Thunderstorm Occurrence Research Articles

Cb‑Fusion – forecasting thunderstorm cells up to 6 hours

As a severe weather phenomenon, thunderstorms can cause casualties and economic loss to the human society. A reliable forecast of these weather events would help to avoid or at least mitigate this damage. To date, the forecasting of thunderstorms however is still a challenge, especially for lead times of one hour and beyond. In this study we present a methodology to forecast deep-convection for several hours lead time: Cb-Fusion estimates the likelihood of thunderstorm occurrence for up to 6 hours in advance over a part of Central Europe, using a data fusion technique that blends data of multiple sources from observations, nowcasts, and numerical weather predictions with a high update rate. The Cb-Fusion is set up to operate in near real time. The skill of Cb-Fusion is evaluated based on 1743 hours of thunderstorm observations collected during the months April to October, 2019. Three categories of thunderstorm size have been distinguished: ‘large’ for a coverage area larger than 5000 km2, ‘medium’ for a coverage area between 5000 km2 and 500 km2, and ‘small’ for a coverage area smaller than 500 km2. Compared to thunderstorm forecasts from numerical models alone, the combination of data from various sources by Cb-Fusion results in a significantly better forecast skill. The study reveals that the forecast is reliable for up to 3 hours lead time for ‘medium’ and ‘large’ scale thunderstorms (median POD of 0.6 – 0.9) but little skill is found for ‘small’ scale thunderstorms and lead times between 3 and 6 hours (median POD of 0.05). It is argued that Cb-Fusion provides meaningful improvements in forecasting thunderstorms for various users.

An Investigation of circulation regime variability and dangerous weather phenomena in Russia in the 21stcentury

The variability of Northern Hemisphere circulation types from 1899-2018 are analyzed and the classification of atmospheric circulation types in the early 21st century are shown. These are associated with the frequency of dangerous weather phenomena, natural processes, and disasters within Russia during this time. Against this background, the synoptic conditions for the formation of extreme precipitation and floods, as well as natural fires in 2019 are considered. It is shown that an increase in the frequency of occurrence of stationary anticyclones and blocking flow regimes as well as an extension of the Mediterranean storm track was associated with an increase in the frequency of extreme precipitation and dangerous natural processes. Additionally, the occurrence of thunderstorms, often dry, led to the development of natural fires.

Analysis of various thermodynamic instability parameters and their association with the rainfall during thunderstorm events over Anakapalle (Visakhapatnam district), India

Thunderstorm events usually take place in cumulonimbus clouds which are complemented with intense rainfall and high-speed winds. In general, rainfall parameter has massive significance when compared to other parameters. In this paper, a group of thunderstorm-related stability parameters were analyzed for pre-monsoon season only. Later, we also tried to study the association between thunderstorm-related stability parameters and rainfall parameter in pre-monsoon season over Anakapalle (Visakhapatnam district) during 2001–2010. We have utilized ERA-Interim ECMWF reanalysis daily datasets for this study. We also tried to compare IMD thunderstorm occurrence days with NOAA CPC-calculated rainfall days in pre-monsoon season over Anakapalle region for every year during 2001–2010. Out of those parameters, upward vertical velocity, convective available potential energy, K-index (KI), humidity index and total totals index parameters have shown good thresholds supporting the rainfall activity during pre-monsoon season. Later, we have also attempted the prediction of DCI and KI parameters over Anakapalle region using artificial neural network (ANN) and auto-regressive moving average (ARMA) techniques. In comparison between the two techniques, ANN technique has shown good correlation with ERA-Interim ECMWF reanalysis data.

Projected changes to severe thunderstorm environments as a result of twenty-first century warming from RegCM CORDEX-CORE simulations

Hazardous weather related to the occurrence of severe thunderstorms including tornadoes, high-winds, and hail cause significant damage globally to life and property every year. Yet the impact on these storms from a warming climate remains a difficult task due to their transient nature. In this study we investigate the change in large-scale environments in which severe thunderstorms form during twenty-first century warming (RCP2.6 and RCP8.5) in a group of RegCM CORDEX-CORE simulations. Severe potential is measured in terms of Convective Available Potential Energy (CAPE) and vertical wind-shear during the severe seasons in three regions which are known to currently be prone to severe hazards: North America, subtropical South America, and eastern India and Bangladesh. In every region, environments supportive for severe thunderstorms are projected to increase during the warm season months in both the RCP2.6 and RCP8.5 scenarios during the twenty-first century. The number of days supportive for severe thunderstorms increases by several days per season over the vast majority of each region by the end of the century. Analyzing the CAPE and shear trends during the twenty-first century, we find seasonally and regionally specific changes driving the increased severe potential. Twenty-first century surface warming is clearly driving a robust increase in CAPE in all regions, however poleward displacement of vertical shear in the future leads to the displacement of severe environments over North America and South America. The results found here relate that severe impacts in the future cannot be generalized globally, and that regionally specific changes in vertical shear may drive future movement of regions prone to severe weather.

Towards understanding the impact of convective weather on aircraft arrival traffic at the Hong Kong International Airport

Convective weather (thunderstorm clusters with horizontal extents from tens to hundreds of kilometers, and vertically up to 10 - 15 kilometers in the troposphere) along aircraft flight trajectory is a key factor influencing air traffic management decisions and aviation safety. For an airport, aircraft arrival performance is typically more sensitive to weather conditions than departures. The occurrence of thunderstorms along or around flight routes may force the arriving aircraft to hold/vector in the terminal maneuvering area (TMA) following air traffic control (ATC) instructions, leading to additional fuel burn and delays. Such effects could cascade and bring further negative impacts to air traffic inside TMA. In this work, we perform a joint analysis of high-resolution radar rainfall distribution with 2-D aircraft flight trajectory data surrounding the Hong Kong International Airport (HKIA) to uncover the potential spatio-temporal impact of thunderstorms on air traffic. Results indicate that aircraft arrival traffic is sensitive to the location of convective weather in TMA, which also introduces a time lag effect on subsequent air traffic on the order of hours.

A rare winter thunderstorm event near Hong Kong

A rare winter thunderstorm event near Hong Kong is documented in this paper. The thunderstorm occurrence was confirmed by the observation of one of the authors and lightning location information systems (both regional and global). It developed over the frontal cloud band associated with the winter monsoon. Observing the upper air measurements, however, the thermodynamic conditions did not seem favorable to the occurrence of the thunderstorm. The convective available potential energy could not be determined in the cool and dry lower troposphere. There was only a moister layer in the middle troposphere. The weather forecaster did not anticipate the occurrence of thunder and lightning. However, some features could have changed dynamically, which suggested the possibility that a thunderstorm would take place. For instance, waves could be identified in the middle troposphere and a broad troughing flow could be analyzed in the lower troposphere. The integrated wave vapor was also of a rather high level. It is hoped that this paper can serve as a useful reference for forecasters in assessing the possibility of thunderstorms occurring during winter over subtropical latitudes in association with frontal cloud bands.

A STUDY OF THUNDERSTORM FEATURES OVER SRIKAKULAM REGION ON 26TH APRIL, 2020

Thunderstorms are the most destructive activities that pose a serious threat to mankind. The rainfall associated with these events cause heavy damage to crops, property and human lives. Studies analyzing these parameters have great significance. Srikakulam is one of the frequently exposed regions to intense thunderstorms during pre-monsoon season. An attempt has been made to study the thunderstorm features over Srikakulam region (18.34oN 83.9oE) using the ECMWF ERA5 reanalysis data. We examined the hourly data for the entire Andhra Pradesh (AP) region for the particular date (26th April, 2020). We used parameters like Lifted Index (LI), K-Index (KI), Humidity Index (HI), Convective Available Potential Energy (CAPE), Convective Inhibition (CIN), Thunderstorm Prediction Index (TPI) and Total precipitable water (TPW) for analyzing the thunderstorm activity. High rainfall activity was seen over Srikakulam region. Almost 16 cm rainfall occurred during thunderstorm activity. LI, HI, KI, CAPE, CIN, TPI and TPW parameters were indicated high threshold before 4 hours at the region of thunderstorm occurrence. CAPE values were greater than 3500 J/kg.

Determining future thunderstorm-prone environments in Southern Ontario by using statistical downscaling to project changes in convective available potential energy (CAPE)

Considering the potential risks associated with thunderstorms, to date, there has been limited analysis on the projection of thunderstorm occurrence trends in Canada. The small spatial and temporal scales of thunderstorms are not resolved in global climate models (GCMs). In this study, the relationship is established between thunderstorm observations from nine weather stations across Southern Ontario, Canada, and daily maximum convective available potential energy (CAPE). The results from the correlation analysis between CAPE and thunderstorm days suggested that the probability of observing a thunderstorm increases as maximum daily CAPE increases. We then utilize the novel approach of applying statistical downscaling (SDSM) to CAPE. After regenerating CAPE over a 30-year reference period (1981–2010) at each weather station, it was determined that the SDSM-modeled CAPE values well compared to observed CAPE values. Future CAPE values up to the end of the current century are then projected using the SDSM models for each station in combination with three GCMs for future climate. The forecast from the downscaling suggested large increases, as much as tripling, in annual mean CAPE, summer mean CAPE, and number of days exceeding a 50% probability and 80% probability of observing a thunderstorm at all weather stations under SRES business-as-usual and RCP 8.5 scenarios for the study period of 2011–2100. All else being equal, this suggests an increase in the number of days with conditions favorable for thunderstorms under a warmer climate.

Long-term analysis of thunderstorm-related parameters over Visakhapatnam and Machilipatnam, India

In this present work, an attempt has been made to analyze various thunderstorm-related parameters and their influence over the two stations Visakhapatnam (VSK) and Machilipatnam (MTM). The thunderstorm-related parameters used in the present study are convective available potential energy (CAPE), lifted index, K-index, total totals index (TTI), humidity index, convective inhibition, thunderstorm prediction index (TPI), deep convective index (DCI) and updraft vertical velocity. This analysis was carried out using NCEP NCAR reanalysis monthly data for the time period from 1948 to 2012. These parameters have given good guidance for studying the thunderstorm event. We also analyzed IMD thunderstorm occurrence days reported at two stations, i.e., VSK and MTM with NCEP NCAR (daily data) calculated CAPE, TTI, TPI and DCI parameter threshold days in pre-monsoon season for every year during the time period 2010 to 2019. Out of those four parameters, TTI has shown good correlation with the IMD recorded days. So we have attempted the prediction of thunderstorms using artificial neural network (ANN) and auto-regressive moving average (ARMA) techniques for TTI parameter. While using these techniques, we have experimented in three training sets, i.e., 90%, 80% and 70%. Another attempt has been made to assess the skill of ARMA and ANN techniques in forecasting the occurrence of thunderstorm activity at VSK and MTM stations. The present study suggests that ANN has high skill than ARMA. From this study, we can understand that VSK has more chances for thunderstorms than MTM.

CORRECTION

Meteorological ApplicationsVolume 27, Issue 3 e1922 CORRECTIONOpen Access CORRECTION This article corrects the following: A deep-learning model to predict thunderstorms within 400 km2 South Texas domains Hamid Kamangir, Waylon Collins, Philippe Tissot, Scott A. King, Volume 27Issue 2Meteorological Applications First Published online: April 13, 2020 First published: 09 June 2020 https://doi.org/10.1002/met.1922AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat In Kamangir et al. (2020), the following errors were present in the original version of the article published: The article title was originally “Deep-learning model used to predict thunderstorms within 400 km2 of south Texas domains” but should have been “A deep-learning model to predict thunderstorms within 400 km2 South Texas domains”. In the Abstract, the first sentence originally read “…thunderstorm occurrence within 400 km2 of south Texas (USA) domains…” but should have read “…thunderstorm occurrence within 400 km2 South Texas domains…”. The word “denoising” was incorrectly written as “deionizing” throughout the text. In section 2.3.1 Shallow neural network, the sentence “The present study is an adjustment of Collins and Tissot (2015)…” should have read “This study was an adjustment of Collins and Tissot (2015)…”. These errors have now been corrected in the article online. REFERENCE Kamangir, H., Collins, W., Tissot, P. and King, S.A. (2020) A deep-learning model to predict thunderstorms within 400 km2 South Texas domains. Meteorological Applications, 27, e1905. https://doi.org/10.1002/met.1905. Volume27, Issue3May/June 2020e1922 ReferencesRelatedInformation

A deep‐learning model to predict thunderstorms within 400 km2 South Texas domains

AbstractA deep‐learning neural network (DLNN) model was developed to predict thunderstorm occurrence within 400 km2 South Texas domains for up to 15 hr (±2 hr accuracy) in advance. The input features were chosen primarily from numerical weather prediction model output parameters/variables; cloud‐to‐ground lightning served as the target. The deep‐learning technique used was the stacked denoising autoencoder (SDAE) in order to create a higher order representation of the features. Logistic regression was then applied to the SDAE output to train the predictive model. An iterative technique was used to determine the optimal SDAE architecture. The performance of the optimized DLNN classifiers exceeded that of the corresponding shallow neural network models, a classifier via a combination of principal component analysis and logistic regression, and operational weather forecasters, based on the same data set.

Thunderstorm dynamics through wind profiling features: the first observation by stratospheric-tropospheric radar of GU

The wind profiling features received from recently installed stratospheric-tropospheric (ST) radar of Gauhati University (GU) ($$26.2^{\circ }$$ N, $$91.75^{\circ }$$ E) are presented in the paper along with the validation process of the data by using radiosonde-derived wind parameters. The final focus of the paper is directed onto the thunderstorm-related changes in wind fields covering both vertical and composite components. Starting with the seasonal thunderstorm occurrence character over Guwahati through eight years of observations, the importance of the work on this meteorologically sensitive subtropical zone is highlighted. For identifying the process of growth of thunderstorm, the magnitudes of updrafts and downdrafts are extracted from the ST Radar-derived storm-time wind profilings. For evaluation of atmospheric system dynamics, the structure constant $$C_n^2$$ is calculated through radar Bragg-scattered signals and the resultant modifications induced by the thunderstorm on this parameter are presented and brought into ambit of discussions. The paper finally identifies the storm-induced effects in the wind field-derived variabilities, for possible adaptation as inputs for a thunderstorm predictive model.

Comparison of Lightning Forecasts from the High-Resolution Rapid Refresh Model to Geostationary Lightning Mapper Observations

Abstract The ability of the High-Resolution Rapid Refresh (HRRR) model to forecast the location of convective storms is of interest for a variety of applications. Since lightning is often present with intense convection, lightning observations from the Geostationary Lightning Mapper (GLM) on GOES-East are used to evaluate the performance of the HRRR lightning forecasts from May through September during 2018 and 2019. Model skill is presented in terms of the fractions skill score (FSS) evaluated within circular neighborhoods with radial distances from 30 to 240 km. Case studies of individual events illustrate that the HRRR lightning forecasts FSS varies from storm to storm. Mean FSS is summarized for the months with peak lightning activity (June–August) for the west, central, and east United States. Our results suggest that forecasters should use HRRR lightning forecasts to indicate general tendencies for the occurrence, region, and timing of thunderstorms in a broad region rather than expect high forecast accuracy for lightning locally. For example, when FSS is evaluated within small neighborhoods (30-km radius), mean FSS drops sharply after the first two hours of model integration in all regions and during all hours of the day. However, when evaluated within larger neighborhoods (60-km radius and larger), FSS in the western United States and northern Mexico remains high for all lead times in the late afternoon and early evening. This result is likely due to the model capturing the tendency for convection to break out over higher terrain during those hours.

The Role of Atmospheric Processes Associated with a Flash-Flood Event Over Northwestern Turkey

This study investigates the atmospheric mechanisms triggering flash-flood event in Thrace Basin of Turkey on November 27, 2018. Underestimation of this extreme precipitation amounts by NWP global and regional models (i.e. ECMWF, ALARO, WRF) and other meteorological difficulties (i.e. complex topography, land-sea interactions) in weather forecasting disabled disaster risk reduction before the event occurred. Detailed synoptic, thermodynamic, in-situ, and remote sensing analyzing results showed that significant amount of moisture during the afternoon times of the day was transferred to the atmosphere (from ground to 300-hPa) as a consequence of the excessive heating of sea surface temperatures (SSTs) of the Aegean Sea (16.5 °C in Ayvacik-Gulpinar place, 0.9 °C above its long-term normals). Strong southwesterly wind speeds associated with slow meridional movement of mid-latitude cyclone from its origin to the Eastern Mediterranean (EM) enabled transferring of relatively warm moist air to the land areas of Thrace Region (> 300 km fetch distance). Strong updraft and instability conditions under developed a supercell resulted with lightning (totally 63 cloud-to-ground and 59 intra cloud) and heavy rainfall especially Suloglu, Kofcaz, and Edirne settlements with the 12-hour total amounts 160.0, 123.0, and 97.4 mm (rainfall return period ~ 100 years), respectively. Flash-flood event caused numerous injuries and the death of a person and damaged, automobiles, houses, crops, and infrastructure of the Edirne and its neighboring settlements. From Showalter, K, Total of Totals, SWEAT, and CAPE instability indices; SWEAT is most appropriate to represent high possibility of occurrence of severe thunderstorms over the Edirne province owing to low-level moisture, warm air advection and low and mid-level wind speed terms in its equation.

Spatiotemporal variations of thunderstorm frequency and its prediction over Bangladesh

Understanding the spatiotemporal variations of thunderstorm (TS) frequency over Bangladesh under the changing climate is of immense importance but remains a challenging task due to the lack of persistent and reliable weather data set. This study demonstrates a statistically significant increasing trend in the occurrence of monthly, seasonal, and annual TS frequencies except for decadal periods across Bangladesh during the past 6 decades. Further, land use influences and seasonal variabilities in convective available potential energy (CAPE) are also explored in various regions to understand the spatial variabilities among different parts in Bangladesh. The results show the highest occurrence of TS in monsoon and the lowest occurrence in the winter season. The results of Ward’s Hierarchical Cluster analysis (WHCA) illustrate four clusters for the whole country, where the higher TS frequency was observed in northeastern and northern regions and the lower in the southern part of Bangladesh due to the variation of CAPE and development of the low-pressure system. The occurrence of TS frequency maxima and lower coefficients of variation are found in the pre-monsoon and monsoon seasons. The variation coefficient of the winter season is high, while the pre-monsoon season is low. The results of Mann–Kendall (MK) and Spearman’s rho (SR) tests reveal the highest increase and decrease trends in M. Court and Tangail stations. The ARIMA model outcomes are consistent with the findings of MK and SR tests. Overall, this study reveals that elevated CAPE and climate warming may be reasoned of increasing the TS frequency over Bangladesh.

Lightning Changes in Response to Global Warming in Rio de Janeiro, Brazil

Physical concepts based on the Clausius-Clapeyron relation and on the thermodynamics and aerosol characteristics associated with updrafts, global climate models assuming different parametrizations and lightning-related output variables, and lightning-related data (thunderstorm days) are being used to infer the lightning incidence in a warmer planet, motivated by the global warming observed. In all cases, there are many gaps to be overcome making the lightning response to the global temperature increase still unpredicted. Values from almost 0% (no increase) to 100% have been estimated, being 10% the most common value. While the physical concepts address only part of the problem and the global climate models need to make many simple assumptions, lightning-relate data have strong time and space limitations. In this context, any new evidence should be considered as an important contribution to better understand how will be the lightning incidence in the future. In this article, we described new results about the occurrence of thunderstorms from 1850 to 2010 (a period of 160 years) in the city of Rio de Janeiro, in the Southeast of Brazil. During this period thunderstorm days were recorded in the same location, making this time series one of the longest series of this type available worldwide. The data support an increase of 21% in the mean annual thunderstorm days during the period, while surface temperature increased by 0.6°C during the period. Considering that the mean annual number of thunderstorm in the beginning of this period was 29, we found an increase of one thunderstorm day per 0.1°C of increase in the surface temperature. Assuming that the number of lightning flashes per thunderstorm remains approximately constant during the period, this number corresponds to an increase in the lightning flash rate of approximately 35% per °C of increase of temperature. In addition, considering that the increase of the global temperature during the period was almost the same that observed in Rio de Janeiro, we can conclude that this increase in the lightning flash rate is due to the global warming with no effect of urban activity. Finally, we found that monthly thunderstorm days and monthly mean surface temperature show a linear correlation with a coefficient of 0.9 along the period.

Probabilistic thunderstorm forecasting by blending multiple ensembles

In numerical weather prediction models, point thunderstorm forecasts tend to have little predictive value beyond a few hours. Thunderstorms are difficult to predict due largely to their typically small size and correspondingly limited intrinsic predictability. We present an algorithm that predicts the probability of thunderstorm occurrence by blending multiple ensemble predictions. It combines several post-processing steps: spatial neighbourhood smoothing, dressing of probability density functions, adjusting sensitivity to model output, ensemble weighting, and calibration of the output probabilities. These operators are tuned using a machine learning technique that optimizes forecast value measured by event detection and false alarm rates. An evaluation during summer 2018 over western Europe demonstrates that the method can be deployed using about a month of historical data. Post-processed thunderstorm probabilities are substantially better than raw ensemble output. Forecast ranges from 9 hours to 4 days are studied using four ensembles: a three-member lagged ensemble, a 12-member non-lagged limited area ensemble, and two global ensembles including the recently implemented ECMWF thunderstorm diagnostic. The ensembles are combined in order to produce forecasts at all ranges. In most tested configurations, the combination of two ensembles outperforms single-ensemble output. The performance of the combination is degraded if one of the ensembles used is much worse than the other. These results provide measures of thunderstorm predictability in terms of effective resolution, diurnal variability and maximum forecast horizon.

Impact of Atmospheric Stability Indices on Convective Systems Over India and Srilanka

Thunderstorms are real-time global phenomena, as their occurrence can take place at anytime at any place. Though their duration is less when compared to large scale processes, their damage is devastating to human life. Thunderstorms are linked with damage factors such as lightning, damaging wind, hails and rain. Real-time satellite data provide atmospheric data which is useful for prediction of thunderstorms. In this paper, an attempt is made to analyze the statistical based stability indices from INSAT-3D, MODIS and ECMWF satellites for the now casting of thunderstorms. The occurrences of severe thunderstorms over India and Srilanka during the month of October 2013, 2014 and 2015 have been analyzed. In these three years, five severe thunderstorm cases were identified using Insat-3D cloud images and thunderstorm reports. Atmospheric stability indices such as K Index (KI), Lifted Index (LI), Total Totals Index (TTI), Total Precipitable water (TPW), Humidity Index (HI) related with severe convection system over India and Srilanka during October month were identified to provide guidance for the study of convection and thunderstorm activity. These indices give us a clear indication of development of convective system before 3 to 4 hours. Results of this study indicate the importance of satellite data for studying the development and decay of convective systems.

Frequency of severe thunderstorms across Europe expected to increase in the 21st century due to rising instability

We show that the frequency of damaging convective weather events including lightning, hail and severe wind gusts will likely increase over Europe until the end of this century. We apply a set of additive regression models to an ensemble of 14 regional climate simulations and find that convective instability will increase as a result of rising humidity near the earth’s surface. Even though a slight decrease in thunderstorm occurrence in southwestern and southeastern Europe is projected, the probability of severe weather will increase throughout Europe, in particular for very large hail. It might be expected that Arctic amplification would lead to a weaker jet stream and, thus lower vertical wind shear, but we find instead that the jet changes little or even increases in situations with convective instability. To cope with the rising hazard probabilities, risk models will need to be adapted, while investment in sturdier structures, like the use of hardened glass in greenhouses or solar panels, will become more cost-effective. Furthermore, the need will grow to advise the public on loss prevention by taking precautionary measures as storms approach.

Thunderstorm occurrence at ten sites across Great Britain over 1884–1993

AbstractThe UK Met Office's Daily Weather Reports (DWR) contain extensive logs of UK thunderstorm activity. To date, only a very small fraction of these data have been digitized as part of the MIDAS dataset, and exclusively after 1950. Using the recently‐scanned UK Met Office Monthly Weather Reports (MWR), which are based on a subset of the observations that form the DWR, we here provide digitized data and a summary of thunderdays from 10 long‐running British stations over the period 1884–1993. The data are presented ‘as is’, with no attempt to provide any corrections or calibration. For 4 of the 10 stations, thunderday observations were discontinued at various times between 1949 and 1964, and it is necessary to switch to a neighbouring station in order to continue the series. Approximately half the series exhibit sharp drops in thunderdays at various points between 1960 and 1990, although none are coincident with known station changes. Comparison with nearby MIDAS stations suggests the low thunderdays are the result of changes in observing practice, rather than genuine changes in thunderstorm occurrence. These potential data issues limit interpretation of the long‐term trends. However, it can nevertheless be concluded that none of the stations show the expected increase in thunderdays as a result of the rise in surface temperature over the 20th century. In order to provide more quantitative determination of the long‐term trends in thunderstorm occurrence, we advocate further digitization efforts to recover the data from the numerous stations in the MWRs, and subsequent analysis of the common signals across neighbouring stations.