Earth Network Research Articles

Probing the D-region ionosphere globally with Earth Networks Total Lightning Network data

An existing technique to use broadband lightning waveforms to probe the D-region ionosphere (60–90 km altitude) is shown to be extendable to a global scale using the Earth Networks Total Lightning Network (ENTLN). This paper demonstrates the technique in detail on a region of the Southeastern United States. This demonstration shows that diurnal D-region height variation and smaller time-scale variations on the order of tens of minutes to hours are evident in the measurement. The technique is then extended to three additional global regions on this same day: Northeastern U.S., India, and Japan. The diurnal behavior between these different regions is compared to a D-region model from the International Reference Ionosphere.

Evaluation of Detection Efficiency of World Wide Lightning Location Network in Southeast Asian Region

AbstractThe detection efficiency (DE) of the World Wide Lightning Location Network (WWLLN) is evaluated in Southeast Asia by comparing WWLLN data with the Earth Network Total Lightning Network (ENTLN) data taking into account time, distance, and peak‐current parameters. The performance of WWLLN is evaluated in the months of November and December in two different years (2020–2021). These periods are selected to assess the change (if any) in DE overtime and the inclusion of new stations. The strokes between the two networks were considered matched if they fell within a time criterion of 100 µs and a location difference of 25 km. Using this criterion, 5.2 × 106 WWLLN strokes were matched with ENTLN cloud‐to‐ground (CG) lightning strokes in November‐December 2020, resulting in a DE of 32.9%. Similarly, 4.6 × 106 WWLLN strokes were found to match in November‐December 2021, yielding a DE of 36.5%. Analysis of the peak‐currents reveals that DE is lowest (<10%) for a peak‐current below ±10 kA. However, for peak‐current exceeding ±50 kA, the DE increases to ∼60%. During November‐December 2021, WWLLN reported 38.95 × 106 lightning strokes globally; amongst them, Dhaka station detected 0.5 × 106 strokes, contributing to a 1.3% increase in the global DE. Dhaka station detects most lightning strokes within 8 × 103 km, which diminishes to zero after 10 × 103 km. The Dhaka station recorded a larger number of strokes at longer detection distances during midnight (22:00–02:00) than during noon (10:00–14:00). The results signify a positive impact of the Dhaka station on WWLLN's DE during the mentioned period.

Research and Investigation of High Frequency Profile of Step and Touch Voltages Due to Fragmentation of the Grounding Grid

The increasing demand and the benefits of electricity energy have led to the creation of huge power plants, so high voltage substations are necessary for electrical energy. Installing different small power plants to prevent consumption in different parts of a country requires reservation units and high costs for repairs, maintenance and fueling. The necessity of these substations can be searched in the lack of uniformity of different costs, the costs of generating electrical energy, the lack of consumption centers near power plants, and generally the existence of a national grid and the possibility of timely disconnection of the faulty parts of the power grid. The priority in the rehabilitating of a strong substation is to meet the system's operation needs. In all electrical installations, especially in industrial complexes, land connectivity is one of the most important and fundamental measures to protect individuals and equipment and improve the performance of the system. The purpose of ground equipment system is to potentialize the metal body of all electrical equipment with the ground to prevent electrocution and protect personnel. This type of connection is not normally a current carrier, but if one of the phases of the device is connected to the fuselage, the fault current enters the ground through this connection and closes its path through the trans neutral or the generator connected to the ground. Earth networks due to different structures of the earth in different regions may cause problems over time, one of them is the fragmentation of the earth's network due to rot of conductors or due to overvoltage. In this research at different and high frequencies, we use 3 frequencies of 50, 100 kilos and 1 MHz, to check the ground network factors and the security of equipment and people in high voltage substations, we use CDEGS software which is a specialized program for designing and investigating the earth network. Objetivo: Research and investigation of high frequency, profile of step and touch voltajes of grounding grid. Metodología: Design and modeling with Comsol software. Resultados: As the frequency increases, both the step voltage and the touch voltage at the corner points increase more. Conclusiones: By comparing the voltage profiles based on the simulation results, it was found that by increasing the frequency of the contact voltage in each profile and mesh, it is noticeably increased. The point to consider after reviewing the simulation results is that by increasing the frequency of both step voltage and touch voltage in the corner points, they are more increased than other parts.

Bayesian Analysis of the Detection Performance of the Lightning Imaging Sensors

Abstract Identical Lightning Imaging Sensors (LIS) aboard the Tropical Rainfall Measuring Mission satellite (TRMM LIS, 1998–2015) and International Space Station (ISS LIS, 2017–23) have provided over two decades of lightning observations over the global tropics, with ISS LIS extending coverage into the midlatitudes. Quantifying the detection performance of both LIS sensors is a necessary step toward generating a combined LIS climatological record and accurately combining LIS data with lightning detections from other sensors and networks. We compare lightning observations from both LIS sensors with reference sources including the Geostationary Lightning Mapper (GLM) and ground-based Earth Networks Total Lightning Network (ENTLN), Earth Networks Global Lightning Network (ENGLN), National Lightning Detection Network (NLDN), and Global Lightning Dataset (GLD360). Instead of a relative detection efficiency (DE) approach that assumes the perfect performance of the reference sensor, we employ a Bayesian approach to estimate the upper limit of the absolute DE (ADE) of each system being analyzed. The results of this analysis illustrate the geographical pattern of ADE as well as its diurnal cycle and yearly evolution. Reference network ADE increased by ∼15%–30% during the TRMM era, leading to a decline in TRMM LIS ADE. ISS LIS flash ADE has been relatively consistent at 61%–65%, about 4%–5% lower than TRMM LIS at the end of its lifetime.

Inter-Comparison of Lightning Measurements in Quasi-Linear Convective Systems

Data from four lightning networks collected during three quasi-linear convective systems (QLCS) are used to understand the differences in detection for optimizing their combined use. Additionally, using unique aspects from each network provides a more complete picture of lightning in a thunderstorm. The four lightning networks examined include a Lightning Mapping Array (LMA), the Earth Networks Total Lightning Network (ENTLN), the Geostationary Lightning Mapper (GLM), and the National Lightning Detection Network (NLDN). The data from each network are inter-matched and locations where each network uniquely detected a flash versus all are analyzed in reference to three QLCSs, including two QLCSs that occurred in the Southeast (22 March 2022 and 30 March 2022) during the Propagation, Evolution, and Rotation in Linear Systems (PERiLS) field campaign, and one case from Oklahoma (26 February 2023). Unique aspects of the lightning provided by each network are examined, including flash initiation altitude, size, type, and energy. Lightning flash trends and characteristics for each QLCS are similar between networks in general, but deviate in certain conditions and locations. Times of decreased matching between networks were associated with localized increases in lightning rates, smaller flash sizes, and lower-energy flashes. The differences in each network’s performance across the QLCSs demonstrates the importance of understanding the limitations in each and the advantage of using multiple networks.

Evaluate WRF-based Lightning Potential Index (LPI) lightning parameterization over Sri Lanka during second inter-monsoon in 2018

It is important to develop accurate and reliable lightning prediction system that can be contributed towards the safety of life, both concerning forecasting for the public safety and safety of aviation and electrical power. This study aims to evaluate WRF-based Lightning Potential Index(LPI) lightning parameterization and its applicability for predicting lightning over Sri Lanka during the second inter-monsoon. The WRF-ARW model 3.9.1 was used to produce predictions for three lightning events with various physical parameterization schemes with two nested domains with a resolution of 12km and 4km respectively. The model simulated LPI values were evaluated using the Earth Networks Global Lightning (ENGLN) dataset. Results show corresponding lightning simulations were produced with spatial distribution aligned with ground-based lightning data. Results consistently show a high correlation of the LPI index with an hourly CG flash rate over the three cases. Moreover, the WRF model was able to capture the lightning using LPI in Sri Lanka, suggesting that it can be used operationally to predict lightning potential region.

Machine-learning-based investigation of the variables affecting summertime lightning occurrence over the Southern Great Plains

Abstract. Lightning is affected by many factors, many of which are not routinely measured, well understood, or accounted for in physical models. Several commonly used machine learning (ML) models have been applied to analyze the relationship between Atmospheric Radiation Measurement (ARM) data and lightning data from the Earth Networks Total Lightning Network (ENTLN) in order to identify important variables affecting lightning occurrence in the vicinity of the Southern Great Plains (SGP) ARM site during the summer months (June, July, August and September) of 2012 to 2020. Testing various ML models, we found that the random forest model is the best predictor among common classifiers. When convective clouds were detected, it predicts lightning occurrence with an accuracy of 76.9 % and an area under the curve (AUC) of 0.850. Using this model, we further ranked the variables in terms of their effectiveness in nowcasting lightning and identified geometric cloud thickness, rain rate and convective available potential energy (CAPE) as the most effective predictors. The contrast in meteorological variables between no-lightning and frequent-lightning periods was examined for hours with CAPE values conducive to thunderstorm formation. Besides the variables considered for the ML models, surface variables and mid-altitude variables (e.g., equivalent potential temperature and minimum equivalent potential temperature, respectively) have statistically significant contrasts between no-lightning and frequent-lightning hours. For example, the minimum equivalent potential temperature from 700 to 500 hPa is significantly lower during frequent-lightning hours compared with no-lightning hours. Finally, a notable positive relationship between the intracloud (IC) flash fraction and the square root of CAPE (CAPE) was found, suggesting that stronger updrafts increase the height of the electrification zone, resulting in fewer flashes reaching the surface and consequently a greater IC flash fraction.

Feasibility of forecasting convective rain by diagnosing intracloud lightning jumps

This study conducted a comparison of the data quality of Earth Networks (EN) and Taiwan Power Company’s Total Lightning Detection System (TLDS) and evaluated the feasibility of using intracloud (IC) lightning to issue convective rain warnings. The results indicate uncertainties in the TLDS positioning of IC lightning. When forecasting convective rain on the basis of IC lightning jumps using EN data, the mean prefigurance and postagreement scores were 0.8 and 0.67, respectively, which were more favourable than the respective TLDS scores of 0.65 and 0.47. In regions with high TLDS positioning uncertainties, the use of EN data increased the number of prefigurance and postagreement hits in each analysis zone and raised the prefigurance and postagreement scores to 0.3 and 0.5, respectively. This indicates that analyses using EN data can reduce the risk of missed convective rain warnings when diagnosing IC jumps and can reduce the false alarm rate. In this study, IC jumps preceded convective rains by a maximum of 27.5–39.3 min on average in all analysis zones across Taiwan. The results suggest that diagnosing IC jumps to forecast convective rain is feasible, but until the uncertainties in the positioning of IC lightning using TLDSs have been remedied, EN data are the more suitable diagnostic choice.

Observation of Blue Corona Discharges and Cloud Microphysics in the Top of Thunderstorm Cells in Cyclone Fani

AbstractBlue corona discharges are often observed at the top of thunderclouds. They are bursts of streamers, but the cloud conditions that enable them are not well known. Here we present observations by the Atmosphere‐Space Interactions Monitor (ASIM) of 92 discharges during its ∼1 min pass over tropical cyclone Fani in the Bay of Bengal from 20:10:55 to 20:12:05 UTC on 30 April 2019. The discharges were observed in convective cells forming in the rainbands of the cyclone where Convective Available Potential Energy reached ∼6,000 J kg−1. The Cloud‐Aerosol Lidar and Infrared Pathfinder Satellite Observation satellite passed over one of the cells ∼12 min after ASIM from 20:23:58 to 20:24:14 UTC. It measured the cloud microphysics related to the discharges and indicated they occurred in a convection region with the cloud top overshooting for over 20 min. The updraft lifted ice particles to lower stratospheric altitudes and formed the gullwing‐shaped cirrus. The discharges are found at an average altitude of ∼16 km where the cloud environment contained ∼2 × 107 m−3 ice particles with ∼50 µm radius, resulting in a photon mean free path of ∼3 m. Around 20% of the blue corona discharges coincide with Narrow Bipolar Events indentified from the Earth Networks Total Lightning Network. Our observations suggest that the overshooting cloud top formed by deep convection and a surge in lightning activity facilitated conditions for the blue corona discharges. This work provides the first‐ever estimate of important microphysical parameters related to blue corona discharges based on data measurements, establishing a reference for future empirical and theoretical studies.

Toward Network‐Based Planetary Biosignatures: Atmospheric Chemistry as Unipartite, Unweighted, Undirected Networks

AbstractPrevious examinations of astrophysical chemical reaction networks found that Earth's atmospheric network was distinct in its hierarchical organization and scale‐free nature. If Earth's unique atmospheric network structure is due to the coevolution between the biosphere and atmosphere, it may hint at a novel planetary‐scale biosignature. Here, we use updated chemical reaction networks of planetary atmospheres and explore their topologies using a plethora of diagnostic techniques from network science and graph theory, including global metrics, centrality metrics, community detection, and cluster analysis methods. We show that the topologies of atmospheric chemical reaction networks of different planetary bodies in the Solar System are distinct from one another. While we find that model networks of Earth's atmosphere do not display scale‐free topology, Earth's chemical reaction network is nonetheless distinguishable from those of other Solar System atmospheres through various other metrics. Earth's network has the most nonrandom topology of all the planetary networks and, in some global metrics, is more similar to biological networks than are the other planetary networks. Finally, we discuss how further investigations of atmospheric chemical networks using more advanced network representations and novel network metrics may lead to the development of a network‐based biosignature applicable to exoplanets.

Monte Carlo Simulations for Evaluating the Accuracy of Geostationary Lightning Mapper Detection Efficiency and False Alarm Rate Retrievals

Abstract Performance assessments of the Geostationary Lightning Mapper (GLM) are conducted via comparisons with independent observations from both satellite-based sensors and ground-based lightning detection (reference) networks. A key limitation of this evaluation is that the performance of the reference networks is both imperfect and imperfectly known, such that the true performance of GLM can only be estimated. Key GLM performance metrics such as detection efficiency (DE) and false alarm rate (FAR) retrieved through comparison with reference networks are affected by those networks’ own DE, FAR, and spatiotemporal accuracy, as well as the flash matching criteria applied in the analysis. This study presents a Monte Carlo simulation–based inversion technique that is used to quantify how accurately the reference networks can assess GLM performance, as well as suggest the optimal matching criteria for estimating GLM performance. This is accomplished by running simulations that clarify the specific effect of reference network quality (i.e., DE, FAR, spatiotemporal accuracy, and the geographical patterns of these attributes) on the retrieved GLM performance metrics. Baseline reference network statistics are derived from the Earth Networks Global Lightning Network (ENGLN) and the Global Lightning Dataset (GLD360). Geographic simulations indicate that the retrieved GLM DE is underestimated, with absolute errors ranging from 11% to 32%, while the retrieved GLM FAR is overestimated, with absolute errors of approximately 16% to 44%. GLM performance is most severely underestimated in the South Pacific. These results help quantify and bound the actual performance of GLM and the attendant uncertainties when comparing GLM to imperfect reference networks.

Analyzing lightning characteristics in central and southern South America

Lightning data has become an integral part of weather observation and has been shown to be an effective tool for alerting and nowcasting. The Earth Networks Total Lightning Network (ENTLN) is a global lightning detection network established in 2009. ENTLN consists of roughly 1800 broadband electric field sensors globally that detect intra-cloud (IC) and cloud-to-ground (CG) lightning, with over 140 sensors in South America. In this study, we analyze the geographical variation of lightning characteristics such as density, peak current, and polarity, with the goal of better understanding the meteorological processes that produce these thunderstorms. Results indicate significant variation in average peak current per flash, with a maximum occurring throughout Northern Argentina, and a localized strong minimum North of São Paulo, Brazil. This minimum is coincident with a region dominated by negative IC flashes. In Central and Northern Argentina, 40%–50% of CG flashes were observed to be positive. Conventional thunderstorms typically have around 10% positive CG flashes. The large-scale patterns observed in this study support the findings of previous case studies regarding inverted polarity storms, mesoscale convective systems, and transient luminous event production, most of which had more limited scopes.

The “Weekend Effect” in Lightning Activity during Winter Thunderstorms over the Tel-Aviv, Israel, Metropolitan Area

We report the analysis of nine winter seasons (December, January, February (DJF)) aerosol and lightning data over the Tel-Aviv metropolitan area, Israel’s largest and densest urban region. Hourly averaged aerosol concentrations (PM2.5 and PM10) were obtained from 27 automatic air quality stations operated by the Ministry for Environmental Protection. Lightning data obtained from the Israeli Lightning Detection Network (ILDN) and by the Earth Network Total Lightning Network (ENTLN) for four overlapping seasons showed an irregular pattern, with lower activity during Sunday–Tuesday, and maximum activity on Wednesday and Thursday, but also on Saturday, when less pollution was present. The accepted explanation for the weekend effect is that increased amounts of particles tend to prolong the lifetime of storms and their total lightning amounts and change their intracloud/cloud-to-ground flash ratios. However, our results suggest that lightning is not directly related to the level of ambient pollution, but likely depends on the synergistic effects of desert dust and urban pollution particles in the area that affect the electrical structure of winter thunderstorms. Since the source of desert dust is independent of any regularity that can be attributed to anthropogenic activity (e.g., increased levels of pollution due to the traffic-load in Tel-Aviv), it obscures any weekend–weekday patterns that are often detected in other large metropolitan areas.

First comparisons of ball-lightning report website data with lightning-location-network data

Ball lightning (BL) is known primarily through observations of untrained eyewitnesses who see it fortuitously. Because this mode appears to be the primary source of data on ball lightning up to now, we established a website in June of 2020 designed to collect detailed reports of ball lightning sightings from eyewitnesses. As a first step toward reinforcing such reports with objective instrumented data, we have performed successful correlations of selected reports of ball lightning with lightning-location-network (LLN) data provided by Earth Networks. We linked ball lightning cases to an Earth Networks database of cloud-to-ground (CG) strokes by selecting strokes that were close, both geographically and in time, to the ball lightning incidents. The seven individual cases presented fit the previously established characteristics of ball lightning well, including demonstrating that BL does not necessarily form at the lightning ground strike location. Reports also emphasize certain features which are not always considered in theoretical treatments of ball lightning, e. g. its ability to pass through closed glass windows. We conclude with recommendations for both theorists and observational scientists.

3D triangulation of Transient Luminous Events over Africa

We present the first 3D triangulation of Transient Luminous Events (TLEs) over Africa. The 6 TLEs were simultaneously observed in the middle atmosphere from Sutherland and Carnarvon in South Africa, separated by 192 km, during the 2019 sprites campaign. These two distinctive locations have low radio interference and are free from light pollution. The lightning times, locations, peak current, and polarities, which initiated the observed TLEs, were obtained from the South African Lightning Detection Network and Earth Networks Total Lightning Networks. We investigate the TLEs' altitude and horizontal displacement from their parent lightning strokes. TLEs appear approximately 12.5 to 49.3 km away from their parent lightning strokes. We found that TLE altitudes range from 29 to 92.6 km. The lightning electric field and peak current may be related to the displacement of TLEs and the TLEs' horizontal spread.

A Statistical Analysis of Total Lightning Flashes and Peak Current from High-Resolution ENTLN Measurements in South China during 2017

Abstract In this study, the spatiotemporal characteristics of total lightning, including cloud-to-ground (CG) and intracloud (IC), flashes and the peak current of CG flashes are investigated using the observations from the Earth Networks Total Lightning Network (ENTLN) during 2017 over Guangdong Province, China, and its adjacent offshore regions, which are the capital of lightning activity and convective rainfall in China. The ENTLN detects 2 825 558 IC and 569 104 CG flashes, accounting for 83.2% and 16.8% of annual-total lightning flashes, respectively, and they are distributed during 224 lightning days or 3316 lightning hours over the regions. The lightning activity peaks in June (IC) and July (CG) with a diurnal peak at 1500 LST. On average, about 70 lightning days with 20 flashes per kilometer squared per year occur inland, and double that many occur offshore. High lightning days and flash density are found in the central and southwestern mountainous region and the Guangdong–Hong Kong–Macao Greater Bay area, where convective activity is high. An analysis of the CG observations reveals that negative CG (NCG) produce an average of 1.96 strokes, with 5% of them having more than six return strokes. The peak currents of NCG strokes increase with increasing multiplicity. Results also show that the peak current and accumulated value of NCG flashes are the appropriate indicators to characterize the intensity of lightning activity in terms of spatiotemporal variations, especially in the mountainous region. It is concluded that IC observations provide more distinct spatial distribution of lightning activity than CG flashes; and those appropriate indicators should be used to represent the peak current of multistroke NCG flashes.

Quantification of lightning-produced NO<sub><i>x</i></sub> over the Pyrenees and the Ebro Valley by using different TROPOMI-NO<sub>2</sub> and cloud research products

Abstract. Lightning, one of the major sources of nitrogen oxides (NOx) in the atmosphere, contributes to the tropospheric concentration of ozone and to the oxidizing capacity of the atmosphere. Lightning produces between 2 and 8 Tg N yr−1 globally and on average about 250 ± 150 mol NOx per flash. In this work, we estimate the moles of NOx produced per flash (LNOx production efficiency) in the Pyrenees (Spain, France and Andorra) and in the Ebro Valley (Spain) by using nitrogen dioxide (NO2) and cloud properties from the TROPOspheric Monitoring Instrument (TROPOMI) as well as lightning data from the Earth Networks Global Lightning Network (ENGLN) and from the EUropean Co-operation for LIghtning Detection (EUCLID). The Pyrenees are one of the areas in Europe with the highest lightning frequencies, which, along with their remoteness as well as their very low NOx background, enables us to better distinguish the LNOx signal produced by recent lightning in TROPOMI NO2 measurements. We compare the LNOx production efficiency estimates for eight convective systems in 2018 using two different sets of TROPOMI research products provided by the Royal Netherlands Meteorological Institute (KNMI) and the Deutsches Zentrum für Luft- und Raumfahrt (DLR). According to our results, the mean LNOx production efficiency in the Pyrenees and in the Ebro Valley, using a 3 h chemical lifetime, ranges between 14 and 103 mol NOx per flash from the eight systems. The mean LNOx production efficiency estimates obtained using both TROPOMI products and ENGLN lightning data differ by ∼ 23 %, while they differ by ∼ 35 % when using EUCLID lightning data. The main sources of uncertainty when using ENGLN lightning data are the estimation of background NOx that is not produced by lightning and the time window before the TROPOMI overpass that is used to count the total number of lightning flashes contributing to freshly produced LNOx. The main source of uncertainty when using EUCLID lightning data is the uncertainty in the detection efficiency of EUCLID.

Is a Penny a Month a Basic Income? A Historiography of the Concept of a Threshold in Basic Income

Abstract Does a penny per month constitute a Basic Income? Were that penny to be paid individually, universally, and unconditionally, the answer would be ‘yes’, following the definition of Basic Income given by some of its leading advocates, be it organisations like the Basic Income Earth Network (BIEN) or prominent scholars such as Philippe Van Parijs. Some might be puzzled as to how this could be ‘a capitalist road to communism’, or give us ‘freedom as the power to say no’, both of which have been advocated by prominent researchers. The purpose of this paper is not to argue for or against their definition, but rather to situate it historiographically, enabling fruitful discussion. The paper will show how there was a widely shared assumption in the 1970s and 80s, at the early stages of both academic articulations and public discussion, that Basic Income comes with some notion of a threshold or level to be taken as a minimum or as adequate. The paper goes on to outline three issues that arise once the concept of a threshold is dropped from the definition. Examined in addition are five justifications for doing so. Much like any other idea, the concept of Basic Income is a social construct. By situating it here within a historical perspective, we wish to facilitate academic discussion regarding both the achievements and erasures that have occurred as a consequence of the concept’s academic refinement—refinement that is in itself majorly indebted to BIEN and Van Parijs.

Upgrades of the Earth Networks Total Lightning Network in 2021

The Earth Networks Total Lightning Network (ENTLN) launched a new processor (P2021) in December 2021. Some major upgrades were made in the new processor, including a new classification algorithm, a new propagation model, and regional data processing architecture. Ground-truth datasets of natural and rocket-triggered lightning acquired in Florida were used to evaluate the performance characteristics of the new processor. Compared to the last processor launched in 2015 (P2015), the stroke classification accuracy increased from 91% to 94% for natural lightning and from 86% to 88% for rocket-triggered lightning. The location accuracy improved significantly with the median location error decreasing from 215 m to 92 m. On a global scale, we found the number of pulses detected by the ENLTN increased in all regions with an overall detection gain of 149%. One can see modest gains in detection in regions with a fairly dense network of sensors and significant gains in regions where sensor density is much lower. Each of the major upgrades as well as their influences on the performance characteristics of the ENLTN are discussed.

Convective‐scale numerical weather prediction and warnings over Lake Victoria: Part I—Evaluating a lightning diagnostic

AbstractRecently a new lightning diagnostic was added to convective‐scale configurations of the Met Office Unified Model (UM). The characteristics and skill of the hourly total ‘flash origin density’ in the 4.4 km Tropical Africa UM were evaluated against Earth Network (EN) lightning observations specifically deployed in the greater Lake Victoria region as part of the HIGHWeather impact lAke sYstem (HIGHWAY) project. Lightning poses a significant risk to fishermen on Lake Victoria, and better forecasts over the lake would be advantageous for the development of better warning systems. The model lightning density was evaluated in two ways: a classical gridded categorical data analysis and the spatial coverage–distance–intensity method. Given the spatial sparseness of the observation type, Gaussian kernel dressing was applied for the latter, to increase the horizontal ‘footprint’, partly to reflect that lightning can travel horizontally for more than 10 km, and to reduce the representativeness mismatch between the lightning observations and the model forecast lightning density. Considering hourly forecasts, lightning shows a marked diurnal, geographical and seasonal variation over to the west and east of the lake. The model is not producing enough lightning flashes over the lake, nor does it produce enough spatially, especially overnight. The forecast location of lightning shows little skill, with timing differences in the peak, especially over the lake. For warnings, the removal of timing offsets through the creation of a ‘maximum‐in‐the‐day’ field is considerably less biased and more skilful than using the hourly flash densities.