of the onset of a blue jet into the stratosphere. Nature, 589(7842), 371–375. https://doi.org/10.1038/s41586-020-03122-6 Offroy, M., Farges, T., Kuo, C. L., Chen, A. B.-C., Hsu, R.-R., Su, H.-T., etal. (2015). Temporal and radiometric statistics on lightning flashes observed from space with the isual spectrophotometer. Journal of Geophysical Research: Atmospheres, 120(15), 7586–7598. https://doi. org/10.1002/2015jd023263 Pérez-Invernón, F. J., Gordillo-Vázquez, F. J., Smith, A. K., Arnone, E., & Winkler, H. (2019). Global occurrence and chemical impact of stratospheric Blue Jets modeled with WACCM4. Journal of Geophysical Research: Atmospheres, 124(5), 2841–2864. https://doi. org/10.1029/2018JD029593 Rison, W., Krehbiel, P.R., Stock, M. G., Edens, H. E., Shao, X.-M., Thomas, R. J., etal. (2016). Observations of narrow bipolar events reveal how lightning is initiated in thunderstorms. Nature Communications, 7(1), 10721. https://doi.org/10.1038/ncomms10721 Schumann, U., & Huntrieser, H. (2007). The global lightning-induced nitrogen oxides source. Atmospheric Chemistry and Physics, 7(14), 3823– 3907. https://doi.org/10.5194/acp-7-3823-2007 Soler, S., Gordillo-Vázquez, F., Pérez-Invernón, F., Luque, A., Li, D., Neubert, T., etal. (2021). Global frequency and geographical distribution of nighttime streamer corona discharges (blues) in thunderclouds. Geophysical Research Letters, 48(18), e2021GL094657. https://doi. org/10.1029/2021gl094657 Soler, S., Pérez-Invernón, F. J., Gordillo-Vázquez, F. J., Luque, A., Li, D., Malagón-Romero, A., etal. (2020). Blue optical observations of narrow bipolar events by ASIM suggest corona streamer activity in thunderstorms. Journal of Geophysical Research: Atmospheres, 125(16), e2020JD032708. https://doi.org/10.1029/2020jd032708 Tilles, J. N., Liu, N., Stanley, M. A., Krehbiel, P.R., Rison, W., Stock, M. G., etal. (2019). Fast negative breakdown in thunderstorms. Nature Communications, 10(1), 1648. https://doi.org/10.1038/s41467-019-09621-z Wescott, E. M., Sentman, D., Osborne, D., Hampton, D., & Heavner, M. (1995). Preliminary results from the Sprites94 aircraft campaign: 2. Blue jets. Geophysical Research Letters, 22(10), 1209–1212. https://doi.org/10.1029/95GL00582 Wescott, E. M., Sentman, D. D., Heavner, M. J., Hampton, D. L., Osborne, D. L., & Vaughan, O. H. (1996). Blue starters: Brief upward discharges from an intense Arkansas thunderstorm. Geophysical Research Letters, 23(16), 2153–2156. https://doi.org/10.1029/96GL01969 Wiens, K. C., Hamlin, T., Harlin, J., & Suszcynsky, D. M. (2008). Relationships among narrow bipolar events, “total” lightning, and radar-inferred convective strength in Great Plains thunderstorms. Journal of Geophysical Research, 113(D5). https://doi.org/10.1029/2007jd009400 Wu, T., Yoshida, S., Ushio, T., Kawasaki, Z., & Wang, D. (2014). Lightning-initiator type of narrow bipolar events and their subsequent pulse trains. Journal of Geophysical Research: Atmospheres, 119(12), 7425–7438. https://doi.org/10.1002/2014jd021842 Zahn, A., Brenninkmeijer, C., Crutzen, P., Parrish, D., Sueper, D., Heinrich, G., & Heintzenberg, J. (2002). Electrical discharge source for tropospheric “ozone-rich transients”. Journal of Geophysical Research, 107(D22), 4638. ACH–16. https://doi.org/10.1029/2002jd002345 Zhu, Y., Stock, M., Lapierre, J., & DiGangi, E. (2022). Upgrades of the Earth networks total lightning network in 2021. Remote Sensing, 14(9), 2209. https://doi.org/10.3390/rs14092209 21698996, 2022, 24, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022JD037535 by CSIC Organizacion Central OM (Oficialia Mayor) (Urici), Wiley Online Library on [06/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License