Lightning is a critical hazard affecting infrastructure, public safety, and operational sectors, including aviation and energy. Accurate nowcasting of lightning remains a challenge due to the rapid development of convective storms and limitations in current forecasting systems. This study proposes a novel, high-resolution lightning nowcasting model that leverages satellite-derived data from the Geostationary Lightning Mapper (GLM) and the Advanced Baseline Imager (ABI), both aboard NOAA’s GOES-16 and GOES-19 satellites. By utilizing machine learning techniques, particularly linear models, the proposed method aims to provide interpretable and operationally feasible forecasts up to 120 minutes in advance. This research addresses the need for globally scalable and radar-independent forecasting systems, especially in data-sparse regions. GLM provides continuous, full-disk lightning detection, while ABI captures multispectral imagery indicative of cloud dynamics. Key features such as cloud-top cooling rates, brightness temperature differences (BTDs), and lightning flash density are extracted and processed on a unified 10 km spatiotemporal grid. Feature engineering incorporates both temporal evolution and local spatial context to enhance model sensitivity to convective initiation. The study prioritizes interpretable models, such as logistic regression and regularized linear classifiers, over deep learning methods, which, while powerful, are often computationally intensive. These linear models are trained to classify the likelihood of lightning occurrence and predict flash rates using a spatiotemporal block cross-validation strategy, ensuring robustness across different regions and meteorological conditions. The results include probabilistic nowcast maps, performance comparisons, and a detailed analysis of feature importance. By isolating the contributions of each satellite-derived variable, the study aims to clarify the physical processes associated with lightning occurrence and enhance early warning capabilities. This research proposes a scalable, explainable, and efficient nowcasting tool that enhances global lightning risk management. It aligns with international initiatives for satellite-based severe weather monitoring and promises significant operational benefits, particularly in regions with limited radar or ground-based lightning detection coverage.

How to cite: Beneti, C., Castelo Branco, K., Pavam, L., and Calvetti, L.: Lightning Nowcasting Using GLM and GOES-ABI Data in a High-Resolution Machine Learning-Based Predictive Model, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-42, https://doi.org/10.5194/ecss2025-42, 2025.

The Lightning Imager (LI), onboard the newly launched Meteosat Third Generation (MTG) satellite, represents a major advancement in spaceborne optical lightning detection over Europe, Africa, and partly also South America. Although an estimate of the detection efficiency of the LI instrument is theoretically known, its validation under realistic conditions is very necessary and desirable for future calibration purposes. This study presents the first validation of LI detections using data from the Earth Networks Total Lightning Network (ENTLN), a ground-based system that observes both intra-cloud and cloud-to-ground lightning. The validation will be performed for the period of October 2024 over eastern South America, an area with frequent and diverse convective activity. To ensure accurate spatial comparison, LI flash locations will be corrected for parallax displacement prior to validation. A spatiotemporal coincidence matching algorithm will be employed to associate LI and ENTLN flashes, allowing for small spatial and temporal offsets rather than requiring exact matches. Subsequently, various verification metrics, such as detection efficiency, false alarm rate and detection ratio will be computed based on the matched events. In addition, a temporal analysis will be performed using hourly total lightning flash counts, including a comparison of diurnal cycle to assess temporal consistency between the two datasets. Based on the simplifying assumption that brighter optical flashes generally correspond to stronger electrical activity, the study will further investigate the relationship between LI-measured radiance and peak current derived from ENTLN.

How to cite: Bliznak, V. and Sokol, Z.: First validation of the Lightning Imager on board Meteosat Third Generation with Earth Networks Total Lightning Network, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-50, https://doi.org/10.5194/ecss2025-50, 2025.

 In 2024, a total of 145,784 cloud-to-ground(CG) lightning strikes were observed over inland regions of the Korean Peninsula, representing a 43.83% increase compared to the 10-year average from 2015 to 2024 and nearly a 99% increase compared to the previous year (Lightning Annual Report, 2024). Lightning events predominantly occur during between June and September, when atmospheric instability promotes vigorous convective cell development. These lightning events, often associated with rapidly evolving convective systems, can cause significant human casualties as well as substantial socio-economic damage.
 To monitor lightning activity in real time, the Korea Meteorological Administration (KMA) operates 21 LINET (Lightning NETwork) observation systems around the Korean Peninsula. In support of operational forecasting, KMA has also developed and operates several lightning prediction models. Two major lightning nowcasting models are currently in use. The first model is a lightning nowcasting model based on radar motion vectors using MAPLE (McGill Algorithm for Precipitation Nowcasting by Lagrangian Extrapolation), which generates lightning forecasts at 10-minute intervals for up to 6 hours. One-hour forecasts are delivered through a mobile application that provides users with location-based lightning alerts, enabling them to be informed of lightning threats within the next hour regardless of their location.
 The second model detects lightning initiation signals using data from three-dimensional dual-polarization radar, hydrometeor classification, and objectively analyzed temperature fields. This Real-Time Radar-Based Lightning Risk Alert Service radar-based detection system identifies areas with a high probability of lightning occurrence by analyzing dual-polarization parameters (Z, ZDR, KDP, VIL) and hydrometeors (graupel, hail) at altitudes below 0°C. The results are provided to forecasters in two-dimensional form every 5 minutes at a 500-meter spatial resolution. These data are useful for identifying high-risk areas 5 to 20 minutes before lightning occurrence, enabling preemptive response actions. KMA continues to improve the accuracy of lightning prediction and enhance service capabilities, with the goal of minimizing lightning-related damage and ensuring public safety.

How to cite: Kim, H. L., Son, M., and Suk, M.-K.: A Real-Time Radar-Based Lightning Prediction and Risk Alert System, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-95, https://doi.org/10.5194/ecss2025-95, 2025.

Lightning is commonly associated with both severe and non-severe convective storms. Numerous studies have documented that lightning occurs more frequently over land than over ocean, a contrast that is clear in the Australian lightning record. Many hypotheses have been proposed to explain the differences. For instance, deeper convection over land increasing the opportunity for microphysical interactions responsible for lightning, the presence of greater number of fine aerosols over land enhancing the number of cloud condensation nuclei; or coarse sea salt over the ocean promoting drop coalescence.

To investigate this contrast in Australia’s tropics, lightning strike data from ground-based lightning observations from the Weatherzone Total Lightning Network were analysed in conjunction with satellite-based overshooting top (OT) data. The results reveal a pronounced land—ocean difference in lightning frequency, particularly in northern Australia. However, the OT data show no corresponding contrast in tropopause penetrating convection, suggesting that the lighting difference may be associated with weaker storms.

Further analysis of convective parameters from the BARRA-R2 reanalysis indicates that lightning-producing storms over land typically occur in environments characterized by lower convective available potential energy (CAPE), higher convective inhibition (CIN), steeper lapse rates, lower cold cloud depth and lower freezing level heights than lightning-producing storms over the ocean. These findings suggest that lightning initiation is more easily achieved over land than over ocean, showing that the causes of the observed land-ocean contrast is more complicated than storms being stronger over land and giving insights into the observed differences in lightning activity.

How to cite: Patel, T., Raupach, T. H., Sherwood, S. C., and Warren, R. A.:  Understanding the land—ocean contrast in lightning strikes across the Australian tropics, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-149, https://doi.org/10.5194/ecss2025-149, 2025.

This research identifies termodynamic and kinematic indices that positively influence elevated values of cloud-to-ground lightning flashes (CGs) in Poland. The analysis used data from the PERUN lightning detection system from IMGW-PIB and ERA5 (ECMWF) reanalyses for the period 2002-2020. In addition, a spatial-temporal distribution analysis was carried out for the period 1940-2022, covering the key parameters necessary for the appearance of convection. Results showed that thunderstorms most often occur in the summer, but also that there are increasingly favorable conditions for the appearance of organized multicellular systems in the spring. CG flashes most often form in a most-unstable convective available potential energy (MU CAPE) environment of about 1300 J/kg along with deep layer shear (DLS 0-6 km AGL) of 13-14 m/s. Using the WMAXSHEAR parameter, it was possible to conclude that overlapping CAPE and DLS values of about 500 m2/s2 imply increased electrical activity. At the same time, a high correlation with the Hail Size Index (HSI) parameter implies a positive relationship between the occurrence of hailstorms and an increased number of CGs generated in the case of supercells. The research also found a gradual increase over evaluated timeframe in air temperature, MU CAPE, MU Mixing Ratio and the MU WMAXSHEAR parameter for the area under study.

How to cite: Sulik, S. and Taszarek, M.: The kinematic and thermodynamic environment during cloud-to-ground lightning occurrence in Poland, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-212, https://doi.org/10.5194/ecss2025-212, 2025.

The Lightning Imager (LI) instrument aboard the first Meteosat Third Generation (MTG) satellite provides valuable data for detecting and monitoring convective storms by capturing localized pulses of visible light emitted by lightning. These detections have a spatial resolution of approximately 10 km and occur at a high temporal resolution of 1 kHz, or one detection per millisecond.

Following a sophisticated filtering process developed at EUMETST to eliminate false positives, lightning detections are identified as pixels that illuminate for at least 1 millisecond. These are then grouped based on spatial proximity into “groups,” which are subsequently combined into “flashes”, composite lightning events lasting at least 1 ms. Each group is assigned a central location, weighted by the intensity of radiation detected at each contributing pixel or event. This method allows for an effective spatial resolution finer than the nominal 10 km.

At ESSL, we have developed a visualization technique in which individual lightning groups are connected by line segments. The algorithm begins with the most centrally located group from the first time step of a flash. It then progressively links the nearest groups to the existing network, one millisecond at a time. This sequential process builds a branching, dendritic structure that resembles the complex geometry of actual lightning strikes. While the resemblance to real lightning structures cannot be quantitatively verified, notably large and long-duration flashes tend to appear in regions where such events are commonly observed by ground-based systems or Lightning Mapping Arrays (LMAs) such as stratiform precipitation areas of mesoscale convective systems.

We will show examples of this visualization and explore its potential utility in forecasting and warning operations. We will also share insights from its use atg the EUMETSAT–ESSL Testbeds.

How to cite: Groenemeijer, P., Holzer, A. M., and Pucik, T.: Visualizing the geometry of lightning detected by MTG’s Lightning Imager, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-297, https://doi.org/10.5194/ecss2025-297, 2025.

We investigated the link between satellite-based lightning detection and severe weather occurrences, focusing on the hypothesis that severe thunderstorms with strong updrafts would produce a high quantity of relatively small lightning flashes. This study utilizes data from the Meteosat Third Generation Lightning Imager (MTG-LI) and ground reports from the European Severe Weather Database (ESWD), encompassing 22.7 million lightning flashes and 26,000 severe weather reports from July 2024 to May 2025.

While over half of all severe weather events were not accompanied by ‘small’ lightning flashes, size boundary defined by the 25^th percentile of the lightning size distribution of the data, a distinct subset of events, particularly large hail, exhibited an exceptionally high concentration, with some instances exceeding 1900 small flashes within the 45- minute window and 20km radius. Specifically, 19.7% of all severe weather events, with and without lightning, were associated with a ‘high’ density of small flashes, a threshold determined by the 75th percentile of the distribution of a density grid specifically designed for this research. Results showed that hail events are notably more correlated with a higher concentration of small lightning flashes compared to other severe weather types, with this correlation increasing significantly with hail size, reaching up to 83.3% for hail over 8cm.  We will report on our efforts to geographically and statistically distinguish regions less prone to reporting, aiming to improve the reliability of hotspot-to-event correlations. This study highlights the potential of MTG-LI data as a valuable indicator for nowcasting severe weather, especially for large hail events.

How to cite: Waymel, C., Groenemeijer, P., and Púčik, T.: Correspondence between MTG-LI satellite-based lightning detections and severe weather reports across Europe, 12th European Conference on Severe Storms, Utrecht, The Netherlands, 17–21 Nov 2025, ECSS2025-322, https://doi.org/10.5194/ecss2025-322, 2025.